利用 Impedance Control 執行人類適應性工作之技巧轉移

Abstract

本文提出一個能複製人類適應性工作之技巧的演算法則.首先, 我們假設 人類在執行適應性工作時, 其所用的控制策略只有兩種, 即切線移動控制 策略和法線施力控制策略. 其次, 我們認為人類的控制策略的本質和 Impedance Control 相似. 最後,我們設計一個結合主軸觀念和時間分割 觀念的法則來複製人類作過的適應性工作. 此演算法則首先建立工作座標 系統, 然後它將針對認意的適應性工作, 找出該適應性工作在切線移動方 向和法線施力方向的阻抗參數, 此阻抗參數將送給機器臂的 Impedance Controller, 用以控制機器臂, 進而複製該適應性工作. 模擬結果顯示我 們的演算法則是可行的. 對三度空間的複製工作而言, 我們的方法只須處 理 8 個參數即可, 而傳統的 Impedance Control 卻必須處理 30 個參 數. This thesis issues a concept to replicate human compliance control strategies. First, we assume that two control strategies, tangential moving control strategy and normal tool holding control strategy, are enough to emulate how human being manipulates compliant task. Second, we consider that human control strategy can be implemented by impedance control. Third, we develop an algorithm that adopts the concepts of principal axis and time division to replicate human compliant task. This algorithm will first construct the task coordinate system. It then identifies impedance parameters for an arbitrary compliant task along the tangential moving direction and the normal tool holding direction for use in the robot impedance controller. The algorithm only needs to process a small number of parameters and can be extended to deal with high dimensional system. The simulation results show that ours suggestions about human compliant task are feasible. The algorithm needs only to process eight parameters to replicate human compliant tasks in 3-D space while the formal impedance control algorithm needs to process 30 parameters for the same case.