Global optimization of functional redundancy in a 6R robot for smoothing five-axis milling operations

Abstract

Functional redundancy optimization is one of the most effective ways to improve robotic posture and performance. However, owing to the trade-off between smoothness of the robot's motion and other important aspects of performance, the optimization problem is too nonlinear to be solved in robotic milling processes. In this article, a derivation is shown that the robot's motion will be approximately smooth when the functional redundancies for each cutter location are smoothing variables. On this basis, a Legendre surrogate model is constructed to plan the overall optimal functional redundancy by remarkably reduced computation. Moreover, a trimmed sequential linear programming method is proposed to improve the solution further. From the outcomes of a 3D S-shaped grooves real milling experiment, the robot motion paths planned by the proposed methods achieve more stable tool-tip milling performance than that of comparable methods, while being several times or even 10 times faster in computational efficiency.