Home

Progetto e realizzazione del sistema di gestione autonoma del volo

1. Coll Throttle Curve Figure 87 Onboard Control Loop 117 The controller structure is the one shown in figure 86 In order to understand the PID FPGA implementation it 1s necessary to define the following unity of measure and scale factors AHRS Euler Angle Output al16 Y al16 180 2 AHRS NED Velocity Output vI16 m s v116 256 2 AHRS Latitude Longitude Output LI32 J LI32 1 80 27 servo s116 180 2 PID Attitude servo angle out 116 sf16 And us s176 180 2 10 104 3 s116 1737 2 VzO vI16 m s us VzO 256 2 10 104 3 VzO 2455 2 PID Vz output And VzO us 13 3504 Initial collective 132 c132 VzO 2 1s 1709 K_Vx and K_Vy Gains Attitude m s Attitude m s al16 v116 1 42 Table 52 Unity of Measures and scale factors used in the control code Each PID implementation will be described in the next sub sections Vx theta Pl A schematic of the Vx Theta nested PI software is reported in figure 88 Input Parameters Input Trim Long ps Parameters vx SP vI16 PI vx SP al6 waa T H L i Pio ong s116 WM Long z al16 Vx v16 calculated from AHRS data Figure 88 Schematic of the FPGA Vx Theta nested PI The first PI implement the forward velocity control along the trajectory This PI calculate the reference theta attitude to b
2. 22 Mission Simulation Environment iessen a O dee 26 AV amp NGCS Simulink model sesssses assess nescatasncivencn puadeneetaicteanineeramasiseansdtcraceshe 28 Control Modes and Force Feedback Law cccccceeseessscsceeceeeeeeeeeeeeeeseeseeeeeeseens 30 Max Acceleration at Constant Altitude A in figure is that of equation 2 1 42 31 Figure 9 Max Acceleration along a descending flight path A is that of extended approach CS a seta casera E Stitt E EI E te E na ert core tata ee ee eee E 31 Figure 10 Forward speed vs Time during the max acceleration phase cceeeeeeeeeees 33 Figure 11 AV Guidance Simulink blocks v ceececcsssesssssessssesesesesecscssevsvsvsvsesecasassvavsvseseeeeees 34 Fig re 1257 V lateral rack control strategy cgccaceds st aiciastiadewsaerisnun E A ae 35 Figure 13 Simulation of lateral track control 43 cccccsessesssseeeeeseeeseeeeeeeceeeeeeeeeeeeeeees 36 Fieu TAGES Tab View COC ao a seein ae aan te eee ees 37 Figure 15 SIl Connection Manager Schematic cjecixeeuiio ined eae es 39 Figure 16 Microsoft Sidewinder Force Feedback II Joystick c cccccscssessessseesseeeeseseeee 40 Figure 17 Spring Forces Programmed in Joystick cccccsssscsccceceeeeeeeeeceeeseeseseeeeessessenaeaes 41 Figure 18 Comparison of Commanded and Actual Stick Position 000000000onnnnnneeeeseeeeeeens 4 Figure 19 Description of Typical SOmtstOp cccccccccccccccc
3. Figure 47 NAV420CA on Test Rig Reference points in terms of angular rates and attitude were given by means of the test rig control unit NAV420 responses were recorded by means of the NAVView software provided by CrossBow 53 Diagrams of some experimental results reported in figure 48 confirmed the good quality of the NAV420 measurement capabilities i u E amp u h sb 8 amp amp Bod Asis cen Se ra kaidi B a w Fipa Ana ibe a E H J Feti Angie ete 5 i 2 aa BT aa ag Time 9 Figure 48 NAV420 Test Rig Measurements 71 NAV 420 Settings The LabView software designed to change the NAV420 default settings is reported in the enclosed CD The CrossBow NAV420 provides information to the user by means of a RS 232 protocol Therefore the developed software was used to set the transmission baud rate the packet output rate and the output packet type This can be done by using an appropriate command list reported in the NAV420 user manual The command list was written directly to the NAV 420 EEPROM using the power up configuration field so that the configuration settings are used always as default values by the system For the onboard software to work properly the NAV420 default settings must be as follows Baud Rate 57600 bps Packet Output Rate 100 Hz Packet Type NAV mode AHRS Mounting and Alignment The CrossBow NAV 420 was installed inside the avionics box of the UNIBO RUAV
4. S Tsach D Penn and A Levy Israel Aircraft Industries IAI Advanced Technologies and Approaches for Next Generation UAVs ICAS 2002 R Pretolani G Saggiani B Teodorani RUAV Ground Support Basic Requirements amp First Layout Internal Report CAPECON Report UNIBO ID 6 1 1 31 July 2003 R Pretolani Progetto e realizzazione del sistema di navigazione guida e controllo per un elicottero con capacit di volo autonomo II School of Engineering University of Bologna PhD Thesis 2007 B Mettler Identification Modeling and Characteristics of Miniature Rotorcraft Kluver Academic Publishers Boston MA 2002 R Pretolani G Saggiani B Teodorani Development of a mission simulation environment for Rotary Wing UAV CAPECON Technical Report Report ID 6 4 September 2004 M L Preatoni AGUSTA Varese Italy R Pretolani G M Saggiani B Teodorani DIEM University of Bologna An Integrated Simulation Environment as a Strategy in Rotorcraft UAVs preliminary design Presented at AHS meeting Phoenix Arizona January 2005 Leonard J Systems engineering fundamentals Defense Systems Management College Press Fort Belvoir Virginia 1999 Lee J Nugent J and Taylor B Advanced rotor control concepts Georgia Institute of Technology AE 6370 Team Project American Helicopter Society RFP Grinsell C Thompson B O Brien P Senga M and Schrage D P Gtmars Georgia Institute of Te
5. 55 56 57 58 59 60 61 P S Anderson Development of a UAV ground control station M SC Thesis Link ping University 2002 C Munzinger Development of a Real time flight simulator for an experimental model helicopter Georgia Institute of Technology School of Aerospace Engineering Atlanta Dec 1998 A Boccalatte F De Crescenzio F Flamigni F Persiani 4 training environment for aircraft pilots by means of virtual reality techniques Proceedings of the XII ADM International Conference Napoli June 4th 6th 2003 ftp edcsgs9 cr usgs gov pub data srtm Preatoni Marzio Agusta Rotary Wing UAV Configurations Final Technical Report CAPECON Technical Report Report ID 6 4 May 2004 J F Boer F Fresta H Haverdings M L Preatoni R Pretolani G M Saggiani B Teodorani AGUSTA NLR DIEM University of Bologna Small Mini Rotary Wing UAV Configuration CAPECON Report ID 10 5 November 2005 Cooper G E Harper R P The Use of Pilot Rating in the Evaluation of Aircraft Handling Qualities NASA TN D 5153 April 1969 www ni com Joerg S Dittrich Design and Integration of an Unmanned Aerial Vehicle 3 Navigation system School of Aerospace Engineering Georgia Institute of Technology May 2002 NAV 420 CrossBow User Manual www xbow com SRFOS Ultra sonic range finder Technical Specification www robotitaly com The I2C BUS Specificat
6. 19 Attitude dynamics 2 2 L A iong Ong Te 1 B lat Onp S T SELT SHO ST S ELITSE H 5 1 6 Long rad ie 6 Lat rad He ere Identified Parameters Along rad rad a rad sec Big vad rad Te sec Table 7 Attitude dynamics identified parameters Velocity dynamics H Hy 5 2 56 rad TH vx 8V Erm 5 rad 5 Vy mis Identified Parameters g m s Xyx 1 s Yy 1 s Table 8 Velocity dynamics identified parameters Heave Dynamics Sa Vz 113 Identified Parameters Leon m s rad Z Vz 1 s Table 9 Heave Dynamics identified parameters For Single Input Single Output SISO systems the controller that is most commonly used in industrial process control is the PID controller This controller has the following transfer function in the Laplace domain 67 K G s K K s 5 4 S where Kp K and Kp the proportional integral and derivative gains respectively Equation 5 4 is often rewritten in terms of time constants K O lt TAa lps with Kc Kp 5 5 por This controller is termed a PID controller because Equation 5 4 has a proportional integral and derivative term Although these controllers are simple they are quite robust simple to tune and often provide sufficient control 68 72 PID controllers have well known tuning methods such as the Ziegler Nichols Step and Ultimate Gain Methods Whilst these tuning methods are u
7. Once data are acquired they are saved in a user defined file for post processing In the software illustrated above only one channel of the DAQ card is configured as example but other channels can be easily added depending on the experimental set up Tables 5 6 report the experimental set up configuration used for the purpose of the vibration tests 103 Accelerometer Name Sensitivity mV g G Scale Factor g V 89158 10 23 Eg 97 752 89160 10 91 1 91 659 89161 10 19 98 135 89162 10 40 96 154 Table 5 Accelerometers Characteristics DAQ Card Settings Sample Frequency Hz 51200 N of Samples per Channel 307200 N of Channels Channel Max Min V different according to test Table 6 DAQ Card Settings By denoting with X measured signal G amplifier gain S transducer sensitivity mV g the scale factor in g V can be calculated through the following formulas x e 63 g 1o00 2 z x e 1 1o00 ZE or VJ 4 4 fe The DAQ Card settings were chosen as follows the sample rate depends on the DAQ card and the channels number The chosen acquisition rate is the maximum selectable for the DAQ card using four channels Since the maximum for the DAQ Card is 210000 Samples s with four channels the maximum sample rate per channel is 52500 Samples s we choose a value a bit smaller the number of samples per channel depend on the acquisition time By fixing an acquisition time of 6 s the ne
8. packaged and interfaced placing attention to vibration isolation electromagnetic interference and accessibility see chapter 4 2 Acquisition Software Development if a RUAV has to fly autonomously information about its states 1s needed which must be used by the control and navigation system Therefore following the hardware set up sensor data acquisition software must be developed and tested in flight in order to validate the acquisition software and ensure measurement reliability see chapter 4 3 Software In The Loop SITL parallel to the hardware set up simulation plays an important role in the development of an autonomous helicopter At this aim a series of flight tests must be also done in order to collect experimental data for identifying the helicopter dynamics characteristics and develop a reliable vehicle simulation model After 61 AS NEEDED that before actual autonomous flight test can take place control and navigation algorithm must be design using the identified model of the helicopter dynamics see chapter 5 4 Hardware In The Loop HIL once the previous tasks are completed the onboard autopilot software must be developed After that the onboard hardware and software must be integrated into the simulation loop For that a Hardware In the Loop HIL simulator was developed in the NI LabView environment In this scenario performance and possible errors of the onboard software can be detected during intensive g
9. were its accuracy range and cost which was a very limiting issue driving the sensor choice Radar altimeter would have higher AGL measurement capabilities at a comparable resolution but at a price out of the project budget Sonar sensors however worked very good to test the system operating capabilities at very low 79 altitude which can be easily extended to higher altitude once future system improvements will be undertaken 4 4 2 1 Description The sonar sensor chosen for the UNIBO RUAV was the SRF08 see fig 53 which can deliver helicopter altitude till six meters with a resolution of 2 cm and has a minimum altitude measurement limit of 3 cm It has a very low voltage and current consumption respectively of 5 V and 12 mA Figure 53 Sonar Sensor SRFO8 Usually ultrasonic sensors use transducers to radiate sounds in many different types of patterns from omnidirectional to very narrow beams For a transducer with a circular radiating surface vibrating in phase as is most commonly used in ultrasonic sensor applications the narrowness of the beam pattern is a function of the ratio of the radiating surface diameter to the sound wavelength at the operating frequency The larger the diameter of the transducer as compared to a wavelength of sound the narrower the sound beam Figure 54 Example of Three Dimensional Representation of the Sonar Beam Pattern As can be seen the sonar sensor produces a narrow conical beam a
10. 2 mM Theta deg 150 160 140 120 100 E or eee Bi eto AA BB E ap RCo ar Be A AB aan nical Be ato sie cea age lated BM cay ales es Be etek etn ae E 8 nee nities ered od 4 G z 10 12 14 16 Time sec 20 25 30 35 40 46 Time sec Figure 102 Heading Control flight test 50 SetPoint Simulation Flight Test 55 134 Latitude Vx m s Vy m s Longitude Figure 103 Flight Path Experimental Responses PID ON OFF Setpoint 120 ee nz 20 40 60 80 time s Figure 104 Flight Pattern Test 100 135 7 1 CONCLUSION AND OUTLOOK An RUAV platform was set up using commercial and cost effective technology Both the hardware and the software were integrated placing attention to modularity growth potential versatility and possibility for ease reconfiguration and software implementation HIL simulations and experimental flights were performed in order to test the feasibility to use the selected hardware and the developed software for helicopter control The controller architecture was developed based on a simple nested PID structure Results demonstrated that the RUAV system was able to provide accurate flight data measurements and good helicopter control capabilities The research activities carried out at the University of Bologna opened several and new research directions concerning the following major fields RUAV Platform future developments HIL improvemen
11. 4 I V gt 0 dh 00 41 Scmax Octrim gt 0 sin y gt 0 lt 0 This means the entire extra term itself is always 0 lt extra term lt 1 This has the effect of reducing the needed A0 in a descent flight path which makes sense because acceleration is now aided by gravity and it is not necessary to command the same nose down pitch attitude as in horizontal flight 42 32 Equation 2 4 was implemented in the acceleration block see figure 6 and was used as input for the helicopter controller The reference attitude provided to the controller was then Ore y AQextended approach 2 5 Simulation tests showed the rotorcraft UAV accelerations along body x axis to be quite brisk between 0 2 and 0 3 g at the beginning and reducing to about 0 1 g as the acceleration phase ended see curve slope in figure 10 Phase Ill Forward Speed Speed m s End of Acceleration Phase 0 5 10 15 20 25 30 35 40 45 Time s Figure 10 Forward speed vs Time during the max acceleration phase The acceleration decreases with forward velocity due to the reciprocal term of the velocity in the very beginning of equation 2 4 The reciprocal term eventually drives AQextended approach tO zero with increasing airspeed at least theoretically To aid the pilot the force feedback joystick is backdriven to a forward position such that the pitch attitude follows the Energy Management law but with an extension whic
12. 60 Servo Actuators Control Circuit The needed input are The first input cables of each switch are connected together and plugged into channel 7 of the RC receiver so that the pilot can switch from to manual autonomous flight mode The second input cable of each switch is connected to the radio receiver channel which receive signal from the RC pilot via radio Five switches were used since five commands are necessary to control the helicopter The third cable of each switch is connected to the related CRIO digital output channel to receive computer input when the helicopter flies in autonomous mode The CRIO digital output channels are the Slot4 cRIO 9411 DO 0 for the lateral cyclic the Slot4 cRIO 9411 DO 1 for the longitudinal cyclic the Slot4 cRIO 9411 DO 2 for throttle the Slot4 cRIO 9411 DO 3 for tail the Slot4 cRIO 9411 DO 4 for collective The switch output signals are send directly to the servo actuators apart from the PWM tail commands The tail command is sent to the tail actuator passing across the onboard hely gyro Actually the onboard hely model gyro 1s coupled with a control unit provided by the factory and contains an Heading Lock Angular Velocity Control System HL AVCS 88 which is used either to stabilize the helicopter in heading or to control the helicopter heading From now on the gyro sensor coupled with its control unit will be referred to as gyro system While all PWM signal outputs cont
13. Each was asked to fill out a questionnaire probing their knowledge and opinions about the task The pilot s typically remarked that the joystick cues helped them achieve desired task performance Finally the Pilot s were asked to carefully go through the Cooper Harper decision tree and make a rating The rating procedure 1s shown below ADEQUACY FOR SELECTED TASK OR AIRCRAFT DEMANDS ON THE PILOT IN SELECTED RATING REQUIRED OPERATION CHARACTERISTICS TASK OR REQUIRED OPERATION Excellent Pilot compensation not a factor for Highly desirable desired performance Pilot compensation not a factor for Negligible deficiencies desired performance Fair Some mildly Minimal pilot compensation required for unpleasant deficiencies desired performance Minor but annoying Desired performance requires moderate deficiencies pilot compensation Is it Deficiencies satisfactory without warrant improvement improvement Moderately objectionable Adequate performance requires deficiencies considerable pilot compensation Very objectionable but Adequate performance requires extensive 6 tolerable deficiencies pilot compensation Adequate performance not attainable with Major deficiencies maximum tolerable pilot compensation miei Deficiencies Controllability not in question performance See te attainable with a require Major deficiencies Considerable pilot compensation is 8 tolerable pilot improvement required for control workload Intense p
14. This result can be seen shown in the airspeed time history of Pilot B in the following figure Phase 3 Acceleration Phase a Ce oe AccHerstsr aie Filg Eno active joystick Airspeed mrs a 5 Ji i a z Time s Figure 36 Airspeed Time History Pilot B Phase 3 It can be seen in figure 36 that Pilot B could accelerate more quickly when the joystick was automatically backdriven to its optimal forward position Other results for Pilot B include a time history of airspeed during manual mode in Phase 2 Here the pilot was tasked with taking over control from autonomous mode and turning the UAV towards the object to be identified This task could not be automated because it inherently is dependent on the object cues in the pilot view and no object recognition software is assumed onboard As aid to the pilot the joystick commands a speed hold mode so that moving the joystick forward commands a faster airspeed The active features held the joystick at this position Of course with the active features turned off the pilot had more difficulty hold the required airspeed and the pilot required more time to turn the UAV toward the object as shown below in figure 37 55 Adequate Peiomane Phase 2 Manual Mode Pilot B w active joystick l Airspeed m s Pilot B no active joystick Time s Figure 37 Airspeed time History Pilot B Phase 2 It can be seen in t
15. YAW RATE COLLECTIVE amp THROTTLE CYCLICS Vx Vy VZ s GUIDANCE amp VELOCITY ATTITUDE Helico ter NAVIGATION p CONTROL CONTROL CONTROL ATTITUDES VELOCITIES Vx Vy VZ POSITION Lat Long H Heading Figure 86 Onboard Control System Architecture The vertical control either can take the form of vertical velocity regulation or height regulation in the first case the vertical velocity profile to be maintained is given by the user in the second case the reference vertical velocity 1s calculated by the altitude hold regulator in the guidance and navigation system The heading control is left to the HL AVCS onboard gyro system The HL AVCS gyro input is actually a yaw rate calculated on the basis of the helicopter heading error Again the helicopter heading error can be calculated either by the NGCS or can be user defined The control system architecture described above was implemented on the onboard computer Currently only the velocity control system was experimented in flight while the navigation and guidance system must be still tested Therefore the reference velocities and heading profiles were defined by the operator at the ground control station depending on the flight test to be performed The complete onboard software implementation 1s described in sections 5 2 5 3 112 5 1 HELICOPTER DYNAMICS IDENTIFICATION AND SIMULINK PID DESIGN RESULTS The identified helicopter dynamics transfer function are reported below
16. accelerations of between 8 and 15 times the normal force of gravity whereas the isolated accelerometer only saw around a 0 3 to 0 8 increase in the force of gravity during the same flight test 62 107 4 8 HARDWARE AND SENSORS DAQ FLIGHT TESTS The UNIBO RUAV avionics hardware was successfully tested in flight Flight data were acquired by means of the data acquisition software described in the previous sections In order to conduct a flight tests the vehicle avionics must be first powered using the avionics box power panel Compiled flight code must be uploaded from the ground station onto the flight computer and started remotely The ground control station constituted by a simple laptop computer connects to the air vehicle displaying its status When everything and everybody on the test team are set up for the flight the engine is started allowing continuous flight for approximately 15 minutes limited by the on board fuel capacity then helicopter re fuelling must be made For the purpose of flight data acquisition tests the helicopter was flying in RC mode while onboard data logging was started and stopped by the ground control station operator High rate on board data recording is independent from data communication and data display on the GCS Data are recorded in a file on the volatile CRIO RAM This file can be downloaded to the GCS using the WIFI data link during the flight tests even without stopping the flight code or after
17. all the task modes The following definitions were used Camera View is the view from the onboard camera if it is being slewed around and Pilot View is the view from the onboard camera if it is locked in a forward looking position 52 Phase 1 Phase 2 Phase 3 Phase 4 Autonomous Waypoint Manual Mode Acceleration Mode Hover Hold Mode Flight object found in camera airspeed held 1 m s object centered object centered within view altitude held 5 m within red square red square drawn on t all ti ilot vi t all ti object laterally centered Ae a e a a total time for Phase 2 Performance acceleration in Phase 3 lt 50 sec object found in camera airspeed held 3 m s object visible in object visible within view altitude held 10 m pilot view at all pilot view at all times i object visible in pilot SS total time for Phase 2 view at end of phase Adequate Performance acceleration in Phase 3 lt 60 sec Table 2 Definition of Desired and Adequate Performance The two most important performance parameters were the centering of the object and the total time for Phase 2 and acceleration in Phase 3 The centering of the object is a classic tracking problem and it is know that precise tracking increases pilot workload But object tracking is critical for this task as it is an object identification task and the object must be held visible and recorded on c
18. angle The blade pitch control system is based on a swashplate mechanism The purpose of this mechanism is to vary the blade pitch both in magnitude but also as a function of the blade angular position around the hub Using the collective control input the pilot controls the average blade pitch angle 22 The blade pitch angle as a function of its angular position is controlled by the longitudinal and lateral cyclic controls The blade pitch angle as a function of its angular position around the hub is zero when the blade is above the tail and it is assumed that the blades rotates counter clockwise see figure 3 is described by O Y Oo Acos B sin 1 1 where o is the average blade pitch angle which is set by the collective control input dco A and B4 the coefficients of the cosine and sine terms are the amount of blade pitch the blade undergoes when it is located above the tail x body direction and on the right hand side y direction respectively 20 A and B are functions of the longitudinal and lateral cyclic controls on and dOrat respectively It is possible to rewrite them as a function of linear gearing coefficients which transform the pilot stick input into angular blade root pitch change A1 Buat Stat B long Olong 1 2 Usually on small scale hobby helicopters a stabilizer bar is also present The stabilizer bar does not produce thrust it has no collective blade pitch setting Inste
19. be performed on the I2C bus by using the above defined standard sequences device addresses and register addresses 84 Sonar acquisition software The sonar acquisition software is shown in figure 59 while details of the implemented subVI are reported in the enclosed Figure 59 FPGA Sonar Data Acquisition Loop The software is basically constituted by two main subVI the first one performs an initialization procedure to set the sonar range and gain to appropriate values the second one read information when the sonar is commanded to range from the CRIO For the SRFO8 to start ranging in cm the following instruction must be implemented in sequence Initialization procedures i2c_start 12c_tx 0xE0 12 get ack 12c_tx 0x02 12 _get ack 12c_tx 0xFF 12 get ack i2c_stop send start sequence SRFO8 I2C address W bit get acknowledgment SRFO8 range register address get acknowledgment set range to appropriate level determined experimentally This value must be set to Hex FF if we want the sonar to range till 6 m get acknowledgment send stop sequence 85 i2c_start i2c_tx OxE0 12 get ack i2c_tx 0x01 12 get ack i2c_tx 0x10 12 get ack i2c_stop Sonar data read i2c_start i2c_tx OxE0 12 get ack i2c_tx 0x00 12 get ack i2c_tx 0x51 12 get ack i2c_stop send start sequence SRFO8 I2C address W bit get acknowledgme
