Research On The Algorithm Of Local Path Planning And Track Tracking Control Of Intelligent Vehicle

At present,autonomous driving is an important part of today’s technological development.The intelligent car obtains the surrounding environment information through the environment awareness system,obtains the trajectory that the intelligent car can travel by using the planning module on the environment map constructed,and tracks the trajectory points in real time through the control system,so as to achieve the autonomous driving of the car.In this paper,with the surrounding environment known,the local path planning and trajectory tracking control problems are studied.The specific research contents are as follows:(1)This paper studies the local path planning method of intelligent vehicles in complex environment such as park,port,parking lot and urban structured road environment.RRT and astar algorithm are used to search the path in complex environment.RRT algorithm searches the path zigzag without constraints.It is only a path connecting the starting point,which is not optimal,and the randomness of scattered points leads to blind search.Therefore,the solution time of the path is long,but RRT algorithm has completeness.When there is a path connecting the starting point,RRT algorithm can solve the problem completely,RRT algorithm can be solved.Astar algorithm is heuristic,it can quickly get a path in the grid environment,but astar can only move towards the surrounding four or eight grids,can not add vehicle kinematics constraints,resulting in the path is not conducive to vehicle tracking.In order to make RRT(fast search random tree)algorithm widely used in the field of intelligent vehicles and solve the problems of low search efficiency and unreasonable nearest neighbor search function of RRT algorithm,this paper proposes an improved RRT algorithm based on astar algorithm.Astar algorithm is used to generate the shortest path in the low resolution grid map to construct the guidance domain to solve the randomness of RRT scattering points.At the same time,vehicle kinematics constraints are considered in the nearest neighbor search function of RRT algorithm to make the planned path meet the normal vehicle tracking.When new nodes are generated,the parent node and rerouting are re selected to make the search path approach the optimal.The simulation results show that the algorithm is effective and practical.(2)According to the characteristic of curvature continuity of cubic B-spline,this method is used to smooth the path points,and the final executable path is obtained.The practicability of the smoothing algorithm is verified by simulation.Since the improved RRT is only effective for static obstacle avoidance,and there are generally dynamic obstacles in the obstacle environment,dynamic obstacle avoidance is required.Using five times polynomial in ST figure(planning trajectory longitudinal displacement and time of twodimensional diagram)to build ST curve,the location of the obstacles in the future time on the trajectory of the projection as obstacle area of ST figure,with the target vehicle speed constant current through and maximum acceleration through planning and track the time used as a benchmark,create a group of candidate velocity planning target time points,from which a series of ST curves can be planned,and use evaluation function to select the optimal ST curves.Finally,the simulation experiment of trajectory planning under the urban structured road environment is designed.(3)This paper designs a trajectory tracking controller based on MPC.The vehicle position,longitudinal and lateral velocity,yaw Angle and yaw angular velocity are taken as the controller state variables.The control quantity is the front wheel Angle.At the same time,a series of constraints such as control quantity and control increment are set to ensure the stability of the controller.At last,the simulation is carried out under the condition of double line shifting at different speeds.In order to verify the effectiveness of the planned path and the stability of the controller,the planned path and control algorithm were simulated and verified by the co-simulation platform built by Car Sim and Matlab/Simulink.The results show that the proposed planning algorithm and trajectory tracking control algorithm have strong practicability.