本研究利用類神經網路建立小型巡防艇在海上航行的運動類型，模型中包含滾轉、俯仰、偏航三個旋轉角度的船舶運動姿態。藉由量測實際的巡防艇航行於海上受到海浪之影響所產生的各種船體運動情況的數據資料，並利用時間序列以及類神經網路進行建模與測試，提供航海模擬機重現船舶在海上運動的基礎模型。經由實際海上實驗所獲得的資料與本文提出的類神經網路運動模型進行比較後發現，本文所提出的類神經網路模型具有相當高的精確度，可以正確預測船舶在波浪中運動的姿態變化狀態。綜合本研究所得到之成果可以提供新進航海人員熟悉、瞭解並及早適應船舶在海上航行的運動模式與感受。藉以提高航海人員對於暈船的適應能力。
In this study, the neural model of small patrol boat is investigated. The model contains three rotate motions, namely, roll, pitch and yaw, respectively. The dynamic data of those patrol vessels are measured and arranged via time series under real situation in sea trail. The reconstructed neural model can be used to represent the approximate dynamic by using ship handling simulation. The original sea trial data and the trained neural model seem identical in trend. Based on this research results, the trained neural model of patrol vessel used to be a training tool for new seafarers.
