Development of a Collision-Free Path Planning Method for a 6-DoF Orchard Harvesting Manipulator Using RGB-D Camera...

Area of Science:

• Robotics
• Agricultural Engineering
• Computer Vision

Background:

• The agricultural sector faces challenges with an aging workforce, necessitating automation in fruit harvesting.
• Robotic manipulators with high degrees of freedom (DoF) are crucial for navigating complex orchard environments.
• Existing path planning methods struggle with efficiency and computational cost in unstructured settings.

Purpose of the Study:

• To develop an efficient and collision-free path planning method for 6-DoF orchard harvesting manipulators.
• To address limitations in current path planning techniques regarding environmental mapping and computational demands.
• To enable autonomous fruit harvesting in complex and unstructured orchard settings.

Main Methods:

• Utilized an RGB-D camera to reconstruct 3D obstacle maps from point cloud data.
• Developed a manipulator simulation model in URDF format integrated with the 3D obstacle environment.
• Implemented the Bi-directional Rapidly-exploring Random Tree (Bi-RRT) algorithm for efficient path planning.

Main Results:

• The Bi-RRT algorithm successfully generated reliable collision-free paths in all experimental trials.
• The proposed method achieved an average path generation time of 0.806 seconds and 12.9 nodes.
• Demonstrated superior efficiency compared to the Sparse Bayesian Learning (SBL) algorithm (0.870 s, 15.1 nodes).

Conclusions:

• The developed collision-free path planning method is effective and fast for 6-DoF orchard harvesting manipulators.
• This approach offers a viable solution for autonomous fruit harvesting in complex agricultural environments.
• The findings provide significant guidance for the practical implementation of robotic harvesting systems.