Home

fulltext - DiVA Portal

1. Jo Jc Find gt Jo D 100 f L Create gt 2 do i cu LL 2 7 place gt HAT J Jc Data Operations gt Reinitialize to Default Value 0 Jc Advanced b Make Current Value Default Fit Control to Pane i j Cut Data nnel graph of rec Scale Object with Pane ONIS een from node ty Add Dimension Remove Dimension Empty Array Add Element Gap Export gt Properties Figure 17 Use the right click dialog to change the default values of your controls 55 Figure 18 The reset button of the Zolertia Z1 sensor node Sensor node ap pearance and circuit board layout may differ between manufacturers A 1 5 Stopping execution To stop the execution of the emulator simply press the stop button amp A 2 Emulator Host Controls and Indicators These are the functions of the controls and indicators of the channel emulators host interface grouped by tab A 2 1 Signal Monitoring This tab gathers controls and indicators for monitoring the emulators input and output channels Pictured in figure 19 56 Signal Monitoring Channel Monitoring Channel Information Status Hardware Setup Errors current time 16 23 46 967 2013 03 08 Sensor Output Input channels Update graphs Save signal data to file Trigger ON Trigger level Plot width s Save file C 005 932 D Exjobb 130301_Arbetsfiler VIs 4 130306_Omstrukturering test dat Packet history 11 19 E98 12 20 INR 5 23 N
2. 30 31 31 31 31 33 33 34 34 38 38 39 41 Al Al Al Al 42 42 42 46 46 46 49 51 54 56 56 56 58 61 62 65 67 B The Channel Data Import subVI C The Signal Data Export subVI 69 71 List of Abbreviations ADC ASIC BER DAC FPGA IEEE IP LabVIEW MAC MCU NI PC PCle PRR RDC RF RFSA RFSG RSSI SINR UART VI Analog to Digital Converter Application Specific Integrated Circuit Bit Error Rate Digital to Analog Converter Field Programmable Gate Array Institute of Electrical and Electronics Engineers Internet Protocol Laboratory Virtual Instrumentation Engineering Workbench Media Access layer Microcontroller Unit National Instruments Personal Computer Peripheral Component Interconnect Express Packet Reception Ratio Radio Duty Cycle Radio Frequency Radio Frequency Signal Analyzer Radio Frequency Signal Generator Received Signal Strength Indication Signal to Interference and Noise Ratio Universal Asynchronous Receiver Transmitter Virtual Instrument List of Figures 1 10 11 12 13 Signal and data monitoring flow of the system setup Via separate interfaces the PC sets up the behavior of the emulator as well as any sensor nodes connected to it Both the emulator and the individual nodes are able to relay control information back to the PC Both the input signals to and the output signals from the emulator are combined into a single signal by the combiners The
3. individual signals are extracted by tuning a receiver to the carrier frequency and bandwidth of the desired signal The Zolertia Z1 sensor node uses the IEEE 802 15 4 compliant radio transceiver CC2420 0 2 020000 A close up view of the Zolertia Z1 s RF components Note espe cially the capacitor marked as C62 here soldered to provide the output signal on an external antenna via the U FL connector The back to back connection of the power splitters combiners al lows the simultaneous connection of up to eight sensor nodes to both the input and output connectors of the emulator The internal signal and data monitoring flows of the emulator The transmissions are converted and sampled by the signal ana lyzer RFSA after which the sampled signal is processed by the FPGA before being output again by the signal generator RFSG The resulting PRR based on different waiting times for acknowl edgement reception Note the periodical decrease in PRR that occurs every 0 3 0 4 mS 4 s cl volo Bosse Room vov The behavior of two nodes A and B as they attempt to com plete a packet transmission under different conditions The upper case is over the air transmission the middle case is transmission through the emulator without the modified protocols and the bottom case is through the emulator while using the modified protocols Note that relative sizes of the different blocks are ex aggerated to illustrate the no
4. 8 The results of the second measurement of PRR using a higher reso lution scale compared to the first test Based on these results a guard time of 70 us was chosen to minimize the overall latency of the system 37 6 Adding interference noise sources This section describes a way to introduce noise and interference sources to the emulator without modifying the existing signal processing done by the emu lator Section 6 1 describes an experiment conducted in relatively noise free environments Section 6 2 describes a similar experiment where a sensor node is configured to output a noise floor of increasing power By programming a sensor node as an interferer or noise source i e to output data that disturbs the transmissions of other sensor nodes the base tests can be expanded with the addition of different type of disturbances This is helpful when attempting to design more robust sensor networks that need to maintain performance under changing conditions This approach has been investigated previously for example by Boano et al in 18 Other types of hardware with greater power output and better spectrum control could also be interesting as sources of interference 6 1 Establishing a baseline PRR To measure the effects of transmitting with an interference source a baseline is first established by measuring the PRR for transmissions under different con ditions Baseline PRRs were measured for simulated over the air and throug
5. be changed during runtime e If desired configure the Trigger channel and Trigger level to suit your needs and turn triggering on e To enable saving of signal data to file toggle the Save signal data to file input and specify an output file Note it is possible to append new data to an already existing file provided the data of the existing file is of the same format as the new data to be written to the file e You are now ready to start the emulator host interface by pressing the run button gt A 1 4 Initializing the sensor nodes Depending on the setup and configuration of your experiment you may need to finalize the setup of the sensor nodes by doing the following steps e If you wish to monitor the serial output of any sensor node make sure the node s are connected to your PC via USB and open up a connection to the sensor nodes serial interface Note make sure to use the correct baud rate when connecting to the sensor nodes typically 115200 e You may wish to restart the sensor nodes when initializing your exper iment Nodes typically have a reset button integrated on their circuit board see figure 18 for an example 54 atrix Channel Selector Channel mode ed Read data from MATLAB fil n E Data file E 0 Jic Visible Items gt is Jo HM J V 4 E Exjobb 1303 130306_Or 2 Z Change to Indicator 2 moo Jc Make Type Def Afo do Jc Description and Tip Ho Number of Samples to read
6. node is pro vided in table 2 For more details refer to the IEEE 802 15 4 standard 3 or to the Zolertia Z1 data sheet 4 able 2 Zolertia Z1 specifications MCU MSP430F2617 Transceiver CC2420 Antenna Embedded or external via U FL connector Built in sensors 3 Axis accelerometer Temperature sensor Power source Battery or USB connector 17 3 1 1 Over the air transmissions In order to verify that the emulator is producing reasonable results it is im portant to understand how the sensor nodes work during normal over the air transmission and to determine a way to measure the performance and quality of a wireless channel To estimate the quality of wireless channels the Package Reception Ratio PRR has been used for both over the air transmissions and transmissions made through the emulator PRR indicates the proportion of suc cessfully received packages over time Assuming any bit errors are uncorrelated and the probability of bit error Per or Bit Error Rate BER is known PRR is calculated as PRR 1 pe 1 where n is the number of bits transmitted The packages transmitted during these tests were all of the size n 22 bytes package 8 bits byte 176 bits package including both header and payload bits In IEEE 802 15 4 see annex E of 3 the probability of bit error is a function of the Signal to Interference and Noise Ration SINN R of the received signal and is given by 1
7. q IQMAX OUT MAX Out In 0 Io IQMAXIN MIN Out In jo jo IQMAX OUT IQMAX IN 132768 FPGA P2P Delta Send Receive 14 Amplitude Peer to Peer Endpoint Size RFSA Jo Peer to Peer Most Samples in P2P Endpoint jo Peer to Peer Endpoint Size RFSG jo Peer to Peer Most Space Available in Endpoint jo Address Value Figure 22 The status tab of the emulator e RFSA state The state of the RFSA e Number of Elements for Writing Elements free for writing in the P2P pipe for the RFSA and RFSG respectively e Number of Elements for Reading Elements available for reading from the P2P pipe for the RFSA and RFSG respectively 63 Pict BENZ Writer overflow Indicates if an overflow has occurred in the RFSA or RFSG Reader underflow Indicates if an underflow has occurred in the RFSA or RFSG Time to underflow Indicates the running time before the first underflow occurs Peer to Peer Endpoint Size RFSA The maximum number of samples that can fit in the P2P endpoint of the RFSA Peer to Peer Most Samples in P2P endpoint The maximum number of samples in the P2P endpoint of the RFSA since start Peer to Peer Endpoint Size RFSG The maximum number of samples that can fit in the P2P endpoint of the RFSG Peer to Peer Most Space Available in Endpoint The maximum number of free space in number of samples in the P2P endpoint of the RFSG since sta
8. sensor node is down mixed and then filtered through a band pass filter in order to avoid any aliasing in the analog to digital converter ADC The sampled signal is then passed through the PCIe link to the NI PXIe 7966R FlexRIO FPGA module where the channel gain matrix can be applied NI PXIe 7966R FlexRIO FPGA An FPGA isa type of integrated circuit that can be configured by the user after manufacturing This ability offers great flexibility when compared to more traditional application specific integrated cir cuits ASIC The drawbacks of implementing your circuits as an FPGA rather than an ASIC are generally increased physical size which in turn affects cost levels decreased performance and increased power consumption In 11 Kuon and Rose present data comparing these metrics for FPGAs and ASICs of equal transistor size They find that an FPGA that uses only lookup tables and flip flops is on average 40 times bigger 3 2 times slower and uses 12 times more power than an ASIC The differences decrease somewhat if specialized hardware such as multipliers adders and dedicated memory blocks that would be needed by practically any conceivable implementation is used in the FPGA Despite the increased cost and decreased performance compared to ASICs FPGAs can still be a very useful and viable alternative especially for prototyping or other cases where you might want to alter the configuration at a later time The NI PXIe 7966R FlexRIO F
9. splitter combiner to the emulators input and output ports 5 Connect up to eight sensor nodes to the free ports of the 8 1 power com biner splitter using an U FL compatible connector see figure 13 Figure 13 Each sensor node must be connected to a port on the 8 1 power combiner splitter 6 Optionally you may connect any number of your sensor nodes to your PC over USB This will give you the ability to capture any serial output from 48 the sensor node s 7 Power on the emulator chassis 8 Restart the PC A 1 2 Required software Make sure you have working copies of the following National Instruments soft wares e LabVIEW 12 0 e NI DAQmx e PXI platform services e LabVIEW FPGA module e LabVIEW Real time module e NI RIO e NI RFSA This is most easily done by using the Measurement and Automation Explorer from Nation Instruments see figure 14 49 File Edit View Tools Help 4 Q My System f Show Help gt Bl Data Neighborhood Il Devices and Interfaces 8 Scales National Instruments Measurement amp 4 3 Software Automation Explorer gt W CompactRIO 120 ELLE MM 4 gt FieldPoint 6 0 11 Measurement amp Automation Explorer MAX provides access to your M Compliance Package 4 5 National Instruments products b 18 Labview 2010 2 4 T Labview 2012 3 What do you want to do EP Analog Modulation EP Control Design and Simulation Module 4 Database Connectivity Toolkit 3 Digital F
