Safe Obstacle Avoidance Maneuver for Autonomous Vehicles

This paper focuses on trajectory planning for autonomous vehicles, specifically addressing the challenge of obstacle avoidance during driving. The proposed approach generates a safe and efficient maneuver in real time when a static obstacle is detected in the vehicle's path. The method consists of a path generation algorithm that computes a collision-free trajectory to avoid the obstacle while guiding the vehicle back to its initial lane. Geometric and dynamic constraints are integrated to maintain vehicle stability and ensure the feasibility of the generated trajectories. The algorithm has been validated through an IPG CarMaker professional simulator and an experimental campaign using an instrumented vehicle. Trajectory generation and path tracking are embeded in the vehicle software platform tested on a track. The experimental results demonstrate the effectiveness of the proposed planning method in generating safe avoidance maneuvers, enhancing the reliability and safety of autonomous driving systems.