利用尾隨載具的側線壓力陣列於被動式前導機器魚定位之研究

Abstract

本論文利用水下載具側身的壓力計陣列以量測機器魚尾鰭振動所造成的流體壓力變化，來估測前行機器魚的位置。根據勢流理論建立出前導者（機器魚）之尾鰭振動壓力場受尾隨者（水下載具）影響下的壓力場數學模式。首先藉由非線性觀測矩陣計算前導者和尾隨者在不同相對位置下的觀測度，並根據所建立之壓力場模式估測前導者的所在位置，並討論其不確定性。定位方法與其誤差分析之數值模擬結果顯示前導者縱向與橫向的定位誤差大小隨著與尾隨者間的相對位置而改變，且當前導者在尾隨者前方扇形區域內有較小的定位誤差，此結果與觀測度所預測之結果相符；水槽試驗亦獲得類似的結果。自然界的魚群編隊型態亦可見尾隨者在後方的群游樣式。

This thesis investigates the estimator design for following a vibrating tail of a robot fish by pressure measurement behind the fish. The pressure field model of the leader (a robotic fish) and the follower (an underwater vehicle) is formulated based on the potential flow theory. Empirical observability gramian is utilized to find the relative position that the leader and follower have better observability. An iterative method based on inversion of a linearized measurement equation is used to estimate the position of the leader. The uncertainty of the estimation is plotted as an uncertainty map to show best tracking positions for the follower. The experimental results show that the leader will have a better estimation in front of the follower. The positional uncertainties are relative-location dependent between the leader and the follower. We conclude that tracking of a robot fish through hydrodynamic pressure variations is possible. The study suggests some better positions for the follower where the estimation of the leader position uncertainties is smaller. The suggested formation pattern generally can be found in a natural fish schooling.