TW397748B - System and method for compensating for compliance of a hexapod positioning device - Google Patents

Abstract

A system and method for compensating for compliance distortions in a positioning device, such as a hexapod machining center, is disclosed. The position device has a movable platform mounted on a plurality of actuators at a plurality of pivot locations. The method comprises the steps of: calculating compliance information of the positioning device: compensating actuator movement commands according to the collected compliance information; and moving the positioning device actuators according to the compensated movement commands.

Description

Translated fromChinese

五、發明說明（1) 發明背景 1. 發明範嘴 本發明大致有關於機器，如機器工具，機器人，連桿 組，協調測量機器，及定位器。本發明尤其有關於將固定 或可移動精密機器結構位置中的誤差去除，該誤差導因於 結構的彈性撓曲。機器結構的彈性撓曲是機器致動器及負 荷施力在結構上，及移動元件加速的結果。尤其是，這些 機器包含六腳機器人中心，其中工具可以在3個線性方向 及3個旋轉方向中移動。 前案說明 1996/07/23授予Kirkham的美國專利5,538,373號，及 1 9 9 6 / 0 9/ 1 7授予Sheldon的美國專利5, 5 5 6, 24 2號，都讓渡 給Giddings & Lewis公司並以提及的方式併入本文。這些 專利揭露一種六腳機器人工具，包含一基底平台藉由許多 腳或致動器而接到工具平台。基底平台配置成支樓一工作 件，並將工具平台配置支撐並驅動一工具。藉由操控致動 器的長度，工具可以沿著相對於工作件的各線性及旋轉軸 而定位。六軸定位提供操作者精密工具的功能，其優於習 知機器工具的功能。 任何機器工具的正確度是依校正及設定時決定的機器大 小的正確度而定。以六腳機器人工具為例，1 2個樞軸點的 每一個的正確位置，其連接致動器到基底及工具平台，必 須正確的判定以精密的界定各致動器的長度。各致動器的 正確長度及樞軸的正確位置可界定工具相對於工作件的正V. Description of the invention (1) Background of the invention 1. The invention is generally related to machines, such as machine tools, robots, linkage sets, coordinated measuring machines, and positioners. The invention is particularly concerned with removing errors in the position of a fixed or movable precision machine structure, which errors are due to the elastic deflection of the structure. The elastic deflection of the machine structure is the result of machine actuators and loads exerting force on the structure and the acceleration of the moving elements. In particular, these machines include a six-legged robot center in which the tool can be moved in three linear directions and three rotation directions. The previous case states that U.S. Patent No. 5,538,373 issued to Kirkham on July 23, 1996, and U.S. Patent No. 5,5 5 6,24 2 issued to Sheldon from 1 9/6/0 9/17, were all transferred to Giddings & Lewis The company is incorporated herein by reference. These patents disclose a six-legged robotic tool that includes a base platform connected to the tool platform by a number of feet or actuators. The base platform is configured as a work piece of the branch building, and the tool platform configuration supports and drives a tool. By manipulating the length of the actuator, the tool can be positioned along various linear and rotational axes relative to the work piece. Six-axis positioning provides the operator with the capabilities of precision tools that are superior to those of conventional machine tools. The accuracy of any machine tool depends on the accuracy of the machine size determined during calibration and setting. Taking a six-legged robot tool as an example, the correct position of each of the 12 pivot points, which connects the actuator to the base and the tool platform, must be correctly judged to precisely define the length of each actuator. The correct length of each actuator and the correct position of the pivot can define the tool's relative position with respect to the work piece.

第5頁 五、發明說明（2) 確位置，以及機器的總正確度。 一種判定機器精密度的標準方法是測量其各組件。例如 在習用壓榨機中，與機器移動結合的導執必須精密的測 量，加工或刮削，監控，並維持在已知公差極限内。為了 減少公差的增加（即各組件的誤差相加而得到的較大總誤 差），各元件的公差值必須減到極小。在上述例子中，導 軌線性中的誤差直接影響機器的總正確度。為了增加機器 的正確度，導軌必須更正確的建構及維持。這會導致極高 徑的製造成本及/或結構複雜性，如各組件與其他要求者 相比必須製造及維持極高的正確標準，俾不發生公差的增 < 加。 此外，判定各組件誤差如何使得總機器誤差增加的過程 對於六腳機器人中心而言是極困難的。在以下著作：J . A. Soons, "ERROR ANALYSIS OF A HEXAPOD MACHINE TOOL," Lambda Map ’97， Third international Conference and Exhibition on Laser Metrology and Machine Performance，1997/07/15-17中，作者分析國家標準及技 術中心的六腳壓榨機所產生的可能測量誤差，及判定這些 誤差的困難。 在六腳機器人工具中，當工具平台移動時，各致動器的 寬度必須也支援變動。平台移動時的加速也將變動力施加 在機器的不同元件上。因此，這些力量在任何已知時間都 影響機器中的彈性撓曲，因此機器的幾何一直會稍微改 變。這些彈性撓曲更減少六腳機器人中心的正確性。Page 5 5. Description of the invention (2) The exact position and the overall accuracy of the machine. A standard method for determining the precision of a machine is to measure its components. For example, in conventional presses, the guidance combined with machine movement must be precisely measured, machined or scraped, monitored, and maintained within known tolerance limits. In order to reduce the increase in tolerance (ie, the larger total error resulting from the addition of the errors of the components), the tolerance value of each component must be minimized. In the above example, errors in the linearity of the guide directly affect the overall accuracy of the machine. In order to increase the accuracy of the machine, the guide rails must be constructed and maintained more correctly. This can lead to extremely high manufacturing costs and / or structural complexity. For example, each component must be manufactured and maintained to a very high level of accuracy compared to other requesters, without increasing tolerances. In addition, the process of determining how each component error increases the total machine error is extremely difficult for a six-legged robot center. In the following works: J. A. Soons, " ERROR ANALYSIS OF A HEXAPOD MACHINE TOOL, " Lambda Map '97, Third international Conference and Exhibition on Laser Metrology and Machine Performance, 1997/07 / 15-17, author analysis Possible measurement errors produced by the six-leg press of the National Standards and Technology Center, and the difficulty of determining these errors. In a six-legged robot tool, when the tool platform moves, the width of each actuator must also support changes. The acceleration as the platform moves also exerts variable power on the different elements of the machine. As a result, these forces affect the elastic deflection in the machine at any given time, so the geometry of the machine always changes slightly. These elastic deflections further reduce the accuracy of the six-legged robot's center.

第6頁 五、發明說明（3) 一種補償機器變動位置及加速所導致的誤差的方式是直 接測量輸出位置的誤差（如六腳機器人工具中在工具本身 的誤差）。這些誤差儲存在機器控制系統中，並用以調整 機器位置。 藉由測量機器輸出位置以補償誤差的方法稱為位置補 償，其需要測量每一合併位置的誤差。需要有限的測量次 數，尤其是當其他軸的位置強烈影響位置中的誤差時，如 在六腳機器人工具中。 因此存在需要來提供一種改良方法以補償因定位裝置的 位置變化及加速而導致的誤差，或者補償其他具有多重自 由度的機器尤其是六腳機器人中心的誤差。 發明目的及總結 因此，本發明的主要目的是提供一種補償定位裝置中順 應性扭曲之改良系統及方法。 本發明的另一目的是提供一種方法及裝置用以將機器製 成為一種彈性系統，其可扭曲以回應施力，並計算系統的 扭曲且用以調整機器工具的命令位置。 本發明的又一目的是提供一種方法及裝置，其中測量機 器工具的扭曲·並藉由使用扭曲模型而與施力相關連，並 接著藉由將施力插入模型而計算機器操作期間其他位置發 生的扭曲。 為了完成這些及其他目的，本發明提供一種在定位裝置 用以補償撓曲之方法，定位裝置具有一可移動平台裝在眾 多樞軸位置之幕多致動器上，該方法包含以下步驟：收集Page 6 V. Description of the invention (3) One way to compensate for the error caused by the variable position and acceleration of the machine is to directly measure the error of the output position (such as the error of the tool itself in a six-legged robot tool). These errors are stored in the machine control system and used to adjust the machine position. The method of compensating the error by measuring the output position of the machine is called position compensation, which requires measuring the error of each combined position. A limited number of measurements is required, especially when the position of other axes strongly influences the error in position, such as in a six-legged robotic tool. Therefore, there is a need to provide an improved method to compensate for errors caused by the position change and acceleration of the positioning device, or to compensate for errors of other machines with multiple degrees of freedom, especially the center of a six-legged robot. OBJECTS AND SUMMARY OF THE INVENTION Therefore, the main object of the present invention is to provide an improved system and method for compensating for compliance distortion in a positioning device. Another object of the present invention is to provide a method and device for making a machine into an elastic system, which can be twisted in response to a force, and calculate the distortion of the system and adjust the command position of the machine tool. It is a further object of the present invention to provide a method and apparatus in which the distortion of a machine tool is measured and associated with the application of force by using a distortion model, and then occurring at other locations during computer operation by inserting the application of force into the model Distortion. In order to accomplish these and other objectives, the present invention provides a method for compensating deflection in a positioning device having a movable platform mounted on a curtain multi-actuator at a plurality of pivot positions. The method includes the following steps:

苐7頁 五、發明說明（4) 定位裝置之順應性資訊；根據收集之順應性資訊而補償移 動指令；及根據補償移動指令而移動定位裝置致動器。 本發明也提供一種補償方法，更包含以下步驟：發展定 位裝置之有限元件模型；以及藉由測量因施力及加速而導 致有限元件模型之撓曲以收集靈敏度資料。 本發明也提供一種方法，其中有限元件模型包含眾多節 點，而收集靈敏度資料之步驟更包含：（1)固定有限元件 模型之輸出；（2)在'—^氏方向施加一單位有限元件模型 中眾多節點之一；（3)在各卡氏方向於有限元件模型中眾 多節點之至少一者記錄產生之撓曲；（4)在各卡氏方向重 覆步驟（2 )至（3 ) ; ( 5 )在眾多節點之每一者重覆步驟（4 ); (6)在一卡氏方向施加一單位加速至有限元件模型；（7 )在 各卡氏方向於眾多節點之至少一者記錄產生之撓曲；以及 (8)在加速之各卡氏方向重覆步驟（6)至（7)。 本發明也提供一種方法，其中補償移動指令步驟更包 含：將樞轴位置製模；計算致動器力；計算平台加速；在 致動器施力並加速至收集之順應性資訊以判定枢軸位置中 之撓曲；根據撓曲而調整樞軸位置；以及根據樞轴位置之 調整而補償移動指令至定位裝置。 本發明也提供一種補償順應性系統，包含：一定位裝置 及一微電腦配置成控制定位裝置，其中微電腦配置成收集 關於定位裝置之順應性資訊，並根據順應性資訊而調整定 位裝置之控制。 本發明也提供一種補償順應性系統，其中定位裝置更包苐 Page 7 V. Explanation of the invention (4) Compliant information of the positioning device; compensation of the movement instruction based on the collected compliance information; and movement of the positioning device actuator according to the compensation movement instruction. The present invention also provides a compensation method, which further includes the following steps: developing a finite element model of the positioning device; and collecting sensitivity data by measuring the deflection of the finite element model due to force and acceleration. The invention also provides a method in which the finite element model includes a plurality of nodes, and the step of collecting sensitivity data further includes: (1) fixing the output of the finite element model; (2) applying a unit of the finite element model in the direction of '-^' One of the many nodes; (3) Recording the deflection generated by at least one of the many nodes in the finite element model in each Kelvin direction; (4) Repeating steps (2) to (3) in each Kelvin direction; ( 5) Repeat steps (4) at each of the numerous nodes; (6) Apply a unit acceleration to the finite element model in one Kelvin direction; (7) record at least one of the many nodes in each Kelvin direction Deflection; and (8) repeating steps (6) to (7) in each direction of the Karst which is accelerated. The invention also provides a method, wherein the step of compensating the movement instruction further includes: modeling the pivot position; calculating the actuator force; calculating the platform acceleration; applying force to the actuator and accelerating to the collected compliance information to determine the pivot position Adjust the pivot position according to the deflection; and compensate the movement instruction to the positioning device according to the adjustment of the pivot position. The invention also provides a compensation compliance system including a positioning device and a microcomputer configured to control the positioning device, wherein the microcomputer is configured to collect compliance information about the positioning device and adjust the control of the positioning device based on the compliance information. The invention also provides a compensation compliance system in which the positioning device is more inclusive

第8頁 五、發明說明（5) 含眾多致動器，連接在眾多枢軸位置之上平台及下平台之 間；而其中微電腦更包含：一機器控制配置成控制定位裝 置；機器框軸模型配置成將定位裝置之幕多樞轴位置製 模；一致動器力計算器配置成使用定位裝置之眾多特性以 判定眾多致動器在定位裝置上之施力；一計算器配置成提 供平台加速；以及一機器靈敏度模型配置成將關於定位裝 置之收集順應性資訊製模，以判定因施力與加速而導致眾 多致動器之撓曲。 本發明也提供一種補償六腳機器人中心撓曲之方法，在 眾多樞軸位置，該中心具有一上平台藉由眾多致動器及眾 多平衡而接到一下平台，該方法包含：（a)發展六腳機器 人中心之上平台及下平台之獨立有限元件模型；（b )固定 上平台有限元件模型之轉軸匣，並固定下平台有限元件模 型之工作台；（c )選擇上平台上眾多樞軸位置之一；（d )在 一卡氏方向施加一單位力至上平台上眾多樞轴位置之一， 並且在各卡氏方向記錄上平台上眾多枢軸位置之位移； (e )在各卡氏方向重覆步驟（d ) ; ( f)在上平台上眾多樞軸 位置之每一者重覆步驟（c)至（e) ;(g)在下平台上選擇眾 多樞轴位置之一；（h)在·—^氏方向施加一單位力至下平 台上眾多樞軸位置之一，並且在各卡氏方向記錄下平台上 眾多樞軸位置之位移；（1 )在各卡氏方向重覆步驟（h ); (j)在下平台上眾多樞軸位置之每一者重覆步驟（g)至 (1 ) ; ( k)在平衡上選擇眾多樞軸位置之一；（1 )在一卡氏 方向施加一單位力至平衡上眾多樞軸位置之一，並且在各Page 8 V. Description of the invention (5) Contains a large number of actuators connected between the upper and lower platforms of many pivot positions; and the microcomputer includes: a machine control configured to control the positioning device; a machine frame axis model configuration The multi-pivot position of the positioning device is modeled; the actuator force calculator is configured to use the characteristics of the positioning device to determine the force exerted by the many actuators on the positioning device; a calculator is configured to provide platform acceleration; And a machine sensitivity model is configured to model the collected compliance information about the positioning device to determine the deflection of many actuators due to force and acceleration. The present invention also provides a method for compensating the center deflection of a six-legged robot. At a plurality of pivot positions, the center has an upper platform connected to a lower platform by a plurality of actuators and a plurality of balances. The method includes: (a) development Independent finite element models of the upper and lower platforms of the six-foot robot center; (b) fixed the pivot box of the upper platform finite element model, and fixed the table of the lower platform finite element model; (c) select the many pivots on the upper platform One of the positions; (d) applying a unit of force to one of the many pivot positions on the upper platform in a direction of Karst, and recording the displacements of the many positions of the pivot on the platform in each direction of Karst; (e) in the directions of each Kelvin Repeat steps (d); (f) Repeat steps (c) to (e) at each of a plurality of pivot positions on the upper platform; (g) select one of the plurality of pivot positions on the lower platform; (h) Apply a unit of force to one of the many pivot positions on the lower platform in the direction of ^^, and record the displacements of the many pivot positions on the platform in the directions of each K.K .; h); (j) Each of the pivot positions repeats steps (g) to (1); (k) chooses one of a number of pivot positions on the balance; (1) applies a unit of force in the direction of a Kelvin to the many pivots on the balance One of the locations and

