Research On Lane Change Trajectory Planning And Tracking Control Considering Driver’s Style

With the development of automobile intelligence and electrification,the application of intelligent automobile in life is more and more recognized and concerned by people.The design of autonomous lane change system of intelligent vehicle is mostly based on the lane change style of the public to realize lane change of intelligent vehicle,and the influence of individual driver factors on the whole lane change process is rarely considered,which cannot meet the needs of personalized lane change.How to consider the driver’s style,meet the track tracking accuracy of lane change and ensure the comfort of lane change is the inevitable requirement of intelligent and personalized automatic lane change system in the future.Based on the driver’s lane change track data,this paper studies the lane change track planning and tracking control of intelligent vehicles.In this paper,effective lane change data is extracted based on NGSIM trajectory data,and the characteristic parameters of driver lane change are utilized to perform cluster analysis on drivers by K-means algorithm,which can be divided into three types: cautious type,general type and aggressive type.Statistical analysis and time-frequency analysis were carried out to verify the differences of all types of drivers,and effective parameters were selected as the input of the identification model.The particle swarm was improved based on the strategy of nonlinear change of inertia weight and nonlinear decline of maximum particle velocity,and the driver style identification model based on improved particle swarm optimization SVM was established.The parameters of the SVM model were optimized and updated,and the established identification model had a higher recognition accuracy,and compared with a variety of commonly used machine learning algorithms.Lay the foundation for the construction of the following lane change model.Secondly,based on the driver’s lane change track data,a personalized lane change track model based on hyperbolic tangent function was established,and the sensitivity of the model’s style factor parameters was analyzed to analyze its influence on the lane change style.Gaussian process regression method was used to identify and calibrate the parameters of the personalized track changing model,and the corresponding track changing model parameters were obtained,so as to realize the personalized and anthropomorphic track planning.The model is compared with the traditional lane changing trajectory model,and the results show that it has good fitting accuracy.Finally,based on the two-degree-of-freedom model,the driver preview model was introduced,and the expected yaw velocity was calculated based on the assumption of steady circular motion.In order to solve the uncertainty in the lateral control of vehicle,a sliding mode control method with high robustness and the desired yaw velocity as input was used to design the trajectory tracking lateral controller.In order to weaken the buffeting caused by the sliding mode control,a fuzzy sliding mode controller was designed combined with the fuzzy control theory.Based on the co-simulation platform of MATLAB/Simulink and Car Sim,the trajectory tracking controller was experimentally verified under three different working conditions.The simulation results show that the designed controller can adapt to and track the lane changing trajectory of drivers with different styles,and has high adaptability and robustness.Finally,based on the two-degree-of-freedom model,the driver preview model was introduced,and the expected yaw velocity was calculated based on the assumption of steady circular motion.In order to solve the uncertainty in the lateral control of vehicle,a sliding mode control method with high robustness and the desired yaw velocity as input was used to design the trajectory tracking lateral controller.In order to weaken the buffeting caused by the sliding mode control,a fuzzy sliding mode controller was designed combined with the fuzzy control theory.Based on the co-simulation platform of MATLAB/Simulink and Car Sim,the trajectory tracking controller was experimentally verified under three different working conditions.The simulation results show that the designed controller can effectively adapt and track the lane changing trajectory of different drivers,and has high adaptability and robustness.