20. constant altitude Figure 8 Max Acceleration at Constant Altitude AO in figure is that of equation 2 1 42 During certain mission phases however the helicopter may need to perform a maximum acceleration but not at constant altitude rather in a descent It follows then that the required pitch down attitude will not be as large relative to the flight path because acceleration is also being aided by a component of gravity see figure below Figure 9 Max Acceleration along a descending flight path A is that of extended approach eq 2 4 The Energy Management extension begins with the same basis equation used in Reference 40 the relation between potential energy in the vertical direction and kinetic energy in the flight path direction 31 E mgh 4 mV 2 2 To take acceleration in a descent into account the equation is modified by taking only the horizontal component of the flight path velocity E mgh 4 m V cos y 2 3 If we derive A extended approach from eq 2 3 following the procedure described in 40 41 the results is an extra term multiplied against equation 2 1 a ha AG Opp ror E 7 w ah E T j cel i F Oe SN FI Pogg Fer Farm ELLY extra term 2 4 A extended approach 18 the pitch attitude required to accelerate in a maximal sense along a flight path regardless if the flight path is horizontal or descending Checking the signs in the extra term in equation 2
21. controlled the view of a slewable camera during autonomous flight 40 For the joystick evaluations the control range and maximum forces of the joystick can be measured Based on an optical measurement using a protractor the joystick can be moved 35 deg forward aft and 35 deg left right 42 The maximum forces available from the joystick are 5 N not so large Nevertheless the forces are sufficient to be felt by a GCS potential pilot to backdrive the joystick and to simulate at least partially the behavior of a manned helicopter type sidestick The joystick forces were programmed in the following fashion depicted in the picture below Force N Displacement deg Origin Movable Figure 17 Spring Forces Programmed in Joystick These forces have the effect of holding the joystick at a desired displacement in both the fore aft and left right directions In other words the effect is that the joystick can be backdriven This is a very beneficial effect in that now a backdriven joystick can be used to in turn control the helicopter simulation origin position joystick position time sec Figure 18 Comparison of Commanded and Actual Stick Position The fidelity of the backdriven stick is shown in the above figure Figure 18 shows a comparison between the position of the spring force origin commanded position and the actual position of the joystick assuming of course the pilot does not have his
22. data line virtual channel SDAW Do Not Connect v Ground From CRIO DO To CRIO DI Figure 56 Sonar Acquisition Circuit The core of the sonar acquisition circuit is shown in figure 56 while the final printed circuit is reported in the enclosed CD Particularly a buffer open collector SN7407N was used to protect the devices from short cuts or other problems two pull up resistors 1 8 kQ were used to pull up the SDA and SCL lines This is necessary because both the SCL and SDA lines are open drain drivers What this means is that the chip can drive its output low but it cannot drive it high For the line to be able to go high pull up resistors to the 5v supply must be provided If 82 the resistors are missing the SCL and SDA lines will always be low nearly 0 volts and the I2C bus will not work two pull down resistors 1 kQ were used to pull down the line output from the CRIO DO These were necessary to adjust the low logical level of the CRIO DO to be compatible with the ones of the SN7407N in the final circuit other components were added to stabilize the voltage line and signals so that the sonar acquisition card can be powered with a maximum input voltage of 9 Volts The card provides then 5 Volts terminals to distribute the correct power to the devices FPGA Sonar data acquisition The sonar data acquisition software required deep knowledge of the I2C protocol in order to be developed Th
23. digital input channels are read and a microsecond counter each for one channel is updated if the input signal logical level is high otherwise the microseconds counter is re initialized to zero This way the PWM pulse with is measured with a resolution equal to one microsecond Moreover two further remarks must be taken into account the counter variable actually incremented every microsecond is a value placed in a virtual memory the LabView shift register The corresponding PWM pulse width value is updated only when the first low bit is read This way the PWM commands 94 are all updated with minimum time latency and the software doesn t yield to false transient measurements the program outputs are the 5 radio commands in us the measured pulse width for each channel and a boolean value for the PID on off channel This boolean value will be used to enable or disable the autopilot in the control loop TRUE means autopilot ON while FALSE autopilot OFF EB Figure 68 Actuators PWM Acquisition Software PWM generation software The PWM output signals have been generated using the FPGA digital output channels configured as follows Slot4 cRIO 9474 DO 0 Lateral cyclic Slot4 cRIO 9474 DO 1 Longitudinal cyclic Slot4 cRIO 9474 D2 Throttle 95 Slot4 cRIO 9474 D3 Tail Slot4 cRIO 9474 D4 Collective The software is reported in figure 69 The software core is the PWM generation subVI which is able to generate a PWM s
24. easily identified experimentally v T T 5 5 5 asus 1500us _ 2250us Figure 62 PWM pulse width and servo angle rotation The servo Angle PWM curve was determined by means of the experimental equipments illustrated in figure 63 90 Encoder Output To Actuator Encoder 900 pulse round Figure 63 Experimental Set up for Servo Angle PWM curve determination The actuator was coupled with an optical encoder in order to measure the angular displacement corresponding to a PWM servo command An encoder is a device that can converts a rotary displacement into digital or pulse signals The most popular type of encoder is the optical encoder which consists of a rotating disk a light source and a photodetector light sensor The disk which is mounted on a rotating shaft has patterns of opaque and transparent sectors coded into the disk refer to figure 64 As the disk rotates these patterns interrupt the light emitted onto the photodetector generating a digital or pulse signal output If the actuator arm is connected to the encoder shaft the encoder disk rotates each time the actuator is commanded to rotate Therefore the encoder signal output will be broken when an opaque disk line is between the emitter detector pair It is the monitoring of this on off pattern which allows the actuator angular displacement to be measured Light Sensan s Light Soini ars Actaling Disk 4 Lr m T i
25. first place we would like to thank Professor Ing Franco Persiani who for first supported and believed in the rotary UAV project and Prof Gianmarco Saggiani for the rotary team coordination But we cannot forget also the support of Ing Veronica Rossi support in onboard software implementation Ing Filippo Zanetti rotary team new entry and support in autonomous flight tests Mauro Ricci and hangar technicians Luciano and Ivano fun club amp support in helicopter mechanics Stefano Lucchi fantastic RC Helicopter pilot Antonio Francia enthusiastic support in helicopter engine set up Ing Tiziano Bombardi support in C program language and his precious UDP dll Prof Alessandro Rivola support in vibration tests and analysis Prof Gian Battista Garito for his interesting lessons on helicopter dynamics theory Ing Matteo Zanzi support in flight control and navigation theory Fodias Guys experience exchange talks and test rig Don Diacunu Pietro helicopter flight test football field Ing Stefano Saputo and Ing Fabio Antonini advices in onboard electronics set up Ing Stefano Mazzoni support in NI Hardware choice Ing Alessandro Boccalatte visual system design The CAPECON Rotary Team People Stephen Mouritsen Jan Floris Boer Marzio Luigi Preatoni Cyrille Sevin TABLE OF CONTENTS LISTOF FIGURES pis eesti ce ee ee J HST TOFTABLES ecceri e R R E A R 11 ABSDRAC Docenia E REEE E 12 NOMENCLATURE ovaria a E R E R
26. ground since flight test were performed in open field with no presence of obstacle and building in the vicinity this aspects won t be a real problem for helicopter operation However the SRFO8 has the possibilities to choose altitude information among 16 different echoes starting from the nearest to the farthest which could be used for future work improvements Figure 55 SRF08 beam pattern 54 8 1 4 4 2 2 Sonar sensors data acquisition Sonar sensor output is provided using an I2C protocol In order to acquire altitude with the onboard computer the sonar sensor was interfaced with the CRIO using an in house made interface card After that the appropriate acquisition software was developed Sonar Card Design The I2C protocol uses two lines just two wires to synchronize all data transfer over the I2C bus called SDA and SCL line The first is the data line and the second is the clock line The SCL and SDA lines must be connected to the CRIO digital output and input in order to write and read commands on the 2C bus A third wire is used for the ground and a5 Volt wire for distributing power to the devices see figure 56 Particularly the Slot2 cRIO 9411 DI 1 is used for reading the clock line virtual channel SDKR the Slot2 cRIO 9411 DI 2 is used for reading the data line virtual channel SDAR the Slot3 cRIO 9411 DO 5 is used for writing the clock line virtual channel SDK W the Slot3 cRIO 9411 DO 5 is used for writing the
27. hand on the joystick It can be seen that the stick follows the commanded position with a small delay and chatter The chatter is the result of some dead band existing at the spring force origin 41 It might be possible in a later study to reshape the force curve to minimize the chatter Evaluation trials were conducted to test if the commanded joystick could in turn control a helicopter simulation in real time Despite the delay and chatter the stick had no problem keeping up with the simulation during stabilized flight and aggressive maneuvers During these trials it appeared the simulation was controlled by a ghost pilot A stick shaker can be also programmed with variable frequency and amplitude This feature can be useful in warning the pilot of impending limits or danger but was not utilized in this study A popular feedback cue for manned helicopter sidesticks 1s softstops depicted in the following figure Force N height Displacement deg Figure 19 Description of Typical Softstop A softstop provides a temporary resistance to stick movement at a given location 42 The resistance is controlled through the height and width variables shown in the figure It was attempted to simulate such a softstop with the active joystick The result was found problematic in that joystick dead band make the softstop feel jerky A simple fix for this behavior was not found and therefore softstops were not use
28. is able to discriminate the sense of rotation so that the helicopter will rotate always in the shorter direction to reach the set point The yaw rate calculated from the dead zone block in figure 89 is intended to be a variation with respect to the condition of zero yaw rate Therefore this value must be added to the initialization trim value The obtained command is the PWM high time in microseconds which is used to generate the PWM signal for the gyro AVCS see PWM generation algorithm section 4 5 2 Unity of measure must be scaled as shown in figure 90 taking into account the scale factor defined in table 12 Tables 24 25 summarize the input and output parameters needed for the algorithm to work properly psi dead zone Input Parameters psi SP al16 User defined profile or from Navigation System Current psi all6 from AHRS data Dead zone high limit al16 546 Table 164 psi dead zone Input Parameters psi dead zone Outputs Yaw rate variation with respect to trim us N B Trim value to be added after ouput 1515 us Table 175 psi dead zone Outputs 5 3 COMPLETE ONBOARD SOFTWARE IMPLEMENTATION The complete RUAV onboard software architecture follows the typical CRIO advised programming technique explained in chapter 4 and 1s illustrated in figure below The source code of the full RUAV software is completely reported in the enclosed CD 124 RUAV Complete SW Implementation Re Tx _ af acquisition NAV
29. is about 11 5 If the signal is filtered under 40 Hz this value is reduce to 2 05 Using an high pass filter over 2kHz the rms value is also reduced to 2 22 106 Acceleration PSD AHRS ol ie vei iy Vibration Isolation Acceleration on AHRS id aby T ilk hd E TT Pa PE Maa RA E ARR Acceleration g ik il T Hl wall i iti L I i 02 0 4 0 6 0 8 1 2 14 1 6 L8 2 22 24 26 2 8 4 32 3 4 Time s Se ab ab wo so oo mo ob Frequency Hz Figure 79 Acceleration Experienced on the isolated NAV 420 Acceleration are readily attenuated and the gms value calculated with equation 4 5 is reduced to 0 7281 Figure 80 shows also the acceleration experienced at point B see fig 77 after the first set of shock mounts acceleration are attenuated but using also the neoprene strips much better results were achieved fig 79 After first set of Shock Mounts After first set of Shock Mounts Acceleration g PSD dB 0 0 1 0 2 0 3 0 5 10 15 20 Time s Frequency kHz Figure 80 Acceleration experienced after the first shock mounts For sake of comparison Boeing results for Raptor 60 are reported in figure 81 Hard Mount Z Axis slow flight Isolated Mount Z Axis slow flight Jobe ils Li uii MTENNI TE NATIN a Ere ifti PONR T NANN iiini i Gs ad m Time s Time s Figure 81 Boeing Results for Raptor 60 62 The hard mounted accelerometer showed
30. keeping the object in view for identification The active features of the joystick were used to backdrive the joystick to eliminate transients between the various phases and to provide situational 50 awareness to the pilot during the phases The pilots reported a marked improvement in mission effectiveness when the active features were turned on With the active features disabled some of the task requirements could not be met Active features on the joystick helped to reduce the workload and total time while at the same time helped to increase situational awareness absolutely needed in any ground control station environment To be consistent with the CAPECON program the flight model chosen for the search identification task was the UAV rotorcraft designed by Agusta 48 The design is a 4 bladed standard helicopter with a single main rotor and a tail rotor The design is based on studies made to fulfil a number of rotorcraft UAV civil applications 9 One of which was a search mission requiring 4 5 hours of endurance and a range in excess of 25 km This range requirement is what makes the use of a GCS compulsory At 25 km the UAV is not in direct view and therefore a means must be found to control it remotely During the first sizing of this UAV for CAPECON the maximum take off weight was 260 kg This assumed only a 40 kg payload which reflected the requirement of only minimal components onboard Therefore the search identificat
31. pilot for safety reasons As shown in table 11 the final proportional PI gains find by simulation results were almost correct while the integral gains were increased of an order of magnitude This may be due to effects unmodelled by the transfer funtions By using an higher integral gain these effects can be attenuated 132 Once the attitude controllers were somehow calibrated the nested PI Velocity attitude controllers were tested During these tests collective and tail commands were still left to the RC pilot for safety reasons As shown in table 10 11 the final gains were much closer to the one found by simulations The third step was to test the heading control together with the nested PI velocity controller During these flight tests only collective was left to the RC pilot for safety reasons The value to be calibrated during these flights was the yaw rate output in microseconds to be sent to the gyro HL AVCS system Starting with a very small value equal to Sus the same value was increased till finding an adequate yaw rate for the helicopter The calibrated final value was 20uUs corresponding to about 10 s This value was kept intentionally low for safety reason but can be increased or varied if necessary In the fourth step the full PI controller was tested including the vertical velocity control During these tests no commands was left to the pilot and the helicopter was flying completely autonomously As shown i
32. provides information to the user by means of a RS 232 protocol The RS232 data acquisition was not performed using the real time serial port since acquisition from serial port is well know to be not deterministic and is therefore incompatible with real time critical control loops For that reason the RS232 data packet was acquired by means of a FPGA digital input channel particularly the Slot2 cRIO 9411 DI 0 to guarantee deterministic data acquisition needed for the control algorithms to work properly The full software developed for the NAV 420 is reported in the enclosed CD together with a step by step explanation The NAV420 string is acquired by reading directly the RS232 electrical signal coming to the CRIO digital input channel In order to understand the program a detail knowledge of the RS232 protocol and of the NAV string contents 1s needed as well as of the LabView software packet reconstruction methodology The program works following the flow chart reported below Initialization Procedure Check NAY 420 transmission rate Bytes Read Information Reconstruction N Checksum Contro Y Y lt gt Figure 50 NAV420CA Acquisition Software Flow Chart 1 The program starts with an initialization procedure This is constituted by a while loop that cicles till a low level signal time interval is found whose length is comparable with the time distance between two consecutive NAV 420 packets This ensure tha
33. repect to trim s116 2 fail Total Lateral cyclic action with respect to trim sI16 N B Lat Trim value to be added after PI ouput scale 1564 us Table 120 phi PI Outputs Vz Pl A schematic of the stand alone Vz PI software is reported in figure 90 Input Parameters Scale vz SP vI16 PWM coll coll aS Coll Throttle s curve _ y PWM Throttle Vz v16 from AHRS data Figure 90 Schematic of the stand alone Vz PI oe amp gt The PI implement the vertical velocity control along the trajectory This PI calculates the collective servo rotation angle in order to maintain the desired set point The trim condition servo rotation corresponding to hover condition must not be added since it was taken into account in the PI integrator initialization This commands are then scaled in microseconds of PWM high time which is used to generate the PWM signal for the servo actuator see PWM generation algorithm section 121 4 5 2 The collective PWM high time is used also to find the corresponding PWM throttle high time in order to send commands to the throttle servo actuator At this aim a calibration curve was derived from the one defined inside the radio settings which was implemented on the FPGA by means of a look up table The PWM collective throttle curve is defined so that the rotor maintain constant rpm The values used for the collective throttle look up table are reported in table 23 Unity of me
34. see figure 49 The GPS antenna is mounted on the tail boom and is connected to the GPS receiver inside the navigation platform with a SMA jack The GPS antenna was changed with respect to the one provided by CrossBow a Geohelix S GPS Antenna was installed in order to improve GPS signal reception The Geohelix characteristics can be found in the enclosed manual When mounting the NAV 420 some precautions must be taken in order to ensure proper functioning and measurement reliability The AHRS unit has its own coordinate system as shown in figure below 72 Hely Alignment NAV 420 Alignment with respect to Helv Figure 49 NAV420CA Mounting The axes form an orthogonal right handed coordinate system X axis from face with connector through the NAV 420 Y axis along the face with connector from left to right Z axis along the face with the connector from top to bottom In this reference system the direction of positive rotation for the rate is defined by the right hand rule Pitch is defined positive for helicopter nose up Roll is defined positive when the helicopter rolls to the right Yaw is defined positive for heading right turn The position output form GPS is represented in Latitude Longitude and altitude convention while the GPS velocity output is defined in the North East and Down reference frame 53 The NAV 420 mounting holes can be used as a reference for aligning the NAV420 s
35. the flight During the flights data were transferred from the air vehicle back to the Ground Control Station GCS via wireless data link and monitored by the GCS operator All onboard electronics worked properly while sensor data was recorded at 100 Hz AHRS raw data figure 82 show vibration disturbances 0 5 p radsec q asee 5 10 15 20 5 a 40 45 450 Time sec Time sec Figure 82 Example of pitch and roll rate AHRS raw data 108 Theta rad psi rad N te G Latitude Forward Speed m sec 12 3898 12 3899 12 3899 12 39 12 39 12 3901 30 Longitude Time sec Figure 83 AHRS filtered flight data However thanks to the XBow NAV420 integrated Kalman filter smooth and stable GPS position information velocity and attitude measurements were available which can be used for control and navigation system implementation Figure 83 shows examples of sensor data measurements taken while the helicopter was overflying the test field at low speed conditions Ultrasonic sensors were also tested Recorded flight tests showed good experimental results although they could provide reliable altitude measurements only up to 5 5 6 m see fig 84 Sonar measurement limit Altitude m 0 10 20 30 40 50 60 70 80 90 100 Time sec Figure 84 Sonar sensors measurements 109 4 8 1 FLIGHT DATA RECORD VIRTUAL RE VIEW Based on the work done in the CAPECON project for the ground co
36. was performed in cooperation with DLR German Aerospace Centre 2 3 1 CONFIGURATION EVALUATION The first application of the developed mission simulation environment was the operational capabilities evaluation of the RUAV designed by AGUSTA inside the CAPECON program 48 49 The RUAV platform was tested over a wide spectrum of different mission scenarios defined in CAPECON The results for a standard Fire Surveillance mission figure 24 are reported in this section as an example of mission simulation and performance estimation The intent was to evaluate the AV and NGCS performance using a realistic mission profile Simulation results showed that the NGCS first step design 1s able to stabilize control and guide the AGUSTA configuration After this tests the GCS layout was also improved to the current layout described in section 2 2 3 taking also into account ground control station pilot suggestions The mission scenario and the related AV performances are described below 46 ne gt y A R raco lif saand Figure 24 Mission Scenario The air vehicle was supposed to take off in manual mode and then follow in autonomous mode the flight path described in Figure 25 and in Table 1 FLIGHT PLAN TAKE OFF Manual Mode Latitude North Longitude Est m m s sar ma Mim T o erst ore LAND Manual Mode Table 1 Flight Plan Data 47 Figure 25 Fire Surveillance Mission flight path The operator at the groun
37. with an onboard safety system in event of computer failure Depending on the size and cost of the air vehicle this can include a completely redundant avionics system or simply a minimum safety system cost 1s of course a limiting factor for avionics and airframe selection and for achievable performance 4 2 FLIGHT TEST VEHICLE DESCRIPTION The air vehicle chosen as RUAV platform is shown in figure 41 It is a Hirobo Eagle II 60 hobby helicopter which was modified to accommodate the avionics hardware A more powerful engine longer fiberglass blades longer tail boom and tail blades were mounted in order to increase the helicopter payload carrying capabilities The assembly also includes a Bell Hiller stabilizer bar which augments servo torque with aerodynamic moment to change the blades cyclic pitch and adds lagged rate feedback to improve the helicopter handling qualities 64 Figure 41 RUAV Air Vehicle The main helicopter characteristics are main rotor diameter 1840 mm tail rotor diameter 330 mm total helicopter mass 11 2 kg engine OS 91 Engine 15 cc power 2 9 CV main rotor rpm 1200 1300 tail rotor rpm 5000 6000 payload carrying capabilities 5 6 kg 4 3 FLIGHT COMPUTER The CRIO system from NI was selected as flight computer due to its ability to fulfill many among the stated design requirements Particularly the most important CRIO features that encouraged its usage as onboard computer for the UNI