10. that these components are designed to work well i e minimal distortion and damping within the 2 40 2 48 GHz frequency range used by the emulator The power splitters combiners are designed to provide a 0 phase shift of the signals passed through it They also introduce an insertion loss proportional to the amount of inputs of the specific splitter combiner This would mean that a combination of a 2 to 1 and an 8 to 1 splitter combiner would give a theoretical insertion loss of 10 log 2 10 log 4 8 12 0 dB 6 Measuring the total insertion loss of a signal in the mentioned frequency range passed through the splitter combiner setup has yielded results of about 12 7 dB 20 Please refer to 7 and 8 for further information on the specifications of the splitters combiners used When connected by these components the sensor nodes are essentially di rectly connected to each other with a minimum loss of power in received trans missions caused by the transmitted signals double pass through the split ter combiner setup one pass from sensor nodes to the emulator and one pass from the emulator back to the sensor nodes resulting in a base loss of signal strength of about 25 4 dB 3 3 Emulator The emulator is constructed using commercially available hardware from NI All of the hardware used is part of the NI PXIe series of hardware meaning that the modules are able to communicate using the PCIe communication buses The s
11. to introduce this useful feature without any modifications to the emulator The basic noise floor can be expanded and modified to imitate those of WiFi Bluetooth microwave ovens and more simply by creating artificial channel data to modulate the basic noise floor output of the sensor node 7 4 Further work The emulator in its current state still leaves room for improvement and fur ther work Some of the aspects that might be considered for improvement are outlined below 7 4 1 Further testing and development Although some tests have been performed during the course of this work it is very likely that more will be required before the emulator can be seen as complete A design specification should be established that clearly specifies the desired features of the finished product In order to accomplish this an end user needs to be able to identify the real world applications where this piece of equipment might be useful so that features might be added to support these applications In addition to any further work on the emulator more work can be done on refining the MAC and RDC protocols used by sensor nodes For example moving the acknowledgement implementation and the channel switching asso ciated with it from software to hardware would decrease system latency This might require exchanging the currently used RDC protocol for the ContikiMAC protocol Furthermore developing software for the sensor nodes that can be used for ext
12. uojo3 iuuas e se 3 34 03 eyep jeubis aui OPAN 3 34 JO 14835 203 01 siapeau uuunjo2 33M pue y 338315 JOU JJ sjsixe 3 e3ep pa32s as 24 jou 10 12uja 322 el jjeubis 3ndui sosuas E ejep jeubis anes Figure 26 The signal data export sno continuously appends data to a speci fied file
13. 02 15 4 compliant radio transceiver CC2420 giving the sensor node its means of communication The IEEE 802 15 4 standard defines the properties of the physical layer and the media access control MAC of low rate wireless personal area networks and as such it is the principal standard for a number of energy efficient wireless systems including the Zolertia Z1 A brief summary of the CC2420 radio transceiver is provided in table 1 For more detailed information refer to the complete CC2420 data sheet 5 able 1 CC2420 radio transceiver specifications RF frequency band 2400 0 2483 5 MHz Transmit bit rate 250 kbps Output power 25 0 dBm Receiver sensitivity 95 dBm Saturation level 10dBm The CC2420 radio transceiver can communicate either via an embedded 16 ceramic antenna or via an external antenna attached via a U FL connector 6 Which antenna is used is determined by the placing of an output capacitor on the circuit board of the Zolertia Z1 see figure 3 U FL connector for external antenna md brett antenna Embedded ex Li E F D gt m rertr Antenna selection capacitor C62 RF shield covering CC2420 transceiver Figure 3 A close up view of the Zolertia Z1 s RF components Note especially the capacitor marked as C62 here soldered to provide the output signal on an external antenna via the U FL connector A brief summary of the specifications of the Zolertia Z1 sensor
14. 6 5 3 1 Gee 2 SINR is defined as the ratio between the power of the received signal P and the sum of the interference J and noise N powers i e P SINR I N 3 Both noise and interference power levels are highly dependent on the envi ronment of the receiver A too low SINR will result in a higher rate of failed transmissions and would at a certain point make the sensor network unusable Maintaining a high SINR on all channels is therefore a prime objective when designing a sensor network Under standard operations the sensor nodes would be transmitting over the air The signal strength of these transmissions is inversely proportional to the distance between the two antennae and can be modeled coarsely by a simplified path loss model Let P and P denote the transmitted and received powers at the respective antenna and d the distance between the two antennae The received power is then given as P PK 4 where d is a reference distance and y is the path loss exponent The range of y is usually between 2 for free space transmission and 6 in areas where transmissions are heavily obstructed by walls or other objects When using the CC2420 radio transceiver P can be user configured to values in the range 18 25 dBm 0dBm The combined antenna gain of both antennae K is a func tion of the wavelength A and the reference distance do K 2 5 It is clear that the received signal
15. L 14 22 5 75 6 725 NEU 505 IN 505 IR Trigger channel Amplitude 1 D 1 D 1 D 1 1 1 1 D 2500 5000 7500 10000 12500 15000 17500 20000 22500 25000 27500 30000 32500 35000 37500 40000 42500 0 Time ms Figure 19 The signal monitoring tab of the emulator e Input Output channels Selects whether the emulators input or output channels should be plotted e Update graphs Selects whether the plots should be updated or not Only pauses the plots not the emulator itself e Trigger ON Selects whether the trigger function should be used or not e Trigger level The smallest amplitude required to initialize the trigger function e Trigger length How many samples that are to be plotted The number of plotted samples are 1024 times the trigger length e Packet history Plots all eight of the emulators input or output signals The ampli tude of all plots has an offset added to it in order to make the plot easier to read e Trigger channel Selects the channel to trigger on A 2 2 Channel Monitoring This tab gathers controls and indicators for monitoring and setting the values of the emulated channels Pictured in figure 20 58 Signal Monitoring Channel Monitoring Channel Information Status Hardware Setup Errors Channel Matrix Channel Selector Channel mode uem M did data from MATLAB file lt A 1 1 a file Jo UN Jo 9 _ Jo y i Jo Jo I4 C User
16. PGA module described in detail in 12 uses a Virtex 5 FPGA device from Xilinx 13 The FPGA module applies the appropriate time varying channel gains to each input signal and provides the information that is to be displayed in the LabVIEW control software The input signal generated by the ADC in the NI PXIe 5663E Vector Signal Analyzer is down mixed and separated from its composite form to eight separate signals X k 7 The eight signals are then multiplied by an eight by eight channel gain matrix C11 k nes C18 k C k NA 8 C81 k t C88 k 22 containing the channel gains that are to be applied between each pair of sensor nodes As a sensor node is unable to transmit to itself the elements of the main diagonal are all zero i e ci k 0 i 1 8 Vk Also since the channel between sensor node 7 and sensor node j is identical to the channel between sensor node j and sensor node i the matrix is symmetric i e cj k cj k Knowing this we can simplify C k as 0 C21 k at C81 k Ae 9 2 cg k cgi k i cg k 0 The output signals y k yg k that result from the matrix multiplication are linear combinations of the eight input signals as determined by the channel gain matrix Y k l C k X k 10 Ys k These resulting signals are finally upconverted superimposed into a single signal and passed to the NI PXIe 5673E Vector Signal Generator In order to presen
17. UPTEC F14 002 Examensarbete 30 hp Januari 2014 UPPSAX UNIVERSITET Refinement of an emulator for the physical layer of the wireless communication of sensor networks Emil Eriksson UNIVERSITET Teknisk naturvetenskaplig fakultet UTH enheten Bes ksadress Angstr mlaboratoriet Lagerhyddsvagen 1 Hus 4 Plan 0 Postadress Box 536 751 21 Uppsala Telefon 018 471 30 03 Telefax 018 471 30 00 Hemsida http www teknat uu se student Abstract Refinement of an emulator for the physical layer of the wireless communication of sensor networks Emil Eriksson Wireless sensor networks have applications in many different fields and environments each with their own set of challenges to be overcome If we can measure and compensate for the properties of the wireless envi ronment at an early stage of the deployment we will be able to have the sensor network operational in a smaller time frame This report describes the development of an emulator for wireless sen sor network of up to eight 802 15 4 complaint sensor nodes The FPGA based emulator has been enhanced with several new features to remove some shortcomings of the existing emulator The emulator now provides the ability to import previously recorded data representing time varying gains of wireless links the ability to store the signal data generated by the sensor node for later review as well as a more intuitive user interface t
18. ZigBee ready RF Transceiver http www crew project eu sites default files CC2420 Data Sheet 1 4 pdf April 2006 6 Hirose Electric Group Ultra Small Surface Mount Coaxial Connectors http www hirose co jp cataloge hp e32119372 pdf 7 Mini Circuits Power Splitter Combiner ZAPD 44 http www minicircuits com pdfs ZAPD 44 pdf 8 Mini Circuits Power Splitter Combiner ZB8PD 4 ZB8PD 4 http www minicircuits com pdfs ZB8PD 4 pdf 9 National Instruments NI PXIe 1078 9 Slot 3U PXI Express Chassis http sine ni com ds app doc p id ds 312 lang en 2013 10 National Instruments Vector Signal Analyzer http www ni com pdf products us cat PXlIe 5663 pdf 11 Ian Kuon and Jonathan Rose Measuring the gap between fpgas and asics In FPGA 06 Proceedings of the 2006 ACM SIGDA 14th international sym posium on Field programmable gate arrays pages 21 30 2006 12 National Instruments NI FlexRIO FPGA Modules http sine ni com ds app doc p id ds 366 lang en 13 Xilinx Virtex 5 Family Overview http www xilinx com support documentation data sheets ds100 pdf Febuary 2009 14 National Instruments Vector Signal Generator http www ni com pdf products us cat PXlIe 5673 pdf 44 15 National Instruments Laptop Control of PXI ExpressCard MXI for PXI Express http www ni com pdf products us cat pxie8360 pdf 16 Adam Dunkels The ContikiMAC Radio Duty Cycling Protocol Technical report Swe
19. and data while at the same time decreasing the general clutter of the host interface In addition to the grouping all available components of the host interface have been relabeled in a more consistent and descriptive way along with their functionality being documented in appendix A 4 3 1 Signal monitoring The signal monitoring tab contains simultaneous plots of all input or output signals along with controls for selecting the trigger channel and number of samples to be plotted Also part of this tab are the newly added options for logging signal data to file The plots show the average of the signals absolute value over a number of samples This operation is performed by the FPGA Using the notation of equations 7 10 we can express each point in time of the plotted sensor output and input signal values as X kn EX km l 5 11 and Y kn Y km l 12 respectively Where k iz is a series of time points and the indices m and n represent the bounds of the time period over which the average is taken In 2 the interval is chosen so that m n 64 During the work on this project it was found that this interval produces excessive amounts of data especially when attempting to log signal data to file Ideally increasing this interval would be done by modifying the code that is being run on the FPGA 29 However modifying this code was outside the scope of this project All signals are therefore downsam
