Cooperating Auv teams: Adaptive area coverage with space-varying communication constraints

This contribution addresses the problem of area coverage by a team of autonomous underwater vehicle. The case addressed is the one in which the area map is not known in advance but adaptively estimated on-line by the vehicles themselves as they move along the area. The approach proposed is distributed, i.e., the vehicles in the team exchange information among themselves and autonomously take individual decisions on where to move next on the basis of the available information. Communication constraints among the vehicles, in terms of communication range and bandwidth, are explicitly taken into account by the algorithm, and can be dependent by the spatial position in the covered area. The paper proposes a series of algorithms for adaptive sampling with communication constraints in which the communication constraints are expressed in terms of connectivity of the graph of the vehicles. Depending on the structure of the graph (ordered serial graph, ordered tree graph, etc.) the algorithms take the structure of a distributed dynamic programming approach or of minimum spanning tree graph searching.