五'發明說明（6) 卡氏方向記錄上平台上眾多樞軸位置之位移；（m)在各卡 氏方向重覆步驟（1) ; (η)在平衡上眾多枢軸位置之每一者 重覆步驟（k)至（m) ;(〇)重覆步驟（c)至（f)及步驟（k)至 (η )以一施加之單位加速來取代施加之單位力；（p)在一機 器控制_將眾多樞軸位置製模；（q )計算致動器力；（r )施 加致動器力至步驟（d )，（ h )，（ e )及（〇 )中記錄之資料；（s ) 根據致動器力施加至記錄資料所導致之撓曲而調整眾多枢 軸位置；以及（t )根據樞軸位置調整而補償六腳機器人中 心之移動指令。 本發明也提供一種在定位裝置用以補償撓曲之裝置，定 < 位裝置具有一可移動平台裝在眾多枢轴位置之眾多致動器 上，該裝置包含：裝置，用以收集定位裝置之順應性資 訊；裝置，根據收集之順應性資訊而用以補償移動指令； 以及裝置，根據補償移動指令而用以移動定位裝置。 本發明也提供一種裝置，其中用以收集順應性資訊之裝 置更包含：定位裝置之有限元件模型；及裝置，藉由記錄 施力及加速所導致有限元件模型之撓曲而用以收集靈敏度 資料。 本發明也提供一種裝置，其中有限元件模型包含眾多節 點及裝置用以收集靈敏度資料，更包含：有限元件模型之 固定輸出；裝置，在一卡氏方向施加任一單位力至有限元 件模型中眾多節點之任一者；裝置，在各卡氏方向於有限 元件模型中眾多節點之至少一者用以記錄產生之撓西；裝 置，在一卡氏方向施加任一單位加速至有限元件模型；及Five 'invention description (6) Displacement of many pivot positions on the platform in the Karst direction record; (m) Repeat steps in each Karst direction (1); (η) Each of the numerous pivot positions on the balance is weighted. Steps (k) to (m) are repeated; (0) Steps (c) to (f) and steps (k) to (η) are repeated with an applied unit acceleration instead of the applied unit force; (p) at Machine control _ mold many pivot positions; (q) calculate the actuator force; (r) apply the actuator force to the data recorded in steps (d), (h), (e), and (〇); (S) Adjusting numerous pivot positions based on deflection caused by actuator force applied to recorded data; and (t) Compensating movement instructions for the six-footed robot center based on pivot position adjustment. The present invention also provides a device for compensating deflection in a positioning device. The positioning device has a movable platform mounted on a plurality of actuators at a plurality of pivot positions. The device includes: a device for collecting the positioning device Compliance information; a device for compensating a movement instruction based on the collected compliance information; and a device for moving a positioning device according to the compensating movement instruction. The present invention also provides a device, wherein the device for collecting compliance information further includes: a finite element model of a positioning device; and a device for collecting sensitivity data by recording deflection of the finite element model caused by force application and acceleration . The invention also provides a device in which the finite element model includes a plurality of nodes and devices for collecting sensitivity data, and further includes: a fixed output of the finite element model; and a device that applies any unit force to a plurality of finite element models in a Karst direction. Any one of the nodes; a device to record the generated torsion in at least one of a plurality of nodes in each Kelvin direction in the finite element model; a device to apply any unit acceleration in a Kelvin direction to the finite element model; and

第丨0頁 五、發明說明（7) 裝置，在各卡氏方向於有限元件模型中眾多節點之至少一 者用以記錄產生之撓曲。 本發明也提供一種裝置，其中用以補償之裝置更包含： 裝置，用以將框轴位置製模；裝置，用以計算致動器力及 加速；裝置，用以施加致動器力及加速至收集之順應性資 訊，以判定枢軸位置中之撓曲；裝置，根據撓曲而用以調 整樞軸位置；及裝置，根據枢轴位置之調整而用以補償移 動指令至定位裝置。 附圖之簡箪說明 附圖為說明書的一部分，可說明本發明的目前較佳實矻 例，並較有上述一般說明及以下較佳實施例的詳細說明來 解釋本發明的原理，在各圖中，相同數字表示相同元件， 其中： 圖J是根據本發明較佳實施例的六腳機器人中心校正系 統的立體圖； 圖2是圖1六腳機器人中心的侧視圖； 圖3A與3B分別是工具臂的示意圖與裝在工具臂中的可廷 伸致動器腳的剖視圖，可用以測量圖2六腳機器人中心的 致動器的長度； 圖4是根據本發明的圖1校正系統的控制系統的一般方瑰 圖； 圖5是接在圖2六腳機器人中心的度量套與球板的放大信: 視立體圖， 圖6是接在圖2六腳機器人中心轉軸匣的度量套的底視立Page 丨 0 V. Description of the invention (7) The device is used to record the generated deflection in at least one of the many nodes in the finite element model in each direction of the Kelvin. The invention also provides a device, wherein the device for compensating further comprises: a device for molding the position of the frame axis; a device for calculating the actuator force and acceleration; a device for applying the actuator force and acceleration To the compliance information collected to determine the deflection in the pivot position; the device is used to adjust the pivot position according to the deflection; and the device is used to compensate the movement instruction to the positioning device according to the adjustment of the pivot position. Brief Description of the Drawings The drawings are part of the description and can illustrate the presently preferred embodiments of the present invention. The general description and the following detailed description of the preferred embodiments are used to explain the principles of the present invention. In the figure, the same numbers represent the same components, wherein: Figure J is a perspective view of a six-legged robot center correction system according to a preferred embodiment of the present invention; Figure 2 is a side view of the six-legged robot center of Figure 1; Figures 3A and 3B are tools respectively A schematic view of the arm and a cross-sectional view of a Kotin extension actuator foot installed in a tool arm can be used to measure the length of the actuator in the center of the six-footed robot of FIG. 2; FIG. 4 is a control system of the correction system of FIG. 1 according to the present invention Fig. 5 is an enlarged letter of the measuring sleeve and the ball board connected to the center of the six-footed robot of Fig. 2: a perspective view, and Fig. 6 is a bottom view of the measuring sleeve connected to the center-rotor box of the six-footed robot of Fig. 2

第丨1頁 五、發明說明（8) 體圖； 圖7是接在圖2六腳機器人中心工作台的球板的上視立體 圖； 圖8的流程圖步驟用以執行圖1校正系統的球轉動程序； 圖9的流程圖步驟是圖8球轉動校正程序中控制伺服系統 所需者； 圖1 0的流程圖步驟用以執行圖1校正系統的模擬分析； 圖1 1的流程圖步驟用以模擬圖1 0模擬分析的上平台移動 步驟； 圖1 2是圖4部分控制系統的詳細方塊圖； 圖1 3的一般流程圖步驟用以執行圖1 2反作用前饋計算器 的反作用前饋計算； 圖1 4的詳細流程圖步驟用以執行圖1 3反作用前饋計算器 的致動器力計算； 圖1 5的信號流動圖顯示圖1 2運動計算器的操作； 圖1 6顯示的流程圖步驟用以收集圖1 5機器靈敏度模型的 順應性資訊；以及 圖1 7的信號流動圖顯示圖1 2伺服回饋系統的操作。 較佳實施例之詳細說明 參考圖1，其顯示根據本發明的六腳機器人中心校正系 統1 0 0。校正系統1 0 0包含六腳機器人中心1 1 0，其經由電 源及控制電纜排線1 2 5而接到一排設備箱體1 2 0。這排箱體 1 2 0最好包含：微電腦箱體1 3 0，伺服電源控制系統1 4 0， 及工具電源控制系統1 5 0。將參考圖4來詳細說明微電腦箱Page 丨 1 V. Description of the invention (8) Body view; FIG. 7 is a top perspective view of a ball board connected to the center table of the six-legged robot of FIG. 2; FIG. Rotation program; The flowchart steps of FIG. 9 are required for controlling the servo system in the ball rotation correction program of FIG. 8; the flowchart steps of FIG. 10 are used to perform the simulation analysis of the calibration system of FIG. 1; The steps of the upper platform movement are simulated and analyzed by Fig. 10; Fig. 12 is a detailed block diagram of the control system of Fig. 4; Fig. 13 is a general flow chart step for performing the reaction feedforward of the reaction feedforward calculator of Fig. 12 Calculations; Figure 14 is a detailed flowchart of the steps used to perform the calculation of the actuator force of the reaction feedforward calculator of Figure 13; Figure 15 shows the signal flow diagram showing the operation of the movement calculator of Figure 12; The flowchart steps are used to collect compliance information of the machine sensitivity model of Fig. 15; and the signal flow diagram of Fig. 17 shows the operation of the servo feedback system of Fig. 12. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT Referring to FIG. 1, there is shown a six-footed robot center correction system 100 according to the present invention. The calibration system 100 includes a six-legged robot center 110, which is connected to a row of equipment cabinets 120 through a power supply and control cable wiring 12.5. This row of cabinets 1 2 0 preferably includes: a microcomputer cabinet 1 3 0, a servo power control system 14 0, and a tool power control system 150. The microcomputer box will be explained in detail with reference to FIG. 4

第12頁 五、發明說明（9) 體1 3 0及伺服泰-trPage 12 V. Description of the Invention (9) Body 1 3 0 and Servo Thai-tr

作馬達以旋轉二控人制/統140 : I具電源控制系統W 市的Gi抓二人中心:的！轴而此系統是由烕州方拉 没 L e w 1 s公司製造。固定—_χ_ 11 0腳1 70的氧— 疋在/、腳機器人中心 ^ i 虱動弹黃1 60包含振動隔離系統 詳情如以下戶斤^ ^ Λ α 70〜戶、％例，其 厅·4。然而要注意的是本發明的 腳機器人中心w认认々分廿— 月的技術可以在六 甲匕以外的各種其他定位裝置中使用。 在圖Z中，顯示六腳機器人中心；I丨〇的較 機器人中心"。包含一底部或下平台2。。^ =動；； =而接到上方或工具平台2〇5。致動器機構η。最好包' 3眾^可延伸連桿或致動器215，其在眾多低樞軸位置 2153接到下平台2 0 0，以及在眾多上樞軸位置2151^接到上 平台2 0 5。例如可使用球螺絲機構來建構各致動器2丨5。雖 然較佳實施例包含球螺絲機構’但本發明並不限於此種致 動器。致動器2 1 5也可包含氣動缸體’液動缸體，或齒條 與齒輪裝置。 ^ 上平台205包含轉軸匣220 ’其配置成支撐一工具225。 下平台200包含工作台240 ’其配置成支撐—工作件230。 當致動器機構2 1 0使得上平台2 0 5沿著相對於下平台2 〇 〇的 預設路彳泛移動時’工具2 2 5即與工作件2 3 0互動。六腳機器 人中心1 1 0也可包含眾多平衡2 3 5以反制重力新發生。平衡 235最好是氮氣填充彈簧’其在枢軸位置250接在底部或下 平台200與上方或可移動平台205之間。 六腳機器人中心11 〇也包含不同種類的振動隔離元件。 例如如圖1所示’六腳機器人中心1 1 0藉由2個氣動彈簧1 6 0As a motor to rotate the two-controller system / system 140: I have a power control system. This system is manufactured by Luzhou Fang La Lew 1 s company. Fixed — _χ_ 11 0 oxygen of feet 1 70 — 疋 in /, the center of the foot robot ^ i lice moving yellow 1 60 contains details of the vibration isolation system as follows: ^ Λ α 70 ~ household,% of cases, its hall · 4 . It should be noted, however, that the foot robot center recognition technology of the present invention can be used in a variety of other positioning devices other than Rokko. In Figure Z, a six-legged robot center is shown; I 丨 〇's more robot center ". Contains a bottom or lower platform 2. . ^ = Moving;; = and connected to the top or tool platform 205. Actuator mechanism η. It is best to include 3 extendable links or actuators 215 which are connected to the lower platform 2 0 0 at a plurality of low pivot positions 2153 and to the upper platform 2 0 5 at a plurality of upper pivot positions 2151 2. Each of the actuators 2 and 5 can be constructed using a ball screw mechanism, for example. Although the preferred embodiment includes a ball screw mechanism ', the present invention is not limited to such an actuator. The actuator 2 1 5 may also include a pneumatic cylinder ' hydraulic cylinder, or a rack and pinion arrangement. ^ The upper platform 205 includes a hinge box 220 ′ configured to support a tool 225. The lower platform 200 includes a work table 240 'configured to support the work piece 230. When the actuator mechanism 2 1 0 causes the upper platform 2 5 to move along a preset road relative to the lower platform 2 0 0 ', the tool 2 2 5 interacts with the work piece 2 3 0. The six-legged robotic center 1 1 0 can also contain numerous balances 2 3 5 to counteract new gravity. The balance 235 is preferably a nitrogen-filled spring ' which is connected at the pivot position 250 between the bottom or lower platform 200 and the upper or movable platform 205. The six-legged robot center 11 〇 also contains different kinds of vibration isolation elements. For example, as shown in FIG. 1 ', a six-legged robot center 1 1 0 with two pneumatic springs 1 6 0

第丨3頁 五 _、發明說明⑽ ---- 而與地面隔離，各 6 側。或者如圖2所八〃疋位在機器基底的3個腳1 7 〇的各 腳中的每—個延伸^丨\3個彈性墊245，從下平台2 0 0的3個 :2°5及下平台作相同的功能，即隔離上平 ，元件可包含無減振樓結構的堅固連接物。振動隔 钟，彈性墊等。 &。、彈黃，不同彈簧與減振器的合 、、六腳機器人中 2美國專利5, 53心0 =更詳細說明及其他實施例可從上 。注意’在圖2的機哭人5 5 6, 2 42號得到，其已併供參 :心的方向，下平台中心110中，其類似於垂直機器 用j則下平台會設置在地於上2 0 5之下。、若無振動隔離系 在此例中，以 ▲上，因此當成機器的固定基底 上，或相對於上平a =吶是無差別的，即將它稱為基 市的不同實施；以及：台僅是可移動平台。然而 。的六腳機器人 ，焉國專利5, 538, 373號中，及 、且與地面隔離 “的各種其他模型中’下平台是八 將；：下’下平台…::動隔離系統而分離。在振動隔 方：：台稱為基底是有器的固定基底來使用，而且 相同：：下平台20。甚至不？。而且，在水平機器人中心 平台^水平面支撑它。因此文位於上平台2 0 5之下，而是在 2下平台等名詞不僅t玄記住的是本文使用的基展 在也是指許多種宜他的：指圖2機器中心的特定類型， 的位：卿機器位裝置及機器人中心配置。 ι及方向，如—纟 ，由眾多測量變換器監控機器 又、益與各致動器2 1 5結合。美國專 五、發明說明（u) 利 5，5 3 8，3 7 3 號棍* 其中各工具臂包人路使用6個工具臂，與致動器2 1 5分開， 量變換器是雷^含—長度測量變換器。一種可能的長度測 沾是，可:、干涉器如圖3 A及3B所示。稍後要詳細說明 位在上與下平」 換器配置成分開的工具臂’其接到其 内部如^下所1 ^面的本身框轴，或者可建構在各致動器 在圖3 A中，八2, Q π n，丨v你山刀别由同心外及内管3 1 0及3 1 5形成工具臂 30U，以便嵌入央 ^ Q, ς 疋東325。一雷射光源33 0的雷射光束通過 窗335而進入工遛 到干涉器壯 〜’的密封中空内部，並接著從鏡3 4 0反射 ^ ^ L 其分成2個光束成分。一光束成分在同 心官的整個長户由、， —& # &中丽進到回反射器3 50，其裝在外管310的 岔封鈿。光從營 ^ ^ 0 s向下朝著干涉器裝置345反射。接著在干 涉态裒置3 4 5啦，μ。 ν ^ . 將2個光束成分再合併，而合併的光束成 分會依其相位而 OQn , ^ 、向建設性或破壞性的互相干涉。雷射光源 , α Γ偵測到邊緣，其導因於當回反射器3 5 0相 對於干涉1¾奘5S 〇 θ °衣13 4 5移動時，光束的2個成分之間的干涉。 相位是依干涉克裝置3 4 5與回反射器3 5 0之間的距離而定， 因此咬緣疋工具臂3 0 〇長度變化的指示。再度，有關工具 臂的詳細說明可參考美國專利5 , 5 3 8，3 7 3號。 雖然圖3 Α說明與致動器2 1 5分開的典型工具臂3 0 0 ，圖3 Β 是測量變換器3 6 0的剖視圖，其包含一可延伸致動器腳 2 1 5 ’其在内部裝有一雷射干涉工具系統。在此圖中可看 出致動器2 1 5包含具中空内部的可旋轉球螺桿3 6 5並藉由驅 動機構3 7 5 (如皮帶與滑輪配置）而與馬達3 7 0連接。球螺桿Page 丨 3 Five _, description of the invention ⑽ ---- and isolated from the ground, 6 sides each. Or as shown in Fig. 2, each of the three feet 17 of the machine's base is extended from each of the three feet ^ \ \ 3 elastic pads 245, 3 from the lower platform 2 0: 2 ° 5 It has the same function as the lower platform, that is, isolates the upper plane, and the element can include a solid connection without the vibration-absorbing building structure. Vibration isolation, elastic pads, etc. &. , Elastic yellow, the combination of different springs and shock absorbers, in a six-legged robot 2 US patent 5, 53 heart 0 = a more detailed description and other embodiments can be taken from the above. Attention 'obtained at No. 5 5 6, 2 42 of the machine crying figure in Figure 2, which has been provided for reference: the direction of the heart, in the center of the lower platform 110, which is similar to the vertical machine, the lower platform will be set on the ground 2 0 5 below. 2. If no vibration isolation is used in this example, it is ▲, so as a fixed base of the machine, or a = na is indistinguishable from Shangping, that is, it is called a different implementation of the base city; and: Taiwan only Is a mobile platform. However. Of the six-legged robot in Laos Patent No. 5, 538, 373, and in various other models that are isolated from the ground, the "lower platform is eight generals ;: the lower" lower platform ... :: dynamic isolation system. Vibration barrier :: The table is called the base is a fixed base for use, and the same :: the lower platform 20. Not even? ... Moreover, it is supported on the horizontal robot center platform ^ horizontal plane. So the text is located on the upper platform 2 0 5 Below, but under 2 platforms, the terms such as platform not only remember that the base exhibition used in this article refers to many other suitable ones: refers to the specific type of machine center in Figure 2, and the position: Qing machine position device and robot The center configuration. Ι and direction, such as-纟, are monitored by a number of measuring transducers, and are combined with the actuators 2 1 5. US patent 5, description of the invention (u) 5,5 3 8, 3 7 3 Number stick * Among them, 6 tool arms are used for each tool arm, which is separated from the actuator 2 1 5. The measuring transducer is a lightning-containing-length measuring transducer. One possible length measuring contamination is: The device is shown in Figures 3 A and 3B. It will be explained in detail later. The lower arm is equipped with a separate tool arm, which is connected to its own internal frame axis such as the following surface, or it can be constructed in each actuator. In Figure 3 A, eight 2, Q π n, 丨v Your mountain knife is made of concentric outer and inner tubes 3 1 0 and 3 1 5 to form a tool arm 30U so as to be embedded in the central ^ Q, ς 疋 325. A laser light source of a laser light source 330 passes through the window 335 and enters the sealed hollow interior of the interferometer, and then reflects from the mirror 3 40. It is divided into two beam components. A beam of light is contained in the entire long house of the concentric officer, and — &# & Zhongli enters the retro-reflector 3 50, which is installed in the branch seal of the outer tube 310. Light is reflected downward from the camp ^ ^ 0 s toward the interferometer device 345. Then set 3 4 5 in the interference state, μ. ν ^. The two beam components are recombined, and the combined beam components will interfere with each other constructively or destructively according to their phases. The laser light source, α Γ detects the edge, which is caused by the interference between the two components of the beam when the retroreflector 3 50 moves relative to the interference 1¾ 奘 5S θ θ ° 13 4 5. The phase is determined by the distance between the interference device 3 4 5 and the retroreflector 3 5 50, so the bite edge 指示 tool arm 3 0 0 length changes. Again, a detailed description of the tool arm can be found in U.S. Patent Nos. 5,5,38,3,7 and 3. Although FIG. 3A illustrates a typical tool arm 3 0 0 separate from the actuator 2 15, FIG. 3 B is a cross-sectional view of the measuring transducer 3 6 0, which includes an extendable actuator foot 2 1 5 'which is inside Equipped with a laser interference tool system. In this figure, it can be seen that the actuator 2 1 5 includes a rotatable ball screw 3 6 5 with a hollow interior and is connected to the motor 3 70 by a driving mechanism 3 7 5 (such as a belt and pulley configuration). Ball screw