20. cknowledgement has had enough time to reach the intended receiver 5 1 Verifying the latency The default implementation of the ContikiMAC protocol as specified by 16 leaves room for latencies of up to 48 us for protocols that implements acknowl edgments of transmissions Given the previously measured latency of 42 18 us for a signal that passes once through the emulator it is likely that the nodes will experience problems with communication while connected through the emulator and employing any communications protocol that relies on acknowledgments of received transmissions i e protocols that require more than one pass through the emulator and thus experience a latency of at least 84 36 ps To verify that the latency is an issue that needs to be addressed tests must be performed using some standard set of protocols that implements acknowl edgement of received transmissions However further issues arise immediately as the implementation of the channel emulator requires transmission and recep tion to happen on different frequency channels Because of this the transmitting node must switch channel while waiting to receive acknowledgement from the receiver and the receiver must switch channel when transmitting the acknowl edgement This is something that is not supported in the standard protocols as transmissions and reception over the air are physically required to happen on the same channel The available RDC and MAC protocols must theref
21. cy could potentially be reduced by as much as 4696 However because of the financial investment that would be involved in exchanging the hardware used this option has not been explored further The issue was resolved instead by altering the MAC and RDC protocols used by the sensor nodes The solution involved modifying the timings of the protocol in order to allow the transmissions to be completed successfully even with the delay caused by the emulators latency 5 2 1 Altering the MAC protocol In addition to adding the channel switching to the RDC protocol the timings of the protocol must also be altered in order to compensate for the latency caused by the emulator A guard time of 70 us was added before a sensor starts listening for the acknowledgement of successful transmission This time was selected based on a series of tests wherein guard times of different lengths were used in an attempt to find an optimum A total of 1000 transmissions were made for each guard time The plot in figure 6 shows the PRR achieved using guard times between 0 00 1 15 ms 34 PRR as a function of the ACK time PRR 0 0 2 0 4 0 6 0 8 1 ACK listen time delay ms Figure 6 The resulting PRR based on different waiting times for acknowl edgement reception Note the periodical decrease in PRR that occurs every 0 3 0 4 ms The PRR appears to follow a periodic behavior depending on the increase in acknowledgement wait time The precise reason for this b
22. des transmission behavior and do not accurately reflect the relative times it takes to perform the operations uh CRUEL Ee RR E E ete ee vL eL RM The results of the second measurement of PRR using a higher resolution scale compared to the first test A comparison between theoretical and measured PRR The two curves are similar although the theoretical values seem to have a much sharper rise as SINR increases a The ExpressCard controller interface connects the emulator and Make sure to connect the power splitters combiners as pictured Connect the power splitter combiner to the emulators input and output DOEPUS x eR RR RR eh P esee ae ie teg a rd Each sensor node must be connected to a port on the 8 1 power combiner splitter re 17 21 14 15 16 17 18 19 20 21 22 23 24 25 26 Use the Measurement and Automation Explorer to find installed software from National Instruments In the Devices and Interfaces part of the Measurement amp Au tomation Explorer you should be able to identify the hardware components ofthe emulator sls Default values of the remaining hardware setup configuration op TIONS 922233 2 875 Fh Jede ow qe e oe PE Ee Mole S OO rea d AS Use the right click dialog to change the default values of your Controls Ast ee Ee ae y er oe e ded es The reset button of the Zolertia Z1 sensor node Sensor node appearance and circuit boa
23. dish Institute of Computer Science December 2011 17 National Instruments NI PXIe 5644R 5645R 6 GHz RF Vector Signal Transceivers http sine ni com ds app doc p id ds 422 lang sv 18 Carl Alberto Boano Thiemo Voigt Claro Noda Kay R mer and Marco Z iga Jamlab Augmenting sensornet testbeds with realistic and con control interference generation In Proceedings of the 10th International Conference on Information Processing in Sensor Networks April 2011 45 A Emulator description This appendix is meant to serve as a guide for any end user wanting to use the emulator in its present form A 1 User manual A 1 1 Hardware setup To ensure that the hardware is connected as required please follow these steps 1 Plug in the ExpressCard controller interface in the appropriate slot of your PC see figure 10 Figure 10 The ExpressCard controller interface connects the emulator and your PC 2 Make sure the cable running between the emulator and your PC is secured at both ends 3 Connect the 2 1 and 8 1 power combiners splitters back to back see fig ure 11 46 Figure 11 Make sure to connect the power splitters combiners as pictured 4 Connect the two free ports of the 2 1 power splitter combiner to the output and input ports of the emulator see figure 12 Either of the ports of the power splitter combiner may be connected to either of the ports on the emulator 4T Figure 12 Connect the power
24. e for a completely automated system to reliably process these files and evaluate the data therein Therefore the first task of implementing the data import function is establishing a standard format for the files used to store the channel gain data to be imported to the emulator The information that may need to be stored in the data file is firstly general data and information about the site and collective data on the sensor nodes and secondly specific information about each measured channel Also as the number of wireless channels measured at different sites can vary greatly the data must be in a format that allows it to be expanded to any size The data format suggested here can easily be modified and the corresponding changes to the LabVIEW files can also be implemented quickly if needed The current implementation is illustrated in table 4 27 able 4 The channel data file structure The initial fields are for site and test information followed by an array of structure arrays each element containing the information of a specific channel fs siteInfo receivingNode transmittingNode rssi channelInfo receivingNode transmittingNode rssil channelInfo The data pertaining to every channel is contained within an array s of Mat lab structure arrays and each channel can be uniquely identified by the combina tion of transmitter and receiver node IDs transmittingNode and receivingNode respecti
25. e largest input signal that can be represented in the emulator Script Controls how samples should be moved over the PCIe link May affect the emergence of underflow errors RFSA Power Level The power level at the RFSA RFSG Power Level The highest possible power generated by the RFSG 66 A 2 6 Errors This tab collects error messages produced by the emulator Pictured in figure 24 Figure 24 The errors tab of the emulator e Channel setup errors Errors that occur while the channel coefficient matrix is being initi ated e Exit errors Errors that occur during the shutdown of the emulator hardware i e shutdown of the P2P pipes RFSA device and RFSG device 67 FPGA channel write errors Errors that occur during the writing of channel data to the FPGA FPGA low channel read errors Errors that occur during the read of emulator output FPGA high channel read errors Errors that occur during the read of emulator input FPGA setup errors Errors that occur when the FPGA is initialized RFSA RFSG setup errors Errors that occur when the RFSA and RFSG are initialized RFSA run time errors Errors that indicate issues with the data acquisition of the RFSA RFSG run time errors Errors that indicate issues with the data output of the RFSG 68 B The Channel Data Import subVI The channel data import subVI operates by iterating through the data fields discover
26. e long been the norm However using wireless sensor networks rather than traditional wired sensors will greatly increase the flexibility of the sensor network especially when considering the issues of network topology and the deployment reconfiguration and maintenance of the physical network Pro vided that the overall performance of the sensor network is still adequate for a given application these advantages could be a great benefit to the network For example most modern industrial facilities use control engineering to ensure consistent quality of products or to minimize waste of materials and energy In these settings networks of sensors are used to gather information on the state of the processes by continuously monitor temperatures pressures flows and a myriad of other parameters The reliability of these sensors in terms of both the gathering and relaying of data is imperative for the control process to func tion efficiently Because of the high risk of interference in these settings wired sensors are often used However if it is possible to measure and compensate for the wireless characteristics of the site great benefits could come from choosing a wireless sensor network Many well known techniques for handling the drawbacks of wireless networks exist but to determine the feasibility of using a wireless sensor network in a spe cific setting field tests are necessary to become familiar with the characteristics of the available
27. e not been de signed to handle such delays This has been one of the main problems to solve in order to make the emulator more usable and it is discussed further in section 5 together with a proposed solution 3 4 Software tools A number of software tools have been used during the course of this work both for development and simulation The possible applications of these tools are extensive and not limited to what is described here This chapter is meant to be a summary of the work done using these tools 3 4 1 LabVIEW LabVIEW is NI s development environment for the G programming language G is a data flow programming language and as such it describes how things connect This can be compared to the way imperative programming languages like for example C C or Java describe how things happen Because of the different design of data flow programming languages certain tasks are much more easily accomplished than with imperative programming languages An application created with LabVIEW is referred to as a virtual instrument VI LabVIEW also promotes designing your application in a modular way by making a previously developed VI a part of another application This is done in a black box manner where inputs and outputs are made available from the contained VI usually referred to as a subVI to the containing VI and the internal workings of the subVI are mostly obscured to anyone running the application The creation of the user inte