38. with the maximum accelerations achieved through the use of the Energy Management equations This means that when the active joystick features were turned on the desired performance for total time was easily achieved When the active features were turned off the pilot had a more difficult time optimizing his flight path and the total time increased The following figure 35 illustrates this result Total time Phase I Acceleration to 33 ms Fr Forme Feedback Ate Panira Pilot 4 Plot B Figure 35 Total Time Comparisons for Pilots A B and C Specifically total time equals the time from the moment the pilot switches on manual mode onset Phase 2 to the moment he reaches 55 m s in Phase 3 It can be seen in figure 35 that there was a consistent improvement in the total time for all three pilots when the force feedback features are turned on To understand this better during the acceleration phase the joystick is automatically backdriven to the forward position corresponding to Energy Management equations The Energy Management equations compute the required AO for maximum acceleration The control system in attitude control mode backdrives the joystick input until this attitude is matched This relieves the pilot of the task of trying to optimize the acceleration while keeping the object 54 centered in the pilot view But with no active joystick cues the pilot needs more time to accelerate to the desired airspeed
39. yaw rate is smoothed to zero If the RC pilot or the CRIO send to the gyro control unit a PWM signal different from the one read during the initialization process this is perceived by the control unit as a new reference yaw rate to be maintained the control unit will send PWM signal to the tail servo in order to maintain the commanded yaw rate 89 The commanded PWM sent to the tail servo is generated by the gyro control unit based on a gain settable by the user 4 6 1 PULSE WIDTH MODULATION SERVO ANGLE CURVE Pulse Width Modulation PWM is a technique commonly used to represent an analogue signal using digital circuitry It involves the switching on and off of a digital output at a fixed frequency switching frequency fs but with varying times of either on or off The ratio of on time to the total period T 1 fs is called the duty cycle d d T T 4 1 where T denotes the PWM on time RC equipments such as servos typically use PWM signals for their control input As opposed to standard PWM signals where the signal value is dependant upon the duty cycle RC equipment use the actual pulse width in seconds to represent the signal Furthermore the RC PWM signals usually has a standard frequency range between 20Hz and 200Hz The servo actuators used on the helicopter operate at a PWM frequency of 50 Hz Depending on the PWM actual pulse width the servo actuators rotates of a certain angle see figure 62 which can be
40. 04 Table 7 Attitude dynamics identified parameters cccccccccecccecceceeceeceeeeeaeeeeeeesseeseeeeeeeeees 113 Table 8 Velocity dynamics identified parameters cccccccccccceceeeceeceeceeeeeaeeeaeeeaseseeeeeeseeees 113 Table 9 Heave Dynamics identified parametefrs cccccccccccccececeeceeceeceeseeaseeeseeaseeseeeeseeeens 114 Table 10 Attitude Controllers PI Gains scott cece ate catesdecaeie diate dieteaseamnaeecaienlesteetdetss desascneed 115 Table 117 Velocity Controllers Pr GANS sermoe a O 115 Table 12 Unity of Measures and scale factors used in the control code 0eeeeeeeeees 118 Table e VA Pl ImputiParamicters aiae E N 119 Table 4 NAPLO einar a a a O 119 Table t5 theta Pr Input ParametetS sopiri a AS 119 Table Ao tie tald ly Qu pts e e A lasnacassauaaaaae aaa tesa eens 119 Tablet Vy Pl input at amie a E A es ecuaeeaceateaues 120 fe 0 ea ic aie od OF 510 E ee ere et ten Se eRe ee ee eee rece E ee eee eee 121 Table 19 phi PL Input Paramete Smii A A 121 Male 20 PM PT TOUDI oaa E A E uabegasacaw 121 Table 2 le Vz PLINput Paramete Siia E E E 122 Tale DN PEO E E E A tan ataeeey 122 Table 23 Collective Throttle Curve Look up table ccccccccccccccccceceeceeceeeeeseeeeeeaeeeeeesseeaees 123 Table 24 psi dead zone Input Parameter Sience ii e a N a R 124 Table 252 psi deaG Z0ne Outputs uriini aAa ee I NTE A 124 11 ABSTRACT This PhD thesis presents the results achieved at the Aeros
41. 370 bit 17 36 us bit 6 4 ms Low level signal time Wait time 10 ms 6 4 ms 3 6 ms 76 _ Ee _ _Pmsipwilevay JLT OTL 10 ms Figure 51 NAV420CA data Packet length time By default the wait time is set equal to 3 2 ms to be sure that the program is ready to read the first signal rising edge corresponding to the start bit of the packet first byte 2 During the second step the actual NAV 420 packet rate 1s read by calculating the inverse of the time difference between two consecutive received packets actually this procedures requires two packets to be read so at the program first call this value is not used 3 4 In the third and fourth steps information are read each bit at one time which are used first to build 1 byte and then respectively a 2 byte or a 4 byte data type This procedure is performed taking into account that NAV420 data transmission is a standard RS 232 protocol with 8 data bits 1 start bit l stop bit no parity and no flow control The 8 bit data transmission starts from the least to the most significant bit and uses inverted logical levels high signal level corresponds to 0 low signal level corresponds to 1 On the contrary the RS232 protocol byte transmission is done from the most significant byte to the least significant byte So for example to transmit a 2 byte information the RS232 protocol first sends the most significant byte and then the least significant In turn each b
42. 420 serial i en signal acquisition gt VNED to VTrack_ Sonar 12C Acq GPS signal che k FPGA software TETE Dre gt TCP IP RT FPGA reat j sts tidal communication 4 i wr loop s Data loggin Interface software Figure 92 RUAV Complete Software Implementation The complete RUAV software is divided in three main parts The FPGA software is constituted by six independent loops in order to increase determinism It includes nested PI control loop as described in the previous sections NAV 420 data acquisition loop as described in section 4 4 1 radio PWM signal acquisition loop as described in section 4 5 sonar sensor data acquisition loop as described in section 4 4 2 PWM signal generation loop as described in section 4 5 a V track calculation loop which transform the NED velocity coming from the NAV 420 into velocity along the trajectory which will be used by the controller 74 The same loop perform also a GPS signal check If the GPS has poor signal the NAV 420 velocity data are not reliable any more therefore the velocity control is automatically disabled and the controller become a merely attitude control The RT Real Time software is constituted by two independent loops the time critical loop timed at 10 ms which perform high rate read write communication with the FPGA software The time critical loop acquires sensor 125 data from the FPGA which are e
43. A environment allows programming only by using integer values and the sensor data output are all 116 values with their own scaled data Therefore controller output and input values must be adjusted with some scale factors in order to provide the correct servo commands value in PWM microseconds high time 5 2 1 DISCRETE PID IMPLEMENTATION Consider the ideal PID controller written in the continuous time domain form 67 u t K e t a fea SKI e Die 5 6 1 0 where e t is the process variable error defined as e t SP PV SP being the Set Point and PV the Process Variable 5 7 115 and u t is the PID output To discretise the controller we need to approximate the integral and the derivative terms to forms suitable for computation by a computer From a purely numerical point of view if Ts is the loop cycle time we can use de t e t e t dt T S fea T S0 5 8 The general discrete PID algorithm can be therefore 73 u t Kelt Yeti Tn e S U 5 9 which is now in the form of a difference equation suitable for coding in an appropriate programming language This particular form of the PID algorithm is known as the positional PID controller because the control signal is calculated with reference to a base level uo which can be known experimentally and must be set up correctly inside the algorithm Actually the PID integral action is calculated by using a trapezoid integrat
44. ALMA MATER STUDIORUM UNIVERSITA DI BOLOGNA I FACOLTA DI INGEGNERIA Dipartimento delle Costruzioni Meccaniche Nucleari Aeronautiche e di Metallurgia DOTTORATO di RICERCA IN DISEGNO E METODI DELL INGEGNERIA INDUSTRIALE CICLO XIX Progetto e realizzazione del sistema di gestione autonoma del volo e controllo in remoto per un velivolo UAV ad ala rotante S S ING IND 03 MECCANICA DELVOLO Coordinatore Chiar mo Prof Ing Franco PERSIANI Relatore Prof GianMarco SAGGIANI Dottorando Ing Barbara TEODORANI Esame Finale Anno 2007 On the Development of a Rotary Wing UAV Platform Avionics and Onboard Software Set Up PhD Thesis by cand Barbara Teodorani S S ING IND 03 FLIGHT MECHANICS Coordinator Chiar mo Prof Ing Franco PERSIANI DIEM University of Bologna Advisor Prof GianMarco SAGGIANI DIEM University of Bologna ALMA MATER STUDIORUM Department of Mechanical Nuclear and Aerospace Engineering II Faculty of Engineering DIEM University of Bologna 2007 ACKNOWLEDGMENTS Expert people in UAV development know well that it is above all a science of integration of different disciplines skills and know how No successful results could be achieved without the precious contribution of numerous people working together That s why we me and Roberto Pretolani are very grateful to all people working with us during the years we spent at the Hangar Laboratories of the University of Bologna In the
45. BO RUAV system were modular and versatile architecture easily reconfigurable with minimal time investment ultrahigh performance and low power consumption relatively low cost system ease and open access to low level hardware resources rapid embedded control and acquisition system development that rival the performance and optimization of custom designed circuitry possibility to use LabView graphical programming tool to develop a variety of different applications 65 relatively small size and weight compared to its control and data acquisition capabilities The CRIO platform includes the CRIO 9004 real time controller endowed with an industrial Penthtum 200 MHz floating point processor a four slot reconfigurable chassis featuring three million gate FPGAs chipset and a wide variety of analog digital I O module types Figure 42 shows the CRIO configuration currently mounted on the UNIBO RUAV system Real Time Core E Penthium 200 MHz a FPGA Modules E Serial Port my mi ty E 16 DO Channels E Ethernet 100 Mb s ee D2 LO 12 D1 channels Figure 42 National Instruments CRIO Onboard Computer The real time controller also features a 100 Mb s Ethernet port for network communication with an host computer and a 9 PIN serial port The FPGA module currently used are CRIO 9411 mounted in slot 1 having 6 digital input channels another CRIO 9411 mounted in slot 2 having 6 digital input channels CRIO 9474 mounted in
46. Figure 71 RUAV Schematic Wiring 98 Redundancy The installation of a completely redundant avionics system was of course prevented from the small size of the UNIBO RUAV Therefore only a minimum safety system was installed which was anyway enough for the purpose of this project As discussed in section 4 5 the core of the RUAV minimum safety system is the electronic switch used for disabling the onboard computer in event of system failures In this perspective two separate radio receivers were mounted on the helicopter one inside the avionics box second receiver in figure 71 whose channels are connected to the CRIO digital input channels for data acquisition one mounted on the helicopter airframe outside the avionics box first receiver in figure 71 which 1s fully electrically separated from the other avionics and is used only by the RC pilot when the helicopter flies in manual mode The first receiver power system is also fully independent from the one of the other avionics box equipments Power system amp Electromagnetic Interference Shielding All modules are powered by means of onboard batteries The CRIO and the AHRS requires a 11 12V DC power connection 3200 mAh Lithium Polimer Battery were used which combines very light weight with long time power supplies this battery package allows almost two hours autonomy at a price of 150 gr For the same reason a 7 4 V Littum Polimer battery package was used to power the
47. Force Feedback WayPoint US Geological Survey Shuttle Radar Topography Mission Digital Elevation Map Pulse Width Modulation Compact Reconfigurable Input Output Attitude Heading and Reference System Software In The Loop Hardware In the Loop Radio Frequency Radio Control Electro Magnetic Interference Electro Magnetic Field Programmable Gate Array Inertial Measurement Unit Global Positioning System Above Mean Sea Level Above Ground Level Sensor Data line Sensor CLock line Most Significant Bit Least Significant Bit Heading Lock Angular Velocity Control System Direct Current Single Input Single Output Set Point Process Variable Digital Input Digital Output Graphical User Interface 14 15 Chapter 1 MOTIVATION AND BACKGROUND 1 1 OVERVIEW The increasing interest in military Unmanned Air Vehicles UAVs 1s fuelling an equally ambitious build up in the civil community It is well known that UAVs may represent a promising and cost effective alternative to manned aircraft for a large number of civil applications 1 Compared to traditional air vehicles UAVs may offer significant advantages in terms of human safety especially in dull dirty and dangerous missions operational cost reduction and work rate efficiency Nevertheless while research activities in UAV or Rotary Wing UAV systems are very advanced in the United States UAV interest in Europe has begun only in the last years As a result in yea
48. ad the main blade pitch receives both the cyclic pitch servo command and a major component imposed by the stabilizer bar Hence Bia and Along are the effective cyclic control derivatives taking into account the effect of the stabilizer bar The primary function of the four principal rotorcraft commands are the following the main rotor lateral and longitudinal cyclic inputs control the roll and pitch moments produced by the main rotor collective input controls the magnitude of the main rotor thrust the tail rotor collective input controls the tail rotor thrust by varying the tail blade pitch Hence the commands have a direct effect on the rotorcraft roll and pitch attitude rate vertical velocity 23 and heading rate respectively The pilot does not control the helicopter position or velocity directly but via a chain of effects that can be summarized as follows The cyclic control inputs result in control moments about the rotor hub via a tilting motion of the rotor disc rotor disc refer to a simplified representation of the combined effect of individual blade motion The rotor control moments produce a fuselage rolling or pitching motion If the helicopter is hovering changing the fuselage s roll and pitch angle will result in a tilting of the rotor thrust vector producing horizontal thrust components that acts as propulsive force For example by holding a constant pitch angle the helicopter will accelerate until the propulsive
49. additional redundant systems on an ultralight helicopter will be also investigated 136 References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 U S Air Force Scientific Advisory Board Unmanned Aerial Vehicles in perspective Effects Capabilities and Technologies SAB TR 03 01 July 2003 CAPECON Consortium Annex 1 Description of Work CAPECON Project No GRD1 2001 40162 Starting Date May 2002 Mouritsen S Boer J F Operational concept for local surveillance UAV missions DLR and NLR CAPECON report January 2003 Sevin C Internal Report on RW UAVs requirements and equipment CAPECON Report ECD D 2 4 1 23 January 2003 Mouritsen S Rotary wing requirements interim technical report DLR CAPECON report ID3 3 4 February 2003 Gian Marco Saggiani Barbara Teodorani 4 matrix method for defining potential applications of a multirole Rotary UAV RUAV in Italy CAPECON Project Internal Report no 1 January 2003 Saggiant G M Teodorani B Rotary wing UAV potential applications an analytical study through a matrix method University of Bologna on Aircraft Engineering and Aerospace Technology vol 76 No 1 2004 pp 6 14 Basset P M Operational concept for rotary wing UAV out of line of sight mission ONERA CAPECON report February 2003 Sevin C Internal Report on survey of potential applications for rotary wing UAVs ECD CAPECON Report ID 2 1 2 25 Se
50. amera at all times Furthermore the tracking is vital so that the UAV does not drift into possible nearby obstacles In this case tracking is used as a substitute for local navigation This means the same task can be utilized not only to identify a boat on the ocean as it is in this simulation but the task can also be applied in a mountainous region The second important performance parameter is the total time for Phase 2 acceleration in Phase 3 Because our onboard equipment is being held very simple we assume no radar altitude or laser range sensors This means that as the pilot is accelerating towards the object he has a difficult time sensing depth from the 2D video monitor In other words the pilot needs a criteria at which to engage hover hold mode In the absence of such a 53 criteria it was decided to define a minimum time to complete Phase 2 and the acceleration to 55 m s in Phase 3 In practice it may become necessary to give the pilot better depth information to prevent him from engaging hover hold mode too late which could result in a collision with the object Nevertheless results from these piloted simulations showed that hover hold could be engaged at a safe distance while allowing a detailed view of the object Varying the desired and adequate performance for the total time of Phase 2 acceleration in Phase 3 had the effect of varying the aggressiveness of the maneuver The final values were selected to be consistent
51. arch identification task 2 3 2 3 Cooper Harper Rating Evaluations The procedures in support of the Cooper Harper ratings involve a number of issues which now will be discussed Step one of course is that the engineers need to design a task which is repeatable by different pilots and has performance criteria which can be measured This was done and the results are documented in the previous sections 56 Next pilots need to be chosen who fulfill minimum qualifications Not just anyone should be invited to evaluate the task The minimum requirements for the pilots were 1 they have an aviation background 2 they have time in either manned aircraft or UAVs and 3 they have a serious attitude when evaluating the task When the task the Ground Control Station the active joystick and the pilots are ready then each pilot takes time to train on the task until he is familiar with it Each is allowed to fly the task several times until an official run is made Official runs were made for both cases with the active features of the joystick turned on and with the features turned off The official run time histories are recorded on computer and then the pilot fills out immediately a questionnaire and makes a rating He 1s obliged also not to tell the other pilots his opinion or rating until all have finished their evaluations This insures that the ratings are objective These procedures were followed with three pilots A B and C
52. ard Christian S Jensen Stefan L Jakobsen Martin Siegumfeldt Autonomous Helicopter Modelling and Control Aalborg University May 2005 T D Talbot B E Tingling W A Decker and R T Chen 4 mathematical model of a single main rotor helicopter for piloted simulation Technical Memorandum 84281 NASA 1982 A R S Bramwell Bramwell s Helicopter Dynamics AJAA Reston VA 2001 G D Padfield Helicopter Flight Dynamics The Theory and Application of Flying Qualities and Simulation Modeling AIAA Education Series Reston VA 1996 V Gravilets B Mettler E Feron Dynamic model for a miniature acrobatic helicopter MIT LIDS report LIDS P 2580 2003 Einthoven P Morse C Energy Management presented at the AHS Flight Controls and Crew Systems Design Specialists Meeting Philadelphia PA October 9 11 2002 Prouty R W Helicopter Performance Stability and Control Krieger Publishing Company Inc 1990 S Mouritsen DLR Braunschweig Germany R Pretolani G M Saggiani B Teodoran DIEM University of Bologna Application of an Active Joystick in a Rotorcraft UAV Ground Control Station Presented at AHS meeting Phoenix Arizona January 2005 M Niculescu Lateral Track Control Law for Aerosonde UAV AIAA 2001 0016 University of Washington Seattle January 2001 139 44 45 46 47 48 49 50 51 52 53 54
53. as coated with 2 layers of aluminium foil in order to prevent any EM interference from the pilot radio transmitter or other external disturbances All aluminum parts were also electrically connected and common grounded The two radio receiver antennas were left hanging under the helicopter which was found to be their best position after several flight tests 4 7 1 VIBRATION ISOLATION Electronic circuits and sensors can be affected by harmful vibrations from the engine and rotors Particularly the AHRS GPS antenna the onboard computer and the sonar altimeter are likely to produce faulty readings with inadequate vibration isolation or may be subjected to damage if their operational vibration range is overcome Therefore the avionics box was appended under the landing gear by means of four elastomeric silent blocks at its corners which can be seen in figure 72 The dampers were mounted symmetrically with respect to the avionics box centre of gravity in order to optimize the load distribution on the isolators Moreover the AHRS the CRIO and the other electronics were isolated inside the avionics box by means of neoprene strips As the GPS antenna and the helicopter gyro are mounted on the airframe structure they needed separate protection from harmful vibrations They were isolated by using short pieces of special hely model rubber that effectively attenuated vibrations The sonar sensor was appended at the bottom of the avionics box by
54. asure must be scaled taking into account the scale factor defined in table 13 Tables 21 22 summarize the input and output parameters needed for the algorithm to work properly Vz PI Input Parameters Vz SP vI16 User defined profile or from Navigation System Current Vz vI16 AHRS KcVz 2 coll m s 2560 Output High Vz O 25593 Output Low Vz O 16221 Initial Collective cI32 Coll init us 1790 coll init not less than 1420 us PI Reset TRUE first PI call otherwise FALSE performed automatically by the program Table 131 Vz PI Input Parameters Vz PI Outputs Proportional action collective cI32 Total Collective action Vz O Table 142 Vz PI Outputs 12 Collective us Throttle us Collective us Throttle us 1218 1893 1578 1414 1238 1857 1598 1404 1418 1549 1778 1269 1438 1529 1798 1249 Table 153 Collective Throttle Curve Look up table Heading A schematic of the heading control software is reported in figure 91 Input Trim r_tail us Parameters PWM tail yw SP al16 7 PWM tail 7 r us O Gyro Dead gyro AVCS a zone os w al16 Figure 91 schematic of the heading control Heading control is achieved using the onboard GYRO HL AVCS system Therefore an algorithm is implemented which gives only a reference yaw rate to the gyro HL AVCS based on the heading error calculated with respect to the reference heading set point The 123 algorithm
55. ce Kalman filter algorithm implemented on an internal digital signal processing module Velocity data includes aiding from the inertial instruments such reducing the latency associated with stand alone GPS measurements Particularly the NAV 420 uses the latest in solid states sensor technology and consists of the following subsystems see figure 45 1 Inertial Sensor Array This is an assembly of three accelerometers three gyros rate sensors and four temperature sensors 2 A three axis fluxgate magnetometer board used to compute heading 3 A WAAS capable GPS receiver for position and velocity measurement 4 A digital signal processing DSP module which receives the signals from the inertial sensors and magnetometers This unit converts the signals to digital data filters the data computes the attitude solution monitors and processes all BIT data and transmits the results to the user The NAV420 analog sensor signals are sampled and converted to digital data at 1 kHz The sensor data 1s filtered and down sampled by a DSP Digital Outputs X Y I Z Acceleration Roll Pitch Yaw Rate X Y Z Magnetic Field Roll Pitch Yaw Angle Position Velocity Built In Test High Speed Sampling amp DSP X Y Z Sensor Accelerometers Compensation MEMS 7 Full State X Y Z Kalman Filter Magnetometers Flux Gate GPS Antenna Temperature GPS Receiver Power Power Input Conditioning 9 TO 30 VDC a Fi
56. cedecsasessseededestioeaiiedesseetvacateateineerseateadee 50 Figure 34 Phase 4 Hover Hold Mode 42 oo eeeceee ec eeeeeccesssscsceceeeeeeeeeeeeeeeesesessseeesessessenaaaas 52 Figure 35 Total Time Comparisons for Pilots A B and C cceccecescssccccceceeeeeeeeeseeeeseeeeeeeees 54 Figure 36 Airspeed Time History Pilot B Phase 3 ccccceesesssssssecceeeeeeeeeeeeeeeeeeeeseeseeeeens 55 Figure 37 Airspeed time History Pilot B Phase 2 ccccesesssssssccceceeeeeeeeeeeeeeseeseeseeeeeees 56 Fig re 38 Cooper Harper Decisi n TreCresmicciiscmosnissnrne nee e a E EE Pa a 57 Figure 39 RUAV System set up and Architecture ccccccssccsscccecceeeeeeeeceeceessessseessessesssaeaes 60 Figure 40 RUAV Avionics Design FloW 00000ooonooosoooeeseereeeesessssssssssssssososorereereeeesessesssssssss 6l Figure AIE RUA Ait Vc WI Geneen ion e EE E E AEE E N 65 Figure 42 National Instruments CRIO Onboard Computer c cccccccccccccceceeceeeeeeeeeeeeeeeees 66 Figure 43 CRIO Field Programmable Gate Array FPGA Structure 51 eee 67 Figure 44 CRIO Programming Structure 51 c essessssssssseessesssessesssessesseeeseesseeseeseeseeees 68 Figure 45 NAV420CA System Architecture ccc s sscssecseesecececcceceeeeeeeseeseeessseecscessnseeses 69 Figure 46 NAY 420 Seip Proce ures ooreis E ates 70 Pigure4 INAV AI0OC AO Test Ri cinirenen E T 71 Figure 48 NAV420 Test Rie Meas mre miei tc ce states ta