第15頁 五、發明說明（丨2) 365是旋轉地裝在2 上。螺帽管3 8 5裝在萬^接頭組件3 80，382的第—者的軛 388而於操作時與球桿3 65外圍’並藉由眾多旋轉圓球 接頭組件382的軛上1、杯3 65連接。螺帽385裝在第二萬向 致動器2 1 5的樞軸仇因此萬向接頭3 8 〇，3 8 2的中心界定 帽385中旋轉時，^，215&及2151^因此當球螺桿365在螺 轉方向而進一步分n接頭組件380，382即依球螺桿的旋 萬向接頭組件3 8 2的4或更接近。管狀外殼3 9 0接到第二 在圖3B的較佳實於扼並圍繞球螺桿365如圖所示。 中，干涉器裝置34 5^Λ，/射光源33 0裝在管狀外殼390 q β ζ八μ ^ 裝在官狀外殼3 9 0的遠端並與球螺桿 ^ ^ ^ f夯3 5 〇則裝在球螺絲的另一端。此合併 的干涉致動态配置沾4味 51的5手情可參考德國公開專利 DE 1 9 6 1 78 8 0 號’申社 λ Γι•心·巧 ^ 甲。月人Giddings & Lewis公司，並根據 美國專刺I請案0 8 /4 3 〇,39〇號，其也在此併供參考。 因此’雷射干涉器測量變換器3 6 0能精確測量二點之間 的^離_’即干涉器裝置345與回反射器3 50在球螺絲驃動致 動态2 1 5上的位置。此距離可以用數學方式轉換以判定框 轴2 1 3 a及2 1 5 b的位置。因此此合併的干涉器與致動器是較 佳貫施例’所以本發明的剩餘部分將使用枢軸位置來說 明’這疋參考致動器2 1 5的樞轴，其與分開工具臂的極轴 相對立。然而’本發明適動任—種測量配置，因為如以下 所述’它是實際定位裝置的校正球操作距離，以判定工具 臂的估計枢轴位置’而且與工具臂是否和致動器2丨5相交 無關。Page 15 V. Description of the invention (2) 365 is mounted on 2 in a rotating manner. The nut tube 3 8 5 is installed on the first yoke 388 of the first joint assembly 3 80, 382, and is in operation with the outer periphery of the club 3 65, and on the yoke of the many ball joint assembly 382 1. 3 65 connections. The nut 385 is mounted on the pivot of the second universal actuator 2 1 5 so when the center of the universal joint 3 8 0, 3 8 2 rotates in the cap 385, ^, 215 & 2151 ^ therefore when the ball screw 365 is further divided into n joint assemblies 380, 382 in the direction of screw rotation, that is, 4 or closer to the rotary universal joint assembly 3 8 2 of the ball screw. The tubular housing 390 is connected to the second in Figure 3B. The preferred embodiment is to choke around the ball screw 365 as shown. In the interferometer device 34 5 ^ Λ, the light source 33 0 is installed in the tubular casing 390 q β ζ eight μ ^ is installed at the distal end of the official shell 3 9 0 and is compacted with the ball screw ^ ^ ^ f 3 5 〇 Install on the other end of the ball screw. The combined interference caused by the dynamic configuration of 5 flavors can be referred to the German published patent DE 1 6 1 78 8 0 ’Shenshe λ Γι • heart · Qiao ^ A. Yueren Giddings & Lewis Company, and filed a case No. 0 8/4 3 0,39 0 according to the United States Special I, which is also here for reference. Therefore, the 'laser interferometer measuring transformer 3 600 can accurately measure the distance between two points', that is, the position of the interferometer device 345 and the retroreflector 3 50 on the ball screw 3D 2 15. This distance can be mathematically converted to determine the positions of the frame axes 2 1 3 a and 2 1 5 b. Therefore this merged interferometer and actuator is a preferred embodiment 'so the remainder of the invention will use the pivot position to illustrate' this reference to the pivot of the actuator 2 1 5 and its separation from the pole of the tool arm The shafts are opposite. However, the present invention is suitable for any kind of measurement configuration because it is described below. It is the correction ball operating distance of the actual positioning device to determine the estimated pivot position of the tool arm. 5 Intersect has nothing to do.

苐16頁 五、’务明說明（13) " 词，ΐ參考圖4 ’其顯示圖1校正系統的微電腦箱體130及 ==电源與控制系統14〇。控制系統40 0包含微電腦41〇， 含輸入與輸出器所4 12，至少一中央處理器414，隨機 记憶體4 1 6，及唯種記憶體4 1 8。微電腦4丨〇更包含一 $部記憶體4 20例如用以儲存校正系統40 0的控制軟體與有 3於六腳機器人中心11 〇校正及控制的資料。—操作介面 4 2 5接到微電腦4 1 〇，操作介面4 2 5例如能包含：顯示^ ’ 鍵盤，印表機或其他輸出裝置，及/或輸入指向裝置。 微電腦4 1 0接到處理器4 3 0，處理器4 3 0提供卡氏座標系 統指令至致動器長度的數學轉換。處理器43〇接到一 ^二 位，號處理器（DSPs)44〇，各接到各致動器21 5的眾多伺服 控制級4 4 3。各控制級4 4 3包含：伺服放大器4 4 5，馬達 =〇，編碼器4 5 5，致動器驅動器4 6 0 (如圖2致動器215的球 不、·、糸核構），及測罝變換器4 6 5 (如圖3 B的雷射干涉哭.·則暑 變換器3 6 0 )。 j里 伺服放大器44 5提供正確的電源給馬達45〇，豆包 =4 5 5+ '編碼器4 55將馬達速度的信號指示送回以提供致動 為控制系統中的封閉速度迴路。編碼器4 5 5的回饋信號接 到適當數位信號處理器4 4 0。當伺服放大器“5收到信號以 ，始新馬達45G時，馬達即移動致動器驅動器偏，並延作 ，縮回致動器215 ’因此提供電方法移動在上平台2q5。到 置變換器465(如上所述的操作），㈣動器215長度的精穿 =化达回數位信號處理器4 4 0,其接著將該資訊通過處理 窃4 3 0而傳送到微電腦4 1 〇 3页 Page 16 V. ‘Description (13) " words, ΐRefer to FIG. 4 ′, which shows the microcomputer box 130 and the power supply and control system 14 of the calibration system of FIG. 1. The control system 400 includes a microcomputer 4110, including an input and output device 412, at least one central processing unit 414, a random memory 416, and a unique memory 418. The microcomputer 4 丨 〇 further includes a memory 420, for example, used to store the control software of the calibration system 400 and the calibration and control data of the three-foot robot center 110. —The operation interface 4 2 5 is connected to the microcomputer 4 1 0. The operation interface 4 2 5 can include, for example, a display keyboard, a printer or other output device, and / or an input pointing device. The microcomputer 4 1 0 is connected to the processor 4 3 0, and the processor 4 3 0 provides a mathematical conversion of the Cartesian coordinate system instructions to the actuator length. The processor 43 is connected to one or two bits, and the number of processors (DSPs) 44 is connected to a plurality of servo control stages 4 4 3 of each actuator 21 5. Each control stage 4 4 3 includes: servo amplifier 4 4 5, motor = 0, encoder 4 5 5 and actuator driver 4 6 0 (as shown in Fig. 2, actuator 215, ball, ..., core structure), And the measurement transducer 4 6 5 (as shown in Fig. 3 B, the laser interference cries. · Zeshu converter 3 6 0). The servo amplifier 44 5 provides the correct power to the motor 45. Bean bag = 4 5 5 + 'Encoder 4 55 sends the motor speed signal back to provide actuation as a closed speed loop in the control system. The feedback signal from the encoder 4 5 5 is connected to the appropriate digital signal processor 4 4 0. When the servo amplifier "5 receives the signal to start the new motor 45G, the motor will move the actuator driver biased and extend, and retract the actuator 215 '. Therefore, the electric method is provided to move on the upper platform 2q5. 465 (Operation as described above), the length of the actuator 215 is precisely converted to the digital signal processor 4 4 0, which then transmits the information to the microcomputer 4 1 0 3 by processing the steal 4 3 0.

五、發明說明（14) 在較佳實施例中，微電腦4 1 0是由威州方拉市的 Giddings & Lewis公司製造的CNC 8 0 0 0控制器；處理器 43 0 —般是Intel PENT IUM( I NTEL公司的商標）處理器；數 位信號處理器440是麻州諾伍市的Analog Devices公司製 造的2 1 8 1 DS Ps型；伺服放大器4 4 5是維州拉福市的V. Description of the Invention (14) In a preferred embodiment, the microcomputer 4 1 0 is a CNC 8 0 0 0 controller manufactured by Giddings & Lewis Company of Fangla, Weizhou; the processor 43 0 is generally Intel PENT IUM (trademark of I NTEL) processors; digital signal processor 440 is a 2 1 8 1 DS Ps type manufactured by Analog Devices, Nowa, MA; servo amplifiers 4 4 5 are from Lafort, Victoria

Kollmorgen Industrial Drives 公司製造的BDS4 型放大 器；馬達450是Kollmorgen公司的B604型無刷DC馬達；及 致動器驅動器4 6 0與雷射測量變換器465是先前圖3B所述的 那一種。上述硬體上執行的軟體的詳情將參考圖12_i7來 說明。 在六腳機器人中心1 1 0的正常操作中，伺服系統包含 DSPs 4 40及伺服控制級4 43，其提供用以將機器定位的封 閉回饋迴路。惟，在校正轉動（如以下所述的球轉動）中， 來自校正裝置4 7 0的資訊（如以下所述的度量套）包含在較 大的回饋迴路中’該迴路包含數位信號處理器4 4 〇。因此 在板·正模式中，词服系統更包含處理器4 3 0及校正裝置 47 0。 、 雖然上述裝置及電腦包含本發明的較佳實施例，熟於此 技術者也可使用許多習知的替代品。例如若處理器4 3 0的 功此併入一個功能更強的微電腦4 1 〇 ,則可省去處理器 430。事實上，616(^叩5&1^以15的(：1^ 8000控制器實際 上包含I BM相容個人電腦以執行它的許多功能，尤其是使 用新介面。雖然圖4所示各電腦或處理器執行的工作將參 考—特殊硬體裝置來說明’但是不該將本發明的技術解#Kollmorgen Industrial Drives BDS4 type amplifier; motor 450 is Kollmorgen B604 type brushless DC motor; and actuator driver 460 and laser measurement converter 465 are the ones described previously in FIG. 3B. Details of the software running on the above hardware will be described with reference to Fig. 12_i7. In normal operation of the six-legged robot center 110, the servo system includes DSPs 4 40 and servo control stage 4 43, which provide a closed feedback loop for positioning the machine. However, in the correction rotation (such as the ball rotation described below), the information from the correction device 470 (such as the measurement set described below) is included in the larger feedback loop 'the loop contains the digital signal processor 4 4 〇. Therefore, in the board-positive mode, the word server system further includes a processor 430 and a correction device 470. Although the devices and computers described above include preferred embodiments of the present invention, those skilled in the art can also use many conventional alternatives. For example, if the function of the processor 430 is incorporated into a more powerful microcomputer 410, the processor 430 may be omitted. In fact, the 616 (^ 叩 5 & 1 ^ 15 (: 1 ^ 8000 controller actually contains an BM-compatible personal computer to perform many of its functions, especially using the new interface. Although each computer shown in Figure 4 Or the work performed by the processor will be described with reference to special hardware devices, but the technical solution of the present invention should not be explained #

第18頁 五、發明說明（]5) ' ' —-—-—-—-- 為只能使用這些裝置。 的Γ士提供接在六腳機器人中心110的度量套5 0 5與球板510 22〇側^ ^體圖。度量套5G5接到上平台205上的轉轴£ ’△而ίΓ板3lQ則接到下平台2GG的工作台240。藉由控制 套二】：2，校正系統即相對於球板510而移動度量 細解釋此校正程序。a。。的精岔致動器資訊。以下詳 參考圖6 ’其顯示校正年统軔伟杏 底禎立栌岡 ^ θ . CA糸、'•无1又侄戶、％例中的度量套50 δ的 •=f V 包含位於圖中單點附近的許多感 測态605。感測器605的配置方式可拈.a, 對於球板51◦上球的位移 式了以债測到上平台205相 圖7顯示球板510的上視立體圖，在較诖實施例中，球板 51〇包含6個精密球70 5，其排列在中心精密球附近，並位 ”基底710的角落。雖然圖6及7的合併顯示配置的 志冽為6〇3可以測量排列在六腳圖樣中的七個精密球7〇 5， 但本發明不限於這種球板及感測器配置。反之，球板5 ι 〇 的目的是要測量的任一種加工品的例子，如任何配置中的 球板，圓柱，面板，步移計，可移動工具球，雷射千涉器 光束等。 ' 由 在較佳實施例中，感測器6 0 5是習知的線性可變差分轉 換器（LVDTs)。各LVDT產生與其可移動核心的位移成正比 的電輸出信號。各感測器6 0 5具有一末端部分6 1 〇其在校正 過程中與精密球7 0 5作機械接觸。末端部分6丨〇是機械式的 與L V D T可移動核心連接，其由磁性材料製造3可移動核心Page 18 V. Description of the invention (] 5) '' — — — — — — — Only use these devices. ΓShi provides a body map of the measuring sleeve 5 05 and the ball plate 510 22〇 connected to the center of the six-legged robot 110. The measuring sleeve 5G5 is connected to the rotating shaft on the upper platform 205, and the 3IQ plate is connected to the work platform 240 of the lower platform 2GG. With control set 2]: 2, the correction system moves the measurement relative to the ball 510 to explain the correction procedure in detail. a. . Information for precise actuators. Refer to Figure 6 below for details. 'It shows the correction chronology of Wei Xing Bing Li stands 栌 ^ θ. CA 糸,' • None 1 and nephew,% = 50 in the metric set in the example. • f V Included in the figure Many sensed states 605 near a single point. The configuration of the sensor 605 can be 拈 a. For the ball plate 51, the displacement of the upper ball is measured by the debt to the upper platform 205. Figure 7 shows a top perspective view of the ball plate 510. The plate 51 includes 6 precision balls 70 5 arranged near the center precision ball and juxtaposed to the corner of the base 710. Although the combined display configuration of FIGS. 6 and 7 is 603, it can be measured and arranged in a six-legged pattern Of the seven precision balls 705, but the present invention is not limited to such a ball plate and sensor configuration. On the contrary, the purpose of the ball plate 5 ι is an example of any processed product to be measured, as in any configuration Spherical plates, cylinders, panels, pedometers, movable tool balls, laser interferometer beams, etc. 'In a preferred embodiment, the sensor 605 is a conventional linear variable difference converter ( LVDTs). Each LVDT generates an electrical output signal that is proportional to the displacement of its movable core. Each sensor 605 has an end portion 6 1 0, which makes mechanical contact with the precision ball 705 during the calibration process. The end portion 6 丨 〇 is mechanically connected to the LVDT movable core, which is made of magnetic material 3 Mobile Core

