Coverage path planning for multiple robotic agent-based inspection of an unknown 2D environment

Abstract

This paper describes a coverage-based path planning algorithm for multiple robotic agents with the application on the automated inspection of an unknown 2D environment. The proposed path planning algorithm determines a motion path that a robotic agent will follow to sweep and survey all areas of the unknown environment, which is enclosed by the known boundary. The 2D unknown environment is decomposed into a union of simplices using the principle of Delaunay triangulation. The area coverage is equivalent to design a path for a robotic agent to follow and visit all simplices subject to certain mission constraints. A hierarchical mission planner is designed to allocate mission tasks among multiple agents in each level and pass information down to the next level along the hierarchy. The proposed path planning algorithm has been tested and evaluated on the problem of planning path for two types of robotic agents - flying agents and crawling agents in a two-tier hierarchical mission planner to cover various unknown 2D environments.