Tube-based model predictive control of an autonomous underwater vehicle using line-of-sight re-planning

This work investigates discrete-time 3D trajectory tracking control of an autonomous underwater vehicle (AUV) subject to input saturation and unknown environmental disturbances. Firstly, a line-of-sight strategy is proposed to achieve local re-planning. The re-planned trajectory is employed in controller design to restrict the magnitude of tracking errors, mitigating abrupt changes in velocity and control input. Secondly, the robustness to environmental disturbances is achieved by employing a tube-based model predictive controller. The control scheme consists of two controllers: one is a model predictive controller based on the nominal AUV model for reference tracking, and the other is a state-dependent feedback controller used to construct time-varying tubes so as to ensure that the perturbed system remains within a tube centered around the nominal trajectory. Under given assumptions, the proposed controller guarantees (local) input-to-state stability of the closed-loop system. Thirdly, a rate of energy consumption metric is formulated to assess the control performance. Simulation studies under realistic ocean environmental conditions demonstrate the effectiveness of the proposed algorithm in comparison with a nonlinear model predictive controller.