Modelling Automatic Lane-Changing Considering Driving Style and its Implementation in an Intelligent and Connected Driving Simulation System

Lane-changing decision planning is a key challenge for large-scale autonomous driving applications, particularly concerning the optimization of driving style parameters. To achieve real-time trajectory planning for intelligent connected vehicles in dynamic environments, this study proposes a dynamic trajectory planning method considering driving style and applies this lane-changing model to an intelligent connected driving simulation platform. First, an interface is developed to obtain real-time dynamic road information and data about the vehicle and neighboring vehicles from the driving simulation software. Second, the driving style is classified based on headway time-distance characteristics, and the highway driving state is divided into three categories: cruise control, front-following, and automatic lane-changing. Then, a decision model based on the Stackelberg game is established to select the driving state and ideal lane of the main vehicle by interacting with vehicles in front and behind the target lane. The target configuration method of the trajectory terminal is designed for the three types of driving states based on the principles of safety, comfort, and efficiency. Subsequently, a fifth-degree polynomial function is used to generate multiple potential trajectories connecting the initial and target terminals. A comprehensive loss function is constructed considering driving style, trajectory comfort, speed change, and trajectory safety. The cost of potential trajectories is evaluated and ranked by combining speed, acceleration, and safety distance checks to output discrete trajectory points. Finally, an independent control system with horizontal and vertical decoupling of vehicle motion is established for trajectory tracking. To verify the algorithm's application capability, the method is applied to a driving simulation test platform, and its performance is tested through simulation experiments. Test results show that the proposed automatic lane-changing algorithm meets the demands of highway dynamic planning and reflects typical driving styles. Through optimization analysis of key parameters, the algorithm can plan trajectories with high safety, stability, and comfort.