基於使用者意圖之行動輔助機器人輔助策略設計

Abstract

近年來，人口老化的問題日益嚴重，照顧銀髮族的生活起居成為一個重要的課題。隨著醫療科技的進步，各式各樣的行動輔具機器人之發展越趨成熟，本實驗室也開發行動輔具i-go以應對銀髮族生活不便的困擾。在本論文中，我們透過建構一意圖偵測模型來擷取使用者的操作意圖，並且因應使用者的狀況，提出兩種輔助控制演算法，其一是路徑導引演算法，主要針對操作能力較差的使用者，預設固定路徑點和方向，可將使用者導引至事先規劃的路徑，其二是行動輔助演算法，適用於操控能力較好的使用者，讓使用者自由地操縱行動輔具到其希冀的方向，行動輔具系統提供合適的輔助力，讓使用者在操作時能更平順且不費力地推動行動輔具。此基於使用者意圖之輔助策略功能經實驗驗證，確實有其一定的功效，未來希望能夠幫助正在復健且行走和操控能力較差的銀髮族，提供一個完善且安全的智慧型行動輔具系統。

Recently, the problem of aging population becomes more and more serious. How to take good care of the elderly is now an important issue around the world. Along with the progress of the medical and robot technology, several robot walking helpers have been developed. It motivates us to also develop a robot walking helper, named i-go, in our laboratory for assisting the lives of the elderly. In the thesis, we propose extracting the human intention via a human intention detection model, along with two assistive control algorithms in response to a variety of situations. They are: (1) the path guiding algorithm and (2) walking assistance algorithm. The former is applied to the user who is with worse control ability by guiding her/him to a pre-determined path, and the latter for the user with better control ability by providing the suitable assistive force during her/his walking on the self-determined path. The proposed assistive strategy design based on human intention has been verified via the experiment. In the near future, we expect the i-go can assist the elderly in real environments.