Autonomous navigation and exploration in a rescue environment

In this paper we present an approach to autonomous exploration of a rescue environment. Exploration is based on unexplored frontiers and navigation on a two-level approach to the robot motion problem. Our method makes use of a motion planner capable of negotiating with very fine representations of the environment, that is used to move in a cluttered scenario. A topological path-planner "guides" the lower level and reduces the search space. The two algorithms are derived from two widely-used probabilistic algorithms, currently successfully deployed in many robot applications, the Probabilistic RoadMap and the Rapid-exploring Random Trees; however, their adaptation to the rescue scenario requires significant extensions.