Song, Lei (Tsinghua University), Zhang, Lingjun (Tsinghua University), Zhao, Chengxiang (Beijing Institute of Technology), Ma, Xiaohan (Beijing Institute of Technology), Liu, Jiaxin (Tsinghua University), Wu, Siyu (Tsinghua university), Wang, Zixuan (Tsinghua University), Yu, Wenhao (Tsinghua University), Wang, Hong (Tsinghua University)

Legal Driving for Autonomous Vehicles through Law-Compliance Potential Fields

Scheduled for presentation during the Invited Session "Safety for Intelligent and Connected Vehicles" (ThBT3), Thursday, September 26, 2024, 14:30−14:50, Salon 6

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 December 26, 2024

Abstract

The widespread adoption of autonomous vehicles necessitates their compliance with existing traffic laws. However, while safety-oriented decision-making methods have matured, shortcomings in law compliance persist. To ensure that autonomous vehicles drive legally while prioritizing safety, a law-compliance potential fields-based approach is introduced. Firstly, this paper discusses the classification and constraints of traffic laws on vehicle states, outlining five principles for legal driving. Subsequently, law-compliance potential fields are established, incorporating factors related to roads, obstacles, and vehicle speeds. The potential fields-based model predictive control (MPC) controller is then employed to generate feasible trajectories, thereby enabling legal driving. Finally, the effectiveness of the proposed approach is verified through three representative scenarios including the intersection, the highway and the urban road. The results demonstrate that the method can ensure law compliance of autonomous vehicles in complex scenarios and can prioritize safety in extreme situations.