Constraint-induced formation switching for adaptive environmental sampling

Water quality monitoring is still mostly done by taking manual water samples and sensor measurements from boats. To enable extensive, efficient and repeatable environmental monitoring, there is a need for `ready to sample' robot systems, which do not require individual vehicle control, or a lot of prior information. This paper describes an approach to decentralized adaptive formation control for environmental sampling. An autonomous surface vehicle (ASV) leads a team of autonomous underwater vehicles (AUVs) to sample a lake environment. The ASV passes a constraint to the AUVs, and the AUVs use this to choose an allowed formation, and solve the assignment problem to determine their position in the formation, in a distributed manner. The approach is tested in simulation and compared to leader-follower formation control. Results show the potential for constraint-induced formation switching in adaptive formation control towards a safe, fully autonomous heterogeneous team of lake sampling robots.