Helilab User Manual: Human Behavior and Flight Data Acquisition
Contents
1. The first part is converting raw eye movement data collected by using converting programs provided by SeeingMachines Since you record participants eye movement data from individual eye tracking devices you will get 7 different files Eye fil Face fll Head fil HeadV2 fil Timing fil World fll and World V2 fil note file names can be changed based on the settings of faceLAB Basically you can transform raw data into txt files using Worldview program except WorldV2 fll see Fig 17 Since WorldV2 fil contains participants eye movement data we need additional converting program called logconverter see appendix B 2 Fig 18 shows about how to convert WorldV2 fl into txt file Since all transformation are done now_is the time to run an in house built transfer program called my Copter_DataTransfer written in MATLAB see appendix A 1 It automatically transfers all human behavior and flight data into DBMS at once and you only need to specify the name of txt files in myCopter_DataTransfer m file Detail information is written in the source code with comments 6 4 Human Behavior Data Analysis Eye movement data collected have to be processed by multiple functions programs You only need to run anal ysis_fixations m which runs all related programs annotation validation time format converting data reduction fixation identification instrument coord2. 5 which was sufficient for determining the specific flight instrument that was referred to that is fixated by the user s gaze at any given time In addition a dispersion threshold identification algorithm was applied to the raw eye movement data to specify the duration and location of meaningful fixations and to remove artifacts 3 Namely we removed eye movements that were faster than 60 sec and short fixations with a duration of less than 150 msecs a Tr a A graphic workstation de b Nine graphic workstations c Time synchronization server ployed behind a cockpit vertical stack and two ad ditional workstations for eye tracking devices Figure 2 Clusters of workstations for visualization eye tracking and time synchronization 4 Issues in Data Acquisition and Analysis In the current settings of our flight simulator in Helilab we can collect three data simultaneously eye movement behavioral data from two set of eye tracking devices control input data from two joysticks for position and throttle control and a set of rudder paddle sending input signals to Flightgear flight data from Flightgear Eye movement data is collected from two workstations linked to each eye tracking device and control input and flight data are collected and can be retrieved from a workstation with Flightgear in it As several independent systems collect different types of data time synchronization is crucial prior to data a
3. a flight simulator established in Helilab located in SpemannstraBe 38 de signed to investigate human behavior in cockpit scenarios The flight simulator is designed and deployed to allow participants the freedom to perform actions similar to the real world environment with high visual fidelity during recordings The current setting of a flight simulator in our laboratory allows us to measure both eye movement behavior and bio signal of participants but this paper focus on human eye movement behavior and flight data collection and analysis without any information of bio signal acquisition and processing This paper illustrates the layout of equipments and its details in section 2 and 3 Issues considered in data acquisition and analysis are described in section 4 The procedure of data acquisition and analysis are illustrated in section 5 and 6 and a simple demo is shown in section 7 The source codes for data analysis are attached in appendix A and supporting materials are listed in appendix B 2 Requirements First the experimental set up should allow the participant the freedom to perform actions similar to the real world environment For example his movements should be unrestrained This allows findings in the laboratory to be generalizable to the real world as well as to minimize irrelevant movement artifacts in the data Second it is important to provide high visual fidelity during recordings This would ensure that the user is referenc
