Research On Vehicle Lane Change Decision Planning And Tracking Control Based On Driving Style

Lane change is one of the most common behaviors when drivers control vehicles.Improper lane change behavior and timing may lead to major traffic accidents.The development of intelligent vehicle lane change function algorithm provides a solution to this problem and has become a research hotspot of autonomous driving technology.Existing lane change trajectory planning and tracking control algorithms are mostly based on vehicle dynamics or kinematic models,and rarely consider the influence of driver factors on planning results and tracking control.In order to improve the driving experience and ride comfort of drivers with different styles,this paper studies lane change decision planning and tracking control model of intelligent vehicles based on driving styles.The main research contents are as follows:(1)Aiming at the problems of large subjectivity and insufficient sample size caused by the questionnaire-based driving style analysis,this paper,based on the perspective of driving behavior,firstly extracted 15 characteristic parameters reflecting driving style from the sample from the NGSIM data set.Secondly,in order to reduce the correlation and redundancy between different characteristic parameters,the PCA was used to reduce the dimension.Finally,k-means clustering,an unsupervised learning method,was used to cluster the six principal components with the highest contribution to the principal component results.The clustering results were evaluated according to the contour coefficients of different clustering numbers.The clustering number K was 3 and the contour coefficient was 0.7154,and the driving styles were divided into aggressive,normal and cautious.(2)Most current researches on automatic lane change strategies of intelligent vehicles are based on vehicle models without considering driver factors.To solve this problem,this paper first establishes the minimum lane change safe distance model based on the established driving style combined with the elliptical vehicle model,and further analyzes the feasibility of lane change and makes different lane change decisions.Secondly,the road compound potential field and obstacle potential field with different driving styles were established,the speed factor was introduced to establish the obstacle speed repulsive potential field,and the path planning was carried out by combining the model predictive control idea.In order to ensure the smoothness of the planned path,the quintic polynomial was used to fit the planned trajectory points.Finally,the path planning algorithm based on APF and MPC at different speed was verified by simulation.The results showed that the vehicles could safely drive in the lowest potential field value in the whole prediction time domain,and the planned path met the vehicle driving requirements.(3)According to the research problem of horizontal and longitudinal control of different driving styles,the motion control is decoupled from horizontal and longitudinal.For longitudinal control,based on single PID speed control,the dual PID control of longitudinal position control and speed control is adopted.The simulation results show that the dual PID control method is more comfortable and safer.The simulation results show that the improved MPC control model can adapt to different driving styles.(4)The algorithm proposed in this paper was verified by simulation.The Car Sim and Matlab/Simulink co-simulation platform was built.The results showed that the automatic lane change algorithm based on different driving styles designed in this paper can meet the safety of lane change and adapt to the needs of different driving styles.