28. e used to scale down the amplitude of all coefficients of the channel matrix Channel graph Plots a time series of eight channel coefficients as experienced by signals transmitted from one specified sensor node 60 A 2 3 Channel Information This tab contains information on the channel coefficient data parsed from a data file Pictured in figure 21 Figure 21 The channel information tab of the emulator e Sampling frequency The extracted sampling frequency e Site info The extracted site info e Channel index 61 The indices of every specific channel extracted from the data file This is the index that is to be specified in the Channel selector matrix in the Channel Monitoring tab e Receiving node The node ID of the receiving node For debugging purposes e Transmitting node The node ID of the transmitting node For debugging purposes e Channel info Extracted information text Can be used to easily identify channels A 2 4 Status This tab contains information on the status of the emulator hardware Pictured in figure 22 62 Signal Monitoring Channel Monitoring ChannelInformation Status Hardware Setup Errors Number of Elements for Writing Writer Overflow 0 Number of Elements for Reading Reader Underflow Unlinked Number of Elements for Writing Writer Overflow jo Number of Elements for Reading Reader Underflow os L Time to underflow 00 00 01 Ireset
29. ed in a comma separated list Interference and noise sources It has been shown that it is possible to implement different interference models in hardware by connecting an in terference source to one of the emulators inputs It has also been demon strated that it is possible to use a sensor node as the source of noise or interference 13 3 System overview This section describes both the hardware and software components that make up the emulator as well as any components meant to interface with it The hard ware used throughout this work is identical to the hardware used by Danielsson and Agren in 2 Also included in this section is a basic introduction to wireless communication and the limitations introduced by each component The system consists of up to eight IEEE 802 15 4 compliant sensor nodes a pair of power splitters combiners and the emulator hardware all of which is monitored and controlled from a PC running a custom made software The sensor nodes can optionally be connected to the PC via a Universal Asyn chronous Receiver Transmitter UART connection allowing the monitoring of sensor node output The PC is connected to the emulator via an ExpressCard interface to the Peripheral Component Interconnect Express PCIe bus The emulator itself has three major components A Radio Frequency Sig nal Analyzer RFSA captures and samples all transmissions from the con nected sensor nodes The sampled transmissions are proce
30. ed in the data file Data files of an unexpected format might produce unexpected result For readers familiar with the LabVIEW development envi ronment see figure 25 for details of how the subVI works 69 all at xe pul a DNE jpuuey gt i mum ii apou Burama22y 0jui jauUeY gt i ii apou upwusues j ouueu y ynoge uoneuuojui snoauoj 22siuu Bururequo2 uiBua Aue jo Hugs e ojujauueu e3ep SSY Pauinbe y Burureuo2 uybua Ajesjique jo Aewe ue issi spou 10suas Durgnusuen 34 jo QI 24 SpoNX1 apou sosuas Butuia221 34 jo QI 21 PONXY sway Buiwoj 0j y surequo pue auueu Ssa aJIM e syuasaidas ui Juawaya u e3 u5us Aresyique jo sjnijs jo Aeue ue aps 3i jnoqe uoneuiojur snoauojjasiwu Bururequo2 y 5u aj Aue jo Bugs e ojupys sjuawaisnseaw Burop uaym pasn fouanba1 Buiduues au sj sway Buimojjoz y Bururequo s P45 auQ sw04 Buimojjo 34 uo aq 03 PAPIA si 314 JYTILYIN 1 SY 1L VIN e Woy sjauuey gt jo jaquinu eyiqie ue ynoge uoneuuojur spepg wed ayy agagogagaagagagaagaagagagagagaagagaaggaagagagagaagagagaugogaaggucgdte Figure 25 The channel data import subVI iterates through the data fields of a higher level VI specified file and outputs that data to C The Signal Data Export subVI The signal data export subVI continuously writes a stream of signal data to the end of a data file The subVI is illustrated in figure 26 71 351 payesedas 7
31. ehavior is likely related to the changes in the timing of the sensor nodes interactions and should be possible to anticipate and avoid with further knowledge of the CX MAC protocol Before each data transmission performed by a node A it will attempt to establish contact with the recipient node B Node A will make several attempts to transmit a short preamble message each time going into receiving mode for a predetermined amount of time to listen for an acknowledgement of successful transmission from node B Once node B has successfully received a the preamble from node A node B will send an acknowledgement of the received transmission addressed to node A Each such acknowledgement is followed by node B waiting to receive the data transmission that is to follow If node A successfully receives the acknowledgement from node B node A stops transmitting preambles and transmits a data package In the CX MAC implementation no acknowledge ment is sent from node B after it receives the data transmission The behavior of a pair of nodes A and B is illustrated in figure 7 35 A B 2 E Transmission E Successful reception A B raus reception Guard time Figure 7 The behavior of two nodes A and B as they attempt to complete a packet transmission under different conditions The upper case is over the air transmission the middle case is transmission through the emulator without the modified protocols and the bottom case is through
32. ensive testing of the sensor nodes performance in certain network constellations and with certain channel data applied would be very beneficial Such software would greatly increase the usefulness of the emulator by allowing a quick and easy way to run the sort of extensive tests that this equipment was initially designed to run It would also be interesting to work further on the disturbance sources This could be done either by creating series of channel data matching the behavior of certain interference sources and later verifying these as reliable sources of 42 interference or by creating or acquiring new hardware specifically for use as disturbance sources 43 References 1 Glenn Judd and Peter Steenkiste Using emulation to understand and improve wireless networks and applications 2005 2 Anton Danielsson and Jakob Agren Utveckling av generisk kommunika tionsplattform f r tr dl sa sensorn tverk Master s thesis Uppsala Univer sity 2013 3 IEEE Standard for Information technology Telecommunications and infor mation exchange between systems Local and metropolitan area networks Specific requirements Part 15 4 Wireless Medium Access Control MAC and Physical Layer PHY Specifications for Low Rate Wireless Personal Area Networks WPANs 2009 4 Zolertia Z1 Datasheet http zolertia com sites default files Zolertia Z1 Datasheet pdf March 2010 5 Texas Instruments 24 GHz IEEE 802 154
33. etworks It offers the ability to perform large scale tests using a variety of sensor nodes either by simulating the nodes or through full hardware emulation Within Cooja in formation on the state of the sensor nodes in the simulation is made easily available This information can then be accessed and used for the purpose of testing and debugging Using a simulation tool when developing and designing a sensor network is invaluable for early testing It needs to be pointed out however that while much information can be gained through the simulation of sensor networks a simulator can never be a complete model of every aspect of the environment The flaws of simulated environments constitute the reason behind the project presented in this report 26 4 Modifying the LabVIEW host interface A host interface is the name given to the user interface of a virtual instrument VI created in LabVIEW It presents acquired data and offers control param eters to the user which enables the user to manipulate the operation of the VI Part of the specification of this work involved reworking the user interface of the previously available LabVIEW host interface This involves both cosmetic changes to the user interface like rearranging relabeling or redesigning the available user interface as well as introducing new functionalities through mod ifications to the existing VI and subVIs or through the addition of completely new subVIs The functiona
34. evious and current versions of the host interface Because of the difficulties in achieving this large scale testing the improvements to the interface have instead been measured by comparing the current state of the host interface with the expectations of the potential future users of this system 31 Through this feedback features such as context sensitive user inputs have been added and labeling and scales of the various controls and indicators have been improved 32 5 Latency This section discusses the problem with acknowledgement transmissions asso ciated with the emulators input output latency and proposes a solution to this issue Section 5 1 describes how the latency problem was verified and section 5 2 describes how it was addressed It was mentioned in 2 that the latency of the emulator may introduce problems with certain types of transmissions The total latency for a single transmission through the emulator was measured to be 42 18 us During the tests performed in 2 simple protocols for radio duty cycling RDC and media access layer MAC were used that kept the transceiver powered at all times and had no means available to detect failed transmissions It was suspected that when using more sophisticated communication protocols this latency will cause the sensor nodes to time out while waiting for other nodes to acknowledge their transmission thus assuming transmission failure and initiating retransmission before the a
35. h the emulator transmissions without any interference source The simulations were performed using Cooja described in section 3 4 3 All three tests were performed using the CX MAC protocol in simulation and over the air experi ments and the modified CX MAC protocol described in section 5 2 1 in the case of the emulator with approximately 10000 transmissions and a transmission rate of 1 Hz During the simulation and over the air experiments sensor nodes were placed as close together as possible approximately 1 cm for over the air When using the emulator the close placement of the sensor nodes is emulated by set ting the channel gain to 1 The PRRs achieved in these measurements are listed in table 5 Environment PRR Simulation 100 0 96 Over the air 98 4 96 Through the 98 3 96 emulator Table 5 PRRs achieved during transmission tests where the only interference source is the background interference present The perfect PRR of the simulation environment is to be expected for the idealized and interference free conditions that exists in the simulator For the over the air transmissions the sensor nodes embedded antennae were used The antennae were aligned with each other and kept at a distance of about 2cm Exactly what the dropped packets in the air and emulator transmissions are 38 caused by is uncertain but it is probable that some of the packet losses are caused by external interference or thermal no
36. he emulator host interface but should not be of any concern to the typical end user 4 3 5 Hardware setup The hardware setup tab is used to configure the behavior of the hardware com ponents and to specify the correct devices to be used It is very important that the correct devices are selected and appropriate values configured 4 3 6 Errors Under the errors tab any errors encountered during the running of the emula tor are presented A number of new error types were introduced during this thesis project To facilitate debugging the host interface is able to catch and distinguish between nine different types of errors 30 4 4 Verification of the host interfaces usability Through bench testing and through feedback from potential end users the function and usability of the host interface in its present form has been verified 4 4 1 Importing channel data The subVI that handles importing channel data from file is described in detail in Appendix B The functionality of the subVI was continuously tested and verified during the development of the feature As long as the data file being processed is of the format described in section 4 1 data sets of any size may be extracted The subVI processes the data until it reaches the end of file or an error occurs The time series stored in the data file need not be of equal length as each time series is parsed and presented individually Data files of an unexpected format will typically ca
