Distributed safety-critical formation control for networked marine surface vehicles with observer-enhanced robustness

This paper proposes a distributed safety-critical model predictive control (MPC) framework for leader referenced formation tracking of networked marine surface vehicles (NMSVs) with 3-DOF nonlinear dynamics, actuator saturation, and unknown matched disturbances. Collision avoidance, connectivity maintenance, obstacle avoidance, and input constraints are enforced as hard constraints with strict priority over tracking performance. A disturbance observer is integrated into a DO-enhanced control barrier function (CBF) design by embedding the disturbance-estimation error energy into the barrier construction, yielding time-varying tightened safe sets whose tightening progressively diminishes as the observer converges. Each NMSV solves a local finite-horizon optimization problem subject to the DO-enhanced CBF constraints, in which the DO-compensated auxiliary input is incorporated as a Lyapunov-based performance requirement, softened via a slack variable so that this performance requirement never becomes an additional source of infeasibility beyond what is already required by the hard safety and actuator constraints. Simulations in cluttered environments with multiple obstacles demonstrate that the proposed method maintains safety and connectivity constraints while enhancing robustness compared with several existing methods.