4. physical measurement which Background images for SID calibration are provided in appendix B 4 Figure 9 Moving direction of each eye tracking device would not be sufficient enough to get accurate data Therefore you should conduct SID calibration for displays see Fig 10 SID calibration has to be done prior to each session of recording to get maximum accuracy Another issue is about the usage of faceLAB LINK faceLAB LINK 1s a software to collect more accurate data in large visual field covered by multiple eye tracking devices see appendix B 3 However we do not recommend to use faceLAB LINK in your experiment in Helilab faceLAB LINK is designed to give better precision during eye movement retrieval with multiple eye tracking devices but it seems only works with a single large screen or multiple displays consists a single plane In our experience faceLAB LINK does not guarantee us to collect eye movement data with sufficient level of accuracy We recommend you to independently collect human eye movement behavior data using two eye tracking devices one deployed on the bottom of an instrument panel display for collecting eye movement data on 9 displays in front of a cockpit and one another device for acquiring data on an instrument panel Figure 10 Calibration menu on built in Worldview software Tn our experience it is required to conduct SID calibration on center center display with an eye tracking device deployed on the
5. provided with this paper e MATLAB R2010b or any release above R2010b e Database toolbox e MySQL b Cockpit side view c Cockpit front view d Two eye tracking devices Figure 1 A flight simulator with eye tracking devices 3 Specification 3 1 Realistic real time simulation of a world environment with a large field of view Eye movement research is ideally conducted in an environment that affords high visual fidelity In our set up nine screens each 102 x 57 cm 1366 x 768 pixels comprised a multi panelled display that afforded a field of view of up to 105 x 100 visual angle see Fig 1 a In addition a dedicated screen provided the user with a heads down flight instrument panel Graphical rendering on all displays were performed simultaneously with clusters of 10 graphics workstations see Fig 2 We employed a flight simulation software Flightgear 2 to simulate the flight dynamics of a Bo105 helicopter to compute the appropriate graphics for the external world environment and the flight instruments 3 2 Unrestrained gaze tracking Two remote eyetrackers faceLAB 5 SeeingMachines are deployed to record eye movements on the flight instrument and front displays see Fig 1 d This system computed the gaze vector of the user in the world environment and from this recorded the intersection of the user s gaze with the flight instrument panel at 60 Hz Eyetracking could be performed to the accuracy of 1
6. top of an instrument display and one additional SID calibration on the instrument panel display with an eye tracking device deployed on the bottom of an instrument display 5 3 Flight Data Acquisition In this section we describe about how we setup a flight simulation program 1 e Flightgear to collect flight data and the procedure of recording in a flight simulator 5 3 1 Settings for Flight Data Acquisition There exist two ways to collect data from Flightgear simulation The first method is to record flight data by clicking record menu after starting simulation and the second way is making a batch file with a XML file enables us to collect data automatically whenever we start Flightgear In current version of Flightgear you only can record less than 10 parameters from the simulator since you use the first way The advantage of the second way is that you can record as many parameters as you want by scripting XML file contains the list of parameters you want to record We recommend you to use the second way and you need to make one batch file and one XML file to use the second way of recording flight data E WilightGearWG_bo 0 H_log_ JV bet Motepads Fe Ch Seach Mew Decido Langues Settings Mayo Bum Pigs Wirga F a Es e os a 4 CRIER FG bot Hea ba Peel fgfs dir bin Winds ots exe si wun Bact Beadinger s Le Bact Pie Este 0 Suet wotisec 0 et potisec s0 4 SEEsehe 1 15 F a i Y I J L be gta glade pa taai
7. Behavior Data Acquisition Following instructions illustrated in a manual offered by SeeingMachines see appendix B 3 you setup your eye tracking system The first step is to input coordinate information of two eye tracking devices in faceLAB software see appendix B 1 As shown in Fig 5 you can set relative coordinates of 10 displays from each eye tracking device The coordinate information is the physical basis of processing eye movement so the information has to be accurate as possible In our flight simulator we physically measured the location of each device using a laser distance meter see Fig 6 Remaining issue is that the coordinate information is about relative distance between each eye tracking device and displays so the information have to be updated whenever the location or position of eye tracking devices are changed Although you did not move eye tracking devices the coordinate information of each eye tracking device can be changed by the body movement of participants or incidents Therefore we need more reliable solution to update coordinate information updating information by physically measured relative distance among devices prior to each recording session would be exhaustive teed Tracking Eve tacking E Create Auto Land Hiii D Create Marna CPU Las 0 Head Trading Confcence ju Lefi Eye hase Tuak Ler Pn EE EE a i I i i i i i i j i i i i A p ar boo D f i WW e I 3 j Frame Rote P
