Set-based collision avoidance applications to robotic systems

Abstract

A robotic system can consist of a single or multiple agents with a fixed or mobile base, with full or under-actuation, and possibly redundancy. Collision avoidance is a crucial task for any robotic system and is necessary to ensure safe operation. In this paper, we use a set-based approach to ensure collision avoidance as a high-priority task of a robotic system while simultaneously defining one or more tasks for the system to achieve. The set-based approach is highly generic and flexible, and we present theoretical results, practical implementation and experimental results of the approach applied to a redundant, fully actuated robot manipulator, a full-scale underactuated surface vessel and a multi-agent system of unicycles.