English  |  正體中文  |  简体中文  |  Items with full text/Total items : 28607/40644
Visitors : 5302649      Online Users : 382
RC Version 4.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Adv. Search
LoginUploadHelpAboutAdminister

Please use this identifier to cite or link to this item: http://ntour.ntou.edu.tw:8080/ir/handle/987654321/13755

Title: 運用隱藏式馬可夫模型於力量回授之插配任務
Application of Hidden Markov Model to Force-Feedback Robotic Peg-in-Hole Tasks
Authors: Wen-Sung Chu
朱汶崧
Contributors: NTOU:Department of Mechanical and Mechatronic Engineering
海洋大學:機械與機電工程學系
Keywords: 力量回授;隱藏式馬可夫模型;插配任務;辨識率;線上辨識
force-feedback;HMM;peg-in-hole;identification accuracy;on-line identification
Date: 2008
Issue Date: 2011-06-30T07:27:02Z
Abstract: 本論文的目的乃是利用隱藏式馬可夫模型(HMM)來辨識接觸的姿態,進而輔助機械手臂進行插配任務。文中首先對三種不同插銷進行力量扭矩訊號的蒐集與k-means分群,再分別設計出適用於本研究之模型,其中狀態轉移機率矩陣A包含了9個狀態數,而觀測符號機率矩陣B包含了15個觀測符號數。 本實驗在不藉由機器視覺的輔助之下,僅使用力量扭矩感測器回授之訊號,進行孔洞位置搜尋實驗、離線辨識、即時線上辨識並加入修正決策、狀態數與觀測符號數之影響等實驗。運用隱藏式馬可夫模型進行接觸型態的辨識中,在離線過程,有九成左右的辨識率,而即時線上辨識則有七成左右的表現,其中線上辨識加入了吾人設計之修正決策,將插銷與子構裝之間的接觸型態修正至有利組裝的姿態,大幅減少組裝過程所需的時間。實驗結果顯示,隱藏式馬可夫模型在組裝過程的辨識率上具有相當的強健性,可幫助我們辨識各種接觸型態,成功完成不同裕度的力量回授插配任務。 關鍵詞:力量回授,隱藏式馬可夫模型,插配任務,辨識率 線上辨識
The objective of this thesis is using the Hidden Markov Model (HMM) to identify the contact states, and then assisting the robot to execute the Peg-in-Hole assembly tasks. We first collect the force/torque information during the assembly and make the data clustered using the k-means algorithm, and then design the model to match this experiment for three different pegs. The model includes ‘State-transition probability distribution,’ contains nine states, and ‘Observation symbol probability distribution,’ contains fifteen symbols. We use the information from force/torque sensor to perform four experiments including: searching the hole location, off-line identification, on-line identification with correction strategies and the effect of changing the number of states and symbols without robotic vision information. In the experiment to identify the contact state using the HMM, the recognition accuracy is about 90% for off-line process and 70% for on-line process, respectively. In addition, the on-line process including the correction strategies can improve the contact state between the peg and sub-assembly and decrease the time for the assembly task drastically. The results of the experiments show that, the HMM has a robust identification capability which can help the robot to identify the contact states and complete the force-feedback peg-in-hole assembly task with different tolerance successfully. Keywords : Force-feedback, HMM, Peg-in-hole, Identification accuracy,On-line identification.
URI: http://ethesys.lib.ntou.edu.tw/cdrfb3/record/#G0M94720048
http://ntour.ntou.edu.tw/ir/handle/987654321/13755
Appears in Collections:[機械與機電工程學系] 博碩士論文

Files in This Item:

File Description SizeFormat
index.html0KbHTML163View/Open


All items in NTOUR are protected by copyright, with all rights reserved.

 


著作權政策宣告: 本網站之內容為國立臺灣海洋大學所收錄之機構典藏,無償提供學術研究與公眾教育等公益性使用,請合理使用本網站之內容,以尊重著作權人之權益。
網站維護: 海大圖資處 圖書系統組
DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback