Path Planning Of Apple Picking Manipulator Based On Improved RRT Algorithm

With the continuous development of society and the improvement of people’s living standards,the fruit and vegetable industry continues to rise,and the apple production and market demand continue to increase.Traditional picking methods are gradually being replaced by automated picking technology due to problems such as high labor costs and low efficiency.As one of the key equipment in modern agricultural production,picking robotic arm is attracting more and more attention.The main contents of this article include:(1)It mainly introduces the general overview of the kinematics analysis and collision detection of the manipulator,and takes the Aubo-i5 manipulator as an example to carry out the kinematics analysis,and uses MATLAB software to model and verify the picking manipulator.Then,the collision detection method of the robotic arm is introduced,and the commonly used bounding box types are explained in detail,and the cylindrical envelope and the bounding sphere envelope method are used to process the robotic arm,thereby effectively reducing the calculation of the collision detection between the robotic arm and obstacles The amount and complexity,and improve the efficiency of detection.(2)In order to solve the problem of fruit positioning of the picking robot arm,a Real Sense D435 depth camera was introduced,and the camera was calibrated and hand-eye calibrated.By preprocessing the collected RGB-D images,the image is processed using a square fitting algorithm to improve the positioning accuracy,and the processed results are converted into point cloud data for point cloud registration.Use the preprocessed RGB-D image to carry out positioning experiments on the target,in order to achieve the purpose of optimizing the precise grasping of the manipulator for picking.(3)For the traditional RRT algorithm,there are problems such as blind search direction and too many search nodes in the obstacle avoidance path planning of the apple picking manipulator.Aiming at these problems,this thesis proposes an improved RRT algorithm,which combines the APF algorithm and the breadth-first search pruning method,conducts simulation comparison experiments between different algorithms in two-dimensional and three-dimensional environments,and analyzes the results to verify the Propose the effectiveness of the algorithm.(4)The overall structure of the apple picking robot is systematically outlined,and a real picking experiment is carried out,and an obstacle avoidance path is successfully generated.Compared and analyzed the improved algorithm with RRT algorithm,RRT* algorithm and RRT-Connect algorithm,in terms of the number of search nodes,average running time,path length,etc.,it shows that the improved algorithm has great advantages.The research in related fields of planning provides important reference and reference.