Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/78154
Title: | 雙機械手臂自我校正與協同操作 Self-calibration and manipulation coordination in a dual-arm robot |
Authors: | Wu-Te Yang 楊武德 |
Advisor: | 林沛群(Pei-Chun Lin) |
Keyword: | 單手臂校正,雙手臂自我校正,順應控制,卡曼濾波器,雙手操作, Calibration of single arm,self-calibration of dual-arm,compliance control,Kalman filter,dual-arm manipulation, |
Publication Year : | 2016 |
Degree: | 碩士 |
Abstract: | 由於工業自動化的發展,機器人應用的需求也日趨提升。而工業機器人擁有良好的重複精度,但是準度(絕對精度)則較差,加上面對未知與日漸複雜的工作環境,機器人性能與控制策略的提升有其必要性,本論文著重於發展單機械手臂校正與雙機械手臂自我校正的方法,以及雙手臂系統的協同操作。
在單手臂校正部分,設計校正桌與校正器等機構,協助機械手臂進行校正,僅考慮幾何誤差。校正桌的功能在於輔助機械手臂取得末端點位置,而非仰賴高單價的量測系統,加上各軸馬達編碼器記錄的角度資訊,再利用最佳化方法,疊代出新的運動學參數,以校正機械手臂的準度。雙機械手臂自我校正的部分,則延續單手校正的概念,以球接頭串聯兩手臂系統,成一封閉鏈,而末端位置固定且已知,利用封閉鏈多餘自由度的特性,選定校正軌跡,配合順應控制,使系統自動執行校正軌跡。將記錄的角度資訊與已知的位置,同樣以最佳化方法,疊代出兩手臂的新參數,提升雙手臂系統的準度(絕對精度)。 在雙手操作的發展上,延續周敬凱學長的研究。在雙手協同架構中,採用主從控制策略,第一部分,改良被動端的力控制策略,提升性能,第二部分,主動端軌跡為使用者輸入,再以卡曼濾波器,混合位置與力資訊,以估測被動端的最佳軌跡,使雙機械手臂能協同操作,夾取多樣化的物體。第三部分,探討雙手操作的影響因素,考量點接觸、線接觸與面接觸,速度、尺寸與重量等因素,以實驗設計法輔助進行與分析。最後,若被夾取物較重,則選擇適當的夾取姿態,以結構輔助撐起物體,而非僅靠摩擦力,以節省手臂出力。 As the development of industrial automation, application of industrial robots is increasingly required. Although industrial robots have excellent repeatability, their positioning accuracy is relatively poor. Also, tasks in production line may become more and more complex, it is essential to improve performance of robot arms. This research pays attention to calibration of single robot arm and self-calibration of dual-arm system as well as manipulation coordination strategy of dual-arm. To calibrate single robot arm, calibration table and mechanism were designed and built. Calibration table can be used to help obtain positions of end-effector instead of relying on expensive measurement system. At the same time, angles of each joint are recorded by encoders attached on every joint. The positions and angles are input to optimization program to iterate new kinematic parameters so that positioning accuracy of robot arm can be improved. The concept of calibration of single arm is extended to calibrate dual-arm. A ball joint is demanded to connect two robot arms and become a closed-chain(or a big arm). The position of end-effector is fixed and known by CAD file. Because of the redundant DOF of the closed-chain, a specific path is selected which dual-arm can follow automatically. During the process, a set of angles is recorded. Similarly, optimization program is used to iterate new kinematic parameters for two arms and their positioning accuracy has been improved. The target of developing manipulation coordination strategy of dual-arm is to manipulate diverse objects stably including rigid and non-rigid bodies. The controller has master-slave structure. Users input desired trajectory to master arm. Then, that trajectory and force interaction between the arm and the object are fused in a Kalman filter to estimate better trajectory for slave arm to manipulate an object stably. Further, to explore manipulability of dual-arm system, a series of experiments were designed which different contact types, motion speed, object size, and weight are considered. Finally, if the object is large and heavy, system can select an appropriate pose to help support its gravity. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/78154 |
DOI: | 10.6342/NTU201601960 |
Fulltext Rights: | 有償授權 |
metadata.dc.date.embargo-lift: | 2026-12-31 |
Appears in Collections: | 機械工程學系 |
Files in This Item:
File | Size | Format | |
---|---|---|---|
ntu-105-R03522818-1.pdf Restricted Access | 4.9 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.