Path Planning And Tracking Of Intelligent Vehicle Based On RRT And LQR

The development of intelligent vehicle technology greatly reduces the influence of human factors in traffic accidents and enhances the safety of vehicle driving.Path planning and tracking control technology are key parts of intelligent vehicle technology,and become study hotspot for scientific research units and social enterprises.The path planning of intelligent vehicle mainly considers two aspects.On the one hand,it is required to quickly plan barrier-free paths to meet the needs of intelligent vehicle to avoid obstacles in time;on the other hand,it is required to plan the shortest path with continuous curvature to satisfy the smooth driving of the vehicle and the shortest driving time.The lateral tracking control of intelligent vehicle mainly considers improving the tracking accuracy while ensuring the stability of the vehicle.the local path planning and lateral path tracking of intelligent vehicle in this paper.Firstly,in order to solve the problems of low path planning efficiency and long tortuous paths of the Rapidly-exploring Random Tree(RRT)algorithm,a new node expansion method biased toward the target point is proposed,which can adaptively adjust the gravity and the target of the random point.The proportion of point gravity in the expansion process of new nodes,together with the obstacle repulsion,enables the random tree to avoid obstacles and expand to the target point faster.A pruning optimization method is proposed to recalculate the distance between of sampling points by adding an evaluation function,and a new effective path is constructed to further optimize the path length.Combined with the Bézier curve,the strategy of inserting control points is proposed to smooth the turning points,and the curve splicing method is used to solve the curvature discontinuity problem at the splicing points.The improved fusion RRT algorithm is compared with the traditional RRT algorithm,the P-probability-based RRT algorithm,and the two-way RRT algorithm.The results show that the improved method is optimized in terms of time,path length,and path smoothing.Secondly,according to the requirements of intelligent vehicle lateral tracking control,the path tracking model with error as the state variable is constructed based on the Linear Quadratic Regulator(LQR)control algorithm and the LQR lateral tracking controller is designed.The feedforward control is used to compensate to control amount,so that the steady-state following error of the lateral control system converge to zero.The effectiveness of the feedforward LQR lateral tracking controller at different vehicle speeds is verified by the Car Sim/Simulink joint simulation platform.Then,the influence of different weight matrix parameters on the tracking performance of the feedforward LQR lateral tracking controller is analyzed,and the adaptive change of the weight matrix parameters of the feedforward LQR lateral tracking controller is studied according to the influence law.The lateral distance deviation and path curvature are taken as the adjustion input of the weight matrix parameters to construct the feedforward fuzzy LQR lateral tracking controller,so that the controller can adaptively output the optimal control amount according to the driving state.The simulation results at different speeds show that the feedforward LQR lateral tracking controller can give consideration to both accuracy and driving stability when tracking the path.Finally,a Hardware in the Loop(Hi L)simulation system is built to verify the effectiveness of the controller.By flashing the control strategy into the real controller and carrying out simulation experiments under the double-line shifting conditions,it is shown that the feedforward fuzzy LQR lateral tracking control strategy can operate effectively in the real electronic control unit.