8. C Run the program and you need to setup ODBC steps for setting anew ODBC connection are illustrated in Fig 16 aloce aR BEY p mi MA COE Ba CH mn AT OSN AABE DSN r DSN cam one ag 2e MEM COE HEU AD HOE AES MUR See MIS ARCS NE Sere 0E HZ 1 Ab mae MySOL GMAC E 1 Driver COTES nA WywEOL AB waecapl hMySUL ODAC 5 0 Driver F SOL Native Client 200a 2 00 Mics gso Connor SOL Server 6 01 760117514 Microgott Como un OODE MEN NOE SEE HAE UMEL AIRIS DB widu mA FE EU MER Abit EE ASE EU AN AE Be mE ca Ore Ht a Step 1 start a new ODBC setting by click b Step 2 select MySQL ODBC connector ing Add DY button r E2 MySQL Connector ODBC Data Source Configuration E COEC Wo 3m HE muse AHSA DEN AAH 06N e DSN Sam a ge ae ae Connector ODBC MS GOE SEU E DM Cat a aac mycopter MySQL ODBC 5 1 Driver Data Source Name ycopter ree MySOL ODBC 5 1 Driver rosecapt MySQL ODBC 5 1 Driver Description TCP IP Server localhost Port 3306 Named Pipe User root GPEC MET HOI SRE URL HQE TITZ DE wash MA Password 90600000 WI TAE QUO AFEA UIE SES ARENEB amp SE A ASH ga Ha sua cow c Step 3 input relevant information to d Step 4 new ODBC setting is created setup ODBC settings Figure 16 ODBC Settings 6 3 Pre processing Pre processing consists of two parts see Fig 14
9. Max Planck Institut fur biologische Kybernetik Max Planck Institute for Biological Cybernetics m MAX PFLANCR LESELLSCHA FT Technical Report No 196 Helilab User Manual Human Behavior and Flight Data Acquisition and Analysis Jung Jin Lee and Heinrich H B lthoff May 2012 This Technical Report has been approved by Director at MPIK Postdoc at MPIK Max Planck Institut fur biologische Kybernetik Max Planck Institute for Biological Cybernetics m MAX PFLANCR LESELLSCHA FT Technical Report No 196 Helilab User Manual Human Behavior and Flight Data Acquisition and Analysis Jung Jin Lee and Heinrich H B lthoff May 2012 Department B lthoff email jung jin lee heinrich buelthoff tuebingen mpg de This report is available in PDF format at http www kyb tuebingen mpg de fileadmin user_upload files publications 201 2 tr 196 pdf The complete series of Technical Reports is documented at http www kyb tuebingen mpg de techreports html Helilab User Manual Human Behavior and Flight Data Acquisition and Analysis Jung Jin Lee and Heinrich H Biilthoff Abstract We here introduce a flight simulator dedicated to collect both human behavior and flight data with high fidelity This paper illustrates the current settings of the flight simulator key issues in using the simulator and the standard procedure of data acquisition and analysis 1 Introduction This paper describes the settings of
10. bled M configure Time Service Time Senice f Use NTP fecommended C Use Windows Time Confiourals This computer is an NTP server Synchronise time with an NTP server Server myrtesarver BE oa a Configuration window b Status window Figure 4 NTP FastTrack configuration and status monitor adapted from Guide to Use NTP FastTrack by SeeingMachines Since you have a server and workstations to sync then you need go to configuration menu of NTP FastTrack and specify whether a machine is an NTP server or not see Fig 4 a You also need to set NTP FastTrack to use Windows time not NTP since the flight simulator is set as a close network only accessible via a network hub device deployed in Helilab Since you set whole NTP FastTrack programs in each machine then you need to wait until jitter is minimized see Fig 4 b According to the manual given by SeeingMachines see appendix B 3 it takes around 24 hours for initial time synchronization but it took only several hours to be stabilized in our experience During recording session time clocks on each computer have to be synchronized so you should have to check the status monitor shown in Fig 4 b on each computer prior to your experiment 5 2 Human Behavior Data Acquisition In this section we describe about how we setup devices to observe eye movement of participants and what is the standard procedure to collect human behavior data 5 2 1 Settings for Human
11. cquisition 7 2 1 Human Behavior Data Acquisition with a Simple Flight Plan with Six Different Tasks Human behavior can be divided into two parts in our system eye movement data and flight control data Eye movement data can be acquired by using faceLAB 5 and flight control data can be collected from Flightgear see appendix B 3 and section 5 3 The important issue in collecting data from faceLAB 5 is annotation During the recording you can annotate and the annotation can be used to separate data based on tasks You have to make sure how to assign annotation ID for each task MATLAB codes provided in appendix A 2 shows how to assign and process annotation ID each straight and level flight are annotated 1 and each designated tasks are annotated 12 Another issue in annotation is input missing In our experience pressing a key or clicking mouse sometimes does not recorded by faceLAB 5 By allowing multiple pressing and clicking can help to reduce the probability of losing annotation information The codes provided in appendix A 2 allows multiple annotation input so you can click or press a key multiple times for each task 15 7 2 2 Flight Data Acquisition with Flight Plans with Three Difficulty Levels Flight data acquisition is simple Since you set parameters to log by modifying a XML file you only need to click a batch file liked to the XML file see section 5 3 1 When you terminate Flightgear simulation logging is also st
