Obstacle Avoidance Path Planning And Tracking Control Of Tracked Unmanned Orchard Platform In Hilly And Mountainous Areas

As the key technologies of orchard intelligent agricultural machinery equipment,obstacle avoidance path planning and tracking control directly affect the safety of intelligent agricultural machinery and equipment during operation,and are also related to the production efficiency of orchards,the increase of fruit farmers’ income and the development of China’s agricultural mechanization level.In order to improve the automation and intelligence of the orchard production equipment after mechanization transformation in hilly areas,this thesis studies the obstacle avoidance path planning and path tracking control problems of crawler unmanned platform in hilly orchards.The main research contents and conclusions of this thesis are as follows:(1)Optimizing the global reference path of the orchard based on quadratic programming.The global pre-stored path collected manually in advance has noise and cannot be directly used for obstacle avoidance path planning and tracking control.In this thesis,a mathematical model of global storage path optimization for orchard was established.Considering smoothness,uniformity and offset constraints,quadratic programming was used to solve the optimized global reference path.Under the premise that hard constraint of the given path offset is 0.1m,the optimized global reference path spacing standard deviation decreases,the distribution was more uniform,and has better smoothness,which accurately provided basic information for subsequent obstacle avoidance path planning and tracking control.(2)Building a dynamic grid map based on the inversion model.Bresenham’s algorithm was used as the algorithm of the inversion model.Two-dimensional projection was performed on the point cloud information around tracked unmanned platform obtained by lidar in real time,and the projection results were rasterized to obtain the occupancy value of the raster map,and the dynamic raster map was constructed to display the obstacle information,which was used for the dynamic planning of the subsequent obstacle avoidance path.(3)Dynamic planning of obstacle avoidance path based on grid map.Based on the Frenet coordinate system,lateral and longitudinal path points were sampled on the basis of the global reference path in dynamic grid map.Path curve smoothness cost,deviation cost of distance from global reference path,and safety cost from the obstacle distance were constructed.Using dynamic programming traverses all sampling points,obtained sampling points sequence of the obstacle avoidance path with the least cost,and used the quintic polynomial curve to connect sampling points to obtain obstacle avoidance path,which ensured the continuity of curvature.Simulation results showed that the obstacle avoidance path generation time of the algorithm was 219 ms,and the obstacle avoidance path met the kinematic constraints of the tracked unmanned operating platform.(4)Improved pure pursuit control algorithm based on preview.Aiming at the problem that the pure pursuit control algorithm based on preview cannot take into account the tracking control accuracy and control stability,a double-layer PID closed-loop feedback control was added.The first layer is yawrate-linear velocity double PID closedloop feedback,which ensured the stable running of the tracked unmanned platform at a constant operating speed in the orchard.The second layer is motor speed PID closed-loop feedback to ensure that the drive motor of tracked chassis accurately executes motor speed instructions.By MATLAB/Simulink simulation verification,after adding feedback control,the lateral error and heading angle error in the path tracking process were better than the original algorithm,the motor speed changed stably,under the external interference,tracked chassis could run at a relatively stable speed.(5)Orchard road experiment.In order to verify the effectiveness and feasibility of the obstacle avoidance path planning algorithm and tracking control algorithm proposed in this thesis,an orchard suitable for mechanization transformation was selected for experiment.The experimental results showed that under the condition of obstacles,the tracked unmanned platform could independently plan a continuous,smooth and safe obstacle avoidance path and realize the obstacle avoidance function.The experimental results of path tracking revealed that after adding feedback control,the average lateral error of path tracking was 8.51 cm and the average heading angle error was 9.1°,which were 24.36% and 20.18% higher than the pure pursuit control algorithm based on preview,and standard deviation of lateral error and course error decreased.Meanwhile,linear velocity was stable,yawrate error was reduced,and the motor speed changed steadily.The experimental results represented that the designed tracked unmanned platform could track the path stably and accurately,and the control precision could meet the requirements of unmanned operation of agricultural machinery in orchard field.