SAI-RRT*: Safety-Aware Informed RRT* for Multi-Joint Manipulator Path Planning in Static Environments

Area of Science:

• Robotics
• Artificial Intelligence
• Motion Planning

Background:

• Path planning for multi-joint robotic manipulators is challenging in complex industrial settings due to high-dimensional spaces and safety needs.
• Conventional Informed RRT* algorithms suffer from low initial sampling efficiency and poor adaptability.

Purpose of the Study:

• To propose Safety-Aware Informed RRT* (SAI-RRT*), a framework enhancing planning efficiency and geometric collision safety for 6-DOF manipulators.
• To address bottlenecks in traditional path planning methods for industrial robotics.

Main Methods:

• Implemented an adaptive constrained circular area sampling technique to improve initial path discovery.
• Integrated a joint constraint framework with kinematic verification and reinforcement learning to prevent node rejection.
• Embedded forward-kinematics-based full-body geometric verification using a capsule model for efficient link-clearance checking.

Main Results:

• SAI-RRT* demonstrated improved planning efficiency compared to conventional methods.
• Enhanced geometric collision safety was achieved through integrated verification processes.
• The proposed method resulted in better path quality in numerical experiments.

Conclusions:

• SAI-RRT* effectively optimizes path planning for multi-joint robotic manipulators in industrial environments.
• The framework offers a robust solution for improving both the speed and safety of robotic motion planning.