Abstract
摘要:為突破傳統錨錠或船載觀測對沿岸海洋環境的取樣限制,本計畫以三年時程開發一自主聲層析海洋測繪系統;藉由鏡射式沿岸水聲層析儀,整合水面無人載具,依觀測環境之時空變異性,自主地調整群隊位置以收集聲線聯網的水聲層析資料,獲得海洋內部溫度與流速場的綜觀影像。本計劃承先前在移動載具水聲層析學、機器人自動控制、海洋物理數值模式等領域的研究成果與自行開發儀器、無人船的經驗,建構此一兼具理論與技術創新的海洋觀測系統。此系統之執行根據調查海域的時/空變化尺度進行。首先自相關資料庫收集環境資訊以海洋模式ROMS(Regional Oceanic Modeling System)建立初步海洋模式模擬,在研究船上預先評估探測環境之高不確定性區域,再依水聲層析資料的積分固有性質與取樣測度,進行自主式無人船路徑規劃。初步規劃之載具速度與點位將透過無線電傳輸至無人船,以進行現場載具對間的聲信號收集。所收集資料即時傳輸送回研究船之電腦系統。藉由資料同化至海洋模式進行預報,以更新不確定性區域,協助下階段自主群隊的路徑規劃改善海洋模式預報,以獲得配合沿岸環境時空尺度的測繪結果。為完成此系統的開發與執行,各階段里程碑如下:1)在硬體開發上,將逐年建置鏡射式沿岸水聲層析儀至自主水面載具的儀器設備以完成自主聲層析觀測系統,並搭載船載流速儀與高精度定位系統,以完整重建海洋流場與精確溫度場;2)在資料處理上,第一、二年研發水下訊號處理技術以提升水聲層析資料的品質,並建立高效率的海洋參數化方法以達即時資料處理的目的;3)在海洋數值模式上,第一年建立平行運算環境以執行海洋模式ROMS,第二年發展結合聲層析資料至海洋模式之系集卡爾曼同化技術,第三年利用ROMS透過系集卡爾曼濾波方法進行資料同化;4)在載具自動化上,第一年建立群隊聯網對環境取樣測度以提供編隊控制與移動軌跡規劃的評估,第二、三年結合海洋模式協助自主群隊的路徑規劃。本計畫之整體效益為:1)整合水聲層析及水面載具技術,突破基礎海洋科研之觀測關鍵技術;2)以海洋環境感知驅動自主式觀測網路之控制理論與水聲層析重建完整流場之逆推理論的學理創新;3)提供近岸小中尺度海洋動力觀測資料對沿岸海洋學、海洋環境監測、海流能評估等有具體貢獻。本研究以發展突破近岸海洋動力現象觀測技術為起始目標,也為後續延伸至大尺度海洋動力之研究鋪路,進而瞭解或解決氣候變遷的重大議題。
Abstract: To overcome the inadequate sampling of the current and temperature fields in coastal seas measured by conventional observations e.g. mooring and shipboard methods, this three-year integrated proposal aims at developing an autonomous tomographic-based ocean mapping system. Synoptic-scale variations of ocean temperature and currents are obtained in real-time using the data collected by a fleet of autonomous surface vehicles (ASVs), equipped with mirror-type coastal acoustic tomography (MCAT) system, that adaptively navigate along the optimum trajectories. The survey mission using this ocean mapping system depends on the temporal/spatial scale of the observation site. First, to identify the uncertain region, initial ocean model simulations are started up at the R/V using prior environmental data collected from various oceanic/atmospheric data resources. Then, the path-planning of ASV fleet is designed based on the sampling metric. After receiving the designed waypoints via the RF, the ASVs conduct the reciprocal acoustic transmissions at the designed waypoints. The collected acoustic data are transferred to the server at the R/V for the data assimilation. The predicted uncertain region provides the path-planning of the ASV fleet at the next stage. Repeating these measurement and prediction steps, the coordinated measurement provides consistent sampling in a dynamic ocean environment. To deliver the system, the milestones are: 1) In instrument development, five MCAT-ASV integrated units are built over three years to construct an autonomous tomographic network. In addition, unique currents and accurate temperature obtained with incorporation of the real-time kinematic positioning system and Doppler velocity log is completed by the end of the second year; 2) In data processing, the signal processing techniques for high-quality tomographic data are developed by the end of the first year. Real-time tomographic inversion via an efficient representation of the ocean fields is achieved by the end of the second year; 3) In ocean modeling, the ocean modeling using the ROMS under a parallel computing environment is completed by the end of the first year. Assimilating the tomographic data to the ROMS via the Ensemble Kalman Filter scheme is developed by the end of the second year. Real-time data assimilation is achieved by the end of the third year; 4) In vehicle automation, the evaluation of ASVs path-planning via an environmental sampling metric is developed by the end of the first year. Feasibility experiment of the ASVs path-planning via the model-predicted uncertainty is conducted by the end of the second year. Demonstration of ASVs path-planning via the model-predicted uncertainty is delivered by the end of the third year. The end-point of this project includes 1) the autonomous tomographic-based ocean mapping system significantly advances the oceanographic measurement techniques, 2) the feedback control scheme driven by the ocean model prediction advances the control theory for coordinating an ASV fleet, and the incorporation of the point-measured currents near the boundary into the tomographic data advances the inverse theory for complete reconstruction of ocean currents, and 3) its outcomes offer efficient and precise measurement in understanding small- and middle-scale physical processes in coastal seas and may be applied to, e.g., environmental monitoring and ocean energy assessment.
Keyword(s)
沿岸聲層析
資料同化
區域海洋模式
參數估測
編隊控制
環境監測
基於模 式的路徑規劃
海洋自主性
移動感測聯網
Coastal acoustic tomography
data assimilation
regional ocean circulation model
parameter estimation
formation control
environmental monitoring
model-based path planning
marine autonomy
mobile sensing network