Resilient Autonomous Exploration and Mapping of Underground Mines using Aerial Robots

Abstract

This paper presents a comprehensive solution to enable resilient autonomous exploration and mapping of underground mines using aerial robots. The described methods and systems address critical challenges related to autonomy, perception and localization in conditions of sensor degradation, exploratory path planning in geometrically complex, large and multi-branching environments, alongside reliable robot operation in communications-denied settings. To facilitate resilient autonomy in such conditions, a set of novel contributions in multi-modal sensor fusion, graph-based path planning, and robot design have been proposed and integrated in micro aerial vehicles which are not subject to the challenging terrain found in such subterranean settings. The capabilities and performance of the proposed solution is field-verified through a set of real-life autonomous deployments in underground metal mines.