Formation Control of Networked Autonomous Underwater Vehicles with Obstacle Avoidance

Abstract

Multiple autonomous underwater vehicles (AUVs) can be deployed at the same time for challenging marine tasks. The group of AUVs usually develops a network and forms a particular structure in space to enhance their capabilities and efficiency at their destination. For the deployment of networked AUVs to a common destination with a desired formation, a decentralized controller is designed by considering the 3D nonholonomic constraints of AUVs as well as collision avoidance with obstacles in the environment. During the traversal of the group, all AUVs are networked via a directed spanning graph that represents leader-follower relationships within the group. Only the leader AUV is aware of the destination and subsequently steers itself to approach it. Through maintaining the leader-follower relationships, the follower AUVs are led to the destination and also build the desired group formation. Each AUV must also abide by its nonholonomic constraints in 3D space and avoid obstacles detected at any time in its neighborhood. The proposed decentralized controller is developed from a special navigation function that generates repulsive and attractive potentials to avoid any collisions and disconnections of the network and further guides each AUV to its desired position as well as achieving the group-level structure. Simulation results are presented to demonstrate that the decentralized controller is able to successfully deploy a group of AUVs to the desired destination and formation with obstacles in the environment.