Model Predictive Control for Autonomous Underwater Vehicles

The Attainable Set Model Predictive Control scheme is discussed and shown to meet the needed system behavioral properties while satisfying real-time requirements underlying the control of Autonomous Underwater Vehicle formations, including the strict on-board resource constraints. More specifically, the proposed approach targets the on-line computational complexity and relies on taking advantage of the control problem time invariant elements, in order to replace, as much as possible, on-line by off-line computation, while guaranteeing asymptotic stability, and promoting the best trade-off between feedback control near optimality, and robustness to perturbations (due to disturbances, and uncertainties), and adaptivity to the environment variability. The data computed off-line is stored onboard in look-up tables, and recruited and adapted on-line with small computation effort according to the real-time context specified by communicated or sensed data. This scheme is particularly important to an increasing range of applications exhibiting severe real-time constraints. (C) 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the 13th International Symposium "Intelligent Systems" (INTELS'18).