Study And Implementation On Path Tracking Control Method Of Unmanned Vehicle

At present,with the development of artificial intelligence technology and Internet application,highly intelligent driverless vehicle will be the inevitable trend of automobile development.Unmanned vehicle can not only improve the efficiency of traffic operation effectively and ensure vehicle safety,but also replace human beings to complete some special missions.Path tracking control of unmanned vehicle is one of the important research contents in the field of unmanned driving,and it is also one of the important conditions for stable driving of vehicles.Therefore,it possesses great practical significance to study on the path tracking control method of unmanned vehicle.In this paper,we firstly analyze the path tracking control methods that have been used of unmanned vehicle at home and abroad,and the advantages and disadvantages of various tracking control methods used in practical application are summarized.Secondly,according to the actual vehicle parameters,a vehicle simulation model is established in Carsim/Simulink,and its kinematics and dynamics characteristics are analyzed.Thirdly,combining with the vehicle model,we study the principles of traditional Pursuit-Tracking and MPC algorithms,comprehensively analyses the feasibility,reliability and accuracy of the two algorithms in practical engineering applications,and improve them in view of the shortcomings of the algorithm itself,then we propose the curvature fuzzy compensation pure tracking control algorithm and the improved model predictive control algorithm based on maximum correlation entropy criterion(MCC),the principles of these two improved algorithms are introduced in detail.Fourthly,according to the relevant mathematical theory of the two improved algorithms,the corresponding simulation controller is designed on Carsim/Simulink platform,and the algorithm is simulated and validated with vehicle model.Fifthly,we use ROS/C++ software framework to complete the algorithm programming,optimization and debugging on the MIC7500 platform,thus completing the design of real vehicle controller.Finally,we build a driverless vehicle path tracking system,and utilize the real vehicle controller designed in this paper combined with the GPS inertial navigation module to achieve the precise tracking of the driverless vehicle and achieve the desired goals.This paper introduced the simulation experiment system platform and the real vehicle experiment system platform for driverless vehicle on Carsim/Simulink platform and E70 BEV platform respectively.The simulation and real vehicle experiment results show: Compared with the traditional pure tracking control method,the curvature fuzzy compensation pure tracking control method proposed in this paper has higher tracking accuracy at low speed.And the proposed model predictive control algorithm based on maximum correlation entropy criterion still maintains good tracking performance and strong anti-jamming ability under high-speed running conditions of vehicles,and it has stronger robustness than traditional model predictive control.