A study on lane change trajectory planning methods considering driving styles

Personalized trajectory planning is essential in autonomous driving, as it adapts the driving path to the driver's style, providing a more comfortable, safe, and energy-efficient experience. In this paper, lane change trajectory planning is decoupled into two core components: path planning and speed planning. For path planning, a target lane change path is designed by integrating driving habits with lane change timing predictions. The cost function is optimized according to driving preferences, while constraints are defined based on the driver's capabilities. Quadratic programming is then employed to optimize the path planning. For speed planning, dynamic programming is combined with an elliptical collision model to identify the convex optimization space in the space-time (S-T) graph. The cost function is optimized based on driving habits, and constraints are also formulated according to driving ability. The effectiveness and accuracy of the proposed lane change trajectory planning strategy are validated in a Matlab/Simulink-CarSim simulation environment under realistic working conditions. The results show that by considering driving styles, the proposed method can fulfill drivers' personalized needs, offering a more precise and comfortable lane-change experience in autonomous driving scenarios.