Lane-Changing and Overtaking Trajectory Planning for Autonomous Vehicles Based on Control Barrier Functions

Lane-changing and overtaking, as common vehicle maneuvers, have long been focal issues in trajectory planning research. In the solutions to the aforementioned problem, the quintic polynomial method has become a widely adopted trajectory planning approach due to its real-time computational advantages. However, trajectories generated by such variable-velocity planning methods impose stringent requirements on velocity tracking at the path-following level, making it challenging to ensure trajectory tracking safety. To address this limitation, this paper proposes a control-based constant-velocity planning strategy utilizing the control barrier function (CBF) methodology. This approach constructs a trajectory point iterative computation framework centered on the CBF-quadratic programming (QP) optimization problem, integrating vehicle constant-velocity kinematic models with constraint requirements specific to lane-changing and overtaking scenarios. Finally, simulation results demonstrate that compared with the quintic polynomial method, the proposed approach exhibits superior safety performance in trajectory tracking. Simultaneously, by leveraging the convex optimization formulation inherent in CBFs-QP problems, this method achieves excellent computational efficiency and demonstrates significant potential for practical applications.