Hardware-in-the-Loop Testing of Connected and Automated Vehicle Applications

Transcription

1 Hardware-in-the-Loop Testing of Connected and Automated Vehicle Applications Jiaqi Ma Assistant Professor University of Cincinnati ITS Midwest Annual Meeting Columbus, Ohio, September 29, 2017

2 Outline Background & Objective HIL System Architecture Q-aware SIAD application Testing and Results Concluding Remarks

3 Background Increasing interests and investment in the research and development of innovative applications of connected/automated vehicles Most of CAV studies apply simulation for evaluation; however, model accuracy and simulation assumptions render limited validity of evaluation results Lack of field data exacerbates the problem of inaccuracy in modeling and simulation because there are no data available for model calibration purpose.

4 Background A few limited CAV field experiments Limited number of test vehicles available for experiments; Larger-scale field operational tests are extremely expensive A relative low cost and more accurate evaluation approach for CAV studies is necessary 4

5 Solutions Use emerging hardware-in-the-loop (HIL) testing tools is the best solution: Allow real test vehicles to interact with virtual vehicles from traffic simulation models Provide an evaluation environment that can replicate actual deployment conditions by using actual hardware and equipment Without incurring excessive costs at early stages of CAV development Categories 5

6 Hardware in the Loop Testing On-board Units Traffic Simulator Driving Simulator Other Hardware, e.g., Signal Controller, powertrain Roadside Units

7 High Level Project Objective Develop hardware-in-the-loop (HIL) testing platform for and set up HIL experimental system at TFHRC Conduct HIL testing to evaluate SIAD 7

8 Signalized Intersection Approach and Departure V2I Communications: SPaT and GID Messages Roadside Equipment Unit V2V Communications : Basic Safety Messages Traffic Signal Controller with SPaT Interface Vehicle Equipped with the Eco-Approach and Departure at Signalized Intersections Application (CACC capabilities optional) Traffic Signal Head

9 Goal Find a set of trajectories to optimize MOEs Travel time, fuel & emission, safety Trajectory smoothing

10 HIL System Setup Two challenges for the HIL testing of EAD: Synchronizing field and simulation traffic conditions on the fly Collecting real-time field traffic data from the testbed and real-time simulation data from a traffic simulator Figure 1 Platform for HIL Testing of EAD Application Figure 2 Data flow Chart for HIL Testing of EAD Application 10

11 FHWA Innovation Research Vehicles Proof of Concept Vehicles Research Fleet Communications - 5.9GHz DSRC, Cellular/LTE, Corrected GPS On-board Technology - Connected Vehicle Data Collection and Processing - Stock Radar and Ultra-Sonic Sensors - Front and rear-facing cameras

12 Vehicle System Data Flow

13 Connected Vehicle Highway Testbed (Intelligent Intersection) at TFHRC CCTV DSRC Signalized intersection with SPaT / MAP Vehicle Pedestrian & Bike Detection Fixed time or actuated traffic signal control with pedestrian / bike displays Dedicated Ethernet & Wi-Fi communications Cabinet space with power & comms, available for future research Cadillac SRX with OBU, GPS, CAN bus integration

14 HIL Architecture

15 Q-aware SIAD Based on the existing SIAD algorithm, additional factors considered: Background traffic: Interactions between the CAV and other vehicles Multiple intersections Two intersections will be considered Different traffic signal control modes Actuated control and traffic coordination will be considered Different penetration rates of CV vehicles The impacts of penetration rates of CV vehicles on the SIAD algorithm will be evaluated 15

16 Q-SIAD Algorithm Y Q-SIAD Start (t=0) The front vehicle is very close? N Basic SIAD Algorithm Y Ta < Gcrm? Scenario 1 Speed Linear approaching & car following (gap regulation) t=t+1 N Vehicle passed the intersection? Y N Estimate Queue Length t=0? or Queue Length Change? Y Basic SIAD Algorithm N Tea< Gcrm <Ta? N (Gcrm <Tea & Tla<Gsn)? (Gcrm <Tea & Tla>Gsn)? Y Y Y Scenario 2 Scenario 3 Scenario 4 v h v c t m t n d 0 0 t arr where, t m = pi/(2m); t n = pi/(2n)+t m ; t arr = d 0 /v h. Time End (a) (b)

17 Experimental Scenarios Scenario 1: Single Intersection with fixed time traffic signal control Case 1-1: Base case without SIAD (Adaptive Cruise Control, ACC) Case 1-2: Q-SIAD algorithm by considering background traffic Scenario 2: Single Intersection with actuated traffic signal control Case 2-1: Base case without SIAD (ACC) Case 2-2: Q-SIAD algorithm by considering background traffic and features of actuated control 17

18 Testing

19 Slowdown scenario Results(1)

20 Speedup scenario Results(2)

21 Results(3) Start Scenario SIAD Scenario Baseline Basic SIAD Queue-aware SIAD 0% MP 100% MP R0 Slowdown % % % R5 Slowdown % % % R10 Cruise % % % R15 Cruise % % % R20 Cruise % % % G0 Cruise % % % G5 Speedup % % % G10 Stop % % % G25 Stop % % %

22 Concluding Remarks Offer a cost-effective approach for quick evaluation of CAV technologies Currently developing HIL for CACC; humanin-the-loop Help public agencies and private sectors to evaluate new CAV technologies

23 Q&A Contact Information Jiaqi Ma Department of Civil Engineering University of Cincinnati