37. he interference patterns of other devices that may be present in the 2 4 Ghz frequency band A simpler way to achieve the same results may be to use the new feature of channel data importing to create a similar behavior Manually create a series of channel data representing the behavior of an interference source and configure a sensor node to output a constant noise floor and applying this channel data to a link between the interfering sensor node and a sensor node operating normally the noise floor created by the interfering node is modulated by the artificial channel data and will appear as an interferer of the desired type It is possible to create any number of interference patterns for use in this way using only a single sensor node as the source of the interference signals This method also removes the need to reconfigure the sensor node before being able to apply a new interference pattern The limiting factor in this setup is the speed at which the emulator can update the channel data Currently the emulator is hardwired to fully update the channel matrix every 10ms As a consequence if a rapidly changing interference pattern is desired it must still be implemented as a hardware solution 40 7 Summary and conclusions The contributions as outlined by section 2 2 closely follow what has been the desired outcome of this project 7 1 Host emulator interface Evaluating the changes to the interface of the host emulator is a difficult
38. his use of the available frequency spectrum has led to further changes in how the sensor nodes must behave in order to conform with the emulator s expectations See table 3 for a clarification on how frequencies are arranged Table 3 The assignment of carrier frequencies for each of the eight sensor nodes attached to the emulator Each frequency channel is 5 MHz in bandwidth Each of these channels are represented by a number specified by the IEEE 802 15 4 standard Sensor Receive Channel Transmit Channel node frequency Number frequency Number 1 2402 5 MHz 11 2442 5 MHz 19 2 2407 5 MHz 12 2447 5 MHz 20 3 2412 5 MHz 13 2452 5 MHz 21 4 2417 5 MHz 14 2457 5 MHz 22 5 2422 5 MHz 15 2462 5 MHz 23 6 2427 5 MHz 16 2467 5 MHz 24 7 2432 5 MHz 17 2472 5 MHz 25 8 2437 5 MHz 18 2477 5 MHz 26 Another issue that relates to the latency between input and output of the emulator is the way the sensor nodes interpret the delay they experience The delay will only cause problems when the sensor nodes of the network are using communication protocols that implement acknowledgments of received trans 24 mission In this case a delay of tens of micro seconds could be interpreted as the distance between two sensor nodes being several kilometers Under normal conditions the signal would be far too weak to travel such distances and with that in mind the sensor nodes communications protocols hav
39. ignal travels through the emulator as illustrated in figure 5 Emulator Chassis To sensor nodes c a P o UO M o 6 a Uu PTT Monitoring Control Signal Path Figure 5 The internal signal and data monitoring flows of the emulator The transmissions are converted and sampled by the signal analyzer RFSA after which the sampled signal is processed by the FPGA before being output again by the signal generator RFSG The blocks labeled RFSA and RFSG refer to the vector signal analyzer and vector signal generator respecitively and are described further in section 3 3 1 3 3 1 Emulator Components As indicated in figure 5 the emulator consists of a signal analyzer for sampling the sensor nodes output signals an FPGA for applying the channel gain matrix a signal generator for outputting the processed signals as well as some other components described further in the following paragraphs 21 NI PXIe 1078 Chassis The NI PXIe 1078 Chassis described in detail in 9 provides a compact basis for connecting the remaining NI hardware components together and provides the PCIe links to the attached modules The maximum theoretical data rate of any module connected through the PCIe link is 200 MB s NI PXIe 5663E Vector Signal Analyzer The NI PXIe 5663E Vector Signal Analyzer described in detail in 10 constitutes the input part of the emulator The sum of all analog signals from any connected
40. ile emulatorHost vi in LabVIEW A 1 3 Configuring the emulator Before running the emulator host interface make sure you configure it to suit your needs e Go to the Hardware Setup tab of the host interface e Make sure the settings for NI FPGA Device NI RFSG Device and NI RFSA Device are correct for your setup You may need to refer to the Devices and Interfaces part of the Measurement amp Automation Explorer see figure 15 51 File Edit View Tools Help 4 C My System Show Help 8l Data Neighborhood 7 a ag Devices and Interf 4 evices al nt aces 4 li NI PXle 1078 Chassis 2 National Instruments Measurement amp lj 1 NI PXIe 8360 Remote Controller Automation Explorer W 2 NI PXIe 5450 PXI2SIot2 ur SS 4 NI PXIe 5611 PXDSlot4 Measurement amp Automation Explorer MAX provides access to your WA 5 NI PXIe 5652 PXDSIot5 National Instruments products E 6 NI PXIe 7966R RIOO 4 8 NI PXIe 5663E PXDSlot8 What do you want to do 7 NI PXIe 5652 PXDSIot7 WP Manage my devices and interfaces W 8 NIPXIe 5601 PXI2SIot8 W 9 NI PXIe 5622 PXDSlot9 aL Network Devices Bl manage virtual channels or tasks for my devices b F Serial amp Parallel 44 Create scales for my virtual instruments Scales Configure my IVI instrument drivers b Software gt i M Drivers b amp Remote Systems 2x Manage my installed National Instruments software Wo Import export my device configura
41. ilter Design Toolkit Manage virtual channels or tasks for my devices Manage my devices and interfaces Manage my installed National Instruments software Eoi Digital Modulation Create scales for my virtual instruments 13 FPGA Configure my IVI instrument drivers T3 MathScript RT Module Real Time 4 Report Generation Toolkit For Microsoft Office B T8 Spectral Measurements Note Some categories are device specific For example the IVI 53 System Identification Toolkit category appears only if you have IVI installed Mais th ES For more information about using MAX select available help categories from the Help menu If you need further assistance or want to know T3 LabVIEW Run Time 821 more about your device visit the National Instruments Technical 8 LabVIEW Run Time 8 5 1 Support Web site 8 LabVIEW Run Time 8 6 1 8 LabVIEW Run Time 2009 SP1 f6 For more information about this version of MAX launch the readme or 8 LabVIEW Run Time 2010 SP1 visit ni com info and enter the following Info Codes 3 LabVIEW Run Time 2011 SP1 f2 LabVIEW Run Time 2012 f3 3E LabVIEW SignalExpress 2012 amp LabWindows CVI Run Time 2010 SP1 Measurement amp Automation Explorer 5 3 3 gt YW Measurement Studio for VS2005 Submit feedback on this topic gt V Measurement Studio for VS2008 gt W Measurement Studio for VS2010 8 Visit ni com support for technical support F NIFIexRIO 12 0 FA NIVO Trace 30 2 Y NIPXI Platform Ser
42. imitations introduced by each of the components Sections 4 to 6 provides the reader with an understanding of what has been done to improve the existing emulator and sensor nodes from the work 10 by Danielsson and Agren Section 4 describes the modification of the user in terface including adding features to import and export data In section 5 a solution to the emulator s latency issue is presented where the Media Access layer MAC and Radio Duty Cycling RDC protocols of the sensor nodes are altered Section 6 describes an experiment conducted to show how interference and noise models can be implemented using the existing hardware In summary by programming a sensor node to emit a noise floor and connecting the sensor node to the emulator as normal it is possible to mimic many different sources of noise and interference Conclusions from the project are listed in section 7 and section 7 4 identifies open issues 11 2 Related work and contributions This section describes how this project relates to some similar projects and outlines how it contributes to a previous project on physical layer emulation 2 1 Related work A proof of concept version of an emulator for the physical layer of the wireless communication of sensor networks was developed by Danielsson and Agren at Uppsala university 2 They used off the shelf hardware and the Laboratory Virtual Instrumentation Engineering Workbench LabVIEW software to create an emu
43. ise in the RF circuits The similar results achieved with both over the air and through the emulator transmissions indicates that the emulator closely resembles the conditions of over the air trans missions 6 2 Connecting a sensor node as a noise source By placing the DACs of the CC2420 transceiver in a test mode it is possible to program a sensor node to emit a constant noise floor see 5 for details A sin gle sensor node was programmed to perform this way and output noise floors of increasing levels At each level 1000 packets were transmitted between two dif ferent sensor nodes through the emulator These transmissions were performed with a transmission power P 12dBm sending one packet every second and using the modified CX MAC protocol The results of these measurements are listed in figure 9 along with the theoretical PRR calculated as outlined in section 3 1 1 The Package Reception Ratio as a function of the SINR PRR 2 4 SINR dB Figure 9 A comparison between theoretical and measured PRR The two curves are similar although the theoretical values seem to have a much sharper rise as SINR increases It is clear that the increased noise floor contributes negatively to the PRR 39 There is a decrease in the measured PRR at an SINR of 3dB and at an SINR of 2 dB all transmissions are effectively blocked By following the work outlined in 18 sophisticated interference sources can be created to imitate t
44. l data within the selected file you can briefly run the host interface by clicking the run button i gt wait a few seconds and then click the stop button amp This will generate a list of all available channel data under 53 the Channel Information tab You can refer to this list during the next step In the Channel Selector matrix specify the index of the channel data you wish to associate with a particular pair of sensor nodes Note the indices should always be entered in the upper triangle of the Channel Selector matrix as the emulator will later mirror this half of the matrix to create a symmetric system of channels Finally use the Number of Samples to read data field to specify how much of the available channel data is to be read from the data file Note once the final sample has been applied the channel gain will remain constant at this value for the remainder of the execution For users wanting to make several runs with the same set of channel data applied may want to use an existing LabVIEW feature to make the entered values the default values Simply right click the control you wish to give a new default value and select Data Operations Make Current Value Default see figure 17 e Next go to the Signal Monitoring tab of the host interface e Use the Sensor Output Sensor Input toggle control to select whether you would like to monitor the connected sensor nodes output or input signals Note this can
45. lator that supports real time two way communication between up to eight IEEE 802 15 4 3 compliant sensor nodes The results of the previous work was encouraging but the following issues remained 2 2 The emulator introduced a latency to any signal fed through it In prac tice this latency caused sensor nodes to time out while waiting for ac knowledgments of received transmissions which was interpreted as a trans mission failure The simple user interface limited what tasks could be performed with the emulator There was only minimal support for on the fly loading of measured channel data If the emulator is to be used as intended a simple way to load channel data is required There was no means to save the results of an emulation session for later review Such a feature is essential in an analysis of the behavior of the emulator throughout a test There was no means to apply external interference Having the ability to apply an external interference to the recorded channel data would make the emulator capable of performing more advanced tests Contributions The purpose of this project was to address the issues presented by Danielsson and gren 2 and outlined in section 2 1 The following goals were achieved Input output latency issues By altering the MAC and RDC protocols of the sensor nodes the latency issues of the emulator have been remedied Previously the input output latency of the emulator caused sensor nodes t