五、發明說明（16) 位於主要及二個次要線圈之中以使主線圈位於二個次要線 圈之間。因此LVDTs將實體輸入，如末端部分61 0的移動轉 成一輸出電壓，其對應從0位置開始的位移。有關LVDTs的 進一步說明可參考Schaevitz Engineering of Pennsauken， N. J.公司於1983 年 10 月出版的 Handbook of Measurement and Control, HB-84(國會圖書卡號 76-24971)。 裝在轉軸E 2 2 0上的感測器6 0 5其功能是藉由判定（a)加 工品的精密本質如面特徵與（b)上平台2 〇 5的位置移動之間 的偏移’而監控或測量各加工品。不同類的加工品可使用 不同類的感測器60 5。例如可使用LVDTs來測量上述的圓 柱，面板，及工具球，並可使用雷射干涉器來監控上平台 在某段距離中的線性移動，該距離等於光波長的積分數。 而且，加工品的類型可決定精密移動的類型。#密移動的 例子包含：沿著一條線移動-精密㈣，平行於一條線的 移動，平行於一平面的移動，績荽 /X Lk ^ ^ t , 、.凡者—條線旋轉，及繞著一 點旋轉。 現在說明流程圖，圖8 - 1 1說明‘ & “丄+ a U 。兄明如何错由電腦模擬來使用 仅正；r、.··充以找出六腳機益人中心丨丨〇的精密大小。 之’本發明中完成的此校正技術是使用一個雙級過程' (2)在機器上執行球轉動’即要求六腳機器人中心110作精 治移動以使度量套5 0 5藉由控制致動 密球7。5移動，以得到一特殊而在繞著精 及⑴在電腦上執行模擬分析，量資料； I 1更用球轉動的測量資料5. Description of the invention (16) It is located between the main and two secondary coils so that the main coil is located between the two secondary coils. Therefore, the LVDTs convert the physical input, such as the movement of the end portion 61 0, into an output voltage, which corresponds to the displacement starting from the 0 position. Further description of LVDTs can be found in the Handbook of Measurement and Control, HB-84 (Cap. 76-24971), published by Schaevitz Engineering of Pennsauken, N. J., October 1983. The sensor 6 0 5 mounted on the rotating shaft E 2 2 0 functions to determine the offset between (a) the precise nature of the processed product such as surface characteristics and (b) the positional movement of the platform 2 05 ' And monitor or measure each processed product. Different types of processed products can use different types of sensors 60 5. For example, LVDTs can be used to measure the above-mentioned cylinders, panels, and tool balls, and laser interferometers can be used to monitor the linear movement of the upper platform over a certain distance, which is equal to the integral number of the light wavelength. Moreover, the type of the processed product determines the type of precision movement. Examples of #close movements include: moving along a line-precision ㈣, moving parallel to a line, moving parallel to a plane, Ji 荽 / X Lk ^ ^ t,,... Whoever—lines rotate, and around With a little rotation. Now explain the flow chart, Figure 8-1 1 illustrates' & "丄 + a U. Xiong Ming how to use computer simulation to use only positive; r, .... to find the six-legged machine benefit center 丨 丨 〇 The precise size of the "The correction technology completed in the present invention is a two-stage process" (2) Ball rotation on the machine is performed, that is, the six-footed robot center 110 is required to make a refined movement to make the measuring sleeve 5 0 5 by Control the movement of the dense ball 7.5 to get a special and perform simulation analysis on the computer around the fine and cymbals to measure the data; I 1 more measurement data using the ball rotation

五、發明說明（17) 而重覆的判定定位裝置的精密大小在規定的公差範圍内。 在六腳機器人中心的較佳實施例中，用精密大小作為致動 器的數個椹軸位置2 1 5 a及2 1 5 b的位置。 圖8提供流程圖形式的校正過程第一部分的詳細解釋， 以說明執行一球轉動校正程序中採取的步驟。如上所述， 必須先從一系列精密機器移動中收集測量資料，稱為球轉 動以決定六腳機器人中心1 1 0的精密幾何或連桿大小。在 較佳實施例中，收集稱為致動器長度變化的參數作為測量 資料如以下所述。 從步驟8 0 0開始，根據機器製造尺寸的已知值而在步驟 8 0 5判定數個致動器2 1 5的枢軸位置的大約估計值。在較佳 實施例中，當機器是六腳機器人中心時，6個致動器2 1 5的 每一者都有2個樞軸位置2 1 5 a及2 1 5 b，即共有1 2個樞轴位 置。這些製造尺寸最好準確到實際值的± 1 / 2英吋以内， 該實際值是最後決定的。 由卡氏座標來界定這些測量樞軸位置，這是從下平台 2 0 0相關的絕對或固定座標系統中測量到的。一致動器2 1 5 的估計枢轴位置例子在下樞軸位置2 1 5 a之中是 [-17.7,-38.2,-23.6]，以及在上樞軸位置215b之中是 [-4 0 . 9，1 0 . 2 3，2 3. 6 ]。此例的單位是英吋，但是也可使同 任何其他單位。注意，因為絕對座標系統是固定在下平 台，因此下枢軸位置是否實際相對於地面而移動並不重 要。 有6個點界定固定在基底或下平台2 0 0的樞軸，而且有6V. Description of the invention (17) The precise size of the repeated positioning device is within the specified tolerance range. In the preferred embodiment of the six-legged robot center, the precise size is used as the positions of the multiple y-axis positions 2 1 5 a and 2 1 5 b of the actuator. Figure 8 provides a detailed explanation of the first part of the calibration process in the form of a flowchart to illustrate the steps taken in performing a one-ball rotation calibration procedure. As mentioned above, measurement data must be collected from a series of precision machine movements, called ball rotations, to determine the precise geometry or link size of the 6-foot robot center 110. In the preferred embodiment, a parameter called a change in actuator length is collected as measurement data as described below. Starting from step 8 0 0, an approximate estimate of the pivot position of several actuators 2 1 5 is determined at step 8 0 5 based on the known values of the machine manufacturing dimensions. In the preferred embodiment, when the machine is a six-legged robot center, each of the six actuators 2 1 5 has 2 pivot positions 2 1 5 a and 2 1 5 b, that is, a total of 12 Pivot position. These manufacturing dimensions are preferably accurate to within ± 1/2 inch of the actual value, which is the final decision. These measurement pivot positions are defined by the Cartesian coordinates, which are measured from the absolute or fixed coordinate system associated with the lower platform 200. Examples of estimated pivot positions of the actuator 2 1 5 are [-17.7, -38.2, -23.6] in the lower pivot position 2 1 5 a, and [-4 0. 9 in the upper pivot position 215b. , 1 0. 2 3, 2 3. 6]. The unit for this example is inches, but it can be the same as any other unit. Note that because the absolute coordinate system is fixed on the lower platform, it does not matter whether the lower pivot position actually moves relative to the ground. There are 6 points defining a pivot fixed to the base or lower platform 2 0 0, and there are 6

五、發明說明（18) 個點界定固定在上平台或可移動平台205的框軸。因為上 平台移動時，必須指定一數值規格給各平台位置。因為上 平台有6維的機械自由度，因此指定6個唯一座標以界定一 位置：3個移動座標及3個旋轉座標。對於上平台的每一位 置，有一個含6個座標的組[,\，丫，2，八，：6，（：](其中乂，丫及2是 線性座標，而A，B及C是旋轉座標）以界定平台位置。因為 6個上致動器枢軸位置與6個上平台抱軸位置重疊，所以有 時以下也稱為平台樞軸。 相對於具有座標[〇,〇,〇,〇,〇,〇]的參考位置來界定上平 台的各位移。在參考位置中，上平台樞軸與上述界定上平 台框轴的6個點重疊，從固定在下平台2 0 0的絕對座標系統 來測量。例如，從參考位置[0，0，0，0，0，0 ]到任一位置[1， 3，5，2，4，6 ]的位移可如下的完成： A. 界定上平台所屬並與其連接的移動座標系統，以便只 有當上平台位於參考位置時，移動座標系統才會與下平台 的固定絕對座標系統重疊； B. 從參考位置開始，繞著移動座標系統的Z軸而將上平 台旋轉6個單位； C. 從參考位置開始，繞著移動座標系統的Y軸而旋轉上 平台4個單位； D. 從參考位置開始，繞著移動座標系統的X軸而旋轉上 平台2個單位； E. 在Z方向移動上平台5個單位； F. 在Y方向移動上平台3個單位；以及5. Description of the invention (18) The points define the frame axis fixed on the upper platform or the movable platform 205. Because when the upper platform moves, a numerical specification must be assigned to each platform position. Because the upper platform has 6-dimensional mechanical degrees of freedom, 6 unique coordinates are assigned to define a position: 3 moving coordinates and 3 rotating coordinates. For each position on the platform, there is a group of 6 coordinates [, \, ya, 2, eight,: 6, (:] (where 乂, ya, and 2 are linear coordinates, and A, B, and C are Rotate the coordinates) to define the platform position. Because the 6 upper actuator pivot positions overlap with the 6 upper platform holding axis positions, they are sometimes also referred to as platform pivots below. Relative to having coordinates [〇, 〇, 〇, 〇 , 〇, 〇] define the displacement of the upper platform in the reference position. In the reference position, the upper platform pivot axis overlaps with the 6 points defining the upper platform frame axis from the absolute coordinate system fixed on the lower platform 2000 Measurement. For example, the displacement from the reference position [0, 0, 0, 0, 0, 0] to any position [1, 3, 5, 2, 4, 6] can be completed as follows: A. Defining where the upper platform belongs And a mobile coordinate system connected to it so that the mobile coordinate system overlaps with the fixed absolute coordinate system of the lower platform only when the upper platform is at the reference position; B. Starting from the reference position, the Z axis of the mobile coordinate system The upper platform rotates 6 units; C. Starting from the reference position, it moves around the coordinate system Rotate the upper platform 4 units on the Y axis; D. Rotate the upper platform 2 units around the X axis of the mobile coordinate system from the reference position; E. Move the upper platform 5 units in the Z direction; F. In the Y direction 3 units on the platform; and

第22頁 五、發明說明（19) G.在X方向移動上平台一個單位。 注意，上述旋轉及移動順序在較佳實施例中不該改變。 因此上平台的這種位移的應用會導致上平台框軸的位置 移動。回想座標[X，Y，Z，A，B，C ]界定上平台位置，樞軸位 置在參考位置[0, 0, 0, 0, 0, 0]界定，而位置座標[X，Υ, Z，A， B，C ]僅在參考位置等於[0，0，0，0，0，0 ]。注意，重要的是 每當已關閉機器時，都能再產生此參考位置。 因為致動器2 1 5接在樞軸位置2 1 5 a與2 1 5 b之間，所以透 過畢氏定理的使用即可大約估計各致動器的長度。在步騾 810，計算上平台參考位置的致動器長度的這些估計值， 並稱為參考位置的致動器長度。 從機器的名目大小可得到移動座標系統中感測器6 0 5中 心的大約估計值。這種感測器位置估計值的例子是上平台 2 0 5的移動座標系統中的[0，0，0 ]。此外，在絕對座標中也 可得到球位置的大約估計值。這種在下平台2 0 0的絕對座 標系統中的球位置例子是[5，0，0 ]。 在步驟8 2 0，微電腦410計算並輸出所需的平台指令資 料，以移動上平台2 0 5到新的位置，以便度量套感測器6 ◦ 5 與一精密球7 0 5作機械接觸。在步驟8 2 5 ’處理器4 3 0使用 致動器回饋資料以控制伺服系統，以使致動器將上平台 2 0 5移動到對應平台指令資料的新位置。如上所述，伺服 系統包含控制系統的一部分，其用以定位機器。以下將參 考圖9來進一步說明控制伺服系統所需的步驟。在步驟 8 3 0，伺服系統控制的輸出記錄在一資料檔。此輸出表示Page 22 V. Description of the invention (19) G. Move the platform one unit in the X direction. Note that the above rotation and movement sequence should not be changed in the preferred embodiment. Therefore, the application of this displacement of the upper platform will cause the position of the upper platform frame axis to move. Recall that the coordinates [X, Y, Z, A, B, C] define the upper platform position, the pivot position is defined at the reference position [0, 0, 0, 0, 0, 0], and the position coordinates [X, Υ, Z , A, B, C] is equal to [0, 0, 0, 0, 0, 0] only at the reference position. Note that it is important that this reference position is regenerated whenever the machine is turned off. Because the actuators 2 1 5 are connected between the pivot positions 2 1 5 a and 2 1 5 b, the length of each actuator can be approximated through the use of Bishop's theorem. At step 810, these estimates of the actuator length of the upper platform reference position are calculated and referred to as the actuator length of the reference position. An approximate estimate of the center of the sensor 605 in the mobile coordinate system can be obtained from the size of the machine. An example of such a sensor position estimate is [0,0,0] in the mobile coordinate system of the upper platform 205. In addition, an approximate estimate of the ball position can also be obtained in absolute coordinates. An example of this ball position in the absolute coordinate system of the lower platform 2 0 0 is [5, 0, 0]. In step 8 2 0, the microcomputer 410 calculates and outputs the required platform instruction data to move the upper platform 2 5 to a new position so that the measuring sleeve sensor 6 5 makes mechanical contact with a precision ball 7 0 5. At step 8 2 5 ', the processor 4 3 0 uses the actuator to feedback the data to control the servo system, so that the actuator moves the upper platform 2 5 to a new position corresponding to the command data of the platform. As mentioned above, the servo system contains part of the control system, which is used to position the machine. The steps required to control the servo system will be further described below with reference to FIG. 9. In step 830, the output controlled by the servo system is recorded in a data file. This output represents

第23頁 五、發明說明（20) 球轉動校正程序的輸出並稱為致動器長度變化資料。 在較佳實施例中，令上平台2 0 5繞著七個精密球7 0 5的每 一者的中心作5 0個精密旋轉式移動，即每球轉動共有3 5 0 個精密移動。惟該注意的是，移動數目是依機器類型，其 特殊應用，期望精密度，加工品類型，及移動類型而定。 雖然在較佳實施例中選定是3 5 0個精密移動，但是只要是 統計上足夠的移動數目即可以。 伺服系統控制器在步驟8 3 0輸出並記錄致動器長度變化 資料後，微電腦4 1 0即在較佳實施例中步驟8 3 5判定上平台 2 0 5的目前位置是否表示目前精密球7 0 5使用的較佳5 0個位 置的最後一個。若不是，則在步驟8 4 0，由微電腦4 1 0所需 的次一位水平台指令資料以移動上平台2 0 5到同一球7 0 5上 的新位置。新平台位置具有一角度方向其與進入同一球的 其他平台位置的角度方向不同。接著在步驟825，重覆上 述資料收集過程。 若微電腦4 1 0判定上平台2 0 5的目前位置表示目前精密球 7 0 5的最後位置，則微電腦4 1 0接著在步驟8 4 5判定目前精 密球7 0 5是否為球板5丨0上七個球的最後一個。若不是，微 電腦4 1 0即在步驟8 5 0傳送所需的次一球平台指令資料以移 動上平台2 0 5到次一精密球，而且在步驟8 2 5，對於次一球 重覆執行上述過程。若微電腦4 1 0判定目前精密球是球板 5 1 0上七個球的最後一個，則在步驟8 5 5完成球轉動校正程 序。完成時，致動器長度變化資料，其表示致動器長度在 參考位置與致動器長度在較佳3 5 0個位置的每一者之間的Page 23 5. Description of the invention (20) The output of the ball rotation correction program is called the actuator length change data. In the preferred embodiment, the upper platform 205 is caused to make 50 precision rotary movements around the center of each of the seven precision balls 705, that is, there are a total of 350 precision movements per ball rotation. It should be noted that the number of movements depends on the type of machine, its particular application, the desired precision, the type of processed product, and the type of movement. Although in the preferred embodiment, 350 precise movements are selected, it is sufficient as long as it is a statistically sufficient number of movements. After the servo system controller outputs and records the actuator length change data in step 8 30, the microcomputer 4 1 0 in the preferred embodiment determines whether the current position of the upper platform 2 5 5 represents the current precision ball 7 in step 8 3 5 0 5 is the last of the better 50 positions used. If not, in step 8 40, the next-order water platform instruction data required by the microcomputer 4 1 0 moves the upper platform 2 05 to the new position on the same ball 7 0 5. The new platform position has an angular direction that is different from the angular direction of other platform positions entering the same ball. Then in step 825, the above data collection process is repeated. If the microcomputer 4 1 0 determines that the current position of the upper platform 2 0 5 indicates the last position of the current precision ball 7 0 5, the microcomputer 4 1 0 then determines whether the current precision ball 7 0 5 is the ball plate 5 丨 0 at step 8 4 5. The last one on seven balls. If it is not, the microcomputer 4 10 transmits the required next ball platform instruction data at step 8 50 to move the upper platform 2 05 to the next precision ball, and at step 8 2 5 iteratively executes the next ball. The above process. If the microcomputer 4 1 0 determines that the current precision ball is the last of the seven balls on the ball 5 1 0, then the ball rotation correction process is completed at step 8 5. When completed, actuator length change data, which represents the difference between the actuator length between the reference position and each of the preferred 350 positions

