以影像為基礎之汽車導引系統設計

Abstract

本論文的研究主旨為利用模糊理論所發展的模糊演算法來控制追蹤目標的行進路線。論文中使用數位攝影機、影像擷取卡、兩個旋轉平台以及追蹤目標（即改裝遙控車）來建構移動物體導引控制之系統，並以Borland C Builder開發相關軟體，搭配所設計的控制電路，作為整個資料傳輸的樞紐，負責將控制資訊傳送給平台的控制電路以及追蹤目標。本研究所建立之模糊控制器演算法是以經驗法則建構出輸入與輸出關係的模糊規則，並以三角函數為歸屬函數，分別將當時及下一次系統處理時間之追蹤目標姿態與參考路徑的夾角，量化成模糊變數，然後對該模糊變數進行模糊運算與模糊推論，最後將模糊化的輸出操作量以高度法作解模糊的動作，而得到實際的操作量，用以操控追蹤目標。經由實驗的結果顯示，我們所提出的導引演算法確實可使被控制追蹤目標沿著所指定的參考路徑完成行進的動作。

In this thesis, we develop an image sensor based platform for mobile vehicle guidance. This is achieved by the construction of a fuzzy rule for vehicle guidance. The designed platform is consisted of two digital cameras, image grabber card, two revolving mechanical platforms and a target vehicle. A set of control circuit and the code Borland C Builder are used to develop the required hardware and software for target monitoring. The microprocessor Intel 8051 is employed to act as a core for controlling the motion of the target vehicle. A heuristic based fuzzy rule is then constructed for target guidance. The experimental data from an example system demonstrates the effectiveness of the proposed design.