Path Following Control And Research Of Unmanned Vehicle Based On Genetic Algorithm

The path tracking of unmanned vehicles is a research hotspot in the field of intelligent driving.The traditional linear time-varying model predictive control(LTV MPC)system has been widely studied as a more mature control system.However,some parameters in the traditional LTV MPC system have a greater impact on the performance of the path tracking system,and unreasonable parameters will reduce the path tracking performance of unmanned vehicles.Moreover,at low adhesion coefficients and high speeds,the path tracking will be poor if the control methods without any reasonable constraint.In view of these problems,based on the principle of LTV MPC,a genetic algorithm is used to optimize the controller's predictive time domain parameters and control time domain parameters.And the soft constraints are added on the front wheel slip angle to improve unmanned vehicles at high/low adhesion coefficients.The tracking error is reduced,as well as the stability of the yaw angle and the front wheel angle are improved.The main research contents of this article are as follows:(1)Four-wheel dynamics model of the vehicle is established,and the characteristics of vehicle tires are analyzed.Then,according to the tire characteristics the model of tires is obtained after linear treatment.Furthermore,design the path tracking control system is designed.Double-shift line and smooth curve are selected as the expected trajectory to verify the performance of the unmanned vehicle path tracking system.(2)Considering the given causality constraints,the four-wheel dynamics model of is simplified to a single-track model.After linear treatment the controller is designed on the basis of the model in the previous chapter.The dual-line-shifting trajectory is used to verify the rationality of the simplified single-track model.Then,the controller can be designed easily,reducing calculation time is decrease,and the real-time performance of path tracking is improved either.At a high adhesion coefficient,the path tracking effect is analyzed at different speeds and different time domain parameters.(3)Based on the genetic algorithm,the prediction time domain parameters and control time domain parameters are optimized adaptively in the MPC system.Set ITAE(Integrated Time and Absolute Error)performance index as the fitness function.Under the two working conditions of double shifting and smooth curve,the path tracking effects of the traditional controller and the optimized controller are compared at different vehicle speeds,and the path tracking error and performance of the optimized LTV MPC controller are analyzed on the unmanned vehicle.(4)On the basis of genetic algorithm,the soft constraint of front wheel slip angle is further added.Under the two working conditions of double shifting and smooth curve,the path tracking performance and driving stability of the traditional controller,the controller optimized by the genetic algorithm,and the controller with the soft constraint of the front wheel slip angle are compared and analyzed.It is verified that the controller optimized by genetic algorithm and added with soft constraint of front wheel slip angle can greatly reduce the path tracking error of unmanned vehicles under low road adhesion coefficient and high speed driving,and improve path tracking performance and driving stability It can effectively solve the problem of poor path tracking effect of unmanned vehicles when driving at high speed under low road adhesion coefficient.