Implementation of a Real-Time Object Pick-and-Place System Based on a Changing Strategy for Rapidly-Exploring Random...

Area of Science:

• Robotics and Automation
• Artificial Intelligence
• Computer Science

Background:

• Autonomous object pick-and-place systems require robust collision-free path planning for robot manipulators.
• Existing path planning algorithms face challenges in balancing success rate and computation time in complex environments.
• Six-degree-of-freedom (DOF) robot manipulators are crucial for versatile pick-and-place operations.

Purpose of the Study:

• To propose an improved path planning algorithm, Changing Strategy Rapidly-exploring Random Tree (CS-RRT), for robot pick-and-place systems.
• To enhance the success rate and reduce the computing time of collision-free path planning for six-DOF robot manipulators.
• To validate the effectiveness of the CS-RRT algorithm in complex environments through simulations and practical experiments.

Main Methods:

• Implementation of an object pick-and-place system using Robot Operating System (ROS), a camera, a six-DOF robot manipulator, and a two-finger gripper.
• Development of the CS-RRT algorithm, an enhancement of the Gradually Changing Sampling Area Rapidly-exploring Random Tree (CSA-RRT).
• Incorporation of a sampling-radius limitation mechanism and a node counting mechanism into the CS-RRT algorithm to optimize path planning.

Main Results:

• The CS-RRT algorithm demonstrated superior performance in terms of success rate and reduced computing time compared to two other RRT algorithms in simulations.
• The sampling-radius limitation mechanism efficiently guides the random tree towards the goal, minimizing time spent searching near the target.
• The node counting mechanism allows the algorithm to adapt sampling strategies in complex environments, preventing search path entrapment and improving adaptability.

Conclusions:

• The proposed CS-RRT algorithm provides an effective solution for collision-free path planning in robot pick-and-place tasks.
• The CS-RRT algorithm successfully enhances both the efficiency and reliability of robot manipulation in complex, real-world scenarios.
• Practical experiments confirmed the robot manipulator's ability to complete pick-and-place tasks effectively using the CS-RRT-based path planning.