Adaptive Grinding Planning for Robotic Arms Based on Parameterized Cost Estimation and Dynamic Hierarchical Optimization

Abstract

In adaptive grinding task, robotic arms are required to autonomously achieve uniform grinding of workpieces. This brings a great challenge to estimation and planning techniques. In this paper, a double-layer planning framework is designed to make industrial robotic arms accomplish adaptive grinding tasks for various tee tubes of arbitrary size, spatial orientation and surface characteristic. Dynamic-hierarchical-ant-colony-system (DHACS) and velocity-vector-decomposition-based (VVD-based) parameterized path planning are designed and applied to calculate the optimal grinding order and generate grinding paths along tee tube surfaces. Experimental results demonstrate that the shortest grinding path along arbitrary tee tube is generated adaptively by the planning framework proposed. Compared to the state-of-the-art methods, the efficiency and effectiveness of the planning framework proposed in this paper are demonstrated. Specifically, in terms of optimal solution quality, the performance advantages are at least 1.224%, 4.629% and 3.673% respectively.