57. chnology American Helicopter Society student design competition Cohen L Quality function deployment Addison Wesley Mass USA 1995 RAO A et al Total quality management a cross functional perspective John Wiley amp Sons Inc USA 1996 Knowles G Advanced quality tools Module Notes WMG The University of Warwick Coventry UK 1997 Raymer D P Aircraft Design A Conceptual Approach AAAA EDUCATION SERIES 138 30 31 32 33 34 35 36 37 38 39 40 41 42 43 B Mettler M B Tischler and T Kanade System identification modeling of a small scale unmanned rotorcraft for control design Journal of the American Helicopter Society 47 1 50 63 January 2002 B Mettler M Tischler T Kanade and W Messner Attitude control optimization for a small scale unmanned helicopter Denver CO August 2000 AIAA Guidance Navigation and Control Conference B Mettler V Gavrilets E Feron and T Kanade Dynamic compensation for high bandwidth control of small scale helicopter San Francisco CA January 2002 American Helicopter Society Specialist Meeting M McConley Draper small autonomous aerial vehicle dynamic model Technical Report E41 98 091 Draper Laboratory August 1998 J G Leishman Principles of helicopter aerodynamics Cambridge University Press New York 2000 Ulrik B Hald Mikkel V Hesselbaek Jacob T Holmga
58. d 2 2 3 GCS RUAV USER INTERFACE The GCS user interface is constituted by three video screens shown in figure 20 which allow the GCS operator to control flight data information by means of the virtual cockpit view to plan or re plan the mission flight path using the mission planning window and to have a good situation awareness during all mission phases by means of the 2D map view of the mission vertical profile view and of the 3D view The virtual cockpit and the 2D map view were developed through simple ActiveX controls such as aircraft instrumentation available from Global Majic Software Inc 42 GMS which can be used as stand alone applications The activeX add ons were chosen in order to simplify the software since they can be easily interfaced with LabView Other secondary displays were created using the indicators library of LabView The virtual cockpit contains six main displays arranged according to the basic T layout as shown in figure 20 They include the air speed indicator attitude indicator altimeter turn coordinator heading indicator and vertical speed indicator As well as the six main displays other various flight control displays and warnings have been created such as main rotor and engine rpm indicators For resource planning and monitoring the Instruments Panel also shows the current fuel level and the mission elapsed time Pilot Camera External View Map 2D View Figure 20 GCS Configura
59. d monitored the surveillance area by means of the simulated onboard slewable camera For evaluating the mission operational capabilities of the RUAYV it was supposed to find the fire at a certain point of the mission path In that case the operator at the ground switched the air vehicle control to manual mode for monitoring the situation The actual flight path and the mission vertical profile are shown in figures 26 and 27 Surveillance Mission Ground Track Latitude North Longitude East Figure 26 Fire Surveillance Mission actual flight path 48 Surveillance Mission Vertical Profile Fire Sighting Altitude m A a gt 2 Land 5000 6000 i OO o 0 0 1000 2000 3000 4l Time s Wp0 Wpi Wp2 Wp3 Wp4Wpd Figure 27 Fire Surveillance Mission vertical profile Once the air vehicle was nearby the fire area a manual descent and loiter were performed in order to have a better situation overview and send fire position and video images to the ground control station Particularly the AV performed a loiter in the area of interest for 40 minutes After a detailed survey the autonomous key on the joystick allowed the operator to redirect the AV to the original flight path If nothing is found the RUAV was supposed to cover the pre planned path 3 times and then land at the base to refuel The surveillance path was performed at the best endurance speed while the fir
60. dance system and the four control mode will be given 2 1 1 CONTROL MODES DESCRIPTION AND FORCE FEEDBACK LAWS Four control modes were defined for mission accomplishment autonomous manual acceleration and hover hold They can be selected by means of the force feedback joystick buttons Depending on the selected flight mode the reference parameters for the autopilot are different as well as the force feedback on the joystick In autonomous mode the reference flight parameters are generated by the guidance system see section 2 3 according to the prescribed flight plan planned at the Ground Control Station during which joystick stick inputs are ignored The force feedback module backdrives the joystick so that it follows the current flight condition see figure 7 In manual mode the reference flight parameters coming from the joystick are forward speed lateral velocity yaw rate and rate of climb descent The force feedback module helps the pilot to maintain the commanded reference speed see figure 7 In acceleration mode the acceleration block calculates the reference flight parameters for the helicopter to achieve the maximum acceleration according to the energy law described in section 2 1 2 The force feedback module gives back to the joystick the reference pitch angle to be maintained calculated using the energy management equations The reference flight parameters commanded by the pilot are the yaw rate and the rate of cl
61. data link access point Since the access point requires 5 V power supply this battery package is connected to a voltage regulator The same was done for the 12 V battery package which supplies also 9 V power to the sonar interface card The two radio receivers are powered by two 5V NiMH separate batteries package to improve safety The NiMH battery used for the second receiver located inside the avionics box is utilized also to power the sonar and to set the voltage level of the digital output channels of the CRIO The power panel is mounted on the side of the avionics box and includes power switch on buttons external interface ports for batteries recharging Special attention was placed also to accessibility of the hardware interfaces the panel comprises also a serial user interface port and easy access to the CRIO Ethernet port for easy connection of the avionics to the planned hardware in the loop simulation system or to the ground control station computer if necessary All battery packages are installed so that they can be easy accessible for package replacement with minimum efforts if long flight tests have to be performed The wires 99 inside the avionics box are tied to several mounting points and to each other in order to prevent any shaving of the insulation The CRIO and the AHRS are already factory endowed with sufficient EMI shielding such that it cannot interfere with other equipments Moreover the full avionics box w
62. des all the airborne systems the basic helicopter platform the onboard computer and sensors the mission payload and all the software necessary to guide navigate and control the air vehicle The ground support system includes all the ground infrastructures and equipments to enable the AV operations such as a mobile ground control station GCS a logistic and maintenance segment and a ground vehicle The data link supports video data and telemetry communications between the AV and the ground support systems while the data distribution is able to transmit annotated significant data collected at the GCS to potential users at remote locations The subsystems both hardware and software equipments can be much or less sophisticated depending on the UAV system size and complexity For the purpose of the RUAV program a small scale hobby model helicopter was used as flying platform which was certainly a significant physical constraint for the RUAV subsystem equipment choice and development The work performed to develop the RUAV platform was carried out following a series of subsequent logical steps first the RUAV hardware including the onboard avionics the air vehicle and the data link system was selected and interfaced placing attention to vibration isolation electromagnetic interference and accessibility following the hardware set up sensor data acquisition software was developed and tested in flight in order to verify se
63. dinates system Xtrack Ytrack the navigation strategy is to turn the ground speed vector V into the direction of the exact track so that the helicopter intercept the track line at point C 34 Figure 12 AV lateral track control strategy The intercepting point C can be tuned by a design parameter k The distance along the track line between the intercepting point C and the Wp2 is at any instant equal to 1 k Xtrack From the geometry similitude of the triangles OAB and OCD a control strategy can be defined following relationship 43 KXtrack 7 Ytrack 2 6 To achieve this objective the error E is computed by E KXtrackYirack YtrackXtrack 0 2 7 The error can be driven to zero using the proportional feedback control law that expresses yaw rate commands as romp KRE KR KXirack Yirack Prat nrak 2 8 The proportional gain K is determined iteratively through simulation until good tracking is achieved with virtually no overshoot The proposed lateral control handles also wind conditions in a simple manner ensuring track stability over a wide set of initial conditions see figure 13 35 Figure 13 Simulation of lateral track control 43 In manual flight mode the guidance system is obviously disabled the operator acts on the joystick in order to control directly body axis forward speed vertical velocity side sleep velocity and yaw rate In this mode the joystick commands are directl
64. e mounted on the landing gear on the avionics structure after the dampers on the CRIO and the AHRS to measure the damping effects at different points Figure 74 Experimental Data Acquisition System The accelerometers were connected by means of BNC cables to a charge amplifier The output signal from the amplifier was acquired by means of a data acquisition card installed on a laptop computer An appropriate software was also developed to acquire accelerometer outputs which is reported in figures 75 76 VoltageO Ww Amplitude g Time SCS A Figure 75 Acquisition Software Front Panel 102 Configure DAQ CARD ie Se P eit a a a Save data to file oO 0 Ooo oo ooo ooo Instance 38 9Clipboard vi E j 00 00 00 00 ia nstance 31 9Clipboard vi EH DDN r Hy 100000 aS Figure 76 Accelerometers Data Acquisition Software The program can be used both for acquiring accelerometer outputs form the DAQ card and for data post processing by pressing the related button at the front panel If it is used for post processing the data file must be selected at the prompt window and frequency analysis can be performed If the program is used for data acquisition the DAQ Card must be configured appropriately through the DAQ assistant indicated in figure above The accelerometer scale in g V must be given as input which can be calculated knowing the charge amplifier gain and the accelerometer sensitivity eq 4 4
65. e 15 SIT Connection Manager Schematic Communication visual system GCS The visualization software was developed in C using OPENGL library and OPENInventor software It runs on the second computer of the GCS The software was provided by the Faculty VLAB and was partially modified to be interfaced with the LabView environment For LabView to communicate successfully with the visual system an UDP communication protocol has been developed in C and integrated in the LabView environment using the LabView call library function node The call library function node are LabView objects that link compiled source code written in a conventional programming language as C to LabView When the call library function executes LabView loads the C code and passes input data in that case the cluster to visual data to the executable code In that way data are sent to the visual system which will display a 3D virtual view on a TFT monitor 39 Communication force feedback joystick GCS The joystick chosen to be installed on the RUAV simulation environment was a Microsoft Sidewinder Force Feedback II Joystick as shown in figure 16 This decision was made rather arbitrarily and should not be construed as an endorsement Other capable force feedback joysticks exist on the market This joystick did have an advantage though It is compatible with a joystick driver written in C available from www Microsoft com For those with extra interest M
66. e area was reached at higher speed The landing maneuver was always done in manual flight mode Other post processed data are reported in the following figures Surveillance Mission Ground Speed Fire sightin Ground Speed m s Time s Wp0 Whpi Wp2 Wp3 Wp4Wwp5d Figure 28 Fire Surveillance Mission ground speed 49 Surveillance Mission Power Required Power Required KW t pen g Wp0 Wp4i Wp2 Wp3 Wp4Wp5 Figure 29 Fire Surveillance Mission power required Figures 28 and 29 show respectively the ground speed and the power required during each mission phase The total fuel consumption was about 33 Its figure 30 If the AV performs 3 times the flying path with no outrun the fuel consumption was 42 Its Surveillance Mission Tank Fuel Fire Sighting Tank Fuel It 3000 Time s Wp0 Wpi Wp2 Wp3 Wp4Wp5 Figure 30 Search Mission fuel consumption 2 3 2 ACTIVE JOYSTICK APPLICATION The task chosen for this study was a search identification mission which originates from an autonomous waypoint flight mode For the task accomplishment it was supposed that when the operator at the ground control station saw an object in the camera downlink he engaged a maximum acceleration mode minimizing the time to the object The pilot then engaged hover hold mode when he was near the object all the while
67. e its abays False Figure 69 Actuators PWM Generation Software 96 4 6 DATA LINK Usually data link are used for unmanned vehicles to send commands and receive telemetry or payload data and can be divided into digital and analog links An example for an analog link is a UHF video signal transmission Digital links provide a way of communicating between ground and vehicle mounted computers The frequency band a data modem operates affects its data rate Typically the higher the frequency the higher the data rate The frequency also affects the range of the data link Lower frequencies typically offer a greater range than high frequencies Furthermore the higher the frequency the greater the Line of Sight problem 1 e the ability to penetrate obstacles like buildings Common data links in the 2 4 Ghz band are more easily blocked than that in the VHF frequencies Also for unmanned vehicle operation a remote pilot data link is often used to steer the vehicle manually for some phases of the flight 52 The radio link used for the UNIBO helicopter works at a frequency of 43 835 MHz Instead a common WIFI access point fig 70 in the frequency of 2 4GHz is used to perform the data link between the onboard computer and the ground control station Depending also on the operating environment the data link range 1s about 200m 300m which is anyway quite enough for the goals of the UNIBO RUAV project Figure 70 RUAV WIFI Access Poi
68. e passed to the theta attitude controller The attitude PI calculates the servo rotation angle variation to be added to the trim value in order to maintain the desired set point 118 This commands are then scaled in microseconds of PWM high time which is used to generate the PWM signal for the servo actuator see PWM generation algorithm section 4 5 2 Unity of measure must be scaled as shown in figure taking into account the scale factors defined in table 12 Tables 13 16 summarize the input and output parameters needed for the algorithm to work properly Vx PI Input Parameters Vx SP vI16 User defined profile or from Navigation System Current Vx vI16 Calculated from AHRS data K Vx T 2 all6A1 6 5 Initial theta a116 2 cc _ oon o o PI Reset TRUE first PI call otherwise FALSE performed automatically by the program Table 63 Vx PI Input Parameters Vx PI Outputs Proportional action theta aI16 2 Total theta action a116 Table 74 Vx PI Outputs ee PI Reset TRUE first PI call otherwise FALSE iat performed automatically by the program Table 85 theta PI Input Parameters theta PI Outputs Proportional action Long cyclic with repect to trim sI16 2 TEREI G i Total Longitudinal cyclic action with respect to trim sI16 N B Long Trim value to be added after PI ouput scale 1544 us Table 96 theta PI Outputs 119 Vy phi PI A schematic of the Vy phi nested PI software is
69. econ meecentesiee aes 13 MOTIVATION AND BACKGROUND 00 oc cccccccccnseeeceecaeeeeeceeaaeeeeeeesaaeeeeees 16 EVOVERVIEW murea a nc ucea uscontact ET cave wales 2c 0 ease eeseanec aed deh cecsioseaees 16 LS UMINARY OF HELICOPTER PRINCIPE SS neroni EAR o a e EE 21 MISSION SIMULATION ENVIRONMENT 0000000 0 cece cccccccceseeceeeeeccecaneeeeeaeees 25 ZASTHE AV amp NGCS SIVIULINK MODEM eg secon clientes EE E E alates 27 2 1 1 CONTROL MODES DESCRIPTION AND FORCE FEEDBACK LAWS oee 29 Ded ENERGY MANAGEMENT EOGALION crenis e tie vacn aca E anaes aes nt coa ses 30 2 1 3 NAVIGATION GUIDANCE AND CONTROL SYSTEM NGCS 0ccccccccccceccceceeesessssseeeeseeeseseeeeeeas 34 LA GROUIND CONTRO LSPA HON audoise e ea e aa aa E desea teeta soolenger 36 2 1 1 COMMUNICATION MANAGER AND SOFTWARE INTERFACING ui ccccccccccccceeettteeteteeeeeceeesenenns 38 222 PORCH EPO BACK JOI SIG rosg ir a ewes bela exe ea ee 40 Doo STC OAV SII TINGE EAC T lta sttasi ints tat uaeutscsape edie a eaten dialed os 42 2 3 SIMULATION ENVIRONMENT APPLICATIONS oo oo eccccsesssccceceececssseessseeeeeeeeesessssesessaeaeesesesenes 46 2 3 1 CONFIGURA TION EVACUATION crecio ones ee ote aan Nana 46 232 ACTIVE JOYSTICK APPLICATION so wisctecnsprcvotastestsstustendnsauead E T OE 50 2 3 2 1 Search Identification Task Description cccseceseseececeececeeeeceseeceacaaaeaesseseseseeeeeceeeseeeeeeeeeaaes 5 TID Poed S nl AOS eoe E deceiasg tet decid a A 52 232 3 Coop r Har
70. ed as a unique control station constituted by two computers the master computer is used for real time mission planning control and for managing communication among the mission simulation environment computers IT is connected with two monitors displaying the mission planning window on the first screen and the flight instrumentations on the other one the second computer is used for providing a virtual view of the mission scenario and for mission payload data display An Electro Optic camera payload was also simulated by means of a Visual system developed at the Faculty VLAB In a preliminary design process this part of the system may be used to improve the GCS human interface design 26 Data Link Communication between the AV and the GCS is simulated of course via local area network LAN Bidirectional communication between the AV and the GCS primary TMTM software master computer 1s done by means of TCP IP protocol managed by a LabView Communication between the two computers of the GCS is done via UDP protocol and is always managed by the LabView software The goals to be achieved with the development of the mission simulation environment were many 22 l to create a simulation environment capable of prototyping a full RUAV system including also the GCS operator into the design process since the preliminary design stage 2 to provide a modular and open environment easy to upgrade by changing a single or a set of c
71. ed in an Inertial Measurement Unit IMU which also provides data from accelerometers Magnetometers are also used to determine heading of the air vehicle by measuring the Earth magnetic field Attitude and position can be then calculated in a state estimator by integrating IMU measurements However the high accuracy simplicity and availability of the Global Positioning System GPS makes it the emerging standard positioning system for UAVs as well as for general and commercial aviation Depending on the quality of the GPS receiver the achievable accuracy and the GPS update rate varies Since common GPS update rate is usually once a second this can 68 result in a limited bandwidth of the UAV controller A common way of solving that problem is to fuse data from all the flight sensors into a navigation filter in a state estimator In addition altitude data coming from a radar or sonar altimeter and magnetometers measurements can be also used to improve the navigation filter Usually an extended Kalman filter approach is used to integrate data from all the navigation sensors 52 An alternative solution to IMU individual gyros and self built navigation filter is to use a complete AHRS like the CrossBow NAV 420 which was chosen as navigation platform for the purpose of this work This kind of unit is able to directly deliver vehicle attitude GPS velocity and position data acceleration and rates at a rate up to 100 Hz thanks to a high performan
72. eded number of samples per channel 1s 6s 51200 Samples s 307200 samples per channel the Max Min Voltage level per channel depends on the expected voltage measurements to avoid overflow and on the desired resolution Since 104 measurements were quite different in the various points of the structure we used different Max Min V level values during the tests 4 7 1 2 Experimental Results A frequency analysis was performed on the experimental data using either LabView software fig 75 76 or Matlab software The most significant results are reported in pictures 78 79 From the power spectral density fig 78 79 it is clear that the major vibration sources are at a frequency of about 200 Hz and come from the engine A very small spectral component is present also at about 20 Hz and 80 Hz which are due to the main and tail rotor but can be neglected if compared to the engine component By defining a gms value as 58 59 g rms 4 std x mean x 4 5 where X is the vector of the acquired data n is the number of samples the x row number std x is the x standard deviation defined as std x CS x y 4 6 M j lI mean x is the x average value over the n sample mean x x 5 x 4 7 i l it is possible to check if the vibration load experienced by the electronic component once they were damped is within the operational vibration load advised by the factory for the onboard avionics compone
73. en via keyboard switch External view key 1 This button allows the operator to view the RUAV from outside fig 21 Figure 21 GCS UNIBO Visual external view Pilot view key 3 This button allows the operator to view the world from the cockpit of the RUAV fig 22 Payload view key 2 This button allows the operator to display the RUAV EO camera viewpoint fig 22 The onboard camera can be then moved by means of the POV joystick button Zoom in and out can be done simply using the computer mouse 44 Figure 22 GCS UNIBO Visual pilot EO view The visualization software was developed in C using OpenGL library and OpenInventor at the University VLAB and further modified to be integrated into the simulation environment It supports also 3D rotorcraft models in VRML format and other elements such as terrain model clouds airports and or buildings in the mission area The rotorcraft model is located in the virtual scenery based on the air vehicle position data received through the GCS primary master computer In order to create a realistic virtual scenery it is necessary to generate a detailed terrain model of the mission area Therefore the visualization software uses a terrain regular grid quadmesh which covers an area of 300 km x 300 km 46 see figure 23 Figure 23 GCS UNIBO Visual terrain mesh The terrain mesh is constituted by about four millions of polygons and nodes The nodes elevatio
74. ensor axes with the ones of the helicopter The NAV420 was installed along the x y z axis of the helicopter Before any flight test can take place it must be also ensured that the NAV420 is not rotated with respect to the helicopter that would cause wrong helicopter attitude measurements At this aim the AHRS is mounted fixed in heading alignment was compared to the one available from a small magnetic compass fixed with the h c in order to have a rough feedback about the heading correct mounting As for the roll and pitch angle the AHRS is mounted on a slew able flat plate around the x and y axis 73 Before any flight the helicopter and the NAV420 are both aligned with the g vector and the earth tangent plane so that they can be considered aligned between each other The helicopter airframe is positioned on a special rack and is aligned by means of a spirit level and of the rack screws see figure 49 After that the flat plate on which the NAV 420 is mounted is adjusted using its own screws so that the measured attitude is zero and the acceleration is parallel to the g vector Usually once the NAV 420 has been aligned with respect to the helicopter only attitude periodic checks are needed if the navigation platform is not moved from its site Another important precaution to be taken 1s that the NAV 420 must be mounted as close as possible to the helicopter Center of Gravity CG If it is not mounted at the CG then rotat
