Optimal Trajectory Planning and Multimodal Force Control for Robotic Large Surface Rapid Grinding

Abstract

Robotic grinding of large walls faces challenges such as difficulty in trajectory planning to meet overall flatness and unstable contact forces. This paper presents a surface geometric mapping-based optimal trajectory planning method to generate grinding paths and optimize wall surface flatness. In this method, wall features are measured using distance sensors to establish a 2D mapping model between the surface and a reference plane. Trajectory points are searched based on the relationship among the feature points, material removal model, and radial height. Then, the surface is segmented into several micro-surface cells, and the posture relationship among them is explored to generate the end-effector’s pose data. We also propose a multimodal force control algorithm that switches force control strategy based on radial height states, dynamically adjusting the end-effector’s pose through force fusion to enhance contact force stability. Comparative experimental results with two representative methods show our method’s advantages, featuring a mean overall wall flatness of 1.997 mm and grinding efficiency improvements by 18.51% and 12.7%, respectively, indicating its potential for grinding large surfaces in real applications.