仿生型自主式水下載具目標追蹤 控制方法之研究

Abstract

本文研究的目的是，利用仿生型水下載具達到定位與追蹤移動中的目標。仿生型水下載具具有省能 、高效率、高操控性、低噪音的特性，所以利用其特性進行前述的兩項動作。首先先建立本身載具的運動模式及目標物的運動模式及估測模式。在估測方面我們利用卡爾曼濾波器作目標的估測和作載具與目標之間誤差的修正，控制器方面我們是利用胸鰭與尾鰭的協調控制方法，使載具操控性能提高，在水中運動更靈活，更容易追蹤到目標物。實作方面定位及自我軌跡位置是利用都卜勒聲納作即時的定位，計算出本身的位置及方向角。而目標是利用單點式聲納偵測目標的相對距離進而進行追蹤。

This study utilizes a biomimetic autonomous underwater vehicle to perform moving-target tracking. A tracker with the structure of a fish-like design has advantages such as high energy efficiency, high maneuverability, and being quiet moving in water compared to conventional autonomous underwater vehicles. Kinematics models of the tracker vehicle, and the target vehicle are established. Measurement models of sensors on-board the tracker vehicle are provided. A Kalman Filter is applied to process measurements of the target and the tracker vehicle. The coordination of a pairs of pectoral fin and tail fin of the tracker is adopted to control the precise and agile motion of the tracker vehicle. A Doppler sonar is used to navigate the tracker and to calculate the position and directional angles of the target. A point sonar is used to detect and examine the relative distances of the target and the tracker in order to make a successful tracking. A series of successful tank experimental data are presented to validate the proposed tracking and control algorithms.