12. cquisition The method for time synchronization 1s illustrated in section 5 1 and the procedural steps of each data acquisition for human behavior Details are illustrated in appendix B 3 Multiple Monitors in Flightgear The flight model can be easily changed to the other model based on the purpose of an experiment For more information please visit the website listed in appendix B The criteria to identify fixations can be adjusted by changing parameters in MATLAB source code given in appendix A data 1 e eye movement data and flight data are duly described in section 5 2 and 5 3 Remaining issues are about data analysis Time series data of eye movement and flight simulation are restored in a special format as given from the vendor of commercial products The vendor dependent format can be converted to txt file for convenience in raw data analysis but it is not an efficient way of observing and analyzing data collected Therefore we developed a data management program with a standard procedure of data collection using MATLAB and MySQL which is a well known and widely used open source database management system Details are illustrated in section 6 5 Data Acquisition Knowing the system architecture of a flight simulator deployed in Helilab will help you to understand contents illustrated in this section As shown in Fig 3 multiple individual systems are set for different purpose Whole system architecture can be divided
13. e appendix B 2 You do not need to change any settings during installation but follow default setting steps Make sure to remember user identification ID and password for MySQL The second step is to create a database You need to open MySQL console and login Using SQL command create database you can create a new database see Fig 15 Remember the name of newly created database in our settings we named the database mycopter This will be used in the next step of ODBC setting E MySQL Console root B x Enter password 2666666666 Welcome to the MySQL monitor Commands end with or Wg Your MySQL connection id is 706 Server version 5 1 41 community MySQL Community Server CGPL Type help or Wh for help Type Wc to clear the current input statement mysql gt create database mycopter Query OK 1 row affected 1 sec mysql gt Figure 15 Create a new database using MySQL console 11 The third step is to install MySQL ODBC connector see appendix B 2 ODBC connector enables us to access DBMS Download and install MySQL ODBC connector After installation you need to set ODBC in Windows control panel By clicking start button of Windows you will find a single line command window for search programs and files on top of start button Since you type ODBC you will find a program located in Control Panel called Set up data sources ODB
14. e easily achieved by using NTP FastTrack You can download the software via URL given in appendix B 2 and you also can download the manual of NTP FastTrack see appendix B 3 We briefly illustrate issues need to be considered in using NTP FastTrack to synchronize clocks of each computers The first thing you have to do is to download NTP FastTrack program and install the program on each computer you want to synchronize In our settings we basically need to install the program in three workstations two workstations linked to each eye tracking device and one graphics workstation has Flightgear in it Then you need one additional computer as a NTP server You might think that it would be possible to use one of three computer as a NTP server but it is not a right way to set up NTP synchronization program During an experiment three workstations constantly collect human behavior and flight data which definitely require a lot of computing resource If you set one of the computer as a NTP server the lack of computing power will highly affect to both synchronization duration and the accuracy of time synchronization and you will never get synchronized time stamp with minimum jitters which is necessary in our experimental purpose It would be a wise decision to make an additional NTP server only dedicated in time synchronization zix Tixi Time Configuration Startup Time Sevice Windows Time NTP Service Status Not Running NTP Configuration NTP is not ena
