Automatic Generation of Intelligent Vehicle Testing Scenarios at Intersections Based on Natural Driving Datasets
- 18 December 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (9) , 5448-5460
Abstract
Turning right at a junction without the aid of traffic signals is a frequent cause of intersection accidents and a situation that intelligent automobiles will need to handle well in the future. This allows this study to choose the traffic circumstances of turning right at junctions and traveling straight through them. It then suggests an automatic approach for creating test scenarios for intelligent vehicles based on Monte Carlo simulation. Through the examination of scene feature parameters and a natural driving dataset, a Markov chain-based right-turn trajectory model was created. Then, test vehicle trajectories and right-turn trajectories were produced using Monte Carlo simulations. By figuring out the average DTW (Dynamic Time Warping) value between the created and actual right-turn trajectories in the four speed intervals with the highest occurrence probabilities, the authenticity of the generated right-turn trajectories was confirmed. The produced trajectories were simultaneously loaded into the software for cooperative simulation. The AEB (Automatic Emergency Braking) system's shortcomings were shown by extracting the relative positions of the two cars before to the incident. The two popular varieties of AEB algorithm models were then compared using safety and comfort parameters. Furthermore, the test scenario's risk level can be managed by changing the scenario feature settings. There are some potential uses for the suggested test scenario generation method in the testing and assessment of algorithms for intelligent vehicles.Keywords
Funding Information
- National Natural Science Funds for Distinguished Young Scholar (52325211)
- National Natural Science Foundation of China (52172399, 52175088)
- Natural Science Foundation of Changsha (KQ2208235)
This publication has 25 references indexed in Scilit:
- Off-road testing scenario design and library generation for intelligent vehiclesGreen Energy and Intelligent Transportation, 2022
- A study on energy distribution strategy of electric vehicle hybrid energy storage system considering driving style based on real urban driving dataRenewable and Sustainable Energy Reviews, 2022
- A Review of Testing Object-Based Environment Perception for Safe Automated DrivingAutomotive Innovation, 2022
- A Computer Graphics-Based Framework for 3D Pose Estimation of PedestriansAutomotive Innovation, 2021
- Calibration and evaluation of the Responsibility-Sensitive Safety model of autonomous car-following maneuvers using naturalistic driving study dataTransportation Research Part C: Emerging Technologies, 2021
- Modeling Trajectories and Trajectory Variation of Turning Vehicles at Signalized IntersectionsIEEE Access, 2020
- Casualty risk of e-bike rider struck by passenger vehicle using China in-depth accident dataTraffic Injury Prevention, 2020
- Simulation-Based Testing Framework for Autonomous Driving Development2019
- Driving to safety: How many miles of driving would it take to demonstrate autonomous vehicle reliability?Transportation Research Part A: Policy and Practice, 2016
- Stochastic approach for modeling the effects of intersection geometry on turning vehicle pathsTransportation Research Part C: Emerging Technologies, 2013