第24頁Page 24

第25頁 五、發明說明（22) 剛器對於處理器4 3 0的電壓差。在+ 定電壓差是否在一預定限制内，^’處理器430決 限制為9 mV ’如果電壓差不在該預=乜貫施例中，此預定 在步驟935將電壓差轉成新的平合指^制内，處理器430 感剩器6 0 5的位移減到極小。例 二貝料，其設計成使 = UU.99,5.QU，“]的二二置;Λ3,5^^ 错由應用一系列習知的座標系統卜十广扣令。此轉換是 似轉換而完成。較佳實施例中要 ，測器軸作位移的 套5 〇 5的感測器6 〇 5軸不是正交的。崩似轉換，因為度量 "、_.，< 人 ° 平乂 1王 n ^ ly'j γ 套5 〇 5的感測器6 〇 5軸不是正交 在步驟940 ’藉由來自LVDTS的積分 積分器以驅動LVDTs完全成為空的並 歹定回饋，即使用 破’可調整新平台指令資料。接著， 心疋的比例增 回步驟910 ’並使用調整的平台指令資：里來種二° :;制％ :步驟91Q至94。中的每—步驟後，即〜二過 〇心令資料直到感測器6 〇 5的輸出带 、為整平 又之内。當此發生在步驟930，在步驟95〇及參考位置圣 圖8的步驟8 1 0 )中’將目前的致動器長度從致動器長度^ 減去。最後’在步驟9 5 5 ’將這些差當成致動器長庶次， 艾化來輸出，並且在步驟9 6 0完成伺服系統的控制。 當收集到致動器長度變化資料的統計上足夠數目時，I 奸是透過上述的球轉動校正程序，即可找出機器的大小= 力ϋ工品的任何未知特徵。後者的例子包含·雷射干涉突.” 夏出線的方向及線上一點的座標，圓枉中心界定的線仅J 及方向，面板界定的平面位置及方向，及工具球界定的Page 25 V. Description of the invention (22) The voltage difference between the rigid device and the processor 4 3 0. If the + constant voltage difference is within a predetermined limit, the processor 430 must be limited to 9 mV. If the voltage difference is not in the pre-set embodiment, the schedule converts the voltage difference to a new level at step 935. Within the fingers, the displacement of the processor 430 sensor 605 was reduced to a minimum. Example two shell material, which is designed so that = UU.99,5.QU, "] 's two and two sets; Λ3,5 ^^ The error is caused by the application of a series of known coordinate systems. The conversion is completed. In the preferred embodiment, the sensor axis of the sensor set, which is displaced by the sensor axis, is not orthogonal. The collapse-like conversion is because the metric ", _., ≪ Ping 1 Wang n ^ ly'j γ Set of 5 〇5 sensor 6 〇5 axis is not orthogonal At step 940 'The integrator from LVDTS is used to drive the LVDTs to be completely empty and determine the feedback, ie Use the broken 'to adjust the new platform instruction data. Then, increase the heart rate back to step 910' and use the adjusted platform instruction data: two to two degrees :; system%: steps 91Q to 94. after each step That is, ~ 2 times the heartbeat data until the output band of the sensor 605 is flat and within. When this occurs at step 930, at step 95 and the reference position of step 8 of FIG. 8) "Subtract the current actuator length from the actuator length ^. Finally, at step 9 5 5 ', treat these differences as actuator lengths, and then output them. And the control of the servo system is completed at step 9 60. When a statistically sufficient number of actuator length change data are collected, the size of the machine can be found out through the above-mentioned ball rotation correction procedure. Any unknown features of the product. Examples of the latter include laser interference. "Xia out of the line and the coordinates of a point on the line, the line defined by the center of the circle is only J and direction, the position and direction of the plane defined by the panel, and the tool ball Defined

第26頁 五、發明說明（23) 轉中心。注意，在許多機器人中心應用中，可期望用雷射 干’步器產生的線而非球板作為加工品，以執行校正轉動。 在亥例子中’匕不疋可決定的精密球中心的位置，而是上 述雷射光束方向’而由雷射干涉器在數個波長範圍内測量 光束上一點的座標。在該例子中，校正轉動可包含雷射轉 動而不是球轉動。事實上’在校正轉動期間可測量到許多 其他類型的加工品。 .如上所述’一旦收集到致動器長度變化資料，最好是從 =轉動校正程序，則藉由在電腦上執行模擬分析即可繼續 ^正^ ί。模擬分析使用致動器長度變化資料以重覆判定 ^機态人中心1 1 〇的許多樞軸位置2 1 5a及2 1 5b。參考圖 10及11的此過程詳情。 ^ 模：在4考圖？半提供含有多個步驟的流程圖以執行平台 完全D且Γ二：可以在微電腦41◦或是在與微電腦410 凡王刀開且不同的電腦系統中 輸到它的致動器長度變化資料^于。模擬分析電腦具有傳 存裝置（如軟碟，CD_R(H丨）在網=，方法疋精由任何大量儲 網路連線（如LAN或麵）來傳*網路或是透過任何習知的 微電腦4 1 0或上述的不同電匕。因此以下將電腦系統（指 佳實施例中，平台模擬器是二殊）稱為平台模擬器。在較 電腦。 δ在上述CNC8000中的個人 從圖1 0的步驟丨0 0 〇開始，4 動器長度變化資料。在步模擬器在步驟1 0 0 5接收致 置的初始估計，如七二;=10,平台模擬器作加工品位 "。這些估計值是根據下平台Page 26 V. Description of Invention (23) Transfer to the center. Note that in many robotic center applications, it may be desirable to use a wire produced by a laser stepper instead of a ball plate as a processed product to perform a corrective rotation. In this example, the position of the center of the precise ball that can be determined is not determined, but the laser beam direction described above, and the coordinates of a point on the beam are measured by the laser interferometer in several wavelength ranges. In this example, the corrective rotation may include a laser rotation instead of a ball rotation. In fact, many other types of processed products can be measured during the calibration rotation. As described above, once the actuator length change data is collected, it is best to turn from the rotation correction program, and then the simulation can be performed on the computer to continue ^ 正 ^ ί. The simulation analysis uses actuator length change data to repeatedly determine the many pivot positions 2 1 5a and 2 1 5b of the human center 1 1 0. Refer to Figures 10 and 11 for details of this process. ^ Mod: Examine the picture in 4? Semi-provided a flowchart with multiple steps to perform the platform complete D and Γ 2: It can be input to the actuator length change data in the microcomputer 41◦ or in a different computer system than the microcomputer 410 where the king knife is open ^ to. The simulation analysis computer has a storage device (such as a floppy disk, CD_R (H 丨) is online), and the method is transmitted from any mass storage network connection (such as a LAN or a surface) through the network or through any known The microcomputer 4 10 or the above-mentioned different electric daggers. Therefore, the computer system (referred to in the preferred embodiment, the platform simulator is two special) is referred to as the platform simulator. In the more computer. Δ Individuals in the above CNC8000 from Figure 1 Starting at step 0, 0, 0, 4 actuator length change data. In step the simulator receives the initial estimate of the placement at step 1 0 05, such as seventy-two; = 10, platform simulator for processing grade " these Estimates are based on the platform

五、發明說明（2c 述 200上精密歧 么 絕對座標系α十先刖位移知識。參考固定於下平台的上 個球位置估舛枯界叱球位置。在絕對Χ-Υ-Ζ座標系統中1 -7.5,0];[4 的例子是：[〇,〇,〇];卜8.5’0’〇];[一%， 7.5,0]。 ’ . 〇] ; [8. 5, 〇, 〇] ; [4. 3, 7· 5, 0 ];卜4· ’ 在步驟1 ίη ς ，的樞輛位二台模擬器分別作下平台215a及上枣J 限中機器尺寸二的：始估計。這些估計值是從製造公％ 所述。 寸的知識中得到。樞軸位置的估計例子如以上 估Ϊ二動〇 1 :台模擬器使用球位置及樞軸位置的初始 各精密移動中模擬球轉動校正程序的 詳細顯示執行此平Γ銘叙士。s此’執行350次模擬。圖11 明。平二動模擬的各步驟，並且將於以下說 410。舌要的是：主立产/、千σ 5的35〇個位置給微電腦 當我；㈣正:Λ/ ’因為在球轉動校正程序期間， 田ybU5正好居中在七個精密球位置之— 3 5 0個致動态長度變化資料點的每一者。 可 在步驟1 0 2 5，丨台模擬器將步驟1010的七個 從步驟1 020的最後3 5 0個球位置中減去以 口估计球位置 著將這些差平方加起來以產生—成本函數于。^值’並接 調整球位置及㈣位置以減少成本函&。在二=30 ’ 中，藉由調整估計球位置及樞軸位置並使用土男、％例V. Description of the invention (2c) The knowledge of the absolute coordinate system of α on the 200-axis is described by α. The first ten displacements are estimated. Refer to the last ball position fixed on the lower platform to estimate the dead ball position. In the absolute X-Υ-Z coordinate system 1 -7.5,0]; [4 Examples are: [〇, 〇, 〇]; Bu 8.5'0'〇]; [One%, 7.5, 0]. '. 〇]; [8, 5, 〇, 〇]; [4, 3, 7, 5, 0]; Bu 4 · 'In step 1, the two simulators of the pivot position are used as the lower platform 215a and the upper jujube J. The size of the middle machine is two: start Estimation. These estimates are obtained from the knowledge of the manufacturing company. The example of the pivot position estimation is as described above. Two actions: The simulator uses the ball position and the pivot position in the initial precision movements. The detailed display of the simulation ball rotation correction program shows the execution of this level. The simulation is performed 350 times. Figure 11 illustrates. The steps of the level two motion simulation will be described below. 35/50 positions of the production /, thousand σ 5 are given to the microcomputer; ㈣ 正: Λ / 'Because during the ball rotation correction procedure, Tian ybU5 is exactly centered among the seven precision ball positions — 3 50 Each of the dynamic length change data points. In step 10 25, the simulator can subtract the seven in step 1010 from the last 3 50 ball positions in step 1 020 to estimate the ball position. Add these squared differences to generate a cost function at. ^ Value 'in parallel to adjust the ball position and the ㈣ position to reduce the cost function & In two = 30', adjust the estimated ball position and pivot position and use Native

Davidon-Fletcher-PowelKD-F-P)非碲炉 非竦私式演算法來減少(Davidon-Fletcher-PowelKD-F-P)

第28頁 五、發明說明（25) 誤差。雖然在較佳實施例中使用習知的D_F_p方法，因為 可：用其速度，其他極小化技術’如最陡峭的下’降或共軛 斜又。可麥考史丹福（Stanford)大學的D. G. 著作的線性及非線性裎式遙給一奎 LuenDerger 書，尤其是參考9. 3章有 關D - 〜ρ方法的進一步資訊。 。 在步驟1 0 3 5，平台模擬器使用調整的Μ轴 以及原始的致動器長度變化資料以再度模擬平△ 位置 上所述，每-時間作一次’在球轉動校正程序動。如 的3 5 0個致動器長度變化資料點中的每一個之 收集到 模擬器計具上平台的移動。在步驟丨〇4〇，使用牛由平台 新近調整的七個球位置及步驟1〇35中35〇個球位1〇3〇中 計算步驟1 0 2 5中所述的成本函數。在步驟1〇45來重靳 成本函數值並藉由減去先前成本函數值（即在第一用目％的 泰程圖時’先前成本函數是步驟1 〇 2 5中計^•的值人執行 是步驟1035中的先前成本函數值）以計算成本烧 否則它 用目前成本函數值來除以此差以決定成本變化;接著 1 0 50，平台模擬器判定步驟1 045中計算的成本變^驟 於一預設極限。在較佳實施例中，預設極限是i 吃否低 差值大於預設極限，則平台模擬器返回步驟1〇3〇 ,〜9。若 步驟1030至1050。 ’连重覆 在步驟1 0 5 0，平台模擬器判定成本變化是否小 限，完成模擬分析及樞軸位置的最後估計值並且t預設柽 1 0 5 5輸出球位置，並且在步驟1 〇 6 〇結束模擬又_在v驟 後估計值表示機器的賞際大小以便在可接受的令 义些最Page 28 5. Description of the invention (25) Error. Although the conventional D_F_p method is used in the preferred embodiment, because it can: use its speed, other minimization techniques such as the steepest descent or conjugate ramping. However, the linear and non-linear equations of D. G.'s work at Stanford University, McCoy, are given to the Kuen LuenDerger book, especially with reference to Chapter 9.3 for further information on the D-~ ρ method. . In step 1035, the platform simulator uses the adjusted M-axis and the original actuator length change data to simulate the flat △ position again, as described above, every 'time', the ball rotation correction procedure is performed. For example, each of the 3500 actuator length change data points is collected by the movement of the platform on the simulator tool. In step 044, the cost function described in step 1025 is calculated using the seven ball positions newly adjusted by the platform and the 350 ball positions 1030 in step 1035. In step 1045, the value of the cost function is repeated and the value of the previous cost function is subtracted (that is, when the Thai chart of the first mesh% is used, the previous cost function is the value calculated in step 10) The execution is the previous cost function value in step 1035) to calculate the cost; otherwise it divides the current cost function value by this difference to determine the cost change; then 1 0 50, the platform simulator determines the cost change calculated in step 1 045 ^ It starts at a preset limit. In a preferred embodiment, if the preset limit is low if the difference is greater than the preset limit, the platform simulator returns to steps 1030 to 9. If steps 1030 to 1050. 'Even repeatedly in step 1 0 50, the platform simulator determines whether the cost change is small, completes the simulation analysis and the final estimate of the pivot position and presets 柽 1 0 5 5 to output the ball position, and in step 1 〇 6 〇 End the simulation and _ After the v step, the estimated value represents the reward size of the machine so that

。、是範圍内 看. Within range

第29頁 五、發明說明（26) 執行球轉動校正程序。 參考圖1 1以更詳細說明步驟1 0 2 0及1 0 3 5的平台移動模 擬。注意平台模擬器不知道或不在乎輸入樞軸位置而球位 置只是估計值。反之，平台模擬器假設表示一個正確的機 器幾何。平台模擬器接著使用致動器長度變化資料來找出 機器的新位置。例如模擬之前的上平台2 0 5位置是參考位 置[0，0 , 0，0，0，0 ]，而且於已發生模擬之後，若致動器長 度變化資料對應該位置，則上平台2 0 5會在一偏移位置如 [1，3，5，2，4，6 ]。因此可以在與本發明所述的校正系統完 全不同且獨立的系統中使用上平台移動的模擬。因此可以 在任何電腦系統中執行平台移動的模擬步驟。 在步驟1 1 0 0開始，平台模擬器先在步驟11 0 5接收一組枢 軸位置及估計的球位置。在步驟六腳機器人中心1 1 0 0使用 樞軸位置以計算樞軸在[0，0，0，0，0，0 ]平台位置（即參考位 置）之間的實際致動器長度，並將致動器長度變化資料加 入期望的特殊平台位置中。結果是位移的平台位置的實際 致動器長度。在步驟六腳機器人中心11 1 5，平台模擬器指 定或接收新平台位置的估計值。例如，參考位置可以是新 平台位置的可接受初始估計值。在步驟11 2 0，平台模擬器 將平台參考位置（及上平台參考框軸位置）轉成估計的新平 台位置。在步驟Π25，用畢氏定理來計算各對下平台致動 器枢軸與上平台致動器樞轴之間的距離。結果是對應估計 的新平台位置的估計致動器長度。 在步驟1 1 3 0，計算致動器長度誤差作為平方差的總和，Page 29 5. Description of the invention (26) Perform the ball rotation correction procedure. Referring to FIG. 11, the platform movement simulation of steps 1020 and 1035 will be described in more detail. Note that the platform simulator does not know or care about the input pivot position and the ball position is only an estimate. In contrast, the platform simulator is assumed to represent a correct machine geometry. The platform simulator then uses the actuator length change data to find the new position of the machine. For example, the position of the upper platform 2 0 5 before the simulation is the reference position [0, 0, 0, 0, 0, 0], and after the simulation has occurred, if the actuator length change data corresponds to the position, the upper platform 2 0 5 will be at an offset position such as [1, 3, 5, 2, 4, 6]. Therefore, the simulation of the movement of the upper platform can be used in a completely different and independent system from the correction system according to the present invention. Therefore, the simulation steps of platform movement can be performed in any computer system. Starting at step 1 1 0, the platform simulator first receives a set of pivot positions and estimated ball positions at step 1 105. In step six, the center of the six-legged robot 1 1 0 0 uses the pivot position to calculate the actual actuator length of the pivot between the [0, 0, 0, 0, 0, 0] platform position (ie, the reference position), and The actuator length change data is added to the desired special platform position. The result is the actual actuator length of the displaced platform position. At step 6 of the 6-foot robot center, the platform simulator specifies or receives an estimate of the position of the new platform. For example, the reference position may be an acceptable initial estimate of the position of the new platform. In step 11 2 0, the platform simulator converts the platform reference position (and the upper platform reference frame axis position) into the estimated new platform position. In step Π25, the distance between each pair of lower platform actuator pivots and the upper platform actuator pivots is calculated using Bishop's theorem. The result is an estimated actuator length corresponding to the estimated new platform position. In step 1 1 3 0, calculate the actuator length error as the sum of the squared differences,