75. eory It was primarily derived from Mettler 20 30 31 32 theory of small scale helicopters the Engine Governor block which changes the throttle settings in order to maintain constant rotor RPM the Navigation Guidance and Control System blocks which are able to provide controls for the air vehicle stabilization and enable the air vehicle to track a set of pre planned flight segments starting from any initial condition the Switch block which is able to change the flight mode depending on a flag input joystick signal coming from the Ground Control Station Four different flight mode were implemented which are detailed in section 2 1 1 the Stability Augmentation System SAS amp Autopilot block works both as stabilization and autopilot system The autopilot gives controls to the helicopter flight dynamics block 28 for the Air Vehicle to maintain reference flight parameters depending on the selected flight mode the Force Feedback FF block gives back to the active joystick at the GCS different force feedback laws according to the current flight mode The active joystick features were used to backdrive the joystick in order to eliminate transients between the various mission phases and to provide situational awareness to a potential ground control station pilot Details of the dynamic model and of the SAS and autopilot can be found in 19 other useful references are in 33 39 In the next section details of the navigation gui
76. erefore some background is provided here below The I2C Protocol If the CRIO has to talk to a slave the sonar SRFO8 it must begin by issuing a start sequence on the I2C bus A start sequence is one of two special sequences defined for the I2C bus the other being the stop sequence The start sequence and stop sequence are special in that these are the only places where the SDA data line is allowed to change while the SCL clock line is high When data is being transferred SDA must remain stable and not change whilst SCL is high The start and stop sequences mark the beginning and end of a transaction with the slave device see figure 57 otat seguente otop sequence JUA JDA alL SCL Figure 57 2C Start and Stop Sequence 55 Data is transferred in sequences of 8 bits The bits are placed on the SDA line starting with the MSB Most Significant Bit The SCL line is then pulsed high then low actually the chip cannot really drive the line high it simply lets go of it and the resistor actually pulls it high For every 8 bits transferred the device receiving the data sends back an acknowledge bit so there are actually 9 SCL clock pulses to transfer each 8 bit byte of 83 data see figure 58 If the receiving device sends back a low ACK bit then it has received the data and is ready to accept another byte If it sends back a high then it is indicating it cannot accept any further data and the master should terminate
77. esssseesseesseesseeeeeeeeeeeeeeeees 106 Figure 79 Acceleration Experienced on the isolated NAV 420 0 0 ccccccccceeeeeeeetessentseeaes 107 Figure 80 Acceleration experienced after the first shock mounts cccceeeeeeeseeeeeeees 107 Figure 81 Boeing Results for Raptor 60 62 seicscacesesiessaleian ieee eee ee 107 Figure 82 Example of pitch and roll rate AHRS raw data oo cccccceceeeeeseeseesesnentneeaes 108 Pisure 83 AHRS filtered lieht Galan wesc ceva eae uae NLE 109 Proure 84 Sonar SEMSOLS measure MEMS errn eet dac are tate rie ieee atta neha aaah tabi 109 Figure 85 Post View Station Architecture c ccssesesesseeceesesecsesecccenceeeeeetescessesensaees 110 Figure 86 Onboard Control System Architecture c ccccccccesessseesseeeeeeeeseeeeeeeeeeeeeeeeeees 112 Figures Onboard Control LO Op cists codes ne E EAE aden ees 117 Figure 88 Schematic of the FPGA Vx Theta nested PI oo cccccccceeeeeeeesseesesssensaeaes 118 Figure 89 Schematic of the Vy phi nested PL ccc ecessssesssssseseseseeeeeeeeeeeeceeeeeeeeeeeeees 120 LIST OF FIGURES Continued Fioure 90 Schematic of the stand alone Vz PL csdedsrcacdtaceh thease neds 121 Figure 91 schematic of the heading Control cccceceeesseescseceeeeeceeeeseeeeeeeeseseseeseeessensaaaes 123 Figure 92 RUAV Complete Software Implementation ccccccccccccccccceeceeeceeeeseeeeeseeeeeens 125 Proure 9e PL fuming t
78. ests OGUN herensia E aessbasatdectaeseeens 126 Fiure 94 Pl tumine tets GUNZ Jeran E A aE 127 Figure 95 Typical Setpoint ProrleS crnkica E RERE A eee 128 Figure 96 Typical Flight Pattern Profile cc cccccceseessssssccceceeeeeeeeeeeeeeeeesessesseesessessaaaes 128 Figure 97 Flight Data Acquisition GUL cccccscssssssessesseaeasaacaneceeeeeeeeeeeeeeeseeseseeseeseseeegs 129 Fiure 9S schematic ot lly Sim Ulan acces osoreccpeseeetorcasnnhtucanendentectonenecaiaieeeeaehstaelaekets 130 Pigure 99 Recorded HI Simul latin sssesssieaisadeinnds at mnacendh E tes 131 Figure 100 Flight Tests Procedure Sisaren Ra 132 Figure 101 Simulate vs Experimental longitudinal controller tracking performance 134 Ficure 102 Heading Control Mohite S erreien aaa E 134 Fure 103 Fugit Paia E NAA N 135 Fore 104 Went Pattern TeSt oA A E 135 10 LIST OF TABLES Tal els Fe Int Fe Veil a aa a sare sagas uae sean tast te semecatonam ern a 47 Table 2 Definition of Desired and Adequate Performance cccccccccceceeeeeeeeeseeseeesesseeeeees 53 Table 3 Cooper Harper Ratings for the Rotorcraft UAV Search Identification Task 58 Table 4 NAV420CA Packet Details NAV Mode 53 cc cccccccseeceesseeessssessssseesseeeeeees 78 Tabled Accelerometers Characteristics spac ccs casnsscedectisssacgansiescdocteeseacecsandannatenteeeeasoateens 104 Table O DAOL ard Se UNG S sie 025 ssc dassadenaencusticeusconcednebetes ooucctisa A 1
79. ete preliminary design of two rotary wing UAV configuration one was a conventional main rotor tail rotor configuration while the other one was a coaxial rotor configuration The culmination of the UNIBO CAPECON work was the design of a mission simulation environment for the rotary wing UAVs The mission simulation environment was used inside the CAPECON for evaluating the two RUAV configuration operational capabilities The work carried out by UNIBO during the CAPECON program will be summarized in chapter 2 Based on the work performed in the CAPECON program an independent RUAV research program was also started at UNIBO laboratories since RUAV systems may represent an alternative to fixed wing UAVs or even a more promising solution for a wide number of civilian applications due to their versatile flight modes maneuverability and vertical take off and landing capabilities The goal to be achieved with the UNIBO RUAV research program was to develop a helicopter capable of autonomous flight which could be used inside the university as platform for researches in control and navigation laws meanwhile it should be proposed as technological prototype to industry interested in UAV development and manufacturing An UAV system is generally constituted by at least four main integrated sub systems see figure2 the air vehicle AV the ground support system the data link and the data distribution 9 18 Figure 2 UAV System The AV inclu
80. eters control Parameters Flight plan Parameters control Wat Tee Tem Vee Target Time Step Vale _ sc j s PWM Trim Tal start of dead sone end of dead zore hiss iE Fot oor TRep3 Tregt Target time TStept Tepe Sev Pare TELS Data logging control Stop GCS PI Velocity Parameters control Figure 94 PI tuning tests GUI 2 Graph control During PI tuning tests the graphical interface was used to monitor the helicopter PI responses by means of dedicated diagrams to enter a given flight plan or a set point time history and to change the PI gains if necessary Six windows of diagrams are available one for each variable under control Moreover the user can start or stop the onboard data logging by means of the front panel button shown in figure 94 The GUI interface can be stopped independently from the onboard software so that if TCP IP communication is lost this will not affect the onboard program and the flight test can be concluded without any data loss Three different type of automatic flight mode are available Normal mode the user sets a velocity or heading profile to be maintained by the helicopter The generated profiles are shown in figure 94 and can be defined by the user setting the wait time the target time the trim set point and the desired step All set points a part form heading are given by means of a ramp to avoid sharp transit
81. fied Onboard Computer CRIO from NI AHRS Crossbow NAV420 Sonar Sensor Ground Station NI Compacthio AHRS Crossbow Real Tim Dro Pa analog ened digital 1 0 RS 232 Fe e i J n mT PWM To 10 100 Mb Be y Ethernet link _SServos OUTPUT 5 Serves OUTPLT commands PWM for js Helicopter Actuators i WIFI Communication SAFETY os Ground ce Station Figure 39 RUAV System set up and Architecture Next section introduces the applied design and integration methodology used to set up the RUAV system Details of the work performed in this thesis for the development of the RUAV system will be given in the next chapters following the design methodology described below 3 1 DESIGN PROCESS The design methodology followed for the RUAV system development is depicted in figure 40 60 Autopilot Design Vehicle Simulation Hardware ete an Selection amp Integration AS NEEDED Onboard Autopilot SY Development Hardware amp Acquisition SW FLIGHT TEST Helicopter Dynamics identification In Flight Autopilot Test ROTARY UAY Figure 40 RUAV Avionics Design Flow It is a multidisciplinary design process which includes six main steps 1 Hardware Selection and Integration this task include the selection and set up of the rotorcraft airframe and of the onboard avionics Once the hardware is selected it must be
82. force is balanced by the aerodynamic drag force Of course some cross axis effect are also present For example in the longitudinal velocity control when the helicopter is pitched the vertical thrust component will decrease requiring an increase in the thrust magnitude to keep the vehicle at level altitude This increase in thrust however will produce a reaction torque at the rotor shaft that in turn will result in a yawing moment for which the pilot will need to adjust the tail rotor thrust Other effects are more subtle such as the roll responses following cyclic and collective control actions and the pitch responses following lateral control actions 20 24 Chapter 2 MISSION SIMULATION ENVIRONMENT This chapter will describe the mission simulation environment developed inside the CAPECON program to evaluate the operational capabilities of the two RUAV configurations designed by the industrial partners AGUSTA and Eurocopter 21 The simulation environment was developed mainly as a tool for supporting the CAPECON industrial helicopter designers in the preliminary design phase of the RUAV systems In this perspective it was taken into account that the design of a RUAV system is quite different from the one of a classical manned rotorcraft since it requires the concurrent definition of its main sub systems the air vehicle the ground support system the data link and the data distribution For example unlike manned helicopter the
83. gure 45 NAV420CA System Architecture 69 The choice of this kind of platform significantly reduced development time in signal processing and sensor fusion greatly improved measurement reliability and guaranteed sensor stability and performance in a high vibration operating environment like the one of a small rotary wing platform 4 4 1 1 AHRS Set Up The NAV420 set up procedures was done following four major steps see figure 46 1 TESTRIG Measurements 2 Setanpoprate eter seinas 3 Irstataton and Final Galbraton 4 Cremation SW Development Figure 46 NAV 420 Set up Procedures first some measurement tests were performed on a certified test rig in order to verify the navigation platform responses a LabView software was then developed in order to change the NAV 420 default settings before using the NAV 420 data inside a control algorithm the update rate the baud rate and the output packet type must be set to appropriate values afterwards the navigation platform must be installed inside the avionics box appropriate procedures must be followed in order to obtain correct states measurements in the end NAV 420 data acquisition software must be developed in order to read sensor information to be used inside the onboard control software Test Rig Experiments Figure 47 shows the NAV420 mounting on the UNIBO test rig during the first experiments performed to verify the platform responses 70 NAV 420 on Test Rig
84. h takes into account also for the extra term This feedback law is used when the helicopter flies in maximum acceleration mode as it is implemented in the force feedback block of the simulator together with the other feedback laws used for the three other control modes see previous section 33 2 1 3 NAVIGATION GUIDANCE AND CONTROL SYSTEM NGCS The NGCS system used to guide the helicopter implements a proportional navigation strategy The same strategy was also used for the onboard navigation software implementation of the rotary wing air vehicle developed in this PhD work see enclosed CD The guidance system is composed by two main parts the lateral track control and the altitude hold 43 Guidance Altitude_Hold Figure 11 AV Guidance Simulink blocks The altitude hold is a simple Proportional Integral PI controller It takes as input the destination waypoint altitude and the current vehicle altitude and gives as output the vertical velocity required to maintain or reach the reference altitude The lateral control strategy guides the helicopter towards the destination waypoints e g WP 2 along a track line defined by two consecutive waypoints e g WPI and WP2 as depicted in Figure 12 by means of a yaw rate command 43 The guidance block first transforms the helicopter latitude and longitude coordinates into a Xtrack Ytrack local system see figure 12 Knowing the AV current position in the local coor
85. he figure that Pilot B could perform the manual mode task easier with a backdriven joystick with an airspeed hold command turned on It should also be noted that during Phase 1 Autonomous mode the joystick is continuously backdriven to match the airspeed hold command Therefore when manual mode is turned on there are no transient inputs The last phase for the pilot is Phase 4 hover hold When the pilot determines that the object is close enough to the UAV he pushes the joystick button for hover hold mode This is an automatic hover stabilization which automatically backdrives the joystick to zero out velocities in all 3 axes The heading is maintained to be the last active heading that is with the nose pointed towards the object During this phase the pilot does not enter any control inputs except that he has to operate the Point Of View switch in the joystick to keep the object centered within the red square on the monitor During the switch from Phase 3 acceleration to Phase 4 hover hold the onboard camera switches from pilot view camera fixed forward to camera view camera is slewable The pilot does not have enough cues to do the hover hold himself that s why it was automated But during the stabilization the pilot still needs to maintain visual contact with the object and this extra workload is reflected later in the Cooper Harper ratings When the aircraft is stable then the pilot officially ends Phase 4 and completes the se
86. helicopter control strategy 1s one of the most critical aspect in the design of a RUAV Besides the ground control station must be also brought into the design space because it plays a crucial role for the RUAV platform operation In order to develop such a complex system at industrial level it was decided inside CAPECON that new design strategies are needed which could be integrated into the design process normally used by the manned helicopter companies 22 28 In this perspective the idea was to create a simulation tool able to merge the contributions of different preliminary design working teams the helicopter preliminary design the NGCS preliminary design the GCS preliminary design and payload integration into a single environment and to test them in cooperative simulation Such a simulation environment was developed at UNIBO Laboratories and is shown in figure 5 25 Figure 5 Mission Simulation Environment The simulation environment is able to emulate the RUAV main sub systems 10 21 It 1s composed by three computers and three monitors figure 5 The Air Vehicle computer it represents the airborne world and contains a Simulink dynamic model of the helicopter and of the NGCS In a preliminary design process this part of the system may allow the designers to test different NGCS solutions The Ground Support System computers for the sake of simplicity the Ground Support System has been simulat
87. icant byte put it into a LabView boolean array and then negate it 2 read the second eight information bit sequence belonging to the least significant byte put it into a LabView boolean array and than negate it 3 put these two boolean arrays into a 16 element boolean array with the first byte in the 8 15 position and the second byte in the 0 7 position For the example reported in figure the information will be interpreted correctly by LabView as follows 1011001110011000 2 27 2 2 27 28 2 2 6349 116 format 5 In the end a checksum control is performed in order to verify the correct data packet acquisition The acquisition continues till the program is stopped 4 4 2 ALTITUDE SENSORS Altitude sensors measure with reference to sea level AMSL Above Mean Sea Level or the local ground AGL above ground level This kind of data is needed in order to control the altitude of the aircraft depending on the vehicle type and on the modes and location of operation Operations within ground vicinity such as landings usually require absolute AGL measurements or a very accurate terrain database 52 Available sensors include e Sonar AGL e Radar AGL e Laser Lidar AGL e GPS AMSL e Barometric AMSL Sonar sensors were chosen to measure the helicopter altitude with respect to the ground Important issues taken into account for the selection of the altitude sensor type besides the type of measurement
88. ich contains the host source code and communicate with the simulation computer by means of a TCP IP protocol an OpenGL visual system computer can be optionally added for rendering the helicopter as it moves around in a virtual scenery The visual system can receive input from the GCS computer using a TCP IP protocol The result of this setup is that the on board computer effectively thinks it is flying the vehicle as all of its configuration data flow is identical to an autonomous flight setup In this scenario performance and possible errors of the onboard software can be detected during intensive ground safe simulations before performing any flight test Figure below shows an example of HIL simulations results SetPont Response Theta deg Dong deg o nH i Time sec Figure 99 Recorded HIL Simulation 131 Chapter 7 FLIGHT TESTS amp PI GAINS TUNING The onboard control software was tested in flight The complete flight campaign was done following five major subsequent steps as show in figure 100 4 Attitude Controller Tests 2 Nested PI Velocity Attitude Controller Test 3 Nested PI Velocity Attitude Heading Controller Test Nested PI Velocity Attitude 4 Heading Vz Controller Test 5 Autonomous Flight Pattern Test Figure 100 Flight Tests Procedures First only the attitude and 0 PI controllers were tested During these tests collective and tail commands were left to the RC
89. icrosoft offers a Software Development Kit SDK for free to support game developers Visit the Microsoft website and download and install the latest version of the SDK 200 to 300 MB The kit includes everything needed to write video games including a joystick interface Find the file called joystick dsw This is a finished project file compatible with the Microsoft Visual C compiler The C code is written using DirectX libraries which enable the generation of force feedback effects for devices that have compatible drivers The code was partially modified to be integrated with LabView so that the force feedback law inputs coming from the simulator can be passed to the active joystick A compiled dynamic link library was generated and implemented in the force feedback manager module of the GCS source code This setup allowed testing of new active control ideas inside the mission simulation environment before turning to more sophisticated simulations Figure 16 Microsoft Sidewinder Force Feedback II Joystick 2 2 2 FORCE FEEDBACK JOYSTICK The joystick has eight buttons and a 4 position coolie hat Button pushes can be recognized within Simulink or National Instruments LabView allowing pre programmed control modes For example one button initiated autonomous waypoint flight while another initiated maximum acceleration flight the other two are left for the hover hold and manual mode The 4 position coolie hat
90. iew Software for Encoder Signal Acquisition and PWM generation ront ING DAON IEG eed cin fc cit alec a a aa ules oeronaaaaenae teviatanatactios 92 Figure 66 Front Panel of the software reported in figure 65 00 0 ccc cecccccccceceeeeeceeeeeeeeeeeseeeeeens 92 Figure 67 PWM on time Servo Angle curve ccccccccecsesesssssessecsecsecesececenceeeeeesatseeeesessess 93 Figure 68 Actuators PWM Acquisition Software ccccccseeeeseesssenceeeeeeeeeeeeeeeeeeeeeeesseeeeens 95 Figure 69 Actuators PWM Generation Software cccceceeeseeessenssscceceeceeeeeeeeeeeeeeeeeeeeeees 96 Fig re 70 ROWAN WIEL ACCESS POU uicit ieaiai E EEEn E Ei EEEa 97 Figure 71 RUA Y Schematic Wirin sarusun na E E aedeea enue 98 Figure 72 Avionics Vibration Isolation System ccccceessessssesceceeeeeeeeeeeeeeeeeseeeeeseeeeeens 101 Figure 73 Typical diagram of resonant transmissibility versus damper frequency 101 Figure 74 Experimental Data Acquisition System cccccccssccscccecceceeeceeceeeeseesesseeessssensaaaaes 102 Figure 75 Acquisition Software Front Pamel ccccccccessssseesesesenseecececcccceeseeeeeseeseeeseesess 102 Figure 76 Accelerometers Data Acquisition Softwar e ccccccsssssssscccececeeeeeeeeeseeeeeeeeeeens 103 Figure 77 Accelerometers mounting points 2 0 ceeeeessesseessesseeseessesseeseessseseseseesseeceeeeees 106 Figure 78 Acceleration experienced on landing Gear eeeseeess
91. ignal based on the PWM period input which is fixed to 20000 us by the servo actuator frequency and the PWM on time input which is provided by the control loop This subVI is used five time one for each helicopter command inside a PWM generation while loop see fig 69 Basically the program creates a virtual clock whose time value ranges between 0 20000 us When the virtual clock time value reaches 20000 Us it is reinitialized to zero Each loop the virtual clock value is compared with the PWM on time input If the virtual clock value is less then the PWM high time a Boolean TRUE output is generated otherwise the output is driven to FALSE PWM Generation subVl Sub for generates signak PAM width modulation This sub amp caled by Hety_PID FRGA wi this subvl is executed only once every time it is called since it is inside the while loop C 9 this is executed every tine the C 9 while loop executed The way the progamis prevented to continue cyding ride miputs 1 PAM Period us U32 period of the PWM signal 12 PWM High Time us 32 width of the PWM signal high level touts D1 PAM signal tis true tl PYM high time i reached It it fake in the period btw the and of PWM high and the PWM period then it comes tue again 1 1 This input value abeays ranges from 0 to 20000 us Its a virtual dock ranitdined to zero each time the value of 20000 us amp reached 1 2 Tt becomes true only during the transition to 20000 us otherwis
92. ilot compensation is required Major deficiencies to retain control 9 Is No Improvement 3 Control will be lost during some portion of it controllable Major deficiencias required operation 40 Pilot decisions Definition of required operation involves designation of flight phase and or subphase with accompanying conditions Figure 38 Cooper Harper Decision Tree The decision tree in figure 38 is well known and every attempt was made in these tests to respect its assumptions 57 Cooper Harper Joystick eee Ratinas Active eee Active pfs os Table 3 Cooper Harper Ratings for the Rotorcraft UAV Search Identification Task Table 3 contains the final results for the handling qualities evaluation for the search identification task The results show quite clearly the impact active joystick features can have on handling qualities With the features turned on control of the rotorcraft UAV for this task is Level 1 normally accepted as quite good But with the active features turned off control of the UAV drops into the lower part of Level 2 which normally means the system requires improvement especially true if it 1s to be certified under civilian airworthiness regulations 58 Chapter 3 ROTARY WING UAV SYSTEM DEVELOPMENT The goal of UNIBO RUAV project is to develop a helicopter platform capable of autonomous flight which could be used inside the University for researches in control and navigation laws man machine in
93. imb descent Moreover also the pitch angle can be changed if necessary see figure 7 29 Feedback Laws Manual gt Reference Speed Autonomous gt Current Speed Acceleration gt Energy Law 0 Hover Hold gt Gurrent Speed Payload Camera Control Manual gt Fwd Velocity Autonomous gt Not Used Acceleration gt Hover Hold gt Not Used Manual gt Yaw Rate Autonomous gt Not Used Acceleration gt Yaw Rate Hover Hold gt Yaw Rate Change Flight Mode Manual gt Lat Velocity Autonomous gt Not Used Acceleration gt Not Used 0 Hover Hold gt Not Used Manual gt Rot Autonomous gt Not Used Acceleration gt RoG Hover Hold gt Roc Figure 7 Control Modes and Force Feedback Laws In hover hold flight mode the hover hold block allows to quickly decelerate the helicopter in order to reach hover flight conditions and to maintain current spatial position The force feedback module backdrives the joystick stick to follow the current speed so that it works as a speed indicator helping situational awareness The reference flight parameters commanded through the joystick are the yaw rate and the rate of climb descent see figure7 2 1 2 ENERGY MANAGEMENT EQUATION Energy management equations were implemented in order to define a maximum acceleration flight mode towards the object for task accomplishment of a potential search m
