Research And Implementation Of Path Planning And Tracking Control For Unmanned Vehicle

Automatic Guide Vehicle is a vehicle that can automatically move goods in complex environments such as smart factories.In the context of automation,Automatic Guide Vehicle is becoming more and more popular.Among them,path planning and trajectory tracking algorithm are the key factors to determine whether the Automatic Guide Vehicle can efficiently complete the work.In this thesis,the Automatic Guide Vehicle in intelligent factory is taken as the research object,and the shortcomings of the path planning algorithm of the Automatic Guide Vehicle in the current storage environment,such as low planning efficiency and unfeasible path,are improved,so as to realize the efficient and collision-free movement of the Automatic Guide Vehicle from the starting point to the target point.The main work of this paper is described as follows.1)An improved A~* algorithm was proposed to solve the problems of low planning efficiency and infeasible path planning in global path planning.Firstly,an exponential decay function is introduced into the heuristic function of A~* algorithm to dynamically adjust the weight ratio,adjust the weight size according to the change of map size,and dynamically adjust the traversal search speed in the process of path search.Then,the information of the parent node of the current node is introduced into the heuristic function to make the search direction approach the end point more purposefully.Finally,the Angle judgment formula is used to judge the global path planned,and the cubic B-spline smoothing curve algorithm is used to smooth the Angle when the judgment conditions are satisfied,so as to enhance the feasibility of the path.2)Aiming at the defects of artificial potential field method in local path planning algorithm,such as easily falling into local minimum point and unreachable target,the gravitational field function is improved according to the characteristics of gravitational field and repulsive field function,and segmented according to the size of environment map.For the problem of unreachable target due to excessive repulsion,the distance between the current node and the target node is introduced into the repulsion function,so that the size of the repulsion field function is related to the distance from the target point.To solve the problem that the vehicle is prone to fall into the local minimum point,escape force is introduced into the total resultant force,and the magnitude of escape force is defined by obstacles to help the vehicle get rid of the local minimum point.3)Aiming at the problem of how to better control the vehicle driving along the reference trajectory,a two-layer local motion control algorithm of "planning + tracking" is designed,which is combined with improved artificial potential field method and model predictive control.According to the characteristics of the warehouse environment,the vehicle model was established and the constraint conditions were set,and the environmental potential field was added to the objective function.Combined with the advantages of the two algorithms,such as good real-time performance,kinematic characteristics and constraint conditions,the vehicle could simultaneously and quickly carry out local path planning and trajectory tracking control movement.4)Matlab/Carsim co-simulation experiment is designed to verify the trajectory planned by the improved algorithm and the effect of controlling the vehicle to drive along the planned trajectory.SnowMan robot is used to build a real storage environment to verify the practicality and innovation of the algorithm involved.Experimental results show that the improved algorithm proposed in this thesis can plan a better reference path,and the improved algorithm has good operation effect when applied to the robot.