第30頁 五、發明說明（27) 該差值為步驟1 1 2 5的估計致動器長度與步驟1 1 1 0的實際致 動器長度之間的差。在步驟1 1 3 5，平台模擬器調整估計的 新平台位置以減少步驟1 1 3 0中界定的致動器長度。此調整 是使用習知的牛頓拉生法在較佳實施例中執行的，雖然也 可使用任何演算法其設計成用以使多變數函數的值減到極 小。可參考John Wiley於1964年出版P. Henrici所著的鏨 值分析基礎（國會圖書卡號64-23840)，尤其是第五章，有 關牛頓拉生法的進一步資訊。 在步驟1 1 4 0，平台模擬器將步驟1 1 3 5中要求的調整量與 預設極限相比。若調整大於預設極限，則平台模擬器返回 控制到步驟1 11 5，並重覆步驟六腳機器人中心1 1 1 5至 1 1 4 0。透過上平台位置或估計新位置的重覆調整，步驟 1 1 4 0中要求的調整最後會變的越來越小。理論上，它可以 小到執行模擬的已知電腦中最大數字的數目，因為當估計 致動器長度趨近步驟六腳機器人中心110 0中找到的實際致 動器長度時，命令位置中的變化最後會趨近0。在較佳實 施例中，步驟1 1 4 0中使用的預設調整極限定義為任何平台 座標X，Y，Z，A，B，C中變化的絕對值必須小於1 . 0 E - 6 = 若調整符合步驟Π 4 0中的極限，則平台模擬器輸出命令 位置作為步驟1 1 4 5中上平台的位置。例如若命令位置在連 續通過流程圖時是[0 · 0 1，- 0 · 2，0 , 2，0，0 ]，則平台模擬器 會輸出該位置作為上平台205的位置，或者如上所述，作 為精密球了 0 5之一的位置。在步错1 1 0 5完成平台移動的模 擬。Page 30 V. Description of the invention (27) The difference is the difference between the estimated actuator length of step 1 1 2 5 and the actual actuator length of step 1 1 10. In step 1 1 3, the platform simulator adjusts the estimated new platform position to reduce the actuator length defined in step 1 1 30. This adjustment is performed in the preferred embodiment using the conventional Newtonian drawing method, although any algorithm may be used which is designed to minimize the value of the multivariate function. Reference can be made to the basis of threshold analysis by P. Henrici, published by John Wiley in 1964 (Congress Book Card No. 64-23840), especially Chapter 5, for further information on Newton's Lassen law. In step 1 1 40, the platform simulator compares the adjustment amount required in step 1 1 3 5 with a preset limit. If the adjustment is greater than the preset limit, the platform simulator returns to step 1 11 5 and repeats the step 6-foot robot center 1 1 1 5 to 1 1 4 0. Through repeated adjustments of the upper platform position or the estimated new position, the adjustment required in step 1 1 40 will eventually become smaller and smaller. Theoretically, it can be as small as the largest number of known computers on which simulations are performed, because when the estimated actuator length approaches the actual actuator length found in step six-footed robot center 110 0, the change in command position It will eventually approach 0. In the preferred embodiment, the preset adjustment limit used in step 1 1 40 is defined as the absolute value of the change in any of the platform coordinates X, Y, Z, A, B, C must be less than 1.0 E-6 = if The adjustment conforms to the limit in step Π 40, and the platform simulator outputs the command position as the position of the upper platform in step 141. For example, if the command position is [0 · 0 1,-0 · 2, 0, 2, 0, 0] when continuously passing through the flowchart, the platform simulator will output the position as the position of the upper platform 205, or as described above As a precision ball took a position of one of 0 to 5. Complete the platform movement simulation at step 1 1 0 5.

五、 發明說明 ：28) 重 要 的 是注意 本 發 明不限於 模擬分析 的此較佳 實施例 5 反 之 y 可 使用任 何 模 擬，其使 用類似的 測量致動 及估計 的 樞 轴 位 置 以找出 新 的 機器位置 0 藉 由 圖 8至1 1的上述程序，當實際機器執行一特殊加工 品 定 義 的 一系列 精 密 移動時， 藉由先收 集致動器 長度變 化 資 料 即 可 校正六 腳 機 器人中心 。接著收 集此資料 並使用 在 模 擬 電 腦 上的重 覆 過 程以找出 樞軸的實 際位置。 因此本 發 明 提 供 一 種定位 裝 置 的校正， 其不使用 可移動平 台的極 精 密 框 軸 位 置。此 外 校 正系統只 要求作出 一組的校 正移動 接 著 即 可 在任何 搖 控 電腦上控 制校正資 料。因此 ，可以 現 场 周 期 性 的再校 正 六 腳機器人 中心以碟 保其長期 正確性 〇 因 此 本 發明的 校 正 技術可以 在沒有其 他習知校 正系統 的 徑 成 本 下確保 極 徑精密的 定位裝置 〇 如 以 上 圖4所述 微電腦4 1 0 控制六腳 機器人中 心以使 上 平 台 2 0 5 > 告著一預設路徑移動 =微電腦4 1 0藉由以 新期望 位 m JL 的 形 式 傳送平 台 指 令資料來 移動上平 台。例如 若上平 台 S 前 位 在 位置[0 ，0 ，0 ,〇,〇,〇] ，則預設路徑會要求移動到 新 的 位 置 [1,2,3 ，5 ，5 ，5 ]。平台指令資料也可以7 !:[ 1，2, 3， 5, 5, 5] 氺 ，其中星號* 表示將該 位置當成 指令或命 令定位 裝 置 將 它 移 動到該 位 置 ，而與目 前位置本 身相反。 此平台 指 令 資 料 必 須透過 參 考 圖4及1 2至1 7中所述的硬體來轉成致 動 移 動 。在以 下 的 圖1 2至1 7 中，要詳 細說明如 何將平 台 指 令 資 料 轉成精 密 的 致動器移 動。 參 考 圖 12，微 電 腦 4 1 0傳送平台指令資料V. Description of the invention: 28) It is important to note that the present invention is not limited to this preferred embodiment of simulation analysis. 5 Instead, any simulation can be used, which uses similar measurements to actuate and estimate pivot positions to find new machines. Position 0 According to the above procedures of Figs. 8 to 11, when the actual machine performs a series of precise movements defined by a special machined product, the center of the six-legged robot can be corrected by first collecting data on the change in actuator length. This information is then collected and repeated on the analog computer to find the actual position of the pivot. Therefore, the present invention provides a calibration of a positioning device, which does not use the extremely precise frame axis position of the movable platform. This external calibration system only requires one set of calibration moves and then the calibration information can be controlled on any remote computer. Therefore, the six-footed robot center can be periodically re-calibrated on site to ensure its long-term accuracy. Therefore, the calibration technology of the present invention can ensure an extremely precise positioning device without the cost of other known calibration systems. The microcomputer 4 1 0 described in FIG. 4 controls the center of the six-legged robot so that the upper platform 2 0 5> moves with a preset path = the microcomputer 4 1 0 moves by transmitting the platform instruction data in the form of a new desired bit m JL On the platform. For example, if the previous position of the upper platform S is at the position [0, 0, 0, 0, 0, 0], the preset path will require moving to the new position [1,2,3,5,5,5]. The platform command data can also be 7!: [1,2,3,5,5,5] ，, where the asterisk * indicates that the position is regarded as a command or command positioning device, and it is moved to the position, which is opposite to the current position itself. The platform instruction data must be converted into actuation by referring to the hardware described in Figures 4 and 12 to 17. In Figures 12 to 17 below, how to turn the platform command data into a precise actuator movement is explained in detail. Refer to Figure 12, Microcomputer 4 1 0 transmits platform command data

第32頁 五、發明說明（29) [X，Y，Z，A，B, C ] *到執線產生器1 2 0 5，其最好在微電腦4 1 0 内部。該注意的是指令[X，Y，Z，A，B，C ]是相對於下平台。 軌線產生器1 2 0 5將此卡氏機器平台指令資料轉成命令位 置，速度及加速度（p，v，a)的一連續串並將三者都輸 出。在較佳實施例中，這是使用S曲線而完成：Page 32 V. Description of the invention (29) [X, Y, Z, A, B, C] * to the line generator 1 2 0 5 which is preferably inside the microcomputer 4 1 0. It should be noted that the instructions [X, Y, Z, A, B, C] are relative to the lower platform. The trajectory generator 1 2 0 5 converts this Karst machine platform instruction data into a command position, a continuous series of speed and acceleration (p, v, a) and outputs all three. In the preferred embodiment, this is done using an S curve:

p⑴=A.t3+B*t2+Ot+D v=dp/dt=3 · A*t:+2*B *t+Cp⑴ = A.t3 + B * t2 + Ot + D v = dp / dt = 3A * t: + 2 * B * t + C

a=dv/dt=6.A«t+2，B 而其中p(t)是時間函數的瞬時位置，v(t)是時間函數的瞬 時速度，a (t )是時間函數的瞬時加速度，而t是時間。上 述計算出的各p，v及a參數由2個向量，一條直線（X，Y， Z)及一角度（A，B ’ C)單數。係數A ’ B，C，D是由執線產 生器依某一方式計算出，即不超過抖動，加速度，及速度 的一些極限。執線產生器產生至少一個S曲線，其係連續 使用供位置[1，丫，2，儿，3，（：]*輸入產生器1205。 計算瞬時速度及加速度的另一種方法是取得位置指令串 p(t)的第一及第二數值差。 再參考圖12，執線產生器1205輸出命令相對位置，速 度，及加速度給反作用前饋計算器1 2 1 0及運動計算器 1 2 1 5。反作用前饋計算器1 2 1 0使用命令相對位置，速度， 及加速度以及施加在下平台上的力以決定各致動器2 1 5上 的致動器力F *，以產生執線產.生器1 2 0 5下令的相對加速a = dv / dt = 6.A «t + 2, B and p (t) is the instantaneous position of the time function, v (t) is the instantaneous speed of the time function, and a (t) is the instantaneous acceleration of the time function, And t is time. Each of the p, v, and a parameters calculated above is singularly composed of 2 vectors, a straight line (X, Y, Z), and an angle (A, B'C). The coefficients A ′ B, C, and D are calculated by the line generator in a certain way, that is, the limits of jitter, acceleration, and speed are not exceeded. The line generator generates at least one S-curve, which is used continuously for position [1, y, 2, er, 3, (:] * input generator 1205. Another method of calculating instantaneous speed and acceleration is to obtain a position command string The difference between the first and second values of p (t). Referring again to FIG. 12, the line generator 1205 outputs the commanded relative position, speed, and acceleration to the reaction feedforward calculator 1 2 1 0 and the motion calculator 1 2 1 5 The reaction feedforward calculator 1 2 1 0 uses the commanded relative position, speed, and acceleration and the force applied to the lower platform to determine the actuator force F * on each actuator 2 1 5 to generate a wire production. Relative Acceleration Ordered by Biomass 1 2 0 5

第33頁 五、發明說明（30) 度。執行此變換器所需的步驟將參考圖1 3來進一步說明。 反作用則饋&十异器1 2 1 0輸出致動器力F *給運動計算器 1 2 1 5及词服回饋乐統1 2 2 5。運動計算器1 2 1 5使用命令位置 及致動器力F*來決定命令致動器長度[*，其包含因六腳機 态人中心1 1 0的彈性杻曲而導致的誤差校正，該誤差是由 致動器力F *及其他慣性力產生。執行命令致動器長度L *所 需的常式信號流程圖將參考圖1 5來進—步說明。 命令致動器長度L*輸出到伺服回饋系統丨2 2 5以及輸出到 微分器1 2 2 0。微分器1 2 2 0提供致動器長度l*變化率（即 d L / d t長度L相對於時間的微分）給飼服回饋系統丨2 2 5。飼 服回饋系統1 2 2 5提供指令扭矩給各致動器2 1 5的控制級4 4 3 如圖4所示。控制級4 4 3包含伺月艮放大器4 4 5，馬達4 5 0，編 碼器4 5 5 ’致動器驅動軸4 6 0，及測量變換器4 6 5 3各控制 級4 4 3提供編碼器4 5 5及測量變換器4 6 5的回饋給伺服回饋 系統1 2 2 5。因此伺服回饋系統1 2 2 5使用命令致動器力f *， 命令致動器長度L* ’命令致動器長度的微分dL*/dt，及各 控制級443的回饋以決定指令扭矩T*。 較佳的’反作用前饋計算器1 2 1 0，運動計算器丨2 1 5，及 微分器1 2 2 0是裝在英代爾的PENTIUM CPU 430上=伺服回 饋系統1 2 2 5裝在數位信號處理器4 4 0上如圖4所示。 現在參考圖1 3 ’其顯示一般流程圖步驟用以執行反作用 前饋計算器1 2 1 0執行的反作用前饋計算。步驟1 3 0 0後，在 步驟1 3 0 5接收執線產生器1 2 0 5的命令相對位置，速度，及 加速度。在步驟1 3 1 0 ’用命令相對位置速度，加速度，及Page 33 5. Description of invention (30) degrees. The steps required to implement this converter will be further explained with reference to Figs. The reaction is to feed & ten different devices 1 2 1 0 to output the actuator force F * to the motion calculator 1 2 1 5 and to convince the music system 1 2 2 5. The motion calculator 1 2 1 5 uses the command position and the actuator force F * to determine the command actuator length [*, which contains the error correction caused by the elastic buckling of the human center 1 1 0 in the six-legged machine. The error is caused by the actuator force F * and other inertial forces. The routine signal flow chart required to execute the command actuator length L * will be further explained with reference to FIG. 15. The commanded actuator length L * is output to the servo feedback system 2 2 5 and to the differentiator 1 2 2 0. The differentiator 1 2 2 0 provides the change rate of the actuator length l * (that is, the differential of d L / d t length L with respect to time) to the feeding feedback system 丨 2 2 5. The feed feedback system 1 2 2 5 provides a commanded torque to the control stages 4 4 3 of each actuator 2 1 5 as shown in FIG. 4. The control stage 4 4 3 includes a servo amplifier 4 4 5, a motor 4 50, an encoder 4 5 5 ′ actuator drive shaft 4 6 0, and a measuring transducer 4 6 5 3 each control stage 4 4 3 provides coding. The feedback from the converter 4 5 5 and the measuring converter 4 6 5 is fed to the servo feedback system 1 2 2 5. Therefore, the servo feedback system 1 2 2 5 uses the command actuator force f *, the command actuator length L * 'the differential of the command actuator length dL * / dt, and the feedback of each control stage 443 to determine the command torque T * . The better 'reaction feedforward calculator 1 2 1 0, motion calculator 丨 2 1 5 and differentiator 1 2 2 0 are installed on Intel's PENTIUM CPU 430 = servo feedback system 1 2 2 5 installed in The digital signal processor 4 4 0 is shown in FIG. 4. Reference is now made to Fig. 1 3 'which shows a general flowchart step for performing a reaction feedforward calculation performed by a reaction feedforward calculator 1 2 1 0. After step 1 3 0 0, in step 1 3 5 receive the command relative position, speed, and acceleration of the wire execution generator 1 2 5. In step 1 3 1 0 ′, command the relative position speed, acceleration, and

第34頁 五、發明說明（31) 下平台力來計算所需的致動器力以提供期望的相對加速 度。藉由同時解出用於6個致動器力F的含6個公式的系統 即可完成此計算。例如在較佳實施例中，該公式系統是以 矩陣形式表示： CPmat *F=B (1) 其中CP mat是耦合矩陣，具有：上方3列的6行填上各致動 器2 1 5的單位向量，及下方3列的6行填上相對單位扭矩與 慣性比，其中單位扭矩是致動器單位向量相對於平台的重 ！ 力中心的矩。計算腳單位向量與重力中心至致動器樞軸的 向量交叉乘積即可完成。 在公式（1 )，F1是含6個致動器力的向量。直線系統公式 中的量（上方3列）是力。角度系統公式中的量（下方3列）是 相對角度加速度。因此公式右側上向量B的前3個元素是以 力單位表示。後3個元素是以加速度單位表示。B向量的前 3個元素是3個方向Χ,Υ及Z的每一者中的直線力，其計算成 為以下公式： M u / ( M u 4- Μ 1 ) · ( Μ 1 · a -f s u m _ f ο ο t) ( 2 ) 其中Mu是上平台的質量，M i是下平台的質量，a是步驟 1305中，上平台相對於下平台的指令加速度。Suni_foot是 步驟1 3 2 5中，振動隔離成分施加在下平台的所有直線力的Page 34 V. Description of the invention (31) Lower the platform force to calculate the required actuator force to provide the desired relative acceleration. This calculation can be done by solving a system containing 6 formulas for 6 actuator forces F simultaneously. For example, in a preferred embodiment, the formula system is expressed in the form of a matrix: CPmat * F = B (1) where CP mat is a coupling matrix, with 6 rows of 3 columns above and filled with actuators 2 1 5 The unit vector, and the 6 rows in the lower 3 columns are filled with the relative unit torque to inertia ratio, where the unit torque is the weight of the actuator unit vector relative to the platform! Moment of force center. Calculating the cross product of the foot unit vector and the center of gravity to the actuator pivot vector can be done. In formula (1), F1 is a vector containing six actuator forces. The quantities in the formula for the linear system (3 columns above) are forces. The quantity in the angle system formula (3 columns below) is the relative angular acceleration. So the first three elements of vector B on the right side of the formula are expressed in units of force. The last three elements are expressed in acceleration units. The first three elements of the B vector are the linear forces in each of the three directions X, Υ, and Z. The calculation is as follows: M u / (M u 4- Μ 1) · (Μ 1 · a -fsum _ f ο ο t) (2) where Mu is the mass of the upper platform, M i is the mass of the lower platform, and a is the commanded acceleration of the upper platform relative to the lower platform in step 1305. Suni_foot is all the linear force exerted by the vibration isolation component on the lower platform in step 1 3 2 5