46. lity of added features have been verified as an ongoing process throughout the development of these features The full documentation and user instructions for these features are available in Appendix B The redesign of the user interface has undergone several iterations each time collecting feedback on the different features The feedback has been incorpo rated in the current design of the user interface This section describes the alterations made to the existing user interface They include both visual changes and added features Section 4 1 describes the feature used to import previously measured channel data from file as well as the file format designed for use with this feature Section 4 2 discusses the feature of signal data export and describes the format of the output file Section 4 3 gives an overview of the new layout created for the user interface Section 4 4 outlines how the verification of the new user interface was done 4 1 Importing channel data The greatest alteration to the user interface has been the addition of the ability to easily import channel data from file Using pre measured data for repeatable testing is one of the main reasons for developing the emulator After measuring the wireless channels at a certain site channel gain data is generally processed in Matlab and stored in a Matlab data file Previous to the work done here no definitive format for these files had been determined meaning it would be impossibl
47. networks 2 This report can be seen as an extension to the work done by Danielsson and gren The goal of this project is to create a physical layer emulator that can be used in the following process e Measure the characteristics of the available wireless channels e save the gathered data e load gathered data onto a piece of equipment that can emulate the physical layer of the wireless communication in real time and e directly connect the antennae of physical sensor nodes to the emulator This process would allow us to perform repeatable tests in a laboratory envi ronment using the same type of hardware that would be used for the deployed sensor network Using this emulator to perform accurate early testing will pro vide the ability to detect and address issues with the network before deploying the equipment to the site This approach would also be beneficial because it could save much time money and effort This report details the work done to improve certain issues of an existing emulator of this type The report is organized as follows Section 2 gives the reader an understand ing of the starting point of this research project and what alterations have been made to the existing setup Section 3 presents an overview of the system setup describing both the hardware and software that makes up the emulator and the components meant to interface with it It also gives a basic introduction to wireless communication and the l
48. ng 4 3 83 Channel information 4 3 4 Status es 4 3 5 Hardware setup 4 3 6 Errors crees 4 4 Verification of the host interfaces usability 4 4 1 Importing channel data 4 4 2 Exporting signal data 4 4 3 General modifications Latency 5 1 Verifying the latency 5 2 Adding latency tolerance to the system 5 2 1 Altering the MAC protocol Adding interference noise sources 6 1 Establishing a baseline PRR 6 2 Connecting a sensor node as a noise source Summary and conclusions 7 1 Host emulator interface 7 1 1 Channel data import 7 1 2 Export of resulting data T 2 L teney ere 5 i090 3 099r ee ve 7T 3 Interference sources 7 4 Further work 7 4 1 Further testing and development Emulator description A 1 User manual 4e A 1 1 Hardware setup A 1 2 Required software A 1 3 Configuring the emulator A 1 4 Initializing the sensor nodes A 1 5 Stopping execution A 2 Emulator Host Controls and Indicators A 2 1 Signal Monitoring A 2 2 Channel Monitoring A 2 3 Channel Information AZA Status amp v ere aom aoee a x3 9 9 oen A 2 5 Hardware Setup A 2 0 EHONO Eoi RO XO Ae Ux Meo 12 12 12 14 15 18 19 21 21 23 25 25 25 26 27 27 29 29 29 30 30 30 30
49. nsceiver is a cheap and compact component that can be integrated in a wide variety of devices One type of device that makes great use of wireless communication is wireless sensors Since wireless self powered sensors are mobile they can be used for a wide variety of applications where a wired sensor network would be problem atic or even impossible to deploy For example wireless sensor networks can be used to track the movement and interactions of animals in the wild or to monitor soldiers on a battlefield Installation of wireless network components is also much less intrusive than wired sensor networks which can be desirable in historical buildings or at other sites where there is a need for the sensor network to provide minimal disturbance Wired and wireless networks provide different benefits and drawbacks and may be preferable in different situations and settings There are some areas where both wired and wireless sensor networks are viable alternatives but his torically wired sensor networks have been favored like local area networks in office environments The main reasons for this are that wired sensor networks are generally less susceptible to electromagnetic interference that may be caused by nearby equipment when compared to their wireless counterpart Wired sensor networks also offer bandwidths that are around an order of magnitude greater than what can be achieved in wireless sensor networks As a result wired sen sor networks hav
50. o make the emulator more useful as a research tool A latency issue that previously prevented the use of any communications protocol depending of transmission acknowledgments has been studied and a solution to the issue is presented A way of introducing external interference to the em ulated system has also been investigated and provides a straightforward way of introducing interference to the system without modifying the ex isting hardware and software Most added features have been designed to be easily modified or ex panded upon Handledare Tomas Olofsson Amnesgranskare Anders Ahl n Examinator Tomas Nyberg ISSN 1401 5757 UPTEC F14 002 Refinement of an emulator for the physical layer of the wireless communication of sensor networks Emil Eriksson Abstract Wireless sensor networks have applications in many different fields and environments each with their own set of challenges to be overcome If we can measure and compensate for the properties of the wireless envi ronment at an early stage of the deployment we will be able to have the sensor network operational in a smaller time frame This report describes the development of an emulator for wireless sen sor network of up to eight 802 15 4 complaint sensor nodes The FPGA based emulator has been enhanced with several new features to remove some shortcomings of the existing emulator The emulator now provides the ability to import previously recorded data rep
51. o time out while waiting for a reply By resolving this issue the analysis and evaluation of more advanced communication schemes has been made possible These communications schemes are a requirement for many ap plications as they provide the means for retransmitting data packets that do not reach the receiver In this report the steps performed in altering 12 the protocols are described and an attempt is made to enlighten the reader about the necessary steps and precautions when doing so User interface To increase the usability of the emulator the software has been modified to be easier to use All controls and indicators have been clearly labeled and their specific functions have been documented Through discussions and test runs with end users the usability of the new user interface has been verified Importing channel data A data format for storing measured channel and site data has been proposed It allows storage of relevant data and can be adapted to suit future needs In conjunction with the newly designed data format a simple way to import previously measured channel data from file and apply these channels in the emulator has also been created As with the data format the import tool can easily be modified to suit a changing need Saving signal data It is now possible to save the transmitted and received data of every sensor node connected to the emulator The data is saved continuously while running the emulator and stor
52. ong with the increased wait time the protocol was also altered to make data reception and acknowledgement transmissions happen on different channels This was necessary due to the design of the emulator where emulator input and output happens on different channels Tests performed with the altered protocol have shown a substantial increase in PRR to a point where transmissions only fail very occasionally If there would be a need to actually reduce the latency not just mitigating it the option of upgrading the hardware is still available Using a module such as the NI PXIe 5644R which integrates the FPGA RFSA and RFSG into a single device could potentially remove nearly half of the existing latency Also as is mentioned in section 5 1 the protocol developed in this work handles 41 acknowledgments in software and not in hardware It is generally true that lower level implementations are faster than their higher level counterparts It is reasonable to assume that it would be possible to reduce the latency by finding a way of implementing the necessary channel switching within a protocol that handles acknowledgments in hardware 7 3 Interference sources It has been shown that a sensor node can be successfully configured to behave as an interference source By configuring the DAC s of the sensor nodes transceiver to a test mode the sensor node will output white noise Implementing added interference in this way is a quick and simple way
53. ore be modified to support this sort of behavior Here the CX MAC protocol has been used as a starting point for creating a protocol that will function with the emulator The main reason for choos ing CX MAC over ContikiMAC is that the former handles acknowledgments in software while the latter handles acknowledgments in hardware Handling ac knowledgments in hardware is faster but modifying the behavior of the protocol to the needs of the emulator would require more work After modifying the CX MAC protocol to add channel switching an exper iment was performed where a total of 1000 transmissions were sent at a rate of 33 5 transmissions per second During the experiment the PRR was used to mea sure the performance of the setup PRR is defined as the number of correctly received data packets divided by the total number of sent data packets The experiment showed a PRR of 0 7 96 which is a clear indication that the latency caused by the emulator is a significant problem that must be addressed 5 2 Adding latency tolerance to the system It was suggested in 2 that upgrading the emulator with the NI PXIe 5644R Vector Signal Transceiver module which integrates an FPGA RFSG and RFSA into a single module could remove most of the latency caused during the in ternal I O of the emulator The details of this module can be found in 17 The latency caused by internal I O in the emulator has been measured to be 19 55 us meaning that the laten