94. ion Version 2 4 Doc N 93983934001 Philips Semiconductors Janaury 2000 Using Quadrature encoder with E series DAQ Cards National Instruments application notes Vibration and Shock Theory www lordmpd com http www mathworks com products matlab The MathWorks Inc ed T M W Inc Natick MA User s Guide for Matlab Simulink Toolboxes 2001 Fundamentals of FFT signal Anlysis and measurements Labview Bookshelf 2004 A V Oppenheim and R W Schafer Discrete time signal processing Signal Processing Series Prentice Hall Englewood Cliffs 1989 140 62 63 64 65 66 67 68 69 70 71 72 73 74 www boeing com Penn State University study on Raptor 60 Good Vibrations Istvan Kollar Frequency Domain System Identification Toolbox MATLAB User Guide L Liung System Identification Prentice Hall 1987 P G Hannel and R V Jategeonkov The evolution of Flight vehicle system Identification Agard Lectures Series on Rotorcraft System Identification 1995 J S Bendat and A G Piersol Engineering Applications of Correlation and Spectral Analysis John Whiley amp Son New York NY 1993 M Tibaldi Progetto di sistemi di controllo Pitagora Editrice Bologna 1995 G Buskey J Roberts P Corke G Wyeth Helicopter Automation using a low cost sensing system IEEE International Conference on S
95. ion around the CG can cause NAV 420 accelerometers to measure acceleration difference equal to the angular rate squared multiply by the distance between the NAV 420 and the helicopter CG This in turn may also affect velocity and position measurements The helicopter CG can be easily determined experimentally The NAV 420 was aligned with the CG in the x and y axes while there is an offset of about five centimeters along the z axis This small offset however will not significantly affect the NAV 420 measurements for several reasons the offset is very small the angular rates are usually small since the helicopter doesn t perform extreme maneuvers the NAV 420 internal Kalman filter updates velocity and position measurements using GPS information so that possible errors can be partially corrected Finally when installing the NAV 420 in a vehicle and the vehicle contains ferro magnetic parts as the helicopter for example it is necessary to perform a magnetometers calibration procedure for hard and soft iron compensation before using it The several steps to be followed for the calibration procedures are described in the NAV420 user manual and can be performed using the NAVView software provided by CrossBow Other calibration procedures are not necessary since NAV420 internal sensors comes already factory calibrated for temperature bias scale factor and misalignments 74 AHRS Data Acquisition Software The CrossBow NAV420
96. ion payload simulation and for a data driven virtual view of the flight vehicle displaying the helicopter current position The basic software was developed through the LabView data acquisition control and visualization software The LabView software has been chosen due to its quick and flexible applications The source code implemented on the primary master computer of the GCS is able to manage communication between the Simulink model of the air vehicle and the master computer of the GCS communication between the visual system developed in C language and the primary master computer of the GCS the graphic interface for mission control and flight data display Figure 14 shows the LabView code which runs on the GCS master computer please refer to enclosed CD for complete software implementation a D i E pre Read from Simulator Figure 14 GCS LabView code The LabView code is constituted by different blocks 2 a read loop which receives data from the RUAV simulator via TCP IP using the LabView simulation interface toolkit 2 blockset a data selection block which is able to split the data received at the GCS primary master computer into four main cluster of data to be displayed on the GCS graphic interface a cluster to visual data a cluster to map data a cluster to virtual cockpit data and a feedback laws data to be sen
97. ion task could be performed using only a datalink videolink and a steerable camera which are well simulated inside the mission simulation environment No sophisticated object recognition software or fancy sensors were assumed The idea was to evaluate the operational capabilities of a system capable to be trucked somewhere set up and flown with minimal personnel equipment and operating cost thus demonstrating also that the task can be performed by an RUAV ata fraction of the cost of a manned aircraft 2 3 2 1 Search Identification Task Description The search identification task had four phases autonomous waypoint flight manual mode acceleration mode hover hold mode The four task phases are shown in Figures 31 34 The most demanding phase was phase 3 the acceleration task A maximum acceleration was required to save valuable minutes taken from the overall search 51 w Phase 2 manual mode turn towards object Phase 1 Autonomous waypoint flight Align object in pilot view Object spotted with Slewable camera Figure 31 Phase 1 Autonomous Waypoint Flight 42 Figure 32 Phase 2 Manual Mode 42 object Phase 3 acceleration mode Keep object aligned in pilot view while performing maximum acceleration Phase 4 hover hold mode Keep object aligned in camera view using the Point Of View button on the joystick View of object from Ground Control Station a Red square is alig
98. ion to avoid sharp changes in integral action when there is a sudden change in PV or SP The integral contribution is therefore express as os A yo 5 10 As for the derivative contribution the derivative action is applied only to the PV in order to avoid effects due to abrupt changes in SP Therefore the following formula represents the Derivative Action up t PV PVU D 5 11 S So finally it is possible to implement the following formula for the discrete PID controller 116 u t K e t t K e t a T S KT yor I KT PV t PY t ei 5 12 Another important aspect is that the use of a summation to calculate the contribution of the integral term can lead to problems causing long periods of overshoots in the controlled response This phenomenon is known as integral windup The algorithm implemented on board provides code for integrator anti wind up 5 2 2 ONBOARD NESTED PI SOFTWARE The Complete nested PI software is shown in figure 87 The control loop runs at 50 Hz taking into account the bandwidth of the helicopter servo actuators The loop performs a series of instructions in three subsequent frames in the first frame the control loop rate is set the second frame is used to read all the input parameters necessary to the controller the third frame contains the PID implementation and calculation of PWM high time to be sent to the PWM generation loop see section 4 5 2
99. ions This setpoint mode is very safe because if communication between the GCS and the onboard computer 1s lost the helicopter will safely complete the task and flight tests have not to be abrupted 127 V SP Step V SP Trim hovering Heading Step i 0 360 Initial Heading 0 360 Figure 95 Typical Setpoint Profiles Fast Mode the user sets a velocity or heading step profile only by giving the step value and no ramp will be used the reference signal stops when the user drives back the set point to trim condition This mode is more risky with respect to the first one since 1f communication between the GCS and the onboard computer is lost the helicopter will not go back to the trim condition after a short period of time The only way to recover from this situation is switching back to radio controlled mode However this mode was used only for preliminary and fast tests Flight Pattern Mode the user can set a complete squared flight pattern to be track by the helicopter as shown in figure 96 c Flight Pattern Initial Heading 0 360 Figure 96 Typical Flight Pattern Profile Inputs defined by the user are the helicopter initial heading the Vx step reference point the target time the time of each velocity step and the time between each velocity step this time is used to rotate the helicopter in heading while the Vx 128 velocity has been driven to zero When the helicopter i
100. ission The maximum acceleration was computed using an extension to an approach documented in reference 40 Reference 40 documents an approach to compute the pitch attitude required to perform a maximum acceleration at constant altitude while not exceeding limits in the vertical axis For the purpose of this work the Energy Management approach was used but it needed a slight extension to take into account also for the rate of descent which could be experienced by the helicopter during a mission descent phase The equation for the pitch attitude leading to maximum acceleration in forward flight and at constant altitude is shown below The energy management extension approach was developed in cooperation with DLR Ing Stephen Mouritsen Ref 42 is the outcome of this work 30 AB evel flight E e 2 1 V EoLA l 5 max On d si For the original derivation please refer to reference 40 The terms in the equation are defined in the notation This equation assumes a helicopter flying in forward flight so that the reciprocal of the forward velocity does not go to infinity Each term of the equation is readily available the forward velocity the maximum collective the trim collective and the stability derivative rate of climb vs collective This equation assumes level flight and computes the nose down pitch attitude such that there is sufficient collective margin to hold altitude see the figure below max acceleration
101. it inside one byte is transmitted from the least to the most significant the LabView program language uses boolean arrays to store bit byte information which uses opposite logic levels with respect to the one of the RS232 protocol Morevover LabView associated the significant bits bytes to the index position inside the array the NAV 420 packet type in NAV mode is composed by 37 byte Data are signed 116 2 byte or 132 4 byte format depending on the information type as show in table below 11 Pitch Angle Degrees 7 8 Degrees 9 10 Roll Rate Degrees second 13 14 Yaw Rate 630 630 nes second 21 22 23 24 Longitude Degrees o g w S o s o OO O Table 4 NAV420CA Packet Details NAV Mode 53 Degrees For example to read a 2 byte information data type the following step are needed as shown in figure 52 First Bit Read Most kera icant byte 4 o transmitted first l koera l al First Array idle Es idle l l l Least Significant Bit Bit transmitted first Negated byte Information Available JE o fo o First Bit Read idle o iae m Start Bit Least significant byte transmitted last Array Rean Negated byte Figure 52 NAV420CA Packet Acquisition Sequence 2 ol o o o o i pom me m a M M ES 2 yna a be ry 78 1 read the first eight information bit sequence belonging to the most signif
102. ither recorded on the CRIO volatile memory for post processing or communicated to the ground control station the normal priority loop which perform TCP IP communication with the ground control station and is timed at 100 ms The Host Software is constituted by two independent loops the communication loop which perform communication with the onboard computer for flight data transmission Communication between the ground station and the onboard computer is bidirectional since the operator at the ground can interact with the onboard software by changing flight parameter values for example the PI gains can be changed during flight test for controller final tuning the user interface loop which contains the code to generate the user interface for flight test control and monitoring The user interface code is made completely independent from all the other code so that different type of Graphical User Interface can be used without need to change any other part of the source code Depending on the flight test to be performed different GUIs were developed figures 93 94 97 show the ones used during PI tuning tests and flight data acquisition tests Phi Theta wy ve ve Pst 3 pop GPS amp AHRS Status Monitoring reve H 1 700 0 800 0 900 0 1000 0 1100 0 1200 0 1300 0 Time 5 e ct mw PI Monitoring Collective 188 189 190 Time AA Figure 93 PI tuning tests GUI 1 126 Param
103. lades rotating about a vertical axis the rotor shaft In order to understand the contents of this work some basic helicopter principles will be introduced The standard helicopter notation that will be used in the next chapters is shown in figure 3 oe Wyr Figure 3 Basic Helicopter Notation 20 Figure 3 shows the helicopter with its body reference frame origin is at the helicopter centre of gravity The principle variables are shown on the x y and z body axes They include the body speeds u v w the Euler angle y the body angular rates p q r The main rotor is represented as a disc that can tilt about the rotor hub in the longitudinal and lateral directions This motion is describe through the angles Bic and Bis measured in reference to a plane perpendicular to the rotor shaft hub plane The actual rotor blade position is described by the angle Y measured from the tail see fig 3 The components of the resultant forces and moment acting at the helicopter centre of gravity are X Y Z and L M N respectively The figure shows also the key forces that contribute to the helicopter motion including the rotor thrust T the longitudinal and lateral rotor moments acting on the hub Lr and Mp the main rotor torque Q the in plane rotor forces Hx and Hy the tail rotor thrust Trp the fuselage aerodynamic forces Far the aerodynamic forces from the tail surface Far 20 21 Rotor blades are attached to the spinning shaft
104. le O04 O86 g 1 1 2 1 4 16 1 6 2 a2 T High ms Figure 67 PWM on time Servo Angle curve The servo angle is a linear function of the PWM on time and the scale factor was found to be 93 BD 210432 s 4 3 ATh which means that one degree servo rotation corresponds to 9 6 us PWM on time 4 5 2 ACTUATORS SIGNAL ACQUISITION AND GENERATION SOFTWARE For the helicopter to fly autonomously PWM signal outputs must be generated by the onboard computer for the servo actuators Furthermore if a dynamic helicopter model must be identified for autopilot design helicopter responses to pilot PWM inputs must be also recorded For that a FPGA software was developed both to generate PWM output signals and to acquire pilot PWM input commands PWM signal acquisition software The PWM radio input signals have been acquired by measuring the corresponding PWM on time in microseconds using the FPGA digital input channels The channels configuration is as follows Slotl cRIO 9411 DI 0 lateral cyclic pitch Slotl cRIO 9411 DI 1 longitudinal cyclic pitch Slotl cRIO 9411 DI 2 throttle Slot1 cRIO 9411 DI 3 tail Slotl cRIO 9411 DI 4 collective pitch Slotl cRIO 9411 DI 5 PID on off radio channel 7 The software is reported in figure 68 Basically the software is constituted by a while loop running at 1 MHz 1 loop every one microsecond The PWM on time is measured by creating a virtual microsecond counter Every microsecond the
105. lead to the definition of formal requirements 5 The formal requirements were used to help defining five fixed wing and two rotary wing configuration which best fulfilled the requirements The configuration were then designed and the related technology were also identified The costs for the configurations were also estimated and final dissemination of results to the European industry was done During the CAPECON project interest in the Small Mini size UAVs increased considerably mainly due to the miniaturization of avionics and onboard systems The project was therefore extended to study also the Small Mini Configurations UNIBO played an important role inside the rotary wing part of the CAPECON program both in the Small Mini and big size Rotary UAVs RUAV The main CAPECON outcomes for the RUAV systems concerned the identification of the most promising applications 6 7 Examples of the identified applications included fire surveillance and fire fighting civilian security monitoring or close inspection of electrical powerline pipeline or dam search part of SAR search and rescue missions agriculture spray etc 8 9 10 17 as far as technology are concerned one important aspect derived from the CAPECON program was the real need to apply already existing proven and cost effective technology to the UAV world Therefore many existing technologies were identified ready for application in the short and mid term 11 18 the compl
106. les FPGA PXI communication with the computer emulating the CRIO Real time core can be performed by means of a FPGA interface card a computer which simulates the helicopter plant through the identified transfer functions reported in chapter 5 and the onboard sensor outputs The helicopter simulation model receives inputs from the PI controllers the software implemented 130 inside the FPGA PXI 7831 is able both to generate PWM electrical signal for the actuators and to acquire them by reading the PWM high time which is sent to the helicopter simulator The PWM high time in microsecond is then translated into degrees of servo rotation see section 4 5 which is the actual input accepted by the identified transfer function In this simplified model states outputs are Vn Ve Vd W p q r which are formatted into a NAV 420 string emulator and are sent to the PXI by means of the computer serial port The NAV 420 FPGA acquisition software running inside the PXI will acquire the helicopter states such closing the loop An hardware interface card was also used for converting computer RS232 output level to TTL level acceptable by the PXI For practicality reasons the helicopter dynamics transfer functions has been developed in the LabView environment moreover the helicopter simulator and the real time code runs on the same machine This is possible because all the source code is organized using independently loops a GCS computer wh
107. liable vehicle simulation model The helicopter dynamics was modelled in nearly hover flight conditions by Pretolani 19 using a transfer function approach 63 66 Based on the helicopter dynamics transfer function identification classical PID controllers were designed in the Matlab Simulink enviroment neglecting cross coupling between the helicopter inputs Results founded by Pretolani are summarized in section 5 1 and were used as starting point to implement the control system on the onboard computer The control system architecture used to control the helicopter is based on a nested PID approach as shown in figure 86 The Vx and Vy track velocities control is implemented using the two level nested loop structure shown in Figure 86 Lateral track velocity Vy errors are used to generate roll demands for the roll control module while longitudinal track velocity Vx errors are used to generate pitch demands for the pitch 8 control module Integral contribution in the outer velocities loop compensate also external disturbances resulting from varying wind conditions The inner attitude controllers generate servo rotation commands for the helicopter to maintain the desired reference condition The reference track velocities Vx and Vy can be generated either by an outer guidance and navigation control system or by a user pre defined reference velocity profile The vertical velocity control uses a stand alone PI feedback control loop 111
108. loating point control signal processing analysis and point by point decision making Normal priority loop for embedded data logging remote panel interfaces and Ethernet serial communication Networked host PC for remote graphical user interface historical data logging and postprocessing 67 Host PC CompactRIO Real Time Controller Normal Priority Loop Communication Datalogging Network TCPAP UDP User Output Interface jee GuI Time Critical Loop Input FPGA Read Write Figure 44 CRIO Programming Structure 51 Depending on the application requirements it s possible to implement one or all of these application components The onboard software currently implemented on the flight computer follows this standard approach 4 4 SENSORS If an UAV is to fly autonomously or needs stability augmentation in remote controlled flight its flight control algorithms need information about its state which can be obtained by means of onboard sensors Depending on the vehicle type and its mission sensors can be different For the purpose of this work sensor types have been split into Attitude Heading and Reference System AHRS and altitude sensors 4 4 1 ATTITUDE HEADING AND REFERENCE SYSTEM AHRS Most common attitude sensors are based on gyros that can be either mechanical piezoelectric or optical A three axis gyro platform measures angular rates along all axes of the vehicle and is usually contain
109. master PC is needed and was therefore developed Moreover another block was developed for communication between LabView and the force feedback joystick Communication Simulink model GCS The LabView SIT V2 0 software was used to interface LabView with the MathWorks Inc Simulink environment The SIT V2 0 provides a way to generate a LabView user interface which can be used to interact with the Simulink model without converting the 38 Simulink model into a dynamic link library The component involved in the interaction between LabView and Simulink are shown in figure 15 LabView exchanges data with MATLAB Simulink using TCP IP For LabView to communicate successfully with Simulink it is necessary to have MATLAB running on the air vehicle computer When MATLAB is launched on the air vehicle PC a Simulation Interface Toolkit SIT Server starts which enables LabView and MATLAB to communicate with each other On the master computer the LabView front panel provides the user interface to the Simulink model By configuring the SIT Connection Manager dialog box it 1s possible to specify the relationship between LabView controls indicators and the Simulink model input output LabView controls are the GCS data sent to the simulator arranged into a 2D array LabView indicators are the GCS data received from simulator arranged into 1D array MATLAB AM AMLAN Ront Panel Adister Umnapwier Airreiricte Comprurter Figur
110. means of a small plastic box isolated with rubber 100 access point bber strips l Crossbow NAV420 a a a ee J Silent Blocks Figure 72 Avionics Vibration Isolation System A good criteria for chosing elastomeric dampers is that the critical frequency of the shock mounts must not be close to any produced by the rotorcraft at its normal operation point Figure 73 shows a typical diagram of resonant transmissibility versus damper frequency If the damper work frequency is higher than its resonant frequency than vibrations can be effectively attenuated 57 Transmissibility T 5 138 20 50 100 200 500 Frequency Hz Figure 73 Typical diagram of resonant transmissibility versus damper frequency With the selected shock mounts and an avionics box weight of 5 kg a natural frequency of 15 Hz can be expected This frequency is far enough from the closest frequency of the system the rotor induced oscillations at about 22 Hz to prevent any adverse effects Moreover the effect of the neoprene strips should help increasing the damping effect as was demonstrated experimentally see section 4 7 1 2 101 4 7 1 1 Vibration Load Experimental Test The vibration isolation system was tested by means of the experimental test bed illustrated in figure 74 During experiments the onboard computer and AHRS were replaced with equivalent metal part to prevent any damage due to unknown vibration effects Accelerometer wer
111. n the previous tables the final calibrated PI gains were higher with respect to the one calculated by simulations This was due to the fact that during simulations the gains were kept intentionally low for the helicopter to maintain a very low rate of climb descent Vertical velocity flight tests can be very dangerous since if the collective throttle curve is not good calibrated or the PI gains are to high the helicopter can crash to the ground without any hope to recover it Therefore the helicopter team decided to keep the gains small at the beginning and increase them once it was sure that the helicopter was flying safely The first test performed with the simulated gains showed that the helicopter was able to maintain hover conditions However the rate of climb descent was quite very low and the PI gains were therefore increased Finally after each controller was good tuned the full control system was tested over a squared flight pattern The distance tracked by the helicopter and therefore the Vx track velocity were kept within the RC transmitter range and pilot good line of sight in order to recover the helicopter if needed Some experimental results are shown in figures 101 104 During all the flight tests the helicopter was brought into hover condition by the RC pilot and then switched into autonomous mode 133 Figure 101 Simulate vs Experimental longitudinal controller tracking performance Heading deq x Imis
112. nd a number of secondary lobes of reduced amplitude separated by nulls The beam angle is usually 80 defined as the measurement of the total angle where the sound pressure level of the main beam has been reduced by 3 dB on both sides of the on axis peak However the transducer still has the sensitivity at greater angles both in the main beam and in the secondary lobes For a symmetrical conical pattern such as that shown in Figure 54 and typical of ultrasonic sensors a simple two dimensional plot known as beam pattern can describe the entire three dimensional pattern The beam patterns of transducers are reciprocal which means that the beam will be the same whether the transducer is used as a transmitter or as a receiver Figure 55 shows the beam pattern for the SRFO8 sonar sensor as a function of angle The beam angle is enough narrow approximately about 45 and 45 54 Of course the presence of secondary lobes may produce unwanted echoes and cause false measurements Therefore the sonar sensor was mounted under the avionics box in order to avoid undesired reflection from the helicopter airframe see fig 39 Other aspects associated with sonar mounting and operation were sonar measurements depend of course on the helicopter attitude for the purpose of this work this effect was neglected since for typical flight conditions the helicopter attitude is not very high sonar measurements can be affected from false echoes at the
