利用超音波訊號分析與模糊控制技術作室內自動車學習與導航之研究

Abstract

本研究提出了一套智慧型自動車系統，利用超音波訊號分析，讓自動車能具有跟隨使用者行走，並同時學習路徑、與人互動、做路徑分析，以及進行智慧型導航等功能。我們利用一台具有超音波感測器的自動車作為實驗平台，使其航行於室內走廊環境中。針對跟隨人物，我們提出了一個利用超音波和模糊控制來讓自動車動態跟隨使用者的方法。另外，為了能讓使用者與自動車互動以下達控制指令，我們提出了一個分析使用者的行為以取得控制指令的方法，先以超音波偵測使用者的行走軌跡和特定姿勢，再利用投票的方式來決定使用者對自動車所下達的控制指令。在學習完路徑後，我們提出了一個自動分析路徑的方法，能經由學習過的路徑分析出導航時所需的轉彎參數等資訊。接著，我們提出了一個利用超音波和模糊控制進行導航的方法，在導航時，利用牆壁資訊作為環境特徵，並使用所設計的模糊控制器結合路徑資訊和超音波訊號，來調整自動車的方向。我們還提出了一個能在導航時利用轉彎參數和超音波以判斷是否抵達轉彎點的方法。最後我們以成功的學習和導航實驗證明了本系統的完整性與可行性。

An intelligent autonomous vehicle system for learning and vehicle guidance in indoor environments is proposed. The system has several capabilities: learning paths by person following, interaction with humans, path analysis, and navigation as a guide. An autonomous vehicle equipped with ultrasonic sensors is used as a test bed. First, a method for person following in an environment is proposed, which has two fuzzy controllers for following a person dynamically. The fuzzy controllers adjust the vehicle’s direction and speed. Furthermore, to interact with humans, a technique for analyzing human behaviors is proposed, which analyzes a person’s walking trajectory by ultrasonic signals and uses a voting technique to obtain human commands implied by the walking postures. After learning, a method for path analysis is proposed, which analyzes the learned path data to estimate certain turning parameters for navigation. In addition, a method for vehicle navigation guidance by the use of the ultrasonic signal sequence and a fuzzy controller is proposed. The main environment feature used for navigation is wall. The fuzzy controller combines some information obtained from path analysis with ultrasonic signals to adjust the vehicle’s navigation direction. A technique for detecting learned turning points is also proposed, which decides whether the vehicle arrives at a turning point or not, using learned data and ultrasonic signals as well. Good experimental results show the flexibility and feasibility of the proposed system for the applications of person following, path analysis, and people guiding in indoor environments.