Research On Tracking Control Algorithm Of Intelligent Car Based On Predictive Control

With the development of science and technology,the car is developing towards the direction of intelligence and information.The driverless technology has become a hot research field.As the simplified research model and control basis of driverless car,intelligent car can realize the basic functions of driverless car,such as automatic obstacle avoidance,path planning and trajectory tracking.It has good development prospects and important practical value.In the aspect of path planning,in order to better plan a collision free and shortest path from the starting point to the target point,the traditional ant colony algorithm is optimized and improved in this paper.In the aspect of trajectory tracking,in order to realize the intelligent car to track the planned reference trajectory with high performance,this paper designs a trajectory tracking predictive controller based on event triggering mechanism.The main contents of this paper are as follows:(1)Through the analysis and derivation of the driving process of the intelligent car,the kinematics model of the intelligent car is established.At the same time,according to the grid map method,the grid environment of the path planning is established.The obstacles are expanded and supplemented.The serial number and coordinates in the map are converted.The coding method of the path is given.(2)This paper studies and analyzes the traditional ant colony algorithm.It designs three kinds of grid environment simulation experiments.The results show that the traditional ant colony algorithm can effectively plan the global optimal path for the grid map with small dimension.But for the complex environment,its path planning effect is inaccurate.The traditional ant colony algorithm exist space for further improvement.(3)Aiming at the problems of traditional ant colony algorithm,an improved hybrid particle swarm optimization and ant colony algorithm is proposed.The main improvements of ant colony algorithm include: the adjustment strategy of pheromone concentration factor and heuristic function factor based on dynamic factor method is proposed.The heuristic coefficient based on virtual shortest distance is defined.The volatile factor with linear decrease and pheromone increment coefficient based on path length ordering are adopted.The penalty function and pheromone concentration restriction mechanism are introduced.On this basis,the reference path is obtained by path pre-planning using particle swarm optimization algorithm.The initial pheromone distribution of the improved ant colony algorithm is obtained according to the pheromone of grid environment and the pre-planning pheromone of reference path transformation.Then,the improved ant colony algorithm is used for second path planning,which greatly improves the speed and accuracy of the algorithm.(4)The trajectory tracking prediction controller based on event triggering mechanism is designed for the kinematics model of intelligent car.Firstly,the error prediction model is obtained by linearization and discretization of the intelligent car model.Secondly,the prediction equation is deduced to obtain the future system prediction output.Then,the trajectory tracking controller is obtained by the design of the target function with constraint conditions,and the optimal predictive control input is obtained by calculating.Finally,the event triggering mechanism is introduced into the controller.The experimental results show that the proposed control method can effectively reduce the calculation times of the predictive controller.At the same time,it realizes the effective tracking of linear trajectory,curvilinear trajectory and the trajectory corresponding to the path planned by the improved particle colony-ant colony hybrid algorithm.