Research On Real-time Obstacle Avoidance Algorithm Of Harvesting Robot Manipulator Based On Joint Space

Agriculture is a typical labor-intensive industry,and the harvest of fruits or vegetables is the key to consuming the largest number of laborers and difficult to implement mechanized operations.Even though agricultural operations in developed countries have reached a high level of automation,their picking processes still rely heavily on manual labor.The automation of picking operations has become a realistic demand for the development of agricultural production.Applying picking robots to actual agricultural production will help alleviate the problem of labor shortage,help reduce the cost of agricultural products,and increase the competitiveness of agricultural and sideline products in the market.In the process of picking,there are randomly growing branches and leaves,and immature fruits hinder the movement of picking robotic arm.This problem makes it difficult for picking robots to be truly applied to agricultural production activities and restricts the development of the picking robot industry.Therefore,research on the planning method of real-time obstacle avoidance path for picking robotic arm is very important for promoting the application of picking robot in agricultural production.The main research content of this article is as follows:1、Kinematic analysis and collision detection algorithm for harvesting manipulator.Through analyzing the charateristics of a variety of different types of spatial geometry envelope models,the collision detection envelope model of the manipulator and space obstacle is determined.The mathematical model of the manipulator was established based on the screw theory,and the forward and inverse kinematics equations of the manipulator were deduced,and the simplified collision detection model was constructed based on the structural parameters of the manipulator.According to the type of the collision detection model between the robotic arm and the space obstacle,the spatial geometry collision detection algorithm is designed.2、Fast configuration space construction algorithm.Based on the analysis of the shortcomings of the traditional joint configuration space construction method,this paper proposes a fast joint configuration space construction algorithm.The algorithm consists of two parts,the offline phase and the online phase.The off-line phase establishes an external and self-collision database by interfering with the workspace unit under an arbitrary configuration of the traversing robot arm.In the online phase,the discrete working space units occupied by obstacles in the real-time scene are calculated,and the mapping model of workspace obstacles in the joint configuration space is constructed based on indexing external and self-collision databases.3、Real-time obstacle avoidance path planning algorithm for high dimensional joint configuration space.Combined with heuristic-based and sampling-based obstacle avoidance path planning algorithm,a Batch Informed Trees(BIT*)was introduced.A two-thread real-time obstacle avoidance path planning strategy is proposed.The planning thread mainly performs the obstacle avoidance path planning task in the current scene.The execution thread is mainly used to communicate with the control process and determine the real-time refreshed planning scenario and the currently executing robot arm obstacle avoidance.Based on the real-time obstacle avoidance path planning strategy,the BIT*algorithm is improved to make it suitable for obstacle avoidance path planning under dynamic environment.4、Using a virtual simulation system built in Matlab to simulate the obstacle avoidance motion the harvesting manipulator.The effectiveness of the proposed algorithm has been validated by changing the spatial position of the obstacles to block the initial collision-free path which produced by the proposed algorithm.At the same time,the proposed algorithm is deployed in the control system of the fruit harvesting robot,and the feasibility and effectiveness of the algorithm in the actual picking process are verified by experiments.The results show that the proposed algorithm can correctly handle the tasks set by the experiment in the design time in both the virtual simulation system and the real robot picking system,indicating that the proposed algorithm can quickly and effectively perform real-time obstacle avoidance path planning for the manipulator.