结构化道路下智能汽车自主换道轨迹规划研究

Automated lane changing plays a crucial role in the advancement of autonomous driving technology. A layered trajectory planning method is present that separates path planning and speed planning into independent processes. The path planning phase involves establishing potential fields for the road, static obstacles, and surrounding vehicles, followed by generating path clusters using the quintic polynomial method. The environmental potential field is determined to derive the optimal lane-changing path. The speed planning process simultaneously considers influencing factors such as lane change efficiency, ride comfort, safety, vehicle dynamics response, time window, and road constraints, and a convex optimization-based method is proposed. To evaluate the proposed scheme, a Prescan-Simulink co-simulation environment is established, and the trajectory planning algorithm is tested under diverse scenarios. The results demonstrate the efficient handling of complex constraints during the lane-changing process using the proposed method, while simultaneously ensuring safety, ride comfort, and lane change efficiency.