第35頁 五、發明說明（32) 總和。B向量的次3個元素是步驟1 3 0 5的指令相對角度加速 度。 藉由先決定耦合矩陣CPma t的值，並接著將其倒轉並乘 上B向量，即可解出公式（1)以求出致動器力向量F，其係 命令致動器力F *，如公式（3 )所示： F = CPmat_1 ·Β (3) 參考圖1 4以進一步說明計算耦合矩陣CPmat及Β向量所需的 步驟。 在步驟1 3 1 0解出公式（3 )以求出命令致動器力F *之後， 即從反作用前饋計算器1 2 1 0輸出致動器力。接著在步驟 1 3 2 0，用命令致動器力F *來決定下平台2 0 0的新位置，速 度，及加速度。例如因為力等於質量乘以加速度，因此將 下平台上的所有施力除以下平台2 0 0的質量即可得到直線 加速度。類似的，施加在下平台上的所有矩的總和乘以下 平台慣性矩陣的反矩陣即可得到下平台加速度。將加速度 積分以得到下平台2 0 0的速度。可以將速度再積分以得到 下平台2 0 0的位置。 在步驟1 3 2 5，用下平台2 0 0的新位置，速度，及加達度 來計算出振動隔離成分2 4 5的位置及速度，其當與隔離成 分的堅硬及減振特徵合併時，即可用以計算出隔離成分施 加在下平台上的力。振動隔離成分在下平台2 0 0的施力， 及下平台2 0 0的新位置，速度，及加速度，在步驟1 3 1 0中Page 35 V. Description of Invention (32) Sum. The next three elements of the B vector are the relative angular acceleration of the instruction in step 1 3 0 5. By first determining the value of the coupling matrix CPma t, and then inverting it and multiplying it by the B vector, formula (1) can be solved to find the actuator force vector F, which commands the actuator force F *, As shown in formula (3): F = CPmat_1 · B (3) Refer to FIG. 14 to further explain the steps required to calculate the coupling matrix CPmat and the B vector. After solving formula (3) in step 1 3 1 0 to obtain the commanded actuator force F *, the actuator force is output from the reaction feedforward calculator 1 2 1 0. Then in step 1320, the command actuator force F * is used to determine the new position, speed, and acceleration of the lower platform 200. For example, because the force is equal to the mass times the acceleration, the linear acceleration can be obtained by dividing all applied forces on the lower platform by the mass of the platform 2 0 0 below. Similarly, the acceleration of the lower platform can be obtained by multiplying the sum of all moments applied to the lower platform by the inverse matrix of the platform inertia matrix below. The acceleration is integrated to obtain the speed of the lower platform 2 0 0. The speed can be reintegrated to get the position of the lower platform 2 0 0. In step 1 3 2 5, the position and velocity of the vibration isolation component 2 4 5 are calculated using the new position, speed, and degree of dada of the lower platform 2 0 0 when combined with the rigidity and vibration damping characteristics of the isolation component. , Can be used to calculate the force of the isolation component on the lower platform. Vibration isolation components are applied to the lower platform 2000, and the new position, velocity, and acceleration of the lower platform 2000 are in step 1 3 1 0.

苐36頁苐 Page 36

五、發明說明（33) 再度使用以決定B向量，並重覆上述過程。 現在參考圖14，在步驟1310中計算致動哭六 驟從步驟1 4 0 0開始敎述。如上所述，必須建 而' CPmat及Β向量以求出公式中的致動器力向量F。在牛 1 4 0 5 ’ PENTIUM處理器43 0計算上平台20 5及下平台2步〇〇驟的矩 臂。藉由將重力位置中心從接到該樞軸位置的致動哭上的 枢軸位置中減去即可計算出矩臂是一向量。在步驟1 4 1 〇， PENTIUM處理器430計算上平台205及下平台2〇〇的單位扭矩 向量。單位扭矩向量等於致動器2 1 5的單位向量與相關矩 臂的交叉乘積。 在步驟1 4 1 5 ’ P E N T I U Μ處理器4 4 0從轨線產生器j 2 〇 5得到 相關的角度加速度。在步驟1420 ，用包含致動器的單位向 量的上方3列來填充耦合矩陣。用相對單位扭矩與慣性比 來填充下方3列。用以下向量公式來計算它： (lui.(R丁·Tugn))-RT*(Rlc*(ll”(Rlcr.Tlgj)) (4) 其中使用此向量6次（各致動器一次）以便用6個向量填充轉 合矩陣C P m a t的下方3列的各行。項1 u i是上平台2 0 5慣性拓 陣的反矩陣’以代替在公式4的分母中輸入一矩陣。類似 的1 1 1是下平台2 0 0慣性矩陣的反矩陣。項TUgn表示6個致 動器的每一者所屬的上平台2 0 5的6個單位扭矩向量。類似 的’項Tlgn表示6個致動器的每一者所屬的下平台200的β 個單位扭矩向量。上平台及下平台的旋轉矩陣r及R丨c分別 說明相對於整體座標系統的平台旋轉。5. Description of the invention (33) It is used again to determine the B vector, and the above process is repeated. Referring now to FIG. 14, the actuation cry is calculated in step 1310, and the description starts from step 1400. As described above, the CPmat and β vectors must be constructed to obtain the actuator force vector F in the formula. The cattle 1 4 0 5 ′ PENTIUM processor 43 0 calculates the moment arms of the upper platform 20 5 and the lower platform 2 steps. The moment arm is a vector by subtracting the center of the gravity position from the pivot position on the actuation cry connected to the pivot position. In step 1410, the PENTIUM processor 430 calculates a unit torque vector of the upper platform 205 and the lower platform 200. The unit torque vector is equal to the cross product of the unit vector of the actuator 2 1 5 and the relevant moment arm. At step 14 15 ' P E N T I U M processor 4 40 obtains the relevant angular acceleration from the trajectory generator j 2 05. At step 1420, the coupling matrix is filled with the upper 3 columns of the unit vector containing the actuator. Fill the lower 3 columns with the relative unit torque to inertia ratio. Calculate it with the following vector formula: (lui. (R 丁 · Tugn))-RT * (Rlc * (ll ”(Rlcr.Tlgj)) (4) where this vector is used 6 times (once for each actuator) so that Fill the rows in the lower 3 columns of the transformation matrix CP mat with 6 vectors. The term 1 ui is the inverse matrix of the upper platform 2 0 5 inertia matrix instead of entering a matrix in the denominator of Equation 4. Similar 1 1 1 Is the inverse matrix of the lower platform 2 0 0 inertia matrix. The term TUgn represents the 6 unit torque vectors of the upper platform 2 0 5 to which each of the 6 actuators belongs. A similar 'term Tlgn represents 6 actuators Β unit torque vectors of the lower platform 200 to which each belongs. The rotation matrices r and R 丨 c of the upper platform and the lower platform respectively describe the platform rotation relative to the overall coordinate system.

第37頁 五、發明說明（34) 旋轉矩陣R的例子是： [cb*cc -cb*sc sb] [ca*sc-sa*sb*cc ca*cc-sa*sb*sc -sa*cb] [sa*sc-ca*sb«cc sa*cc-rca«sb*sc ca.cb] 其中： sa=sin(A) ca=cos( A) sb=sin(B) cb=cos(B) sc=sin(C) cc=cos(C) 出現在公式（4 )的R及R 1 c轉置矩陣可以將量以公式表示在 相同的座標糸統中。 在步驟1 42 5，PENTIUM處理器4 30根據公式（2)而計算B向 量的上方3個值。使用以下公式來填充B向量的下方3個 值： maccc—ang+R丁•Rlc.ll — Rlc'foot—ang—tot (5) 其中maccc_ang是來自產生器1205的指令相對角度加速 度，R及Rlc是先前定義的，而foot_ang_tot是一向量以表 示振動隔離成分在下平台上所有施加的矩總和。 在步驟1 4 3 5，數位信號處理器4 4 0根據上述值而建立B向 量。在步驟1 44 0，PENTI UM處理器43 0將耦合矩陣CPmat反 轉並將它乘上B向量以找出致動器力向量F。接著將各致動 器2 1 5的命令致動器力F * (或是力向量F )到圖1 2的運動計算Page 37 5. Explanation of the invention (34) An example of the rotation matrix R is: [cb * cc -cb * sc sb] [ca * sc-sa * sb * cc ca * cc-sa * sb * sc -sa * cb ] [sa * sc-ca * sb «cc sa * cc-rca« sb * sc ca.cb] where: sa = sin (A) ca = cos (A) sb = sin (B) cb = cos (B) sc = sin (C) cc = cos (C) The R and R 1 c transpose matrices appearing in formula (4) can express quantities in the same coordinate system. At step 1 42 5 the PENTIUM processor 4 30 calculates the three values above the B vector according to formula (2). Use the following formula to fill the bottom 3 values of the B vector: maccc—ang + R 丁 • Rlc.ll — Rlc'foot—ang—tot (5) where maccc_ang is the relative angular acceleration of the instruction from the generator 1205, R and Rlc Is previously defined, and foot_ang_tot is a vector representing the sum of all moments applied by the vibration isolation component on the lower platform. In step 1 4 3 5, the digital signal processor 4 4 0 establishes a B vector based on the above value. At step 1 44 0, the PENTI UM processor 43 0 inverts the coupling matrix CPmat and multiplies it by the B vector to find the actuator force vector F. Next, the command actuator force F * (or force vector F) of each actuator 2 1 5 is calculated by the motion of FIG. 12

苐38頁 五、發明說明（35) 器1 2 1 5 ’而流程在步驟丨4 4 5結束。 參考圖1 5，其說明圖1 2運動計算器1 2 1 5的信號流程圖。 輸入運動計算器1215的是來自執線產生器1205的命令位置 向量P (其包含X，Y，Z，A，B，及C的6個位置指令）及致動器力 F *。運動計算器1 2 1 5使用旋轉矩陣產生器1 5 0 5中的旋轉命 令位置資料[A, B，C ]以產生一旋轉矩陣R。上平台枢軸位置 常數（其係圖1 〇步驟1 〇 5 5中決定的極軸位置向量）則加入加 法益1 5 2 5並輸入到乘法器1 5 1 0以便將框轴向量旋轉R。運 動計算器1 2 1 5接著將軌線產生器1 2 0 5的線性指命位置資料 [X，Y，Z ]加入加法器1 5 1 5中的旋轉框轴位置資料。 運動計算器1 2 1 5使用致動器力F *及拖轴位置計算器1 5 2 0 中的致動器單位向量以計算上平台205及下平台2〇〇的樞軸 位移。這些樞軸位移導因於機器平台的彈性撓西，這是因 為致動器力F *及慣性力。這些扭曲稱為機器的順應性。參 考圖1 6以說明收集此順應性資訊及建立一機器順應性模型 所採取的步驟=將上平台樞軸位移加入加法器1 5 2 5中的上 平台樞軸位置常數。類似於上述上平台處理，下平台樞軸 位置常數’其表示從1 〇 5 5的樞軸位置向量加入加法器1 5 3 0 中的下平台樞轴位移中。現在將上及下平台枢軸位置向量 送入致動器長度計算器1 54 0，其用畢氏定理來計算出致動 器向量的大小。此大小是致動器長度。運動計算器丨2丨5也 使用純量萬向接頭順應性計算器！ 5 4 5以計算致動器接頭硬 度的平銜圈環順應性表示的純量成分’該接頭如圖3 β的萬 向接頭3 8 0及3 82。I!型接頭純量成分的順應性則加入加法苐 Page 38 V. Description of Invention (35) 器 1 2 1 5 ′ and the process ends at step 丨 4 4 5. Referring to FIG. 15, a signal flow chart of the motion calculator 1 2 1 5 of FIG. 12 is described. Input to the motion calculator 1215 is a command position vector P (which includes 6 position commands of X, Y, Z, A, B, and C) and an actuator force F * from the thread generator 1205. The motion calculator 1 2 1 5 uses the rotation command position data [A, B, C] in the rotation matrix generator 15 0 5 to generate a rotation matrix R. The upper platform pivot position constant (which is the polar axis position vector determined in Fig. 10, step 10, 5) is added to the addition benefit 1 5 2 5 and input to the multiplier 1 5 1 0 to rotate the frame axis vector by R. The motion calculator 1 2 1 5 then adds the linear reference position data [X, Y, Z] of the trajectory generator 1 2 0 5 to the rotation frame axis position data in the adder 1 5 1 5. The motion calculator 1 2 1 5 uses the actuator force F * and the actuator unit vector in the drag position calculator 1520 to calculate the pivot displacement of the upper platform 205 and the lower platform 200. These pivotal displacements are due to the elastic torsion of the machine platform due to the actuator force F * and the inertial force. These distortions are called machine compliance. Refer to Figure 16 to illustrate the steps to collect this compliance information and build a machine compliance model = add the upper platform pivot displacement to the upper platform pivot position constant in the adder 1 2 5 5. Similar to the above-mentioned upper platform processing, the lower platform pivot position constant ′ represents that the pivot position vector from 105 is added to the lower platform pivot displacement in the adder 1530. The upper and lower platform pivot position vectors are now fed into the actuator length calculator 1 54 0, which uses Bishop's theorem to calculate the size of the actuator vector. This size is the actuator length. Motion calculator 丨 2 丨 5 also use scalar universal joint compliance calculator! 5 4 5 The scalar component expressed as the compliance of the armature ring for calculating the stiffness of the actuator joint. The joint is shown in Fig. 3 β universal joints 3 8 0 and 3 82. The compliance of the scalar component of the I! Joint is added.

第39頁 命令致動=度中。結果是從運動計算器1 2 1 5輸出的 資訊及建：圖2 ’其顯不樞軸撓曲計算器1 5 2 0收集順應性 貝成及建立一機器順應性 丁 上所述，六腳機器人中心U0:正 的流程圖。如 眾多拖軸位置21 5a及2151)的正確//疋由微電腦4則定 述，當施力在上平台2 0 5及下決定。也如上所 U5b會稍微移動。這些領著力/八20 0化，框軸位置2〗5a及 的Μ固致動器力（伸張及壓 .6個可延伸致動器21 5 (壓縮），·及分散質量的加速度，。來自平衡235的3個平衡力 從圖16的步驟1 6 00開始，使 件軟體上執行而建構上平台的Z 工作站在CAEDS有限元 該軟體是結構動力研究公司的 ^ 70件模型（步驟1 6〇5 )， 2 0 5的轉袖E 220固定在地面i，=在模型t，上平台 曲減到極小。這些有限元件模 ;;目的是使此成分的撓 施力或者記錄位移。在較佳實施：二眾：節點’以便在此 便可在此施力”可在上平台中的6個樞：Si個二節力點以 以及可在上肀台205的平衡樞軸位 軸广置2丨讣把力； 口 15b。有限元件模型也匕平台2〇5中的致 ’ Y及z方向中上平台的加逮度供位移郎點中的撓 施加單位加速度至3個Ϊ氏方向X，Y，z之 。致動器樞軸2 1 5b在上平台2 0 5的撓曲回 ^ 1 b 1 5 Φ ^ . rb ^ η ^ w 的3個卡氏成分中的6個位移節點，包含=二：5己餘撓曲 動器Μ轴位置21 5b。有限元件模型也賠以二:&致 曲以回應 在步驟1 6 1 —的上平台 應接著在少鰥1615中記錄’其中各卡氏成分記錄在抱 五、發明說明（37) 軸2 1 5 b中的每一者。步驟1 6 2 0使得流程在步驟1 6 1 0及1 6 1 5 重覆以便在各卡氏方向獨立的施以加速。 記錄加速的上平台成分的回應後，在步驟1 6 2 5選定一施 力點，並將一單位力施在單·—^氏方向X，Y，Z中的選定施 力點（步驟1 6 3 0 )。致動器樞軸2 1 5 b在上平台的撓曲回應或 樞軸移動則在步驟1 6 3 5中記錄，其中各卡氏成分記錄在上 平台20 5中6個致動器框軸中的每一者。步驟1 64 0使得流程 重覆步驟1630及1635以獨立的施加單位力在各卡氏方向’ 而且仍在選定節點中施力。步驟1 6 4 5使得所有節點都執行 步驟1 6 2 5至1 6 4 0，其中施加單位力在該節點中：6個致動 器樞軸21 5b及3個平衡枢轴2 5 0。 下平台200的分析從步驟1650開始’在步驟1655 ’用上 述IBM工作站在CAEDS有限元件軟體上執行而建構下平台 200的有限元件模型。下平台200的工作台240的模型固定 在地面上，因為目的是使此成分的撓曲減到極小。如同上 平台，下平台的有限元件模型包含眾多節點’以便在此施 力或記錄位移。在較佳實施例中’界定數個力節點以便可 在此施力；可在下平台中的6個樞軸位置2 1 5 a施力。記錄 撓曲的3個卡氏成分中的位移節點’包含下平台2 0 0中的致 動器樞軸位置2 1 5 a。 設定下平台2 0 0的有限元件模型後’流程即執行步雜 1 6 2 5到1 6 4 5，即在單--^氏方向於一力點各施加單位力’ 並記錄下平台2 0 0中所有致動器樞軸2 1 5 a最後撓曲的卡氏 成分。最後，當已記錄所有的樞軸位移後流程即在步靜P. 39 Command actuation = medium. The result is the information and output from the motion calculator 1 2 1 5: Figure 2 'It shows the pivot deflection calculator 1 5 2 0 to collect compliance and build a machine compliance Ding said, six feet Robot Center U0: Positive flowchart. For example, the correct positions of the multiple drag shafts 21 5a and 2151) are determined by the microcomputer 4 when the force is applied on the upper platform 2 0 5 and lower. As mentioned above U5b will move slightly. These leading forces / 8200, the frame axis position 2 5a and M solid actuator force (extension and compression. 6 extendable actuators 21 5 (compression), and the acceleration of the dispersive mass. The three balancing forces of the balance 235 start from step 1 6 00 in FIG. 16, and the Z workstation which is executed on the software and constructed on the platform is in CAEDS finite element. This software is a ^ 70-piece model of the structural dynamics research company (step 16). 5), the sleeve E 220 of 205 is fixed to the ground i, = at the model t, the upper platform curvature is reduced to a minimum. These finite element molds; the purpose is to make the force exerted by this component or record the displacement. Implementation: Two public: nodes 'so that you can apply force here' 6 pivots in the upper platform: Si two joint points and a wide range of 2 pivot points on the balance platform 205丨 讣 Force force; Mouth 15b. The finite element model is also used in the platform 205 to cause the upper and lower platforms in the Y and z directions to apply the unit acceleration to the three directions in X, Y, z. Actuator pivot 2 1 5b in the deflection back of upper platform 2 0 5 ^ 1 b 1 5 Φ ^. Rb ^ η ^ w 6 displacement nodes, including = 2: 5 redundant flexor M-axis position 21 5b. The finite element model is also compensated by: & cause to respond in step 1 6 1-the upper platform should then continue at less鳏 Recorded in 1615 'wherein each Kelvin component is recorded in Bao Wu, description of invention (37) each of axis 2 1 5 b. Step 1 6 2 0 makes the process in steps 1 6 1 0 and 1 6 1 5 Overlap so as to apply acceleration independently in each direction of Karst. After recording the response of the accelerated upper platform component, select a force application point in step 1 6 2 5 and apply a unit force in the single-direction X, The selected force application point in Y, Z (step 1630). The deflection response or pivot movement of the actuator pivot 2 1 5 b on the upper platform is recorded in step 16 3 5 in which each card The constituents are recorded in each of the six actuator frame axes in the upper platform 20 5. Step 1 64 0 makes the process repeat steps 1630 and 1635 with independent applied unit forces in the directions of each Kelvin 'and still being selected Apply force in the node. Step 1 6 4 5 causes all nodes to perform steps 1 6 2 5 to 1 6 4 0, where the unit force is applied in the node: 6 actuator pivots 21 5b and 3 balance pivots 2 50. The analysis of the lower platform 200 starts from step 1650 'at step 1655' using the above-mentioned IBM workstation to execute on the CAEDS finite element software to construct the finite element model of the lower platform 200. The lower platform 200 The model of the workbench 240 is fixed on the ground, because the purpose is to minimize the deflection of this component. Like the upper platform, the finite element model of the lower platform contains many nodes to apply force or record displacement here. In the preferred embodiment, 'a number of force nodes are defined so that force can be applied here; force can be applied at 6 pivot positions 2 1 5 a in the lower platform. The displacement node of the 3 deflected Karst components is recorded to include the actuator pivot position 2 1 5 a in the lower platform 200. After setting the finite element model of the platform 2 0 0 'the process will be executed from 1 2 6 5 to 1 6 4 5 that is, the unit force is applied at a force point in the direction of the single- ^' direction and the platform 2 0 is recorded All actuator pivots in 0 2 1 5 a Karst component of last deflection. Finally, when all the pivot displacements have been recorded, the process is at rest.