15. ed 7 Tite Trur is in at Tres La fam a P Tr rs True iif SEgte dirt ere Eee tt GiPeratl aircral Ble pe mba Rimes Ed fe tl imrok day rien Cr pDESpis ste View O Poet is hee propi aim view Oo poni ig picckh otieec deg pitched ierit prop fein view pr tnntioets preps wuld oont ig y ofteer meSyefiierts prop faim view i rie Te htotierl tepec 80 EST ie apne Le 0 fi BO he liini 5500 ap Bat SRE Poe SS REE ot 2 he Lisi 5500 6 L Een po peongoed a Flightgear folder b A batch file with the location information of a XML file Figure 11 The location and the script of a batch file As shown in Fig 11 you have to make a batch file In the end of the batch file you need to specify the name of a XML file which the batch file will be used as shown in Fig 11 b The next step is to make a XML file which contains the list of parameters of interest As shown in Fig 12 the location of a XML file has to be match to the information written in the batch file illustrated above The last step is to make folder to restore a log file The folder name and location are pre defined in your batch file and you need to manually make the folder in a right location see Fig 13 Flightgear starts to record flight data in the log file since you start recording by clicking the batch file you made and the log file will be saved inside of the pre defined folder as shown in Fig 13 b 5 3 2 Recording Proc
16. edure Since you successfully made two files as illustrated above it is very simple to collect flight data You only need to click the batch file you made and Flightgear will start to run flight simulation and record flight data Recording will be automatically stopped since you switch off Flightgear 6 Data Analysis Since you collect data from multiple recording devices you need to transform data into desirable form enable to effectively manage and analyze data Before starting analysis on human eye movement and flight data we first need to install database management system DBMS to restore data Using database toolbox we will show how to retrieve data for analysis in MATLAB Pre processing is also required to process raw data prior to transfer the data into DBMS The routine of data processing 1s illustrated below M A batch file and an XML file are provided in appendix B 4 The full list of parameters can be found in Flightgear menu on top of the screen since you start Flightgear E p liphiGear deta probal e nl Hole pod Pis DE Bwt Ven Dwpdrg Language betting Mag Eun ARS hR ac sys Fel eapon res moid amil el versions i 0 te Cthrsp ererliser A lt logqieg Loge lt esab Leds litte cna Lede G2 s lenient Log on fig log carci 1 less Cheterve e es inber val mee ide limiters delimiters Fi ON ee ee Pair D nes forums F POSITION DATA JESE di micicuds shove ground level t gt cena Led gt re f
17. era led Coot lesa todeaglft lt tin lex plopecl ys pet ienfaltitede sagl t LC property entry iantry lt t altitwis Et gt Cer Leds trit enadi edh TEA LeeLee ES Lae Lee lt prepert ys pent tionfaltitede t lt property TL Ca Lies ri z jaritide deg Cer Leds ptit cab Lede lt ritieslatitede degefs it les Arable Marti Language A barge 000 lined 7 Wart Case fara a A folder which has a XML file in it b A XML file with the list of parameters of interest Figure 12 The location and the script of a XML file Frit Om 8 Address D FightGearilogi fig log csv CSV Fie 997 KB a A target folder where the log file b A log file which contains flight will be placed in data Figure 13 The location of a log file recorded by Flightgear 10 6 1 Data Process Routine Fig 14 shows the routine process of data transform and transfer Details are illustrated in section 6 3 and DBMS installation required prior to pre process is described in section 6 2 Two Workstations Linked to Eye tracking Devices Vendor Provided Transform Program In House Transfer Module One Graphics Workstation with Flightgear Figure 14 Data Process Routine 6 2 Settings of Database The section is about installation of MySQL More details can be found in the official website of MySQL and here we briefly describe about the installation First you need to install MySQL se
18. fixations m e data_reduction in house script uni home projects helilab data_reduction m e fixation_identifier in house script uni home projects helilab fixation_identifier m e gmt_serial_converter in house script uni home projects helilab gmt_serial_converter m e instrument selection in house script uni home projects helilab instrument selection m 16 0 200 400 BOO o0 1000 1200 1400 1600 a Fixations during straight and level flight 300 HOU Torque Compasstwo Compasstwo V Speed 00 BOO S00 400 300 200 100 i 200 400 e00 a00 1000 1200 1400 1600 b Fixations during left turn Figure 22 Fixations during flight maneuvers 17 a Trajectory of an airplane with flight plan 1 b Trajectory of an airplane with flight plan 2 c Trajectory of an airplane with flight plan 3 Figure 23 Trajectory of an airplane with multiple flight plans 18 A 3 B 1 B 2 Figure 24 Trajectory of an airplane with multiple flight plans plot_fixations in house script uni home projects helilab plot_fixations m Flight Data Analysis PlotTrajectory_SingleFlightPlan in house script uni home projects helilab PlotTrajectory_SingleFlightPlan m PlotTrajectory_MultipleFlightPlans in house script uni home projects helilab PlotTrajectory_MultipleFlightPlans m Appendix Supporting Materials Commercial Software MATLAB http www mathworks com products matlab MATLAB Database T
