A modal-disturbance-observer-based generalized-force model predictive control for attitude stabilization of a wave-adaptive vessel

Unknown wave excitation forces pose a substantial challenge to the attitude stabilization of wave-adaptive vessels, as they cannot be measured directly in real time. Conventional feedback control—reacting only after heave, pitch, and roll motions have developed—introduces an inherent response delay. This delay restricts the disturbance‑rejection capability of active suspension systems under phase‑varying wave inputs. This paper proposes a modal‑disturbance‑observer‑based generalized‑force model predictive control (MDO‑GMPC) method for a seven‑degree‑of‑freedom wave‑adaptive vessel. First, a seven‑DOF dynamic model is established to characterize the coupled motion of the upper platform and its four pontoons. A spatial modal transformation is then introduced to decompose the equivalent wave excitation at four pontoon locations into heave‑, pitch‑, roll‑, and torsion‑related modal disturbances. Based on the transformed model, a modal disturbance observer is designed by augmenting the modal disturbance states, thereby reconstructing the unknown wave‑force information from measurable vessel, suspension, and actuator responses. The estimated modal disturbances are further incorporated into the prediction horizon of the MPC scheme as feedforward information, while the control input is formulated in generalized‑force coordinates to coordinate the four active suspensions and reduce actuator redundancy. The main contribution of this work lies in the integration of online modal disturbance observation with feedforward predictive compensation, enabling the controller to respond proactively to wave excitation before significant attitude deviations occur. Experiments under both regular‑ and random‑wave conditions with four‑channel wave excitation are conducted, comparing the proposed method with passive suspension and conventional MPC. The results show that MDO‑GMPC achieves more accurate reconstruction of wave‑force trends and superior suppression of vertical motion, pitch, and roll responses, confirming its efficacy for attitude stabilization of wave‑adaptive vessels subject to unknown wave disturbances.