- 學位論文
人臉辨識與機器人導航於居家服務之設計與實作
The Design and Implementation of Face Recognition and Robot Navigation on Home Services
摘要
本論文提出一套動態人臉辨識的架構與機器人導航之設計。在人臉辨識的架構上主要分為偵測與辨識兩大階段，在偵測的部分利用Y軸投影的方法進行初步的人臉特徵分析，進而取得人臉候選區域，並且利用主成份分析(PCA) 對取得之人臉候選區域進行最後的檢驗。在確定人臉之後，利用Fisherfaces 來降低資料的維度與取得較具有代表性的特徵，最後利用灰關聯分析(GRA) 來做為人臉分類的依據。在導航設計的部份，為了貼近人類生活習慣，我們使用人員操作帶領的方式與聲納來自動建立地圖，日後依此地圖，透過進化演算法來進行路徑規劃。在機器人朝向給定目標行進之同時，利用模糊控制器與整體行為調整器來進行避障的動作，以確實達到導航之功能。
並列摘要
The thesis focuses on dynamic face recognition and mobile robot navigation. The process of face recognition can be divided into detection and recognition two stages. In detection stage, the principal component analysis (PCA) algorithm is used to verify selected candidate face areas obtained from calculating the corresponding y-axis projection of image data. Once face areas are identified, the Fisherfaces method is adopted to reduce the dimension of face images and extract representative features for the proposed Grey Relational Analysis (GRA) based classification. As to robot navigation, a human-leading strategy with sonar sensors is used to automatically construct the map of moving region in which a robot will be operating. Subsequently, an evolution algorithm and a fuzzy controller with the Global Action Updater are designed to solve path planning and obstacle avoidance problems, respectively to reach the goal of mobile robot navigation.