兩指抓取位置分析

Two-Fingered Grasping Position Analysis

黃振威(Jen-Wei Huang)

指導教授 ： 劉昭華

淡江大學/工學院/機械與機電工程學系碩士班/碩士(2012年)

https://doi.org/10.6846/TKU.2012.01249

摘要

本論文針對兩指抓取剛體的情況，解出兩指與剛體點摩擦接觸下之最佳抓取位置。研究方法是利用逆向動力學中，兩指點抓持剛體為不確定系統，本文指定一正向接觸力 為未定力，並將其他接觸力表示為其未定力之函數，經由庫倫磨擦定律分別解出兩接觸點不產生滑動之未定力 區間，再找出此兩區間之交集，即為兩接觸點之可抓取區域。最後比較剛體表面所有接觸點組合之可抓取區域，當某接觸點組合之可抓取區域下限 值為最小時，此接觸點組合即為最佳夾取位置。分析結果顯示球體、長方體、及圓柱體具有單一軸向線性加速度時，其最佳夾取位置皆落在垂直軸向加速度之平面對蹠點。針對正方體具有兩方向線性加速度的情況，當兩加速度數值接近時，最佳夾取位置會落於垂直兩加速度之平面交線。最後，當長方體受到單一方向角加速度及同方向線性加速度時，發現最佳夾取位置的兩節點連線會垂直此加速度方向。

關鍵字

二指抓取 ； 逆向動力學之不確定性 ； 最佳夾取位置

並列摘要

In this thesis the optimum grasping positions for two-fingered grasping of
a rigid body is determined. It is known that the inverse dynamics problem has
one degree of indeterminacy. In this study a normal force N1 is chosen as the
indeterminate force, and all other forces may be expressed as linear functions
of this force. For each finger Coulomb’s law of friction is then utilized to
determine the region of N1 in which no relative slip occurs. Intersection of
the two regions for N1 , one for each finger, is the graspable N1 region.
Comparing graspable regions for all possible combinations of contact points,
the region with the smallest positive lower limit may be chosen, then the
corresponding contact pairs are the optimum grasping positions. Results show
that for spherical, rectangular parallelepiped, and cylindrical objects, the
optimum grasping positions are antipodal points. When a cubic has a linear
acceleration in two directions, the optimum grasping positions are along the
intersection of the two planes that are normal to the two linear accelerations.
Finally, when a rectangular parallelepiped has both a linear acceleration and
an angular acceleration in the same direction, the optimum grasping nodes
form a line that is normal to the direction of the acceleration.

並列關鍵字

two-fingered grasping ； Indeterminacy of inverse dynamics ； Optimum grasping positions

參考文獻

[12] 陳俊儒，機器人雙指抓取接觸分析，淡江大學機械與機電工程學

[11 陳偉恩，兩指抓取位置分析，淡江大學機械與機電工程學系碩士

Homography-based Grasp Tracking for Planar Objects, Proceedings

- IEEE International Conference on Robotics and Automation, v

Visual Servoing for Gripper-to-object Positioning, 2004 IEEE/RSJ