Research On Adaptive Path Planning And Motion Control For Mobile Vehicles

In recent years,intelligent mobile vehicles have become an important part of modern industrial,military,medical and domestic services.Mobile vehicles can reach environments where humans cannot adapt,perform dangerous and repetitive tasks instead of humans,effectively reducing injuries to personnel and reducing labor costs.As a result,researchers at home and abroad have carried out a great deal of research into the development of intelligent vehicles,and an important problem in intelligent vehicle research is path planning,i.e.generating a path or sequence of motions from an initial state to a desired target state under specific constraints.However,existing vehicle path planning algorithms have the problems of generating a single trajectory,poor adaptability of the solved global path in the face of diverse task demand scenarios,and large deviations when vehicles follow the global path and avoid obstacles leading to incoherent motion states,for which a new intelligent vehicle path planning strategy and motion control scheme is proposed,as follows:1.This paper constructs a vehicle kinematic model and then proposes a new multi-metric evaluation path model based on the planning requirements for global paths in mobile vehicles.Then,drawing on the Frenet coordinate framework in the unmanned path planning method,a description of the position state of the mobile vehicle is proposed,which converts the Cartesian coordinate system to Frenet coordinates and intuitively expresses the relative relationship between the vehicle and the global path.2.For global path planning,multiple collision-free paths to the target are sought based on a reinforcement learning online decision mechanism;then a value assessment system is established based on the path performance characteristics to dynamically update the action generation values among the robot motion nodes;and simulation tests are conducted under different weighting ratios.3.For local path planning,multiple local paths to be selected are generated under the Frenet coordinate framework using a quintic polynomial;the obstacle shape contours are fuzzed with the obstacle potential field method and an expansion interval is set;combined with the sampling points of the paths to be selected,the collision values obtained from the calculation of the obstacle expansion interval and the deviation values of the globally planned paths are used to comprehensively evaluate the obstacle avoidance capability of the paths and filter out the optimal obstacle avoidance paths.4.A physical framework for the whole vehicle is built,and an industrial millimeter wave radar and an open-air Bei Dou navigation system are used for the vehicle movement environment.The experimental results show that the algorithm can generate the corresponding optimal global path planning according to the task requirements,and based on the global path planning,detect the vehicle position state in real-time,dynamically solve the local obstacle avoidance path a with small offset from the global path and coherent motion,and feedback control the vehicle to achieve the global path motion following,to verify the effectiveness and feasibility of the algorithm,and provide a potential solution for the mobile vehicle path planning problem.The algorithm is used to validate the effectiveness and feasibility of the algorithm,providing a potential solution to the mobile vehicle path planning problem.