Reactive planning and control with mobile robots

Abstract

A system that enables robots to perform nonelementary tasks in an environment where unpredictable demands require adaptation of the robots' actions is described. A motion control system dealing with this problem is implemented in the context of a representative motion task. The control system is composed of a planning module (written in the action network formalism) that governs various behaviors implemented with artificial potential fields or other specific control strategies. The hardware employed in the experiments is described. The demands of a reactive architecture for robots are examined in detail. The main features of the action network and artificial potential field techniques are presented, as well as explanations of the way they were integrated with other modules to form a whole control system. A demonstration on a GOFER robot is described.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>