Concise Robust Adaptive Control For Ships Under Phenomenal Sea States

In order to solve the problem that the existed marine control theory is not applicable to the phenomenal sea ("Phenomenal sea" is defined in the appendix 1), the research topic is selected as "path following control for marine ships under phenomenal sea states" and the theoretical exploration has been carried out systematically in this thesis. The research work could provide the theoretical guidance for the marine salvaging and shipping under the phenomenal sea state.This research work includes 5 key aspects:identification modeling, test platform constructing with the simulated phenomenal sea state, guidance principle, ship motion control theory and implementation in the control engineering. By employing robust control, adaptive control and the concise idea, the derived control algorithm is with robustness, adaptability to the varying sea states and the concise form that is easily implemented in the practical engineering. The objective of this thesis is to develop a set of path following control scheme for marine vehicles, which could meet demands of the control engineering and be implemented normally under the practical environmental condition.In this thesis, one concerns with the problem:the ship motion states are with the characteristic of multivariable coupling and the innovation is limited in the conventional maneuvering test. The auto-constructed multi-innovation identification scheme is developed for the ship maneuvering motion identification by employing the multi-innovation identification technology. The ship maneuvering test is conducted by virtue of the scientific research vessel YUKUN of Dalian Maritime University, China, including the turning trail and the symmetrical/asymmetrical zigzag trail. Then one does the full-scale trail data-related modeling works for vessel YUKUN using the auto-constructed multi-innovation identification algorithm. Combining with the physical based mathematical model of marine disturbances (i.e. sea wind, irregular wind-generated wave and ocean current), the test platform is actually constructed that could simulate the phenomenal sea state.Guidance and control are two important modules for the ship to navigate atomically. About the guidance module, the exited LOS steering law cannot be directly applied to path following control of underactuated ships. Motivated by this observation, a novel "dynamical virtual ship" guidance principle is developed by combining the virtual ship’s guidance and dynamical switching mechanisms, where the waypoints-based planning trajectory is set by the navigator. The proposed principle could provide more reasonable guidance for the automatic navigation of ships. As to the control module, some constraints are targetedly considered by the proposed scheme, such as the model uncertainty, dynamics of the actuator, the real-time requirement and the phenomenal sea disturbance. In this part, the online updating of neural networks weights is avoided for the merit of the robust neural technique and the low-frequency gain learning algorithm. Only two gain related learning parameters are required to stabilize the dynamical ship with uncertainty and unknown disturbances. The present algorithm is with some advantages:a concise form, robustness and ease to implementation in the practice.In this thesis, all the experiments are carried on using the Matlab/Visual Basic mixed programming. The results under the simulated phenomenal sea states have been presented to illustrative the effectiveness of the proposed design. This research has the practical significance of improving the maritime safety of ships and promoting the development of the ship automation equipment. That would provide the theoretical basis for implementing the efficient shipping.