Driver Behavioral Analysis in Heterogeneous Traffic Conditions

50 %
50 %
Information about Driver Behavioral Analysis in Heterogeneous Traffic Conditions

Published on October 27, 2016

Author: VibhanshuSingh5

Source: slideshare.net

1. Vibhanshu Singh Supervisor: Dr. Caleb Ronald M Driver Behavioral Analysis in Heterogeneous Traffic Conditions Tathva 2015 Department of Civil Engineering, NIT Calicut

2.  Introduction  Traffic Data  Driver Behavioral Analysis  Conclusion 22 Overview of the presentation CED 2

3. 3 Introduction Why traffic simulation? Core models of a traffic simulator Traffic simulation Homogeneous and Heterogeneous traffic Objectives

4.  Dynamic representation of real world traffic systems by building a computer model (Drew 1968)  Driver, vehicle, environment infrastructure 44 Traffic simulation CED 4

5.  Understand the complex dynamic interactions between the driver, the vehicle, and the environment  Design of traffic systems  Operational quality of the existing facilities  Understand when and how traffic jams emerge?  Road safety 55 Why traffic simulation CED 5

6. 66 Core models of a traffic simulator CED 6  Car-following  Lane-changing

7. 77 Homogeneous & Heterogeneous traffic CED 7 Homogeneous • Mostly single vehicle type • Strict lane discipline • Well defined surrounding vehicles to apply the traditional car-following and lane-changing models Heterogeneous • Different vehicle types • No lane discipline – Irregular position of vehicles across the roadway • Suitability of traditional car-following and lane- changing models???

8. 88 Homogeneous & Heterogeneous traffic CED 8 AMERICA INDIA When Normal When Drunk When Normal When Drunk

9.  To obtain the trajectory data of vehicles  To understand the longitudinal movement behavior of drivers  To analyse the lateral movement behavior of drivers  To analyse the non-lane disciplined movements of drivers 99 Objectives CED 9

10. 10 Traffic Data Trajectory dataset Data processing Site details

11.  Data collected from a 245m long uninterrupted stretch in Chennai, India  30 minutes data from 2:45 PM - 3:15 PM  Flow of 6010 veh/hr  Data source - Kanagaraj et al. (2015) http://toledo.net.technion.ac.il/mixed-traffic-trajectory-data/ 1111 Site details CED 11

12.  Trajectories of 3005 vehicles extracted at a resolution of 0.5s generating a total of 111,629 observations 1212 Trajectory dataset CED 12 Vehicle Number Vehicle Type Time (sec) Length (m) Width (m) Long. Distance (m) Long. Speed (m/sec) Long. Acc (m/sec2) Lat. Distance (m) Lat. Speed (m/sec) Lat. Acc (m/sec2) 1 1 12.00 1.80 0.60 116.97 6.87 -0.09 2.81 -0.15 -0.16 2 1 12.00 1.80 0.60 23.46 5.50 0.38 2.49 0.22 -0.35 1 1 12.50 1.80 0.60 123.43 6.79 -0.18 2.59 -0.14 0.16 2 1 12.50 1.80 0.60 28.62 5.47 0.83 2.50 -0.04 -0.12 1 1 13.00 1.80 0.60 129.82 6.01 -1.06 2.55 -0.01 0.11 2 1 13.00 1.80 0.60 34.72 5.45 -2.22 2.41 0.04 0.26 3 1 13.00 1.80 0.60 33.53 10.34 -0.91 4.97 0.04 0.10 1 1 13.50 1.80 0.60 135.17 5.45 -0.48 2.58 0.07 0.07 2 1 13.50 1.80 0.60 38.52 3.81 -0.16 2.59 0.26 0.17 3 1 13.50 1.80 0.60 43.14 8.84 -2.65 5.06 0.05 -0.07

13. Algorithm for processing the raw data file start read the trajectory data file for all entries get the vehicle number and vehicle type get the time sort according to the vehicle number, vehicle type and time end for for all vehicle types read the lateral shift threshold values for all vehicle types get cumulative differences of lateral coordinates if cumulative difference > threshold value and >0; count as right movement if cumulative difference > threshold value and <0; count as left movement get the corresponding lateral movement durations end for end 1313 Data processing CED 13

14. 14 Driver Behavioral Analysis Longitudinal movement behavior Lateral movement behavior Vehicle composition Lateral movement duration distribution models Non-lane discipline behavior

15. 1515 Vehicle composition CED 15 Vehicle Type Share (%) Motor Bike 56.4 Auto Rickshaw 12.2 Passenger Car 26.6 Heavy Vehicle 04.8

16. 1616 Longitudinal speed CED 16 ANOVA Test Results P-value: 0.00 Null Hypothesis: Reject

17. 1717 Longitudinal Accl. and Decl. CED 17 ANOVA Test Results P-value: 1.62E-12 Null Hypothesis: Reject ANOVA Test Results P-value: 5.00E-59 Null Hypothesis: Reject

18. 1818 Lateral speed CED 18 ANOVA Test Results P-value: 0.00 Null Hypothesis: Reject

19. 1919 Lateral Accl. and Decl. CED 19 ANOVA Test Results P-value: 6.00E-161 Null Hypothesis: Reject ANOVA Test Results P-value: 1.50E-60 Null Hypothesis: Reject

20. 2020 Frequency of lateral movements CED 20

21. 2121 Lateral movement duration CED 21 ANOVA Test Results P-value: 4.96E-79 Null Hypothesis: Reject

22. 2222 Lateral movement duration models CED 22 Lognormal (0.55, 0.62)

23. 2323 Lateral movement duration models CED 23 Lognormal (0.98, 0.70)

24. 2424 Lateral movement duration models CED 24 Lognormal (0.94, 0.74)

25. 2525 Lateral movement duration models CED 25 Lognormal (1.25, 0.50)

26. 2626 Non-lane discipline CED 26

27. 27 Conclusion Conclusions Contributions Summary Future work

28.  Proposed an algorithm to process the raw trajectory data of vehicle and to obtain various driver behavioral characteristics  Analysed the longitudinal movement behavior of drivers in heterogeneous traffic conditions  Analysed the lateral movement behavior of drivers  Developed lateral movement duration distribution models  Analysed the non-lane discipline characteristics of heterogeneous traffic streams 2828 Summary CED 28

29.  Significant differences exist in the driving behaviors of various vehicle type drivers  Motorbike drivers are more aggressive in their behavior  Heavy vehicle drivers are conservative and mostly homogenous in their behavior  Lane discipline is not observed by the drivers in heterogeneous traffic conditions 2929 Conclusions CED 29

30.  Established the fact that ‘vehicle-type’ plays a major role in simulating the drivers’ behaviors observed in heterogeneous traffic streams  Established the fact that ‘lane discipline’ is not followed by the drivers in heterogeneous traffic conditions  Established the need to account for vehicle-type dependency and non-lane discipline while developing heterogeneous traffic simulators 3030 Contributions CED 30

31.  Analysis of the drivers’ motivation and choice making behaviors during lateral movements  Development of vehicle-type dependent longitudinal movement driver behavioral models  Development of vehicle-type dependent lateral movement decision and choice making driver behavioral models  Development of vehicle-type dependent lateral movement execution models  Proposing a simulation which can handle non-lane based movements of heterogeneous traffic drivers 3131 Future work CED 31

32. 32 Thank You !!!

Add a comment