113. ne ect ling hile eth sae abs besa ORD nla 100 AZT ald Vibration Coad Experimental Testees cnoi einna e EE A caviedbnasedsee 102 Ak PX Petite 1A R SUES aoe aE E A A 105 4 8 HARDWARE AND SENSORS DAQ FLIGHT TESTS 2 0 cecscccccceesseccceeaesscceeeeaeecceeeeenes 108 4 8 1 FLIGHT DATA RECORD VIRTUAL RE VIEW 00 ccccccccccccccccvcccccceeeeecccueececeueeesessueseeeuaeseeeeaeeeeseaeed 110 SSL ESIMUFA TION in A ada E 111 5 1 HELICOPTER DYNAMICS IDENTIFICATION AND SIMULINK PID DESIGN RESULTS 113 D2 CNB ARID CONTE is YOTE Morii tht tod U8 nce lan ae Sana a a 115 5 2 1 DOUSCRELE PIDIMPLEMENTA TION ciosijctcetepsseculass sis e ietca Senses auece E 115 9 2 2 ONBOARD NESTED FI SORT AI Tiare i oct iota ete a a E TNE 117 5 3 COMPLETE ONBOARD SOFTWARE IMPLEMENTATION eeeesesssssssssssrersessssssseserreeessssssesssrees 124 eE STMULA TION uria nr nmter ee tree errs tyre ernest errr nT errr 130 FLIGHT TESTS amp PI GAINS TUNING 200 ccccccccccecceceeceeceeecesceecesceeeeeeeees 132 WC OINGTUIS TOI AN OUTLOOK aronet eect 20 ee ce eile dare alee E 136 REFERENC ES sects cette deste Gnade AEA 137 Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 LIST OF FIGURES CAPE COIN STUCUTE ce iin dccnszdestast ccd i bets aciedaeek E E 17 NEA VS ble MM onc rect es et anycacanetseiea a tidcodccuubgnsinaensates 19 Basic Hehcopter Notation 20 serienn a a E 21 Blade Degrees of Freedom Schematic and Rotor Head Mechanization 20
114. ng am F T Eia an an au Hn 7 EE a Figure 64 Optical Encoder Principle 56 9 By counting the number of the encoder output pulses using a DAQ card it is possible to know the rotation angle corresponding to a PWM actuator input 56 At this aim the phase A encoder cable was connected to a counter input channel of a NI PCMCIA shown in figure 63 while a PWM actuator signal was generated using the PCMCIA counter output The software developed for this experiment is reported in figure 65 Encoder Reading Loop Ris ing 7 Arrr EA a Conver sion Number depending on Externally Controlled v Encoder type qJ g L e DAQ mse DAQ mw sl Z iher BE en CI Cnt Edges v o 7 ae Duty Cycle PWM Generation Loop DBU oed 1000 OK message warnings Y a ontinuous Sa les v Be POCO OOO iN omen ahah FAIS OOOO Oommen sa i LEA DA Orase 7 R GE eee EE aaa DAO re Channel E es it CO ia DAO mse s el fa Se a Bee ORE OA Ai Ea el P de State CO Pulse Freq Irrplicit p_CO Pulse Freq urrent Duty Cycle Poet if x lt 5 y 0 0192 x 0 019 if x gt 5 y 0 0192 x CLOSE 0 211 a i ee a time varying Instance 31 9Clipboard vi ei uty dde JA KY Graphi Deg ees Encoder wart Pipa on N T fo create or replace Figure 65 LabView Software for Encoder Signal Acquisition and PWM generation through DAQ Card XY Graph Servo Dev1 ct
115. nlikely to produce an optimally tuned controller they do provide a good starting point for further optimisation Therefore the PID gains were tuned in the Matlab Simulink environment using the transfer function reported above Even if decoupled PID loops were considered this was enough to control the helicopter since such loops should view the coupling between axes merely as a disturbance and should be able to compensate this effect in a robust manner Simulation results showed also that a PI controller derivative term set to zero will provide sufficient control capabilities 19 Therefore a simple PI was implemented on the onboard computer The calculated PI gains are reported below together with the final gain value calibrated experimentally and currently used for the onboard control system 114 Attitude PI Gains K0 Kp servo s servo s Calculated 0 77366 1 0418 0 11346 Table 30 Attitude Controllers PI Gains Kc servo 9 Velocity PI Gains KV KV KcV KV KcVz KVz AAm s 0 m p m s m coll m s coll m l 0 Table 41 Velocity controllers PI Gains 5 2 ONBOARD CONTROL SYSTEM For the implementation of the onboard velocity attitude control system a lot of things must be taken into account the control loop is not in a continuous time domain but cycles at a discrete time interval Therefore a discretise controller must be implemented the FPG
116. nment requirement View of object from Ground Control Station Red square is alignment requirement i 5 ag a ok erarnan DNA ay ra i e ET EEEREN AE EE ie K pre ered Ta v E Rte 7 A p k q ia a p r s Er mc ane Lites RAE as a g FP palette e bii a i a aa Eog i Eo a PENERE ee M a a hinah a T 5 y P oe i iy r E A i aT p 1 ra x Tiea J i ia F i Mi ai aik nth i ra e U AAE ER pe joo AE Ghar Pe dad Ea ern ee Recta ayaa ie ee Fo eae Ta P ae Figure 33 Phase 3 Acceleration Mode 42 Figure 34 Phase 4 Hover Hold Mode 42 2 3 2 2 Piloted Simulations The piloted simulations were organized with the intent of using the Cooper Harper handling qualities rating scale 50 to measure the impact of the joystick active features Therefore the task was designed to be flown with or without the active joystick features to offer a direct comparison It was decided to invite at least three pilots each with UAV piloting experience to do the evaluations The final results were then averaged to show trends The trends and pilot comments showed the active features made a marked improvement in task performance and situational awareness Following the Cooper Harper rating procedures so called desired performance and adequate performance limits were defined These limits were necessary to inform the pilots how much aggressiveness was required and which flight limits to be respected Table 2 shows the limits for
117. ns are based on free DEM Digital Elevation Map data available at the USGS 45 US Geological Survey catalogue internet homepage 47 The DEM data are SRTM Shuttle Radar Topography Mission data type with 3 arc sec resolution that 1s a pixel of 90m x 90m The software incorporates also Gaussian data filtering routines to provide effective noise filtering of the SRTM data Terrain rendering is done using view dependant polygons rendering algorithms This kind of algorithm is able to create quad tree hierarchy structures of the terrain polygonal mesh such providing quick and efficient terrain rendering even if the terrain mesh is constituted by several millions of polygons Nevertheless since the algorithm is view dependant the number of polygons actually rendered every iteration is reduced to about four ten thousands 4000 10000 One or more textures such as satellite or aerial pictures can be also applied to the terrain in order to produce a more realistic virtual environment 2 3 SIMULATION ENVIRONMENT APPLICATIONS In this section the results of two interesting research activities performed using the developed mission simulation environment will be described The first one concerns the evaluation of the rotary wing UAV configuration developed by AGUSTA inside the CAPECON program the second is an investigation on how the active features could be used for successful RUAV mission task accomplishment The last research activities
118. nsor measurement reliability This step plays a crucial role in a RUAV development because if the helicopter has to fly 19 autonomously reliable information about its states 1s needed by the onboard control and navigation system parallel to the hardware set up simulation plays also an important role in the development of an autonomous helicopter A simulation model was developed based on helicopter dynamics identification flight tests to be used for the design of the onboard control and navigation algorithm once the previous task were completed the onboard hardware and software were integrated into the simulation loop using a Hardware In the Loop HIL simulator In this scenario performance and possible errors of the onboard software can be detected during intensive ground safe and risk free tests in the end autopilot flight tests were performed for final verification and tuning of the control and navigation system This thesis will focus mainly on the RUAV avionics package set up and on the onboard software development while the other steps will be covered in reference 19 One important aspect to be taken into account in the development of a RUAV system is that it is actually an aerial robot The set up of a capable task worthy aerial robot is essentially an integration effort and always requires knowledge of several different disciplines and experimentation on new system development In the past years most of the re
119. nt SRFO8 gain register address get acknowledgment set gain to appropriate level determined experimentally This value must be set to Hex 10 if we want the sonar to range till 6 m get acknowledgment send stop sequence send start sequence SRFO8 I2C address W bit get acknowledgment SRFO8 command register address get acknowledgment command to start ranging in cm get acknowledgment send stop sequence Now after waiting 65mS for the ranging to complete the following commands are sent i2c_start i2c_tx OxE0 12 get ack i2c_tx 0x02 12 get ack i2c_start i2c_ tx OxE1 12 get ack send start sequence SRFO8 I2C address W bit get acknowledgment SRFO8 range register address before reading from the sonar it 1s necessary to tell the sonar which of its internal addresses we want to read get acknowledgment send again start sequence SRFO8 I2C address R bit get acknowledgment 86 12 rx high read the most significant byte sonar data are in the U16 format and therefore requires two byte 12_give ack give acknowledgment 12 rx low read the least significant byte sonar data are in the U16 format and therefore requires two byte 12_give ack give acknowledgment i2c_stop send stop sequence data rebuilt builds altitude information in a 16 element Boolean array in a similar way as is done for the NAV 420 I16 da
120. nt 4 7 HARDWARE INTERFACING WIRING AND MOUNTING The RUAV hardware was assembled together placing attention to accessibility flexibility and modularity A commercial off the shelf plastic box was selected to house all avionics components which was installed under the modified landing gear The plastic box cover 97 was suspended to the landing gear by means of rubber shock mounts for vibration isolation see fig 72 The very light weight plastic box can be attached to the cover and easily removed when maintenance or other works needs to be done on the avionics components Moreover this mounting system allows the structure of the plastic box to achieve its rigidity which is of course necessary to perform good flight tests The tail boom provides also installation points for the GPS antenna enabling firm fit and leaving the boom structure unchanged The sonar sensor was appended under and outside the avionics box see section 4 7 1 for vibration isolation A schematic wiring diagram of the vehicle mounted avionics is depicted in figure 71 NiMH SY Ratt atlery z Charge NIMH SV a Battery Charge P 15 PIN HAY 420 Ex Connector LA Sonar By 3 a leo Slot 1 Slot 2 Slot3 Stora ens Di Di DO D 7 me CRIO e Battery SDK SDA SDKR OA Charge SDAR 3 B Sonar Card ax 10V Access i ol z Point L L O e i t n LiPo 7 4 Li Po 11 4 V ie Ls Charge
121. ntrol station a post view station was develop in order to help data record analysis by reproducing the flight tests in a virtual scenery The post view station runs a LabView software which was derived from the one developed for the CAPECON mission simulation environment and is reported in the enclosed CD The program reads the data record file and displays flight information on a virtual cockpit reproducing also a pilot virtual radio stick and on helicopter states diagrams meanwhile data are sent via TCP IP to the visual system which is able to reproduce in real time an external view of the air vehicle The station architecture is based on two computer see fig 85 one is used for the visual system while the other one for the LabView code and user interface Such a system was very useful during post processing analysis since it s possible to have real time and immediate memory of the helicopter behaviour during experimental tests thereby facilitating flight data record interpretation sel Eer son E Te Bz X J f ao Br a s titi Visual Computer Data Record Manager lt Optional A Sa Figure 85 Post View Station Architecture 110 Chapter 5 SITL SIMULATION After the hardware was set up a series of flight tests were done in order to collect experimental data for identifying the helicopter dynamics characteristics and develop a re
122. nts NAV 420 Operating Vibration Range lt 6 g rms 20Hz 2kHz CRIO Operating Vibration Range lt 5 g rms 10Hz 500Hz Experimental results confirmed that the elastomeric dampers efficiently attenuate vibration on the onboard avionics The high vibration load experienced by the landing gear at the engine frequency of 200 Hz may seem somehow surprising above all because very poor literature or better no literature at all exists about that Therefore experimental tests were repeated several times even with a different acquisition system and were compared with the results provided by Boeing for a similar helicopter This data were available at low engine rpm but they seem to confirm the order of magnitude of the measured data 105 Test Eli09 Vibration level along the z axis ae Landing Gear Figure 77 Accelerometers mounting points As example acceleration experienced on the landing gear position P in figure 77 and on the emulated NAV 420 position E figure 77 will be reported Landing Gear 220 Hz Landing Gear 20 2 _ 104 ao D D QO g T A T I 0 5 1 15 20 Time s Frequency kHz Figure 78 Acceleration experienced on landing gear From the PSD 60 61 diagram it can be seen that the main vibration load is caused by the engine spinning at about 200 Hz This is also confirmed by the fact that filtering the signal in the band 0 1kHz the gms value calculated with equation 4 5
123. omponents The open system approach may allow quick tests of multiple design solutions reducing design risk and time with the expectation of life cycle cost reduction 3 to create a tool that improves the RUAV performance estimation and evaluates the RUAV stability and controllability qualities by testing the full system into a realistic operational scenario Details of the mission simulation environment software implementation will be given in the next sections while the most important achieved results will be reported in section 2 3 2 1 THE AV amp NGCS SIMULINK MODEL In order to simulate the behaviour of an autonomous RUAYV it is necessary to model the air vehicle dynamics and the navigation guidance and control system NGCS k The complete Simulink model developed by UNIBO is shown in figure 6 21 CA_Nt_Mission ae Fie Edit View Simulation Forme OTT Helicopter 4 Navigation Dynamics m aage SAS amp E Figure 6 AV amp NGCS Simulink model The model is composed by several different parts two communication blocks for exchanging data with the GCS computer the helicopter dynamics block which is able to simulate the flight dynamics of a classic main amp tail rotor helicopter The model is a non linear rigid blade dynamic model using main rotor first order flapping dynamics steady amp uniform Inflow and combined momentum amp blade element th
124. pace Engineering Department Laboratories of the University of Bologna concerning the development of a small scale Rotary wing UAVs RUA Vs In the first part of the work a mission simulation environment for rotary wing UAVs was developed as main outcome of the University of Bologna partnership in the CAPECON program an EU funded research program aimed at studying the UAVs civil applications and economic effectiveness of the potential configuration solutions The results achieved in cooperation with DLR German Aerospace Centre and with an helicopter industrial partners will be described In the second part of the work the set up of a real small scale rotary wing platform was performed The work was carried out following a series of subsequent logical steps from hardware selection and set up to final autonomous flight tests This thesis will focus mainly on the RUAV avionics package set up on the onboard software development and final experimental tests The setup of the electronic package allowed recording of helicopter responses to pilot commands and provided deep insight into the small scale rotorcraft dynamics facilitating the development of helicopter models and control systems in a Hardware In the Loop HIL simulator A neested PI velocity controller was implemented on the onboard computer and autonomous flight tests were performed Comparison between HIL simulation and experimental results showed good agreement This PhD work wa
125. per Rating Evaluations seisarems a a aiea 56 ROTARY WING UAV SYSTEM DEVELOPMENT 0 cc cccccccseeeeeeeees 59 I DESIGNPROCE S Sapia E seen sacueans osceemies 60 HARDWARE SELECTION AND INTEGRATION 0000000 0c cccc ccc eesseeceeceeeeeees 63 4 1 HARDWARE DESIGN REQUIREMENT S c ceicscdensdnche eddie daliebsvecaldetacndeseuhsdactvidencsh ial Suances 63 42 FLIGHT TEST VEHICLE DESCRIPTION contone AA E 64 AS rAGHECOMPUTE R eoi E 65 TABLE OF CONTENTS Continued 45 4 CRIO REALD TIME APPLICATION TIE SIGN opie eis sii E aancasaeieuae a a 67 LASEN SOURS esis 05 Sees shee ccs pu ees sis sa wishes sae ceca aes nel ph asada sees Rec ees 68 4 4 1 ATTITUDE HEADING AND REFERENCE SYSTEM AHRS 0 ccccccccccccccccc cece cece eee eeeeeeneeeseeeeeeeeeeees 68 AEP EARS OE UD ea a cae eee bales E cash tiara 70 AA 2 Pele UID ESN SOIR wazesaicecoicannecoi E O E A naive 79 AARDE e Oa E E AA E AA TE 80 4 4 2 2 Sonar sensors data acquisition e esesesseseseeseresreseseeeststtststestststeststisistetistetesrstisesrenisresesrsresent 82 AS A EATR Seia ern a a R 87 4 6 1 PULSE WIDTH MODULATION SERVO ANGLE CURVE eeeerreerrreerrrerrrereesrere 90 4 5 2 ACTUATORS SIGNAL ACQUISITION AND GENERATION SOFTWARE ccccccccccccessvesseeseees 94 Fh OU ATE TIN I ccs da assess gtetins deserve A E RE Resta 97 4 7 HARDWARE INTERFACING WIRING AND MOUNTING oo eccccccseccceeeeceeeeseeseeeees 97 AT de VIBRATNONISOLATION sreci cists ued ait aes tet t
126. ptember 2002 Cyrille Sevin Rotary Wing Civil UAVs Concept Of Employment Rochester CAPECON Technical Meeting July 2002 R Pretolani G Saggiani B Teodorani Mini Small Rotary Wing UAV Technologies CAPECON Technical Report Report ID 10 3 November 2005 L A Young E W Aiken NASA Ames Research Center et al New Concepts and Perspectives on Micro Rotorcraft and Small Autonomous Rotary Wing Vehicles S Mouritsen E Rehwald DLR German Aerospace Research Centre Micro UAV Payloads and Data Link Survey Final Technical Report CAPECON Report ID 10 6 10 7 November 2005 R E Weibel R J Hansman MIT Safety Considerations for Operation of Different Classes of UAVs in the NAS ATAA 2004 6421 AIAA s 4th Aviation Technology Integration and Operations ATIO Forum 20 22 September 2004 Chicago Illinois 137 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 B L Aponso E N Bachelder D L Lee System Technology Inc Automated Autorotation for Unmanned Rotorcraft Recovery AHS International Specialist s MeetingOn Unmanned Rotorcraft 18 20 January 2005 Chandler Arizona R E Weibel R J Hansman MIT An Integrated Approach to Evaluating Risk Mitigation Measures for UAV Operational Concepts in the NAS AIAA 2005 6957 AIAA s 5th Infoech Aerospace Conference 26 29 September 2005 Arlington Virginia
127. r 2001 the European Union has sponsored the UAV development program CAPECON to attempt to kick start a civil UAV industry in Europe and try to fill the gap with the United States Since 1999 the University of Bologna UNIBO has carried out several research projects concerning the development and manufacturing of fixed wing UAV systems for the civil aviation market For that reason when the EU decided to start the CAPECON program UNIBO didn t hesitate to take part in CAPECON Civil UAV Application end Potential CONfiguration solution was the first European wide program tying together the resources of eight countries nine industrial organization five aerospace centres and six universities Its main goal was to provide European industry with detailed design and manufacture know how on safe cost effective and commercially viable civil UAVs The program was structured to make a logical progression 16 from customer needs to final products 2 The process see figure 1 started with UAV applications analisys braking them down into discrete missions which were then lumped into multirole missions 3 4 CAPECON Structure 1 Application Survey 2 Operational Concepts 3 Requirements 4 Technology and Configuration Development 5 Cost Analysis 6 Dissemination Program End Figure 1 CAPECON Structure The most promising and commercially viable multirole missions were selected and further translated into operational concepts which
128. rO gt Frequency Hz 750 00 Idle State F 7 Low Encoder k Devijctri Duty Cycle 0 11500 1 0 1 Duty cycle Figure 66 Front Panel of the software reported in figure 65 92 It is constituted by two parallel while loops the first one is used for PWM signal generation and data logging for the purpose of this work a time varying PWM pulse width was generated by changing the PWM duty cycle with time The PWM frequency was left fixed and equal to 50 Hz the second one is used to count the encoder pulses and convert it into an equivalent angle measurement At this aim it must be taken into account that the encoder available at the Hangar Laboratories has different decoding possibilities depending on its usage If only phase A cable is used the encoder is not able to discriminate the sense of rotation which however is not a stringent requirements for this work and work in X2 mode which means X2 resolution multiplication Therefore in order to calculate the angular servo displacement the following formula must be used Amount of rotation Counts 360 2N 4 2 where N is the pulse revolution For the selected encoder N is equal to 900 pulse revolution hence the scale factor is 0 2 Counts which correspond also to the encoder angular resolution The acquired data were processed using Matlab curve fitting tools which yielded to the following PWM width Servo Angle curve Servo Ang
129. refore efforts were concentrated to ease hardware configuration and reconfiguration and allow for future system growth From this point of view important requirements are 63 to provide accurate flight data acquisition for dynamic model development and validation to be versatile enough to enable fast and easy integration of different input output sensors and to allow future system growth in term of payload sensors and interfaces to be as light as possible in order to lower the total platform weight and maintain good helicopter maneuver capabilities Preliminary flight test demonstrated that the helicopter still had good maneuverability with 6 kg payload mass to be able to withstand the high vibration load typical of small scale helicopters The primary sources of vibrations are the engine the main rotor spinning at roughly 22 Hz the tail rotor and the tailboom bending resonance These vibrations must be reduced to fit the operational vibration range of the onboard sensors and to provide accurate flight data measurements Experimental tests performed with commercially manufactured elastomeric dampers showed that vibrations can be effectively reduced to the desired level to be protected against the electromagnetic and RF interference common shielding precautions were used to isolate the onboard electronics from EM interference to allow onboard implementation of feedback control laws and demonstrate good control capability to be endowed
130. reported in figure 89 Input Parameters Input Trim Lat us Parameters vy SP v116 p SP al6 Scale oni Pie Pa a al16 Vy v16 calculated from AHRS data Figure 89 Schematic of the Vy phi nested PI The first PI implement the lateral velocity control along the trajectory This PI calculate the reference phi attitude to be passed to the roll attitude controller The attitude PI calculates the servo rotation angle variation to be added to the trim value in order to maintain the desired set point This commands are then scaled in microseconds of PWM high time which is used to generate the PWM signal for the servo actuator see PWM generation algorithm section 4 5 2 Unity of measure must be scaled as shown in figure taking into account the scale factor defined in table 12 Tables 17 20 summarize the input and output parameters needed for the algorithm to work properly wr So Cae Vy 6 Output High a116 3640 Output Low al16 3640 Initial phi al16 27 74565 PI Reset TRUE first PI call otherwise FALSE performed automatically by the program Table 107 Vy PI Input Parameters 120 Vx PI Outputs Proportional action phi al16 2 Total phi action al16 Table 118 Vx PI Outputs Current phi al16 from AHRS data PI Reset TRUE first PI call otherwise FALSE aa performed automatically by the program Table19 phi PI Input Parameters phi PI Outputs Proportional action Lateral cyclic with
