Finding efficient paths for multi-robot path coverage

This work presents a solution to find an efficient set of paths in order to cover all locations for a robotic search and rescue mission. Assuming that a set of locations with their available robotic paths are given, our problem is to find multiple paths for robots including their start locations and corresponding paths. Our algorithm is based upon all-pairs shortest paths, since shortest paths among all nodes are guaranteed for the optimized solution; however, combinatorial choice among them is not trivial. Our algorithm is heuristic, combining paths according to how small each path cost is and how rare the connectivity of nodes is. For comparison, we also present a method of decomposing paths from a minimum spanning tree. Through the experiments in simulation, we have demonstrated that our shortest-based algorithm finds a set of efficient paths in which the maximum path cost is smaller than the one chosen by related algorithms, a nearest neighbor and a minimum spanning tree.