54. pled by a factor 16 after being extracted from the FPGA and before being presented on screen or written to file It was found that even after this downsampling it was still possible to visually detect packages at least as small as 11 bytes and possibly even shorter Using the Trigger length control it is possible to determine how many such values are plotted at any given time 4 3 2 Channel monitoring The channel monitoring tab gives the controls to select the behavior of the channels as well as a plot to monitor the absolute values of the channel gains In particular when using the new function for importing channel data you now have the ability to select which of the previously measured time series is to be applied to each individual channel 4 3 3 Channel information The channel information tab displays any information that may have been parsed from a data file including both information about site and channels This is particularly useful when determining which channels should be applied between which two sensor nodes The entire concept for this tab was imple mented in this version of the emulator host interface 4 3 4 Status The status tab attempts to present the status information of the hardware especially information pertaining to the signal generator and signal analyzers states This information is used mainly to monitor for over or underflow issues something that may be important to catch during further development of t
55. rd layout may differ between manu facturersis oua d cee mtb E S The signal monitoring tab of the emulator The channel monitoring tab of the emulator The channel information tab of the emulator The status tab of the emulator The hardware setup tab of the emulator The errors tab of the emulator The channel data import subVI iterates through the data fields of a specified file and outputs that data to a higher level VI The signal data export subVI continuously appends data to a Specified fil 2 2 qd RR URS whe gene ge dhs 70 List of Tables N CC2420 radio transceiver specifications Zolertia Zl specifications ee The assignment of carrier frequencies for each of the eight sen sor nodes attached to the emulator Each frequency channel is 5 MHz in bandwidth Each of these channels are represented by a number specified by the IEEE 802 15 4 standard The channel data file structure The initial fields are for site and test information followed by an array of structure arrays each element containing the information of a specific channel PRRs achieved during transmission tests where the only interfer ence source is the background interference present 1 Introduction Today wireless networks are everywhere The technology has evolved rapidly over the last few decades to a point where a wireless tra
56. resenting time varying gains of wireless links the ability to store the signal data generated by the sensor node for later review as well as a more intuitive user interface to make the emulator more useful as a research tool A latency issue that previously prevented the use of any communications protocol depending of transmission acknowledgments has been studied and a solution to the issue is presented A way of introducing external interference to the em ulated system has also been investigated and provides a straightforward way of introducing interference to the system without modifying the ex isting hardware and software Most added features have been designed to be easily modified or ex panded upon Contents 1 2 Introduction Related work and contributions 2 1 Related work 2 2 Contributions e System overview 3 1 Sensor nodes ea wa yo ls 3 1 1 Over the air transmissions 3 2 Peripheral equipment 39 3 Emulator i 245293353443 3 3 1 Emulator Components 3 3 2 Transmitting through the emulator 3 4 Software tools 3 4 1 LabVIEW 5 2352 55x 3 4 2 Contiki 4 4 623 yos 4 9 Qo0 4 3 Gs eed Ww A Ea Modifying the LabVIEW host interface 4 1 Importing channel data 4 2 Exporting sensor signal data 4 3 General modifications 4 3 1 Signal monitoring 4 3 2 Channel monitori
57. rface is integrated in the development of the VI this will be described in more detail in section 4 LabVIEW also contains several toolboxes for integrating with many differ ent types of instrumentation hardware like signal generators data acquisition hardware and FPGAs from both NI and other vendors 3 4 2 Contiki Contiki was first created by Adam Dunkels at the Swedish Institute of Computer Science and is an open source operating system for hardware constrained low power systems using wireless communication Contiki fully supports standard IPv4 and IPv6 networking through the uIP and uIPv6 network stacks allowing communication over the internet In addition to this Contiki also supports a number of additional protocols and standards for low power wireless commu nication making it ideal for use with devices and for applications with limited power supply 25 Contiki supports a range of different hardware platforms from a number of manufacturers including the Zolertia Z1 sensor node used throughout this work and described in further detail in section 3 1 The Instant Contiki development system is distributed in the form of a virtual machine and provides all the compiler tool chains and libraries required to develop a system running Contiki as well as a number of tools useful for debugging large networks of wireless devices such as the simulation software Cooja 3 4 3 Cooja Cooja is a simulation environment for wireless sensor n
58. rt Address he address of the channel coefficient that is currently being updated Should loop from 0 to 31 Value The value that is currently being written to the indicated channel coefficient FPGA P2P Delta Send Receive Plots the MAX out in over time Should remain somewhat constant over time otherwise a timing error is probably occurring between the RFSA and RFSG 64 A 2 5 Hardware Setup This tab contains the controls for configuring the emulator hardware Pictured in figure B poo script p2p repeat forever stream 128 end repeat end script Figure 23 The hardware setup tab of the emulator e NI FPGA Device Specifies the NI RIO device corresponding to the FPGA device e Enable FPGA Enable or disable the FPGA e NI RFSG Device 65 Specifies the PXI slot containing the RFSG device NI RFSA Device Specifies the PXI slot containing the RFSA device NI RFSG Reference Clock Source he clock to be used to control the timing of the RFSG Must be set to PXI_ CLK to avoid timing issues between the RFSA and RFSG IQ Rate The bandwidth sampled by the RFSA Measured IQ Rate Measured bandwidth of the RFSG should match the IQ rate Frequency Center frequency of the emulators input channels When using the frequency channels between 2 445 GHz and 2 480 GHz this should be set to 2 4625 GHz Reference level The power in dBm corresponding to th
59. s WISE 130306_Omstrukturering struct_test_2 mat 9o do io Jo Lo Jo do do Data file using dB false Jo io io ido Io Jo do Jo C Yo A 0 Jo Jo Jo 5 o Lo 0 Number of Samples to read Number of samples to plot i om Ome Jo 1d mi Uo gio 7 7 z A A K K Iterations Channel scaling au NC com CN co co CN ERN L om om CN CON CON ON CN ON m m om NN CON C ON Tab Control 2 Channel graph of receiving channels as seen from node transmitting on channel 19 IW 20 nu 2 ei 22 iW 23 iq 24 iw 25 iw 26 im Figure 20 The channel monitoring tab of the emulator e Channel matrix Shows the current values of all the channel coefficient s e Plotted channels The Row of channel coefficients that are to be plotted e Channel selector Specifies which channel read from data file is to be assigned to each channel Channel mode Selects the mode by which the channel coefficients should be set Current available modes are Read data from MATLAB file Diagonal All constant x Random channels Data file The data file to read channel data from Data file using dB Not implemented Number of samples to read The number of samples to read from the data file Number of samples to plot The number of samples to be displayed in the plot Iterations Number of times the channel matrix has been updated For debug ging purposes Channel scaling Can b
60. sor nodes Under normal over the air transmissions all sensor nodes would typically be configured to transmit and receive on a single frequency channel However as there is some latency in the emulator a sensor node would be able to switch from transmitting to receiving mode in time to hear its own transmission Since all signals both output and input ultimately travel through the same antenna cable there is also the risk of a feedback loop being created between the em ulators input and output ports Some sort of shielding would also be required to prevent sensor nodes from receiving transmissions before they pass through the emulator To solve this issue the sensor nodes have been programmed so that each node transmits and receives on two different channels i e every sensor node uses two wireless channels for communicating with the network This means that before and after transmitting a sensor node must change its frequency channel In addition to this there is also the issue of how the emulator can identify a given transmission as originating from a certain sensor node and output that signal with different channel gains to different sensor nodes The issue was solved by assigning a connected sensor node its own unique pair of frequency channels for transmission and reception leading to a total of 16 frequency channels each with a bandwidth of 5 MHz being used by the emulator to mimic the behavior of over the air transmissions T
61. ssed by a Field Programmable Gate Array FPGA and finally output by a Radio Frequency Signal Generator RFSG allowing the sensor nodes to receive the transmis sions A description of the setup is provided here The data flow of the system set up is illustrated in figure 1 14 a amp dle Splitters Sensor Combiners nodes Emulator Ce FLAME Chassis le use Koenen gt Monitoring Control Signal Path Figure 1 Signal and data monitoring flow of the system setup Via separate interfaces the PC sets up the behavior of the emulator as well as any sensor nodes connected to it Both the emulator and the individual nodes are able to relay control information back to the PC Both the input signals to and the output signals from the emulator are combined into a single signal by the combiners The individual signals are extracted by tuning a receiver to the carrier frequency and bandwidth of the desired signal 3 1 Sensor nodes The sensor nodes used throughout the work described in this report has been the Zolertia Z1 4 depicted in figure 2 15 Figure 2 The Zolertia Z1 sensor node uses the IEEE 802 15 4 compliant radio transceiver CC2420 There is no reason to believe that the same results should not be achievable using any IEEE 802 15 4 compliant sensor node as long as precautions are taken to adapt for any differences in hardware The Zolertia Z1 is equipped with the IEEE 8
62. strength will decrease as the distance and path loss exponent i e the amount of obstruction between the transmitter and receiver increases As the received power decreases so does the receivers SINR and PRR This theoretical model is used in section 6 2 to verify the behavior of the system when a noise source is introduced 3 2 Peripheral equipment In order to connect the sensor nodes to the input and output of the emulator a set of coaxial antenna cables and a pair of power splitters combiners are needed The antenna cables are connected to the U FL connector of the sensor nodes which therefore must have its output capacitor soldered to connect an external antenna The two splitters combiners are connected back to back see figure 4 in order to fully interconnect the eight sensor nodes as well as the input and output of the emulator 19 Figure 4 The back to back connection of the power splitters combiners allows the simultaneous connection of up to eight sensor nodes to both the input and output connectors of the emulator The splitters combiners are both the same type of component but with opposite orientation When used as a combiner they are designed to take the superposition of signals from all ports on one side and place the combined signal on the single port on the opposite side Correspondingly when used as a splitter the input signal on the single port is output simultaneously on all output ports It is important