19. igures can be achieved by running two programs will be illustrated in section 7 with designated flight plans as demonstration 13 7 Demo A Flight Plan with Basic Maneuvering Tasks 7 1 Flight Plan Design We design two sets of flight plan a simple flight plan with six different tasks and flight plans with three difficulty levels A simple flight plan is used to demonstrate human eye behavior acquisition and analysis and multiple flight plans are used to show how to acquire flight data and analyze it 7 1 1 A Simple Flight Plan with Six Different Tasks With the assistance of an expert helicopter pilot ex Military 650 flight hours we developed a flight scenario that consisted of the following basic maneuvering tasks ascending descending acceleration deceleration and turns see Fig 19 Figure 19 A flight plan SL straight and level flight AC acceleration DC deceleration AS ascending DS descending LT left turn RT right turn Such maneuvers are comparatively easy to execute and are within the capabilities of a non expert pilot with experience in flight simulators Our expert pilot recommended the primary flight instrument that he typically relies on for the execution of each basic task in bold as well as what he considered to be secondary instruments see Table 1 Primary Information Altimeter SOOft Acceleration Speed indicator 70 to 100knots Compass East s00r Deceleration Speed indicator 100 to 70
20. inate mapping and fixation plot 12 l une em anA M M um 000000 Figure 17 Transform raw data into txt file a T EFT a Vendor provided logcon b Converting raw data to txt file verter program with a raw data by dropping raw data on top of logconveter program icon Figure 18 How to use logconveter program e analysis_fixations e annotation_validation e data_reduction e fixation_identifier e gmt_serial_converter e instrument_selection e plot_fixations As the result of processing you can discriminate fixations from raw data and you also can use this information to calculate statistics of fixation w r t fixation coordinate on an instrument panel Detail instructions are illustrated in source codes see appendix A 2 6 5 Flight Data Analysis Flight data collected from Flightgear includes many information including trajectory of an airplane As illustrated in section 5 3 1 you can record many different parameters from Flightgear as much as you want by modifying a batch file and a XML file We here provide two different simple programs to visualize flight trajectory e PlotTrajectory_SingleFlightPlan e PlotTrajectory_MultipleFlightPlan These two programs not only show you the trajectory of an airplane with single or multiple flight plans but also let you know how to use flight data restored in DBMS see appendix A 3 The f
21. ing the appropriate information with his eye movements Finally it is crucial to formally define the task routines under investigation in terms of the specific actions and visual information that are expected to be necessary for successful task completion Failing to do so could yield eye movement data that is difficult to interpret meaningfully and in a principled fashion Thus a flight simulator was established in our laboratory with the following attributes e realistic real time simulation of a world environment e large field of view e unrestrained gaze tracking e configurable control interface In section 3 we illustrated the specification of current settings in Helilab considering the requirements above Helilab room no 026 is locate at the inside of Panolab room no 025 on the ground floor The current work is performed within the context of myCopter a project funded by the European Union under the 7 Framework Programme to research technologies that will enable the use of personal aerial vehicles for daily commuting 1 One part of this project focuses on measuring and modeling behavior of a consumer pilot in active flight control tasks Details in the requirements and specification of a flight simulator in Helilab are illustrated in section 2 and 3 2 1 Additional requirements for data processing All source codes for data analysis are written in MATLAB The followings are the requirements in using MATLAB source code
