Please use this identifier to cite or link to this item:
標題: 老年人看護行走輔助器之研製
Design and Implementation of a Nursing-Care Walking Assistant for the Elderly
作者: 黃裕勝
Huang, Yu-Sheng
關鍵字: nursing-care walking assistant;看護型行走輔助器;path planning;self-localization;safe navigation;dynamic programming;EKF;fuzzy hybrid navigation;路徑規劃;自我定位;安全導航;動態規劃;增廣型卡爾曼濾波器;模糊混合導航
出版社: 電機工程學系
本論文的研究目的在發展老年人看護型行走輔助器之系統設計和控制的方法學與技術。用機電整合設計方法來建造一個實用的行動輔助器輔助老年人在平坦的室內環境行走。本文探討三個主要主題,包含路徑規劃,自我定位和安全導航。在路徑規劃方面,動態規劃方法被使用來實現一種基於最短距離為標準的全域路徑規劃。本文提出的自我定位演算法是一種基於精簡化環境模型並經由增廣型卡爾曼濾波器(extended Kalman filter)融合2D雷射掃描器和里程計進行估測。為輔助老年人從一個地點行走到另外一個地點,模糊混合導航方法被用以合併尋標與防碰撞行為來控制輔助器。混合導航結合反應導航的性質與2D雷射掃描器環境辨識技術,以及即時的半環狀超音波感測資料。模擬結果以及實驗數據驗証本文所提之設計方法的有效性。

This thesis develops methodologies and techniques for system design and control of a nursing-care walking assistant. A mechatronic design approach is employed to construct a pragmatic walking mobile robotic system which assists the elderly to walk in a flat indoor environment. Three main topics, including path planning, self-localization and safe navigation, are solved in the thesis. A global path planning algorithm based on the criterion of the shortest distance is presented via the dynamic programming method. A self-localization algorithm based on the EKF approach and the minimalistic environmental model is proposed by measurements obtained from a 2-D laser scanner and an odometer. To help the elderly people walk from one position to another position, a fuzzy hybrid navigation approach is employed to steer the assistant by merging its two primitive behaviors, goal-seeking and obstacle avoidance. The hybrid navigation approach uses the nature of the reactive navigation integrated together with the identification of the environment model utilizing the 2-D laser scanner, and real-time sensing information from semi-ring ultrasonic ranging finders. Numerous simulations and experimental results are conducted to verify the effectiveness of the proposed methods.
Appears in Collections:電機工程學系所

Show full item record
TAIR Related Article

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.