Visual attention and swarm cognition towards fast and robust off-road robots

Abstract

This article argues in favour of using visual attention mechanisms on off-road robots and of exploiting the social insects metaphor for their robust implementation. Visual attention helps these robots focusing their perceptual resources on a by-need basis when dealing with the complexity of unstructured environments. However, focusing perceptual resources is a hard problem given the well known speed-accuracy trade-off and the fact that several foci of attention may need to co-exist and interact with both memory and action selection. The similarity between the task of deploying parallel foci of attention and the foraging behaviour exhibited by army ants motivates the use of the social insects metaphor to solve the problem at hand in a self-organising, and consequently, robust way. All these arguments are phenomenologically supported by experimental work recently published on three foundational aspects of off-road mobility: obstacle detection, trail detection, and local navigation.