Study Of Path Tracking And Path Planning For Intelligent Vehicle

As the core part of the intelligent transportation system,intelligent vehicle is a multi-disciplinary integrated system,and its goal is unmanned driving.The intelligent vehicle drives by itself instead of the driver,so that the driver's fatigue is reduced and the wrong driving behaviors are avoided,which is of great significance to improve traffic safety and reduce environmental pollution.Path tracking and path planning make intelligent vehicle drive along the target path and avoid obstacles,which are the key technologies to realize the intelligent vehicle unmanned driving.Therefore,the intelligent vehicle is studied in this paper,and the research work is carried out on the related technologies of path tracking and path planning.Vehicle lateral dynamical model of two degrees of freedom,preview tracking error model and vehicle position estimation model based on Ackermann geometry were established for the requirements of intelligent vehicle path tracking control.Combined with the non-preview tracking error model,the effect was validated though comparing the results of preview and non-preview simulation,simulation results at several groups of different speeds indicates that the tracking performance based on the preview tracking error model is better.So that the preview tracking error model is selected as the basis of the path tracking controller.The open-loop instability of path tracking system state equation was proved by Lyapunov method,combined with controllability and observability of analysis of the state equation,the LQR controller was presented based on optimal control theory to control vehicle position.Bode method was used to prove that the longitudinal velocity and the preview distance has a great influence on the response characteristics of the system,the high robustness of control system was validated by analyzing its dynamic response characteristics.At present,the parameters of weighted matrix of the LQR controller are selected mainly by manual trial and error,and unable to realize parameters optimization,therefore,genetic algorithm was adopted to optimize the parameters of weighted matrix Q in this paper,then the problem that the LQR controller is unable to obtain the optimal solution was solved.On this basis,the path tracking simulation on three roads which have different curvatures was carried out,the results indicates that the LQR controller optimized by genetic algorithm has faster dynamic response and high tracking precision.In view of the problem that there is local minimum point when the potential grid method is used to plan the path,an improved artificial potential field method was proposed,then the heuristic function of the A*algorithm was constructed by the total potential field function.At the same time,the grid map was established,and the method of finding the location of the obstacles was defined,Finally,The path planning ability and accessibility of the improved algorithm was validated by the simulation on typical obstacles.In order to verify the control effect of the LQR controller on the vehicle system model that closes to the actual vehicle,collaborative simulation was carried out in this paper based on Carsim and Simulink.Simulation results of the double lane tracking and the fixed curvature road tracking represent that the vehicle can track the target path effectively.The road curvature changes largely in a short time,which causes the mutation of the vehicle lateral acceleration and yaw angular acceleration,which reduces the operating stability and ride comfort,a curvature amendment algorithm based on the cosine curve was proposed to correct path curvature mutation,the simulation results show that the algorithm can significantly improve the path tracking performance.