| Summary: | 碩士 === 國立臺灣海洋大學 === 通訊與導航工程學系 === 100 === This work uses a small boat as an experimental platform for comparison study of model-based and model free ship track-keeping autopilots. Specifically, the Internal Model Control (IMC) and Linear Quadratic Regulator (LQR) methods are examined in the model-based approach and the Fuzzy Logic Control (FLC) and Artificial Neural Network (ANN) are investigated in the model-free approach. The two guidance methods adopted are the Line-Of-Sight guidance (LOS) method and the New Line-Of-Sight guidance (NLOS) method. LOS method is vulnerable to environment disturbance resulting in an offset of the ship track, the NLOS method makes an earlier correction to the ship's position, thus reducing the ship path trajectory deviation. No matter with what kind of guidance method, is adopted the model-based controllers is tend to exhibit over shoot behavior. However, the model-free autopilots produce smoother track-keeping trajectories. Generally speaking, the model-based autopilots are easier to design, because there are specific guide lines to follow. However, the model-free autopilot design methods mainly rely on some sort of experiences; hence, less tractable and more efforts are required. In summary, this study uses two guidance methods along with four ship track-keeping autopilot design methods and performances of different guidance law and controller design methods are analyzed.
|
|---|
