Two-layer model predictive formation control of unmanned surface vehicle

Abstract

This paper studies the model predictive formation control problems of unmanned surface vehicles (USVs). A two-layer model predictive control (MPC)-based formation control strategy is proposed. The proposed formation strategy artificially divides the model of each USV into two layers according to its kinematic model and dynamic model. Using the kinematic model, a distributed MPC algorithm is designed to ensure formation and coordination; with the generated command from the distributed MPC as tracking reference, an MPC is designed for the dynamic model to ensure optimal command tracking. Simulation and comparison studies show that the designed algorithm is computational more efficient in comparison with the centralized MPC formation algorithm.