Cooperative flow field estimation using multiple AUVs

This paper presents a cooperative method to estimate the flow field by a group of autonomous underwater vehicles (AUVs). In this paper, it is assumed that each vehicle can detect the relative positions of its neighboring AUVs during the underwater phase. Since AUVs' trajectories depend on the initially unknown flow field, we define the deviation between the actual and predicted trajectories as the motion-integration error, and the difference of the actual and predicted relative positions between an AUV and its neighbor as the relative motion-integration error. Using these integration errors, a system of nonlinear equations for vehicle trajectories and unknown flow fields is constructed. Then the flow field is estimated by solving an inverse problem for these equations with two different types of error constraints. The convergence of the cooperative estimation algorithm is proved. Finally, simulations are provided to illustrate the effectiveness of the proposed algorithm.