Motion Planning of Tree Fruit Robot Picker in Presence of Obstacles Using Heuristic Approach

The paper introduces a heuristic approach for planning the motion of fruit-picking robots operating among various, randomly placed obstacles within the task's operational space. A linear robot, equipped by three degrees of Cartesian-type freedom, has been designed for this task, featuring an additional telescopic arm and a specialized gripper. This gripper is uniquely fitted with two rotational joints at the wrist, enabling precise orientation adjustment of the tool in relation to the fruit on the branches. The planning of the robot's movement leverages machine vision and artificial intelligence algorithms, including artificial neural networks and fuzzy logic for decision making. This heuristic approach to the robot picker's movement planning draws inspiration from human-inspired techniques used in manual fruit picking on fruit plantations. The methodology employed utilizes environmental data to determine the shortest paths for the robot's gripper to reach specific targets, and to assess the relative position (orientation) of the actuator in relation to the target or obstacles encountered along the way (such as branches, clusters of leaves, trees, etc.). The effectiveness of this movement planning methodology has been confirmed through simulation models of the robot and its environment.