Obstacle Avoidance Path Planning for Harvesting Robot Manipulator Based on MAKLINK Graph and Improved Ant Colony Algorithm

摘要

To the problems of real-time obstacle avoidance path planning for apple harvesting robot manipulator in dynamic and unstructured environment, a method based on improved ant colony algorithm is presented. Firstly, Vector description is utilized to describe the area where obstacles such as branches is located as irregular polygon in free space, and MAKLINK graph is used to build up the environment space model. Then, the improved Dijkstra algorithm is used to find the initial walk path for apple harvesting robot manipulator. Finally, the improved ant colony algorithm is applied to optimize the initial path. The experiment result shows that the proposed method is simple and the robot manipulator can avoid the branches to pick the apple successfully in a relatively short time.