ARCANE: towards autonomous navigation on rough terrains

Abstract

An approach to rough terrain autonomous navigation is outlined. The study has been motivated by the fact that an autonomous robot navigating over unknown terrain acquires data continuously as it moves. The approach was based on simple and fast algorithms working on 'robocentric' maps acquired by onboard sensors. The lower layers of the architecture do not model the environment, but rely on the high frequency of the perception-planning cycle to accomplish successful navigation toward a visible goal. These layers have been designed to be robust with respect to perception errors, and the discontinuous nature of the environment's perception. They were also designed to cope with the fact that the robot discovers the environment continuously as it moves.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>