22. into two categories visualization and data collection Ten graphics workstations are dedicated to visualize scenery during flight simulation and two workstations are dedicated to collect eye movement data One of the graphics workstations both work in visualizing a control panel and collecting flight data Human eye movement data and flight data are collected in different independent systems so the time stamps of each data have to be synchronized Time synchronization server is dedicated in time synchronization using NTP FastTrack and the detail information on how to synchronize clocks of each workstation is described in the following subsection below The standard procedures of data acquisition for human behavior data and flight data are also illustrated in this section Nine Graphics Workstations Two Workstations Linked to Eye tracking Devices yg serait ay RE with Flightgear Time Synchronization a Server Figure 3 A conceptual system architecture of a flight simulator in Helilab Black lines represent I O connection between computer and displays and or input devices Gray lines show network connectivity among server and workstations Note this is a conceptual system architecture so the physical organization of devices in a cockpit simulator can be different from the figure given See appendix B You can also download official manual of NTP FastTrack via URL given 5 1 Time Synchronization Time synchronization can b
23. ir Pead hiodel Con Right Eye Gene Quality Level o rr I I Li i Li L F i I 1 I P I D a Cal D JF 49 7 W D 1 J 1 a World Model menu enables to open built in Worldview program in faceLAB b World Model settings of an eye tracking device c World Model settings of an eye tracking device deployed on the top of an instrument panel deployed on the bottom of an instrument panel Figure 5 World Model settings with built in Worldview software i gt a A flight simulator color lines b A color line assists to posi c A color line assists to posi on the floor are used to calculate tion a monitor deck on the left tion a monitor deck on the right coordinate information of each side pointed with a yellow ar side pointed with a yellow ar device row row Figure 6 The organization of devices in a flight simulator The solution we propose is using a fixed reference point based distance measure We set a reference point on the top of an instrument panel display and it is used as the pivot point of physical distance measure see Fig 7 and 8 There are four important coordinate information used to measure distance from the reference point the center point of each eye tracking device the center point of a top center display and the center point of an instrument panel display Using fixed relative distance from the top center display you can calculate relative distance from the reference point to 9 displays In this setting
24. izen and Lewis L Chuang for their assistance in setting up the flight simulator and discussions References 1 F M Nieuwenhuizen M Jump P Perfect M D White G D Padfield D Floreano F Schill J C Zufferey P Fua S Bouabdallah R Siegwart S Meyer J Schippl M Decker B Gursky M H finger and H H B lthoff myCopter Enabling technologies for personal aerial transportation systems In Proceeding of the 3rd International HELI World Conference 2011 pages 1 8 11 2011 2 A R Perry The flightgear flight simulator In 2004 USENIX Annual Technical Conference Boston MA 2004 3 D D Salvucci and J H Goldberg Identifying fixations and saccades in eye tracking protocols In Proceedings of the 2000 symposium on Eye tracking research amp applications pages 71 78 ACM 2000 21
25. knots Compass East Altimeter 400 to 800ft kca Vertical speed indicator 500ft min Speed indicator Compass East Altimeter 800 to 400ft De Vertical speed indicator 500ft min Speed indicator Altimeter 800ft Speed indicator 100knots Turn Left Artificial horizon Altimeter 800ft Speed indicator 100knots Turn Right Speed indicator B North to Eas Artificial horizon Table 1 Different Flight Phases bolded instruments represent the major instrument needed to accomplish each task as defined by an expert pilot 14 7 1 2 Flight Plans with Three Difficulty Levels We design new flight plans with different difficulty levels Using three designated point as illustrated in Fig 20 we construct three different flight plans Detail information of three locations is provided in General Information of Three Locations docx see appendix B 4 Tasks consists each flight plan are illustrated in Time Altitude Chart of Flight Plans pdf and New 3 Column Flight Phases docx see appendix B 4 Annotation is also necessary to separate data based on task and Basic Commands to Record Annotation Ds docx and Key Combinations Correspond to Task docx are ready for the annotation For your convenience during recording check list is also given as Flight Plan Experimental Purpose docx see appendix B 4 LC Figure 20 Aeronautical map for flight plans with three difficulty levels 7 2 Data A
