Real-Time Reconfiguration and Safe Navigation for AUVs Network Using Distributed Nonlinear MPC and Relaxed CBFs: Theory and Experimental Validation

Abstract

The controlled coordination of multiple agents provides an opportunity for increased robotics system efficacy. This could be realised through the execution of more complex tasks, reduction of completion time, or enhanced spatial coverage. However, limited communication between agents necessitates a robust and distributed coordination strategy to maintain operational effectiveness and reliability. This study explores the design and implementation of a novel distributed nonlinear model predictive control scheme that integrates control barrier functions (CBFs) for trajectory tracking and formation control of multiple underwater robots. CBFs act as safety constraints, guaranteeing that system safety and performance objectives are simultaneously considered within a short prediction horizon, reducing the computational burden in real-time implementation. In addition, the relaxed decay rate technique enhances the feasibility of the optimization and system safety at the same time. A series of validation scenarios, including obstacle avoidance, are presented through both software-in-the-loop simulations and physical experiments in the aquatic environment. These scenarios demonstrate the robustness and flexibility of the proposed control strategy, ensuring safe and coordinated operations of multiple underwater robots.