基於距離資訊之行動輔助機器人動態避障

Abstract

近年來，人口老化問題日益嚴重，照顧銀髮族的生活起居成為一個重要的課題，¬而在老人常見的問題中，以行動不便的問題最常發生。隨著機器人科技的進步，各式各樣的行動輔助機器人之發展越趨成熟，本實驗室也開發行動輔助機器人i-Go，以因應銀髮族生活不便的困擾。為了保障老年人在行走的安全，在本論文中，我們建構一個障礙物偵測系統，分析環境中障礙物與機器人的距離資訊，以準確地辨識出障礙物的大小、位置及移動速度等資訊。接著，我們結合Velocity Obstacles和Artificial Potential Field 兩種演算法，提出動態避障導引演算法來規畫參考路徑，根據障礙物和目標位置資訊，導引使用者閃避環境中的動態障礙物，並安全抵達指定的位置。此基於距離資訊之動態避障導引系統功能經實驗驗證，確實有其一定的功效。未來希望能夠將i-Go送入真實的家庭中，幫助正在復健且視力和行走能力較差的銀髮族，提供一個完善且安全的智慧型行動輔助機器人。

The problem of aging population becomes more serious nowadays. With their age growing, the elderly are with the degrading health conditions that may not allow them to walk steadily. Along with the progress of the robot technology, robot walking helpers have been developed. It thus motivates us to develop a robot walking helper, named i-Go for assisting the daily lives of the elderly. To protect the elderly during walking, in this thesis, we first develop an obstacle recognition system which can accurately identify the size, location and speed of the obstacles the distance information between the obstacles and robot. We then propose a dynamic obstacle avoidance and guidance algorithm, which combines velocity obstacles and artificial potential field algorithms, to provide a reference path avoids the obstacles during guidance. Experiments are performed to verify this scheme. In the near future, we expect the i-Go can assist the elderly in their daily lives.