Identification-based controller design using cloud model for course-keeping of ships in waves

Abstract

Course-keeping plays an important role in ensuring navigation safety of ships. This contribution proposed new approaches about system identification and controller design to facilitate the main goal which is to design a controller for course-keeping of ships based on the identified ship plant and the cloud model. The investigated plant as the control model was the first order linear Nomoto model. In order to estimate the parameters of this model, the support vector machines (SVM) optimized by artificial bee colony algorithm (ABC) was applied in combination with simulated data including the rudder and heading angles generated by the dynamic model of a Mariner class cargo ship. Based on the identified linear Nomoto model of Mariner class cargo ship, the cloud model was then applied to design the controller for course-keeping. To well demonstrate the feasibility and effectiveness of the proposed controller, a fuzzy logic PID controller, and a PID controller were also considered as comparison mechanisms. Aiming at validating the robustness of the proposed cloud model-based controller, it was used to compensate for the critical environmental disturbances induced by waves. Finally, simulation results indicate the desirable performance of ABC on optimizing parameters in SVM. Comparison with PID and fuzzy logic PID controllers demonstrates that the cloud model-based controller presents slightly preferable performance on course-keeping. This is due to that the flexible and intuitive modification of the digital characteristics of the cloud model and the structure of cloud inference engines makes the cloud model efficient to satisfy the required mapping between inputs and outputs for a controller.