Smooth path planning for redundant robots on collision avoidance

Abstract

This paper presents a new algorithm for programming secondary tasks for redundant robots under and out of the collision imminence. For the collision avoidance, the new algorithm uses the concept of Assur virtual chains, together with interpolations based on cubic Bézier polynomials, taking advantage of its properties in terms of continuities, for smoothing the transitions between tasks. Out of collision condition, the redundancy is addressed by the method of Adaptive Extended Jacobian. This set of strategies, acting together, results in a stable algorithm and in smooth trajectory profiles for redundant robots avoiding collisions. An application using a redundant P3R-planar robot is developed, and graphical comparative analyzes are presented.