請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86185
標題: | 基於球形連桿之手術機器人的誤差與敏感度分析 Error and Sensitivity Analysis of a Spherical Linkage Based Surgical Robot |
作者: | Yu-Jie Lin 林于傑 |
指導教授: | 蘇偉儁(Wei-Jiun Su) 蘇偉儁(Wei-Jiun Su | weijiunsu@ntu.edu.tw | ), |
關鍵字: | 手術機器人,間隙誤差,敏感度分析,球形連桿, surgical robot,clearance error,sensitivity analysis,spherical linkage, |
出版年 : | 2022 |
學位: | 碩士 |
摘要: | 由於手術機器人的定位精度對病患安全至關重要,因此設計過程中,需要控制各個誤差來源的範圍。在眾多的誤差來源中,接頭間隙對定位精度造成的影響更甚於組裝誤差以及加工誤差,因為接頭間隙造成的誤差受到端效器受力影響,難以利用校準改善,因此本文將著重在接頭間隙對定位精度的影響。間隙由於受到接頭幾何限制,因此可以推算轉軸坐標系與軸承坐標系兩者之間位移以及旋轉的最大值,進一步結合順向運動學以及最佳化,求得端效器在不同方向的最大誤差。但是執行一次最佳化僅能獲取手術機器人在一個姿勢下的最大誤差,無法確保在手術範圍內,手術機器人的設計公差可以達到精度要求。此外,無法達成精度要求時,各個旋轉接頭的幾何限制對於定位精度的影響是未知的,因此需要多次調整公差來改善定位精度。針對上述問題,本文使用最佳化取得不同姿勢下的最大誤差後,將各個幾何限制以及機械手臂姿勢作為輸入,最大誤差作為輸出,利用機器學習擬合兩者的關係函數,並進一步透過機器學習模型分析各個接頭幾何限制以及姿勢對定位精度的影響力,利用敏感度分析結果作為之後調整的依據。本文以球型連桿之手術機器人作為範例,根據分析結果並聯機構的誤差表現並不完全優於串聯機構,與[1]中的分析結果相同,並聯機構並不全然能限制端接器誤差。此外,就工作空間中最大位移誤差而言,肩關節的間隙影響力大於肘關節的間隙。 Positioning accuracy of surgical robots is critical to patient safety during surgery. Hence, it is necessary to control different error sources in design phase. Among various error sources, clearance shows stronger influence on positioning accuracy than assembly error and manufacturing error. The reason is that positioning error caused by clearance is affected by resultant force on end-effector, and it is difficult to resolve this problem with calibration. Therefore, this thesis focuses on how clearance influences positioning error of a surgical robot. Clearance error is constrained by joint geometry. Therefore, the maximum translation and rotation between frames attached to shaft and bearing can be obtained. In combination with forward kinematics and optimization, the maximum positioning error in different directions can be derived. However, conducting optimization once can only provide the maximum error at a certain pose. It doesn’t guarantee that the surgical robot meets positioning accuracy in workspace. Besides, the relation between different joint geometry constraints and positioning accuracy is unknown, and thus, engineers have to make several adjustments to improve positioning accuracy. Due to aforementioned reasons, a method based on machine learning is proposed to find the relation between various joint geometry constraints and positioning accuracy. It takes the value of joint geometry constraints and maximum error as input and output, respectively. Furthermore, the machine learning model is used to analyze the influence of geometry constraints on positioning error, that is, the sensitivity analysis of surgical robot to geometry constraints is presented and it can be viewed as a guideline on joint tolerance adjustment. Spherical linkage based surgical robot is taken as an example in the thesis. On the basis of the analysis result, the positioning error of parallel mechanism in the workspace is not completely lower than serial mechanism, which is same as the conclusion in [1]. This phenomenon indicates that parallel surgical robot can’t constrain end-effector error in some poses. In addition, clearance of shoulder joints affects maximum translation error more than one of elbow joints. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86185 |
DOI: | 10.6342/NTU202203132 |
全文授權: | 同意授權(全球公開) |
電子全文公開日期: | 2022-09-12 |
顯示於系所單位: | 機械工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-0409202218360200.pdf | 4.27 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。