26. ocument as a supplement uni home projects helilab General_Information_of_Three_Locations docx Time Altitude Chart of Flight Plans in house document as a supplement uni home projects helilab Time Altitude_Chart_of_Flight_Plans pdf Basic Commands to Record Annotation IDs in house document as a supplement uni home projects helilab Basic_Commands_to_Record_Annotation_IDs docx Key Combinations Correspond to Task in house document as a supplement uni home projects helilab Key_Combinations_Correspond_to_Task docx New 3 Column Flight Phases in house document as a supplement uni home projects helilab New_3 Column_Flight_Phases docx Experimental Purpose Flight Plan in house document as a supplement uni home projects helilab Flight_Plan_ Experimental_Purpose docx Plane_Calibration in house image file as a supplement uni home projects helilab Plane_Calibration png Control_Calibration in house image file as a supplement uni home projects helilab Control_Calibration png 20 Acknowledgements This research was supported by the myCopter project which is funded by the European Commission under the 7 Framework Program Heinrich H Biitholff was also supported by the WCU World Class University program through the National Research Foundation of Korea funded by the Ministry of Education Science and Technol ogy R31 10008 The authors thank Jakob Kolb Daniel Riedel Hans Joachim Bieg Joost Venrooij Frank M Nieuwenhu
27. oolbox http www mathworks com products database TM faceLAB 5 http www seeingmachines com product facelab Open Source Software Free wares and Scripts TM MySQL http www mysql com downloads mysql MySQL Connector ODBC http www mysql com downloads connector odbc NTP FastTrack http www seeingmachines com support downloads 8 19 B 3 B 4 Worldview http www seeingmachines com support downloads 9 LogConverter uni home projects helilab logconverter exe WorldModelMaker in house script uni home projects helilab WorldModelMaker xlsx World Model for an eye tracking device on top in house script uni home projects helilab World_Model_FL1 w3d World Model for an eye tracking device on bottom in house script uni home projects helilab World Model FL2 w3d Manuals and Unpublished Reports Multiple Monitors in Flightgear Quick and Dirty http www inkdrop net dave multimon pdf Guide to Use NTP FastTrack http www seeingmachines com downloads NTP_FastTrack_Guide pdf faceLAB User Manual uni home projects helilab faceLAB User Manual pdf faceLAB LINK User Manual uni home projects helilab faceLAB LINK User Manual pdf Additional Documentations and Files Bo105 Log in house batch file as a supplement uni home projects helilab boi05 log bat Log Protocol in house XML file as a supplement uni home projects helilab log_protocol xml General Information of Three Locations in house d
28. opped automatically 7 3 Data Analysis 7 3 1 Human Behavior Data Analysis with a Simple Flight Plan with Six Different Tasks As illustrated in section 6 4 we provided several MATLAB program to process and visualize fixations during flight maneuvers Fig 21 shows the layout of instrument panel and Fig 22 shows fixations on the instrument panel by processing raw data using analysis_fixations 4 i 140 7 20 120 100 EMOTS 40_ F 80 oo Wi J a x Figure 21 The layout of instrument panel of a Bo105 model from Flightgear the red circle in the center of the panel represents the 1 5 resolution of the eyetracker not shown during experiments 7 3 2 Flight Data Analysis with Flight Plans with Three Difficulty Levels As described in section 6 5 we provided two MATLAB program to visualize trajectory of flight Fig 23 shows the trajectory of an airplane with each flightplan by executing PlotTrajectory_SingleFlightPlan If you want to see all trajectories at once you can use another program PlotTrajectory_MultipleFlightPlans see Fig 24 A Appendix MATLAB Source Code A 1 Pre processing e myCopter_DataTransfer in house script uni home projects helilab myCopter DataTransfer m A 2 Human Behavior Data Analysis e analysis_fixations in house script uni home projects helilab annotation_validation m e annotation_validation in house script uni home projects helilab analysis
29. s the relative distance from the reference point to 10 displays are fixed since a cockpit module is not moved Since you match the position of the cockpit module to lines drawn on a floor you can get an accurate physical measure However you still need to calculate the relative physical location of the 10 displays w r t the position of eye tracking devices For your convenience we made a simple spreadsheet using Microsoft Excel see appendix B 2 WorldModelMaker You only need to measure an angular distance information from two eye tracking devices from the reference point current World Model for each eye tracking devices are given in appendix B a Relationship between a reference b Relationship among displays the length of yellow lines rep point and key devices the length of resents the distance from top center display to center center and yellow lines represent the distance bottom center display and the length of green lines represents from major four devices to the refer the relative distance among displays on a same horizon ence point Figure 8 Relationship among devices in three dimensional space 5 2 2 Recording Procedure The procedure of eye movement data acquisition is duly described in faceLab user manual see appendix B 3 There are two addition issues we need to address in this section SID calibration and faceLAB LINK Without SID calibration eye movement data collected by using faceLAB will be fully relied on
Download Pdf Manuals
Related Search