63. t data to and enable packet monitoring in the user interface the FPGA module also calculates the absolute values of the input and output signals and makes this information available to the LabVIEW host user interface NI PXIe 5673E Vector Signal Generator The NI PXIe 5673E Vector Signal Generator described in detail in 14 provides the output part of the emulator The digital signal received from the NI PXIe 7966R FlexRIO FPGA module is passed through a pair of digital to analog converters DACs The signal is reconstructed by feeding it through a low pass filter and up mixing it to a center frequency of 2 425 GHz NI PXIe 8360 Laptop Control module The NI PXIe 8360 Laptop Con trol module described in detail in 15 provides a completely transparent link between the emulator hardware and the control software running on a PC The PC and the chassis communicate through a PCIe link with a maximum sustained throughput of 214 MB s Note that all signal processing is performed within the emulator and the PC only provides the controls for properly starting the emu lator and setting the channel gain matrix as well as presenting the signal and channel data to the user 3 3 2 Transmitting through the emulator Using the emulator it is possible to emulate the physical layer of the wireless communications between up to eight connected sensor nodes There are how 23 ever a couple of issues in relation to the frequency channels used by the sen
64. task that requires extensive user testing For various reasons performing these tests has not been a part of this work instead potential end users have been allowed to provide feedback on the user interface and features as the work has progressed 7 1 1 Channel data import The new subVI created for importing data has been verified as working with data files that conform to the data format suggested in section 4 1 Since the fi nal specifications for this data format is yet to be determined the current design of the subVI has been implemented with a minimal amount of additional fea tures For example no safeguards have been implemented to ensure predictable behavior when attempting to parse data files that do not follow the expected data format Instead the aim has been to implement a minimal working subVI that can be further refined once the final specifications of the project have been clearly established 7 1 2 Export of resulting data As with the channel data import subVI the signal data export subVI has been verified to be working and outputs data files of the format outlined in section 4 2 Adding support to this subVI for the export of additional data should be trivial if the need to do so arises 7 2 Latency After investigating different ways of addressing the latency issue the problem was finally solved by altering the CX MAC protocols radio duty cycle to al low for a longer waiting time for transmission acknowledgments Al
65. the data file Any channel parsed from the data file can then be applied between any pair of sensor nodes by specifying the appropriate index in the channel control matrix in the main application More information on the channel data import subVI is available in appendix B 28 4 2 Exporting sensor signal data Another major feature that has been added is the ability to save sensor node input and output to file Signal data will be written continuously to a file specified by the user This file will be in the form of a comma separated list with sixteen values for each row Each row represents a point in time and each value represents the signal strength detected on this channel at this time When writing to an empty file the string values sensorOut0 sen sorOut7 and sensorInO sensorIn7 will be written in sequence to the first row of the file This is to simplify the import of the data file into Matlab or any similar piece of software for data processing The VI handling the output of sensor signal data could easily be expanded to log more data as new needs are discovered More information on the channel data import subVI is available in appendix C 4 3 General modifications For increased overview of the available controls and indicators the different parts of the host interface have been grouped by functionality into separate tabs This grouping provides a more intuitive way of finding relevant controls
66. the emulator while using the modified protocols Note that relative sizes of the different blocks are exaggerated to illustrate the nodes transmission behavior and do not accurately reflect the relative times it takes to perform the operations In the upper case node B wakes up and is instantly able to receive the preamble from node A and after replying with an acknowledgement node B also receives the data transmission In the middle case the emulators latency first causes a delay before node B is able to receive the preamble the acknowl edgement sent from node B also experiences the same latency and node A times out before receiving the acknowledgement Finally in the bottom case the preamble and acknowledgement are still affected by the latency of the emulator but the guard time introduced at node A ensures that node A does not start listening for the acknowledgement unnecessarily early A second test with higher resolution on the time interval 0 05 0 14 ms was also performed to further narrow down the optimum guard time of the acknowl edgement listening The results of this second test is plotted in figure 8 36 PRR as a function of the ACK time PRR Qe Reeeel Sateen bs Reed siweieis EIRENE poudes EEI Qpe PRR SEE TRE MEERES Jm PEA e 0 3 a aeemenhe Reed ies eed Reeeel salto one 0 05 0 06 0 07 0 08 0 09 0 1 0 11 0 12 0 13 0 14 ACK listen time delay ms Figure
67. tion file 3 Note Some categories are device specific For example the IVI category appears only if you have IVI installed For more information about using MAX select available help categories from the Help menu If you need further assistance or want to know more about your device visit the National Instruments Technical Support Web site For more information aboutthis version of MAX launch the readme or visit ni com info and enter the following Info Codes MAXFixList Improvements and bug fixes MAX53Knownlssues Known issues s Submit feedback on this topic a Visit ni com support for technical support Help Figure 15 In the Devices and Interfaces part of the Measurement amp Automa tion Explorer you should be able to identify the hardware components of the emulator e The remaining settings available on the Hardware Setup tab should be configured as shown in figure 16 52 Figure 16 Default values of the remaining hardware setup configuration options e Next go to the Channel Monitoring tab of the host interface e Select the Channel mode you wish to use e Except for the Read data from MATLAB file option all options are self configuring e f you choose to use the Read data from MATLAB file option you need to follow these additional steps Select a valid previously generated data file using the available Data file field If you are unsure of the ordering of your channe
68. use errors as the subVI attempts to interpret the available data It is however possible for the subVI to handle a data file which contains more fields than those proposed in section 4 1 provided the additional data fields are all stored after the expected data fields Any additional data parsed in this way will simply be discarded Designing the subVI in this way enables more flexibility of the file format used without any modification necessary to the subVI The final version of the subVI was tested by converting a number of available previously gathered data sets containing the type of channel information that would be used when performing tests with the emulator Results showed no issues with importing channel data 4 4 2 Exporting signal data The subVI that handles the output of signal data to file was continuously tested during development This testing highlighted the problem with exces sive amounts of data outlined in paragraph 4 3 1 and was the main reason for downsampling the signal before plotting and exporting the signal data It was also verified that the exported signal data could be imported into Matlab This gives the possibility for later review of signal data produced during experiments with the emulator 4 4 3 General modifications Objectively verifying the increased usability of the redesigned host interface is difficult without feedback from end users These end users should also ideally be familiar with both the pr
69. vely A time series of captured Received Signal Strength Indication RSSI values is stored in the rssi array This array can be of arbitrary size Finally the channelInfo field gives the ability to store any arbitrary informa tion that may be relevant to the measured data such as a verbal description of the channel s propagation path or a string that allows you to easily identify a specific channel Outside the array of structure arrays information relevant to the entire test site and the entire system of sensor nodes can be stored specif ically fs is a floating point value representing the sampling frequency in Hz used during the data collection siteInfo is a string containing description and miscellaneous information about the test site where the data was collected The site and all channel information data is kept within an outer structure array p This helps organize the data by gathering it in a single data object The application that parses the data file is implemented as a subVI inside the main LabVIEW application During the parsing the data is expected to be structured as described above and the subVI will keep parsing the data file for channel data until it reaches the end of file At that point the subVI will exit and make any data extracted available to the containing VI The main application will list all channels found in the data file and index them from 0 to N 1 where N is the number of channels successfully parsed from
70. vices 3 2 F NIR Series Multifunction RIO 12 0 4 NI Script Editor 1 34 F Ni System Configuration 5 3 3 gt W NiVision Run Time 2012 WF NI488231 5 NI DAQmx ADE Support 9 5 5 ES NI DAQmx Device Driver 9 5 5 S NI DAQmx MAX Configuration 9 5 5 NI DCPower gt um NI DMM 30 6 amp NLFGEN NI HSDIO 18 3 Ed NI Hws1 5 NEPAL29 gt aka NI RFSA gt Ji NI RFSG F Nr RIO 120 gt WB NI SCOPE J NI Serial 39 gt 8 NESWITCH 45 5 F NrSync 335 H NFTck19 2 Nr usi20 gt FXI NEVISA 5 2 3 NI VISA Runtime 5 2 F Traditional NI DAQ 744 A Vision Builder AI 2009 SP1 Ag Vision Builder AI 2011 SP1 D i MI Drivers gt Remote Systems Import export my device configuration file MAXFixList Improvements and bug fixes MAX53Knownlssues Known issues Figure 14 Use the Measurement and Automation Explorer to find installed software from National Instruments Also ensure that you have a complete version of the emulator host interface The complete version should include the following files e addOffset vi e channelSelectorReciprocality vi channelSort vi e config instruments vi emulatorHost vi generateChannelData vi e importData vi e saveSignalData vi as well as the FPGA bitfile ChannelSimulator PXle 7966R CHMatrix 8Ch lvbitx The VI files should all be placed in a single folder and will be imported into your LabVIEW library when you run the emulator Finally open the emulator host interface by loading the f
71. wireless channels and to realize the consequences these charac teristics have on communication performance However getting physical access to a site and thoroughly measuring the characteristics of the available wireless channels can be both time consuming and problematic Theoretical knowledge about the effect of various aspects of wireless com munications on the performance of the wireless network is ample However the complete interaction of all these aspects as well as the large scale effects of having several nodes interacting with each other is difficult to recreate reliably even when using advanced simulation tools An alternative to simulation is to create an emulation environment While a simulation models the behavior of the system to some degree of accuracy an emulation mimics the outward behavior of the system completely T he topic of physical layer emulation has been investigated previously A large scale setup of a physical layer emulator was created by The Emulator Team at Carnegie Mellon university Their findings showed that the performance measured in an emulated environment is more easily validated than the performance of a sim ulated environment and that the emulated environments aids the development and evaluation of enhanced wireless protocols 1 Another project at Uppsala university by Danielsson and Agren created a proof of concept version of an em ulator for the physical layer of the wireless communication of sensor

fulltext - DiVA Portal

Contents

Download Pdf Manuals

Related Search

Related Contents