Robotic Grinding Path Planning for Unstructured Industrial Scenarios

Currently, the casting grinding operation mainly relies on manual operation. However, manual operation results in problems such as low efficiency and unstable processing quality. This study is dedicated to developing an intelligent grinding system based on industrial robots. For the purpose of generating suitable grinding paths, we propose an adaptive path-planning algorithm based on point cloud information, called Grid-based Grinding Path Planning (GriPP), which can generate optimal grinding paths in the defective regions of castings. To better apply the paths planned by GriPP, an improved inverse kinematics algorithm, the Motion Axis Decomposition-Based Robot Inverse Kinematics Algorithm (MAD-RIK), is designed for grinding operations. Additionally, we have developed a robotic grinding control system that implements the entire process flow, from point cloud acquisition and path planning to motion control. Through experimental verification, GriPP demonstrates good performance in typical casting grinding tasks, and MAD-RIK requires less time than other methods while maintaining accuracy.