僅依視覺導引之小型船舶進港及靠泊控制

Visual Only Guidance and Control of a Small Boat for the Approaching and Berthing Maneuvers

本文旨在探討如何完全基於影像處理技術，提供船位及航向等導航資訊之小型船舶進港及靠泊自行控制系統。首先藉由建置於岸邊之前後疊標形成進港疊標線，經由置於船艏之CCD攝影機擷取疊標影像，並計算其質心求出船舶與疊標影像之航向偏差，即可藉由航向自航器控制小船沿預定之疊標線前進。當船舶進入待泊區範圍，則由架設於岸上之影像處理單元擷取置於船體前後之筒狀目標物質心資訊，藉以判別小船之姿態。而船體與泊席之相對距離則可由船體圓筒目標物與岸上疊標影像資訊估算之。最後將船體姿態與離岸距離迴授至靠泊控制器，即可使小船以平行於岸壁之姿態完成靠泊任務。自航控制器之設計分別採用內模式控制法及模糊控制法完成之。進港階段使用之影像處理單元乃架構於MATLAB/Simulink/xPC環境，而靠泊階段之影像處理單元則採用LABVIEW工作環境。經於海洋大學小艇碼頭進行系列實船測試顯示，本文所提完全依賴視覺資訊導航之小型船舶自動進港及靠泊控制系統確實具有可行性。 This work is on the development of a vision only guidance and control system for a small boat during the approaching and berthing maneuvers. Specifically, for the approaching phase, an onboard image processing unit is used in acquiring the image of two leading marks placed on shore to estimate the heading deviation, which serves as a feedback error signal to the course control autopilot to steer the boat towards the leading line defined by the leading marks. For the berthing phase, a shore-based image processing unit captures the images of a set of targets placed on the boat and another target placed on the berth to estimate the relative heading of the boat and the distance from the boat to the berth, which are fed into the berthing control autopilot so as to steer the boat sidewards to the berth. The approaching phase image processing unit adopts the MATLAB Simulink/xPC framework, and the berthing phase image processing unit adopts a LABVIEW setting. Both the Internal Model Control (IMU) and the Fuzzy control methods are employed in the design of the approaching control autopilot and the berthing control autopilot. A series of successful approaching and berthing maneuvering control experiments conducted in the waterways of the National Taiwan Ocean University small boat harbour indicates the feasibility of the proposed system.