Integrating force-fields methods in a robot planning/programming language

Abstract

In advanced robotic applications which involve unstructured and cluttered environments, highly redundant robots are considered very useful because they allow one to negotiate obstacles and various kinds of constraints. The price, however, is the increased complexity of the planning and control tasks. In this paper the authors present a prototype of a robotic task planning system that integrates an analogical planner, based on relaxation networks (M-Nets), with a robot planning/programming language (NEM++). M-nets are a parallel and distributed architecture which operates via relaxation in a similar way to a Hopfield net for generating synergies of motor commands in polyarticulated robotic structures.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>