Fault Tolerance for Autonomous Robots by Means of Adaptive Filters

Abstract

In this paper we discuss the effect of component faults in the legs of autonomous hexapod robots. This has been carried out by means of the simulation of the hexapod robot OSCAR developed at the University of Lubeck. For each of the considered fault classes, a compensating Adaptive Filter algorithm has been developed. The fault effects and their compensations are illustrated by the examples of slow motor response of a legujoint and an improper orientation of a leg at start position of a robot trajectory. The different algorithms have been integrated to form, inside each robot leg, one coherent Leg Component Fault Tolerance Layer.