Li, Zheng (Tianjin University), Wang, Yijing (Tianjin University), Zuo, Zhiqiang (Tianjin University), Shi, Yang (University of Victoria), Zhao, Rui (Tianjin University)

Trajectory Optimization for Autonomous Driving in Uncertain Environments Integrated with Risk-Based Corridors

Scheduled for presentation during the Regular Session "Trajectory planning II" (ThAT9), Thursday, September 26, 2024, 12:10−12:30, Salon 17

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on October 7, 2024

Abstract

Lane change maneuvers of autonomous vehicles in a dynamic highway environment suffer from various uncertainties, especially those from unknown kinematics of surrounding vehicles. In this paper, an optimization-based trajectory planning scheme is proposed to reliably generate collision-free lane change references. By an inner approximation of risk contour, the probabilistic constraints under a given risk level can be converted to the deterministic form, and then, the risk-based corridors are constructed. Decoupled longitudinal and lateral states are sequentially optimized in terms of such corridors and the preferentially determined boundaries of merging instant. Finally, the optimal trajectory is selected from all fused candidates via a prescribed cost indicator. Through both numerical examples and validations on the naturalistic human driving dataset, the advantages of our scheme in reducing crash risks are verified from both statistical analysis and case studies.