Development of Low-Speed Autonomous Driving System for Electric Vehicles Based on Dynamic Risk Map Construction in Intersections

本研究旨在构建高危交叉口汽车防撞自动驾驶系统。结合车辆动力学、控制理论和机器学习的知识，设计了自动驾驶功能，以完成交叉口通过、左转和右转动作。提出了一种表达交叉口与其他运动物体碰撞风险的风险图，该图包含了潜在的遮挡风险。以装有环境感知系统和用于车辆运动控制的执行器的电动汽车为例，验证了该控制系统的有效性。

This research aims to construct an autonomous driving system for automotive collision avoidance in high-risk intersections. The autonomous driving functionality is designed by integrating the knowledge of vehicle dynamics, control theory and machine learning in order to accomplish intersection passing, left-turn and right turn maneuvers. A risk map which expresses the collision risk with other moving objects in intersections including potential risk for occlusions is proposed. The electric vehicle equipped with environment perception systems and actuators for vehicle motion control is used to verify the effectiveness of the proposed control system