Kinect於英語互動學習系統之應用
摘要
本論文設計了兩個應用Kinect於英語互動學習的系統，第一個系統是擴增實境 (Augmented Reality) 系統，輔助老師的英語單字教學。本系統經由 Kinect 來抓取人體停止、前進、後退、左轉以及右轉的姿勢，透過無線射頻傳送指令到自走車，讓自走車行走於虛擬的英語拼單字遊戲板上，藉此學習英文單字，並且能夠讓孩童藉由遊戲來互相切磋比較，進而互相討論與學習，讓孩童能夠寓教於樂。本系統還藉由直覺化的操作，讓自走車控制能夠更符合人性。 第二個系統是機器人互動學習系統，希望藉由本系統來輔助老師上課時使用全肢體反應教學法(Total Physical Response)來教學。本系統由 Kinect 抓取人體的骨架資訊，透過動態時間校正 (Dynamic Time Warping, DTW) 或是隱藏式馬可夫模型 (Hidden Markov Model, HMM) 來判斷做的動作是否正確。本系統提供老師方便上課教學的編輯介面，並且可以讓學生透過老師的幫助，自行使用本系統讓學生能夠自行探索動作和英語的意義，藉由肢體動作以及口語表達出正確的英文語意，並且輔以機器人來互動學習，讓學習是由真實的口語以及動作來輔助學生學習英文的意義並非單純的填鴨式學習。

關鍵字：數位學習、擴增實境、TPR教學、動態時間校正、隱藏式馬可夫模型

An Application of Kinect in An English Interactive Learning System
Abstract
This thesis designed two applications on Kinect interactive English learning system; the first system is Augmented Reality system, aim to assist teachers in English vocabulary teaching. This system catch skeleton through Kinect for stop, forward, backward, turn left and right, through RF transmission commend to the cars. Cars can drive in virtual English word games to help student learn English words, and allows children to learn from each other by game, and then discuss with each other and learning so that children can be edutainment. This system also represents an intuition car control, can be more consistent with humanity. The second system is Robot interactive learning system, to help teachers in class using the Total Physical Response. This system catch skeleton through Kinect, through Dynamic Time Warping or Hidden Markov Model to determine the action is correct or not. This system provides the teacher convenient teaching editing interface, and allows the students explore their own movements and English meaning himself. We supply the robot to interactive learning, so that learning in real-spoken language and action to assist students to learn English meaning is not simply spoon-fed learning.

Keywords : fallDigital Learning、Augmented Reality、Total Physical Response、Dynamic Time Warping、Hidden Markov Model