Three-dimensional consensus control based on learning game theory for multiple underactuated underwater vehicles

Abstract

This paper proposes a distributed controller for a rendezvous problem of multiple underactuated underwater vehicles (UUVs). The multiple UUVs move in three dimensional space. The formation controller designing process is divided into two parts. In the first part, a systematic distributed optimization method will be established on the basis of a learning game algorithm. This algorithm inherits the learning game theory's nature, that is, simplification and convergence. The property of learning makes the algorithm robust and autonomous. In the second part, based on above algorithm and the line of sight method, the way point tracking controller is designed. In this paper, tracking control problem of position error is converted into the stable control problem of speed error, the heading angle error and the vertical inclination error. The process of controller design is divided into two stages, which avoids the singularity problem. Based on the Lyapunov stability theory and adaptive backstepping iterative design method, precise tracking control of the UUV speed, heading and pitch angle is achieved. Finally, simulation results illustrate the validity of the proposed controller.