131. rol directly the servo motor angle this is not true for the PWM tail signal which actually is used as input for the gyro system When the radio is switched on for the first time the gyro control unit reads the tail PWM signal coming from the radio which correspond to zero yaw rate since the helicopter stands at the ground and the gyro sensor measures zero yaw rate and is initialized The initial PWM tail command or the CRIO PWM tail signal is therefore perceived by the gyro control unit as reference yaw rate to be maintained by the helicopter If the PWM command is equal to the initialization value the reference yaw rate to be maintained is zero otherwise it is perceived as a reference of constant yaw rate whose value depends on the commanded PWM see fig 61 Hely Gyro CU PWM Input PWM tail r measured Hely Gyro Sensor PWM Initialization value Commanded PWM 4 g r 0 heading init r cost Figure 61 Tail amp Helicopter Gyro System Interaction If during the flight the helicopter experiences a perturbation in yaw not due to a pilot or computer command the gyro feels a change in the helicopter yaw rate response which is communicated to the gyro control unit In turn the gyro control unit will send a PWM signal to the tail actuator till the helicopter yaw rate 1s driven to zero moreover since an Heading Lock AVCS is implemented on the gyro control unit the helicopter returns also to the initial heading while the
132. round safe and risk free tests see chapter 5 6 5 In Flight Autopilot Test autopilot flight test must be performed for final verification and tuning of the control and navigation system see chapter 7 At this point improvement within the previous steps can and should be undertaken until a configuration is reached that promises satisfactory results for the final RUAV system set up 62 Chapter 4 HARDWARE SELECTION AND INTEGRATION The synthesis of the RUAV hardware is a trade off evaluation like any other design process An optimal design solution is sought by finding the best compromise to satisfy the design requirements The main requirements driving the hardware selection and integration process are outlined in section 4 1 while sections 4 2 to 4 6 describe the selected hardware and the overall hardware system set up 4 1 HARDWARE DESIGN REQUIREMENTS The main requirements taken into account for the RUAV system design were both operational requirements and physical constraints Numerous requirements were placed on the avionics system design while the air vehicle configuration was somehow freezed to the one already available at he UNIBO laboratories which was modified only to increase performance and payload carrying capabilities The most important design criteria followed for the RUAV testbed development was maximum flexibility i e easy and quick reconfiguration while maintaining good air vehicle performance The
133. s carried out in full cooperation with Roberto Pretolani who was mainly responsible for the helicopter dynamic identification and the control system design His PhD thesis 19 will report this work in details 12 OOe Se t42Ug lt 5 z Far 4 Scoll Stat ine 3h dE A tevel flight NOMENCLATURE helicopter longitudinal speed in body axes helicopter lateral speed in body axes helicopter vertical speed in body axes roll rate in body axes pitch rate in body axes yaw rate in body axes Euler angle for helicopter roll Euler angle for helicopter pitch Euler angle for helicopter heading blade azimuth angle blade pitch angle the main rotor torque tail rotor thrust fuselage aerodynamic forces tail surface aerodynamic forces rotor speed blade flapping angle average blade pitch angle lateral cyclic blade pitch longitudinal cyclic blade pitch lateral stick to cyclic pitch gearings effective lateral control derivatives taking into account the effect of the stabilizer bar longitudinal stick to cyclic pitch gearings gearings effective longitudinal control derivatives taking into account the effect of the stabilizer bar collective control input cyclic lateral control input cyclic longitudinal control input rate of climb vs collective delta pitch attitude in level flight relative to flight path A extended approach delta pitch attitude in the energy extended approach relative to flight path dc ma
134. s running in this mode Vy and Vz are automatically set to zero The second interface available is much simpler and was used for flight data acquisition tests It is composed by two windows the first one reproduces a virtual cockpit together with the radio stick movement and a GPS 2D flight path the second one displays the helicopter states and the commanded inputs see figure 97 Figure 97 Flight Data Acquisition GUI 129 Chapter 6 HIL SIMULATION To allow safe risk free onboard software testing the PI controller was first implemented in a HIL simulator shown in figure 98 PWM High Time us gt es 3 us innan i PXI 7831 CRIO FPGA Equivalent HW amp SW ET Sensor data Acquisition Control Algorithm PWM Commands Generation amp Acquisition T AS PWM High Time us to Servo Degrees FLYING PLATFORM NAV 420 NAV 420 MODEL amp RTSW String Emulator HELICOPTER gt From To Ground Station z one om oon SIMULATION COMPUTER Host Computer em ZRI Visual Computer Optional 2 AJ Ay Figure 98 Schematic of HIL Simulator The HIL test bench includes as much of the flight hardware in the testing loop as possible and is constituted by an exact duplicate of the flight computer the CRIO System and of the onboard software including the PI controllers At this aim a National Instrument PXI 7831 was used which is equivalent to the CRIO FPGA modu
135. search efforts in miniature autonomous helicopter were lost for hardware integration and for obtaining reliable sensor measurements For that reasons taking also into account the outcomes of the CAPECON program it was decided to evaluate the feasibility of using COTS sensors and electronics for the RUAV avionics package Both the hardware and the software were integrated placing attention to modularity growth potential versatility and possibility for ease reconfiguration and software implementation Results achieved in this work showed that the selected hardware and the onboard software were able to provide accurate flight data measurements and good helicopter control capabilities Thanks to its modular architecture and accurate flight data measurement capabilities the developed RUAV system may become a useful research test bench in several different field such as aircraft rotorcraft dynamic model identification researches in control and navigation laws fast and ease software implementation could results also in a speed up of the research time 20 researches in man machine interface and air system integration which is addressed as one of the most critical technology aspect for the future development of the civil UAVs and their integration into the airspace 14 16 1 2 SUMMARY OF HELICOPTER PRINCIPLES As well known helicopters are air vehicles which are able to fly thanks to the lift force produced by lifting surfaces the rotor b
136. slot 3 having 8 digital output channels another CRIO 9474 mounted in slot 4 having 8 digital output channels Each CRIO module contains already build in signal conditioning FPGA devices are very useful and powerful since they combine the versatility of a reconfigurable digital architecture with a matrix of configurable logic blocks surrounded by a periphery of I O channels This way signal can be routed within the FPGA matrix in any arbitrary manner by programmable interconnected switches and wire routes figure 43 66 TO Pe On DO C CO CG nj ngngngninin a SHeoGoS 7 OOOO H pe 0 a PROGRAMMABLE OOoCCoO o OOO000 a OUoocoo a Om Oooo c Soure Xilinx YETTA CONFIGURABLE LOGIC BLOCK CLB Figure 43 CRIO Field Programmable Gate Array FPGA Structure 51 Control loops can be also implemented inside the FPGA environment using while loops up to 40 MHz 25 ns Moreover FPGA modules are ease programmable with NI LabView without need to know specialized hardware design languages such as VHDL the LabView code is directly compiled in VHDL before being downloaded on the FPGA devices 4 3 1 CRIO REAL TIME APPLICATION DESIGN The real time control and acquisition system which is possible to develop with the CRIO system typically contains four main components see figure 44 RIO FPGA core application for input output inter thread communication and control Time critical loop for f
137. ssccssocsoctscntsenssccocetoceseesssssosssnsssesssesseuss 42 Pierur 20 GC SC Om Ot AG OM eee EEEE R E A E AERA 43 Fig re 21 GCS UNIBO Visyal external VIEW sicie arna R a AA Ei 44 Figure 22 GCS UNIBO Visual pilot EO View cccccscecsessesecossessosseesseessevevsvecesesvervsebsssasasoess 45 Fig re 23 GCS UNIBO Visual terrain MGS Is acceocsstsehecsaaecieacbesccenseacastsduceneaiendadenicbeusascedecseeens 45 Fiere 24 IWVSSION SCOMAT O tac sc ctcascrsstiead E cn nietetetn earns 47 Figure 25 Fire Surveillance Mission flight path c cccccscssseseeeeseeeeeeeeeeeeeeeeeeeeeeeeeeeeees 48 Figure 26 Fire Surveillance Mission actual flight path ccccccccssesessesseeeeeeeeeeeeeeeeeeees 48 Figure 27 Fire Surveillance Mission vertical profile cccccccccccccccccccceeceeeeeeesseeesesessssssseaeaes 49 Figure 28 Fire Surveillance Mission ground Speed ccccccceceessseseseeseeeceeeceeceeceeceeeeeeeeeees 49 LIST OF FIGURES Continued Figure 29 Fire Surveillance Mission power required ccceeseesesssssecceceeeeeeeeeeeeeeeseeeeeeens 50 Figure 30 Search Mission fuel CONSUMPTION cceeceeseeseeseeseeseneneceeeeeeeeeeeeeeeeeeeeeeeeeeeeneees 50 Figure 31 Phase 1 Autonomous Waypoint Flight 42 oe cceceeeeeeeeeseeseeseeeeeens 50 Figure 32 Phase 2 Manual Mode 42 K uuussrrrrrrrrrrrrrrrrrrrerrerrereeeereerereeseseen gt 52 Figure 33 Phase 3 Acceleration Mode 42 sccccesc
138. ssios steed E OERE E 71 Fize 49 NA V420C A Noun iin cien ea re EE E E O O 73 Figure 50 NAV420CA Acquisition Software Flow Chart cccccssssscsccccceeeeeeeeeeeeseeeeeeeees 75 Figure 51 NAV420CA data Packet length time s ecesescenscccccceceeeeeesereeesssseseeseeess 77 Figure 52 NAV420CA Packet Acquisition SeCQuence c ccseeseeesesseseeeseeeeeseeeseeeeeeeeeeeees 78 Figures 537 Sonar Sensor SIRE OS ais na ees pal Ose eee tenia 80 Figure 54 Example of Three Dimensional Representation of the Sonar Beam Pattern 80 Figure 35 SRFOS beam pattern 54 inea E sant onctactenatonetadantianiys 81 FPioure 30 Sonar ACQUISITION CIT CUE ainra A A EET AEE OOE O 82 Fiure 57 PC Starland Stop Sequenee 353 kea ATN 83 Froure6 ZC Dit raster 3S oa E ETO 84 Figure 59 FPGA Sonar Data Acquisition Loop ccccccsseesessseseseeeeeeseeeeceeceeeeeeeeeeceeeeeeeeeees 85 LIST OF FIGURES Continued Fig re 60 Servo Actiators Control Cielle a eeraa ETET AnA 88 Figure 61 Tail amp Helicopter Gyro System Interaction cccccccccccccseceeceeceeeessssesessesssseaes 89 Figure 62 PWM pulse width and servo angle rotation cccccsssscssccceceeceeeeeeeeeeeseeeeeeeeeees 90 Figure 63 Experimental Set up for Servo Angle PWM curve determination 00000 91 Figure 64 Optical Encoder Principle 56 cccccccsesesesssssseesceecesecsccececenceneeeeeseseeeesenseeess 9 Figure 65 LabV
139. t mission simulation environment future activities Mission Simulation Environment and HIL Simulator These simulation platform will be further improved More sophisticated dynamics models will be implemented on the HIL simulator including a more accurate model of al the flight sensors Coupled with the developed RUAV platform these simulation environments will provide useful test beds for safe ground pre flight tests or for studying different control and navigation strategy Researches in man machine interface and air system integration could be also be performed which were addressed as one of the most critical technology aspect for the future development of the civil UAVs and their integration into the airspace RUAV Platform The onboard navigation system software will be also tested in flight and further flights at higher speed will be made Automatic take off and landing flight mode will be also implemented and further flight tests will be performed Thanks to its modular architecture and accurate flight data measurement capabilities the RUAYV system may become a useful research test bench in several different field such as aircraft rotorcraft dynamic model identification researches in control and navigation laws fast and ease software implementation could results also in a speed up of the research time support in main machine interface research activities The feasibility to install the designed avionics hardware integrated with
140. t the 75 NAV 420 string is read starting from the first byte of the packet The wait time value can be estimated from the NAV 420 packet rate and from the baud rate During the initialization procedure the time measured in FPGA clock ticks corresponding to one bit and half bit of information is also calculated These two times will be used by the program to correctly read the bit sequence which composes each byte of information in the NAV 420 packet Wait time and FPGA tick counts calculation Since the NAV 420 output rate is 100 Hz and the baud rate is 57600 bps then microseconds corresponding to I bit of information 17 36 us time needed for I packet transmission 10 ms Taking into account that the FPGA works at 40 MHz which is 40 000 000 tick s or 1 tick corresponds to 0 025 us then it is also Time in tick corresponding to I bit of information 17 36 us 0 025 Us tick 694 tick Time in tick corresponding to 1 2 bit of information 347 tick Since the packet in NAV mode is constituted by 37 bytes of information and each byte in the RS232 protocol is constituted by 10 bit 1 start bit 8 information bit 1 stop bit not all the 10 ms contains data information there will be a certain time during which the electrical signal will remain low that can be used as wait time to identify the NAV packet first byte see figure 51 Hence Bit per packet 37 byte 10 bit byte 370 bit Actual packet length in ms
141. t to the active joystick two graphic blocks have been created for generating real time plots of various flight parameters animated map display flight plan window and virtual cockpit a Force Feedback joystick manager which manages input output communication with the USB FF joystick a send loop to the visual system for displaying data on a 3D graphical interface which uses an UDP communication protocol a send loop to the air vehicle computer for sending real time control signal In remote piloted flight mode the control signals comes from the joystick interface while in autonomous flight mode the control navigation signals depend on the flight plan defined at the flight plan graphical interface 2 1 1 COMMUNICATION MANAGER AND SOFTWARE INTERFACING Communication between the Simulink model and the GCS primary master computer was done using the LabView simulation interface toolkit 2 SIT V2 0 A vector of data is sent from to the air vehicle computer to from the GCS via a TCP IP communication protocol Data received at the GCS primary master computer are then distributed to the graphical interface to the active joystick and to the visual system meanwhile control command data must be sent from the GCS to the air vehicle computer for the RUAV control and navigation Since the visual system is developed in C and requires an UDP communication protocol another communication interface between the visual PC and the
142. ta 4 5 ACTUATORS Servo actuators allow accurate helicopter commands thanks to the on board circuit They are controlled by a Pulse Width Modulated PWM signal where the desired servo motor angle is usually proportional to the pulse width see further in Section 4 5 1 Five servo actuators are currently mounted on the helicopter which must be powered at SV 9202 for throttle control two S9405 for lateral and longitudinal cyclics controls 9255 for collective control 9252 for tail control which can provide a 6 kg cm torque The servo actuators control circuit 1s show in figure 60 During the helicopter flight servo actuators are controlled either by the RC pilot via radio or by the onboard computer when the RUAV flies autonomously The core of the onboard actuator circuit are 5 helymodel switches which are used to change the helicopter flight mode They have 3 input cables and one output cable which brings signal commands power and ground to the servo actuators 87 NiMH 5Y oe aa 9 Ground Gyro Power Ground Gyro gain settings via radio 1 LAT 3 THR G To Tail Actuator Tail Hely model 1 Receiver f 4 TAIL 6 COLL 7 PID Switch via Radio 7 PID Switch Power Ground To Actuators CRIO Commands autonomous mode CRIO DO Bring PWM signal to actuators This are 3 wire cables including command signal actuator power ground Figure
143. terfaces and system integration meanwhile it should be proposed as a technological prototype for industries interested in UAV development and manufacturing In order to develop such kind of platform avionic systems are required that enable the helicopter to maintain a stable attitude and follow desired trajectories This avionics package is comprised of sensors computer and data link hardware as well as software to guide navigate and control the air vehicle These aspects are particularly critical for helicopters which are well known to be inherently unstable systems and place numerous requirements on the avionic system design The overall RUAV system architecture developed at UNIBO is show in figure 39 It has the typical UAV system architecture as defined in the CAPECON program 49 but simplified for a small RUAV The modified Hirobo 60 helicopter mechanics was used as flying platform The RUAV avionics is constituted by an onboard computer the CRIO system from National Instruments which acquires sensor data from an Attitude Heading and Reference System AHRS and the sonar altimeter and sends PWM commands to the helicopter servo actuator based on the control and navigation laws implement on it The data link between the onboard computer and the ground control station is performed by means of a simple WIFI access point Details of the RUAV hardware set up will be given in chapter 4 59 RUAV Systems Helicopter Hirobo Eagle 60 modi
144. the transfer by sending a stop sequence 55 spa _ D7 D6 DS D4 D3 D2 D1 DOIACK sc _JiLJ2 LJsLi4lislie Li Jel f9 Figure 58 2C bit transfer 55 In order to manage the I2C protocol appropriate command sequences were defined START Sequence for the CRIO to start communication with the sonar STOP Sequence for the CRIO to stop communication with the sonar TX Sequence for the CRIO to transmit information to the sonar Get ACK Sequence for the CRIO to get acknowledgement from sonar that means the sonar has received the data For the Get ACK Sequence to work properly it must contains also a Clock stretching wait routines see software details The clock stretching is necessary to be sure that the sonar has actually received the CRIO commands and is ready to send the data RX Sequence for the CRIO to read information from the sonar Give ACK Sequence for the CRIO to acknowledge the sonar that means the CRIO has received data from sonar Moreover the SRFO08 sonar has a predefined device address using 7 bits 1 R W bit Read Write bit and register addresses whose values can be found in the SRFO8 manual which must be used for communication with the CRIO Particularly EO it is the sonar device address write bit it is used when the CRIO wants to write the sonar E1 it is the sonar device address read bit it is used when the CRIO wants to read from the sonar Therefore communication Sonar CRIO can
145. tion The Instruments Panel helps the operator to control the UAV flight The mission planning interface right screen in figure 20 was created using also a LabView software Mission planning 1s the first step to do at the beginning of the mission simulation process The first step 1s to decide the targets that represent the mission goals and constraints After that mission planning is accomplished mostly through interaction with the Set Waypoints amp Flight Plan Menu figure 20 The operator enters the waypoints coordinates in terms of latitude longitude and altitude for the UAV to fly towards the targets The user interface shows the waypoints and flight paths in 2D graphics Waypoints can be also added or inserted into an existing flight path for mission re programming while the simulation is running Once the waypoints WP have been defined the operator specifies the flight plan parameters into the flight plan menu in term 43 of waypoint identification number and velocity to be maintained by the helicopter while overflying a specific WP When running in autonomous flight mode the LabView code automatically sends the waypoints coordinates and the flight plan parameters to the helicopter autopilot computer Situation Awareness The visual system is used for viewing the engaging area in 3D and providing the operator with a good situation awareness The virtual world window enables three different visualization modes chos
146. via a rotor head Besides its rotation around the hub speed Q position WY the blade can also rotate about three hinges which are shown in figure 4 Feathering is the motion of the blade about its length and is described by the blade pitch angle flapping is the blade motion in a direction normal to the main rotor disc and is described by the flapping angle B lead lagging is the motion of the blade in the rotor disc plane and is described by the angle Significant variations in the design of the helicopter rotor head exist Blade motion is enabled by mechanical hinges near the blade root articulated rotor by blade root compliance hingless rotor or by combination of both lead lagging amp hub a eie blade flapping B feathering yoke flapping teetering hinge q hinge Pa flapping hinge restraint 4 neotering restrain rotor shaft Figure 4 Blade Degrees of Freedom Schematic and Rotor Head Mechanization 20 In the hobby helicopter used for the development of the RUAV system there are no actual flapping hinges The motion about the three hinges is restrained by elastomer fittings that act both as springs and dampers see fig 4 elastomeric flapping and teetering hinge The spring effect is used to transform the blade flapping motion in a hub moment In most rotorcraft the rotor speed is kept constant by an electronic engine governor The thrust and rotor moments are produced by changing the blade pitch
147. x dc trim ae zZ o lt Vr Vp Vx Vy Vz Zyz maximum collective trim collective forward speed total energy mass gravity acceleration altitude velocities in the north N east E down D reference frame velocities along the trajectory PWM high time PWM frequency PWM period PWM duty cycle natural frequencies of the longitudinal of the fuselage rotor bar modes natural frequencies of the lateral of the fuselage rotor bar modes flapping motion rotor time constant including the effect of the stabilizer bar longitudinal speed derivative lateral speed derivative vertical speed damping derivative 13 Zg Acronyms CAPECON DLR RUAV UAV UNIBO PID PI EU AV DL DD GCS COTS NGCS LAN VLAB SIT SDK FF WP USGS SRTM DEM PWM CRIO AHRS SITL HIL RC EMI EM FPGA IMU GPS AMSL AGL SDA SCL MSB LSB HL A VCS SISO SP PV DI DO GUI vertical control derivative Civil uav APplications amp Economic effectivity of potential CONfiguration solutions Deutches Zentrum f r Luft und Raumfahrt Rotorcraft Unmanned Aerial Vehicle Unmanned Aerial Vehicle UNIversity of BOlogna Proportional Integral Derivative Proportional Integral European Union Air Vehicle Data Link Data Distribution Ground Control Station Commercial Off The Shelf Navigation Guidance Control System Local Area Network Virtual reality LABoratory Simulation Interface Toolkit Software Development Kit
148. y sent to the autopilot 2 2 GROUND CONTROL STATION The work involved in this part of the project was the design and development of a ground control station for real time control and display of the simulated RUAV flight test data The GCS is the hub of an unmanned air vehicle 44 45 It processes the incoming data and sends control instructions to the air vehicle Typically a GCS will envelope three main functions mission planning mission control data processing The level of autonomy of the RUAV and the mission complexity dictate the GCS architecture For the purpose of this work a simplified standard GCS was developed which is able to operate the RUAV in both autonomous or remote piloted flight The GCS was designed to be easily modified for controlling and monitoring of a real RUAV therefore this part of the work was used as starting point to develop the GCS of the small scale helicopter UAV system 36 The GCS includes a visual system which induces a sense of presence in the engagement area provides a multi modal input interface including head tracker and joystick which enables efficient interactions Key problems to be solved during the development of the GCS were see next sections in details the interfacing of the different hardware and software components the development of the graphic interface for mission planning and control and flight data display the development of a visual system for modular miss
149. ystems Man and Cybernetics 2005 K Sprague V Gravilets D Dugail B Mettler E Feron Design and applications of an avionics system for a miniature acrobatic helicopter Digital Avionics Systems Conference October 2001 Daytona Beach Florida C Castillo W Alvis M Castillo Effen K VAlavanis W Moreno Small scale helicopter analysis and controller design for non aggressive flights Center for Robot Assisted Search and Rescue University of South Florida M LaCivita W Messner and T Kanade Modeling of small scale helicopters with integrated first principlesand integrated system identification techniques Monteal Canada June 2002 Presented at 58th Forum of American Helicopter Society V Gavrilets I Martinos B Mettler and E Feron Control Logic for Automated Aerobatic Flight of a Miniature Helicopter Massachusetts Institute of Technology Cambridge MA 02139 LabView PID Control Toolset User Manual J A Farell M Barth The Global Positioning System amp Inertial Navigation Mc GrawHill 1999 141

Progetto e realizzazione del sistema di gestione autonoma del volo

Contents

Download Pdf Manuals

Related Search

Related Contents