Trajectory Planning for Autonomous Overtaking on Curved Roads with Avoidance of Human Driving Vehicles and Static Obstacles

A proposed obstacle avoidance trajectory planning method for autonomous vehicles considers social behavior characteristics to address the interaction problem with human drivers in mixed traffic environments. The method integrates graph theory and centrality functions to classify and predict driver behavior, employing model predictive control (MPC) to plan the autonomous vehicle's trajectory. By combining these social behaviors, the method ensures safe completion of the driving task based on probability. Experimental results from an interactive obstacle avoidance scenario in a curved environment confirm the successful completion of the task, validating the proposed method's feasibility and effectiveness, supporting its potential for real-world application.