Distributed Dynamic Programming for Adaptive On-line Planning of AUVs Team Mission with Communication Constraints

An algorithm for adaptive on-line planning of environmental exploration missions with a team of Autonomous Underwater Vehicles (AUVs) is proposed. The algorithm has the primary goal of determining an estimate of the sampled environmental quantity with an estimation error below a prescribed threshold. The additional degree of freedom of the algorithm is exploited to spread the team over the exploration area, in order to minimize mission time while, at the same time, the communication connectivity of the team is preserved. A distributed dynamic programming approach is employed in order to satisfy these two conflicting requirements.