第41頁 五、發明說明（38) 1 6 6 0結束。 注意，施力的回應撓曲是使用有限元件模型來計算的， 其也以用實驗方式得到。在有限元件模型中執行時實際的 平台結構可以是固定的，力成分可以實體施加在力節點 上，而且可以實體測量位移節點撓曲回應。 圖1 6較佳收集的資訊可使用在六腳機器人中心1 1 0的枢 軸撓曲計算器1 5 2 0中。力節點上施加的力的實際卡氏成分 乘上計算出的位移節點（即致動器枢軸）的對應有限元件模 型撓曲，以回應施加在力節點的單位力。施在上平台的加 速的實際卡氏成分則乘上計算出的位移節點（即致動器柩 軸）的對應有限元件模型撓曲，以回應施加在上平台2 0 5的 單位加速。重疊原理允許各力或加速產生的位移相加，藉 以產生位移節點（即致動器樞軸）的實際撓曲估計值。 在附錄A以上平台的數學形成來說明重疊過程，而附錄A 在此併供參考。C 1矩陣包含計算出的有限元件撓曲以回應 單位力，最好是根據圖1 6來決定。施加在力節點的力成分 包含在F u向量中，其用旋轉矩陣R旋轉成固定下平台座標 系統。產生的P向量包含期望的樞軸位移，其由圖1 5的樞 軸撓曲計算器1 5 2 0輸出。在相同座標系統中提供該位移用 以在參考位置[0，0，0，0, 0，0 ]定義致動器樞軸座標。將位 移加入圖1 5步驟1 5 2 5的參考位置中的致動器極軸座標° 該注意的是以上節點數目及記錄數目只是供較佳實施例 用，本發明並不僅限於此目的。反之，一熟於有限元件模 型者可了解也能使用圖16的變化以提供彈性扭曲的任何統P.41 V. Explanation of the invention (38) 1 6 6 0 End. Note that the response deflection of the applied force is calculated using a finite element model, which is also obtained experimentally. When implemented in a finite element model, the actual platform structure can be fixed, the force component can be physically applied to the force node, and the displacement response of the displacement node can be measured physically. The better collected information in Figure 16 can be used in the pivot deflection calculator 1520 of the six-legged robot center 110. The actual Karst component of the force applied at the force node is multiplied by the corresponding finite element model deflection of the calculated displacement node (ie, the actuator pivot) in response to the unit force applied at the force node. The actual Karst component of the acceleration applied to the upper platform is multiplied by the corresponding finite element model of the calculated displacement node (ie, the actuator's y-axis) to flex in response to the unit acceleration applied to the upper platform. The principle of overlap allows the displacements produced by forces or accelerations to be added together to produce an actual deflection estimate of the displacement node (ie, the actuator pivot). The mathematical formation of the platform above Appendix A illustrates the overlap process, and Appendix A is here for reference. The C 1 matrix contains the calculated finite element deflections in response to a unit force, and is best determined in accordance with Figure 16. The force component applied at the force node is contained in the Fu vector, which is rotated by the rotation matrix R into a fixed lower platform coordinate system. The resulting P vector contains the desired pivot displacement, which is output by the pivot deflection calculator 1520 of FIG. 15. This displacement is provided in the same coordinate system to define the actuator pivot coordinates at the reference position [0, 0, 0, 0, 0, 0]. Add the displacement to the polar coordinates of the actuator in the reference position in step 1 5 2 5 of Fig. 15. It should be noted that the above number of nodes and the number of records are only for the preferred embodiment, and the present invention is not limited to this purpose. Conversely, those familiar with finite element models will understand that any system that can also use the variations of Figure 16 to provide elastic distortion

第42頁 五 發明說明（39) =十上的足夠數目，以準確的建構六腳機器人中心丨丨〇的順 應性板型。例如在機器的任何已知節點中可記錄施力，速 度’或加速的不同合併。 現在參考圖1 7 ’以說明圖1 2伺服回饋系統丨2 2 5的信號流 動圖形。伺服回饋系統丨2 2 5的輸入是命令致動器力ρ *，命 ^致動器長度U及命令致動器長度的微分dL*/dt。數位信 戒處理器4 4 0將命令致動器力F *送入單位轉換器丨7 〇 5以便 $力轉成扭矩。藉由將命令致動器乘以上平台2〇5的 ^程再除以2 *p丨即可完成。這是加法器1 7 1 0中加入的3個 祖矩的第一者以求出輸出指令扭矩T *。數位信號處理器 =〇將命令致動器長度L*送入加法器1 7 1 5 ,以便減去致動 器長度L、’它是從測量變換器465回饋的。透過比例積分微 分（_P ID)濾波器而送入致動器長度誤差，其中誤差的比例 增盈Kp則轉成扭矩’而誤差的微分增益Kd的微分則轉成扭 ，。接者將3個扭矩加入加法器丨7 2 〇，而結果是加入加法 ^ 1710的3個扭矩的第二個扭矩。帛分增益^是用#低# t 4 g <強化’而微分增益Kd是用以增加減振穩定 I B Μ工作站也將命令致私㈣ 將命令致動器長卢…，广長度L¥送入轉換器1 72 5 ’其 接著將指令角二;動器螺絲的指令角位置θ*。 減法器1 7 3 0中減去以11’.碼器4 4 5回饋信號的角位置0在 益KP及微分增益SikKc；=出—角度誤差。接著將一比例增 所述。接著將產生的h玲度誤差，其原因類似於如上 兔加入加法器丨7 3 5，而結果是角度Page 42 5 Description of the invention (39) = a sufficient number of ten or more to accurately construct a compliant shape of the six-footed robot center. For example, different combinations of force, speed 'or acceleration can be recorded in any known node of the machine. Reference is now made to Fig. 17 'to illustrate the signal flow pattern of Fig. 12 servo feedback system 2 2 5. The inputs of the servo feedback system 2 2 5 are the commanded actuator force ρ *, the commanded actuator length U, and the differential dL * / dt of the commanded actuator length. The digital signal processor 4 4 0 sends the command actuator force F * to the unit converter 丨 705 so that $ force is converted into torque. This can be done by dividing the command actuator by ^ of the above platform 205 and then dividing by 2 * p 丨. This is the first of the three ancestral moments added to the adder 17 10 to find the output command torque T *. The digital signal processor = 〇 sends the command actuator length L * to the adder 1 7 1 5 so as to subtract the actuator length L, ′, which is fed back from the measurement converter 465. The length error of the actuator is input through a proportional integral derivative (_P ID) filter, where the proportional gain Kp of the error is converted into torque ′ and the differential of the differential gain Kd of the error is converted into torque. The receiver then adds 3 torques to the adder 7 2 0, and the result is a second torque of 3 torques of the addition ^ 1710.帛 Differential gain ^ is to use # LOW # t 4 g < Strengthening 'and the differential gain Kd is used to increase vibration reduction and stability. The IB workstation will also order the private order. The command actuator will be long ... Into converter 1 72 5 'which then sets the command angle two; the command angle position θ * of the actuator screw. Subtractor 1 7 3 0 subtracts 11 ′. The angular position 0 of the feedback signal of the encoder 4 4 5 is 0 in the gain KP and the differential gain SikKc; = out—the angle error. A proportional increase is then described. Then the resulting h-degree error will be similar to the above. Rabbits add the adder. 7 3 5 and the result is the angle.

五、發明說明（40) 誤差扭矩的總和。將角度誤差扭矩的總和送入另一加法器 1 74 0。數位信號處理器4 4 0也輸入微分致動器長度dL*/dt 到轉換器1 74 5以便將它轉成角速度。接著將角速度乘上增 益以克服馬達4 5 0與球螺絲的黏性摩擦，並且將產生的圖 誤差之也加入加法器1 74 0。也將角速度微分S以得到角加 速，其乘上馬達的慣性J及致動器螺絲，而且將產生的扭 矩加入加法器1 7 4 0。加法器1 7 4 0的最後扭矩是加入的3個 扭矩的第三者，以完成指令扭矩T*至伺服放大器44 5。 在較佳實施例中，本發明的順應性補償技術利用反作用 前饋計算器1 2 1 0產生的清除力資料。惟，也可構思用其他 f 資料來使用相同的順應性補償步驟。例如馬達扭矩回饋資 料可用作此目的，雖然其一般是極吵雜的信號，而且會使 其很難以使用。 為了進一步參考，以下的附錄B說明反作用前饋計算器 1 2 1 0的詳細動作公式，而附錄C說明上述B向量的微分。這 2個後附的附錄都在此併供參考。 雖然已在此顯示並說明本發明的特定實施例，熟於此技 術者仍可作進一步的修正及改良。尤其該注意的是附圖所 述的軟體處理步驟及處理器只是可以使用的軟體常式及硬 體裝置的較佳實施例。本發明可以作各種軟體及硬體的修 正以利於應用在各種其他應用中。只要在申請專利範圍中 並保留上述基本原理，所有的這些修正都在本發明範圍及 精神下。V. Description of the invention (40) Sum of error torque. The sum of the angular error torque is sent to another adder 1 740. The digital signal processor 4 4 0 also inputs the differential actuator length dL * / dt to the converter 1 74 5 to convert it to angular velocity. The angular velocity is then multiplied by the gain to overcome the viscous friction between the motor 450 and the ball screw, and the resulting graph error is also added to the adder 174 0. The angular velocity S is also differentiated to obtain the angular acceleration, which is multiplied by the inertia J of the motor and the actuator screw, and the generated torque is added to the adder 1 7 4 0. The final torque of the adder 17 4 0 is the third of the 3 torques added to complete the commanded torque T * to the servo amplifier 44 5. In a preferred embodiment, the compliance compensation technique of the present invention utilizes the clearing force data generated by the reaction feedforward calculator 1 2 1 0. However, it is also conceivable to use other f data to use the same compliance compensation steps. For example, motor torque feedback data can be used for this purpose, although it is generally a very noisy signal and can make it difficult to use. For further reference, the following Appendix B describes the detailed action formula of the reaction feedforward calculator 1 2 1 0, and Appendix C illustrates the above-mentioned B vector differentiation. These two appendices are here for reference. Although specific embodiments of the invention have been shown and described herein, those skilled in the art can make further modifications and improvements. In particular, it should be noted that the software processing steps and processors described in the drawings are only preferred embodiments of software routines and hardware devices that can be used. The present invention can be modified for various software and hardware to facilitate its application in various other applications. All these modifications are within the scope and spirit of the present invention, as long as they are in the scope of patent application and retain the above-mentioned basic principles.

第44頁 五、發明說明（41)Page 44 V. Description of the invention (41)

附錄A 各種上平台樞轴位移補償 框軸位移向量可以用以下公式來計算出： κ·=.α(Β.ττυ)β 其中Appendix A Pivot Displacement Compensation of Various Platforms The frame axis displacement vector can be calculated using the following formula: κ · = .α (Β.ττυ) β where

Cl=上空間框的27個順應性矩陣的矩陣 R=上平台的旋轉矩陣Cl = matrix of 27 compliance matrices of the upper space box R = rotation matrix of the upper platform

Fu=6個腳力向量的向量，及3個平銜力向量a，b，c P =6個框軸位移向量的向量 C1的形式如下： cie CU Sb σΐ〇 ^21 *-〇3 ^ZS *-» ^ ^31 9jl- - C34 ^35 ^ ^1¾ C^j C^j C“ C" Q c“. cji 〜Cy cJ4 _C« C(a..C5J.C« CiS C6t C4* C» ^icFu = 6 vectors of foot force vectors, and 3 equal force vectors a, b, c P = 6 box axis displacement vectors C1 of the form: cie CU Sb σΐ〇 ^ 21 * -〇3 ^ ZS * -»^ ^ 31 9jl--C34 ^ 35 ^ ^ 1¾ C ^ j C ^ j C" C " Q c ". Cji ~ Cy cJ4 _C« C (a..C5J.C «CiS C6t C4 * C» ^ ic

Ci中的各元素由硬度矩阵單數。例如元素C23是極轴2的順應 ‘ί生以回應锊3上的力3 這些順應性矩障的形式如下：Each element in Ci is singular from the hardness matrix. For example, the element C23 is the adaptation of polar axis ‘ί to respond to the force 3 on 锊 3. These compliance moments are of the form:

第45頁 五、發明說明（42) Δ 素 元 中 其 化 變 力 的 向 方 X 應 回j__、； ο .. _^以 ^πϊιϋ 芒气·曲 •-:^!^玍^'撓iiK 紅14-的 ^ ^-向 rq:.方 X 是 腳力向里Fu的形式如下。各元素是三維腳衝力向量 例如F1是腳1的力向量，而是平衡a的力向量。 4區袖位移向量p s的形式如下： psr：= (?, P2 P, P, P； p5)« 例如P丨是樞軸1的三維撓曲向量。Page 45 V. Description of the invention (42) The square X of its changing force in the Δ element should return j__ ,; ο .. _ ^ to ^ πϊιϋ 芒 气 · 曲 •-: ^! ^ 玍 ^ '挠 iiK Red 14- ^ ^ -direction rq :. square X is the form of foot force inward Fu as follows. Each element is a three-dimensional foot force vector. For example, F1 is the force vector of foot 1, but a force vector that balances a. The form of the 4-zone sleeve displacement vector p s is as follows: psr: = (?, P2 P, P, P; p5) «For example, P 丨 is the three-dimensional deflection vector of pivot 1.

第46頁 五、發明說明（43) 祆铎B :具浮動下二台的各公式 直这系統：Page 46 V. Description of the Invention (43) Bian Duo B: The formulas with the next two floats Straight this system:

闲錄B反作罔前if計算器的動作公式 fy =功率.¾7力廷量 ep=功率爵力單位向量 F;=氣動彈著力向量 M.、=上平台質量 下平台買量 F:-',，安裝吾7力向量（來自數值積分下平台） 1.. =上+台慣性矩連 込=當abc二0時，閔於X，Y，Z的上五台慣性矩淳 R二旋轉矩漳以旋辞上平台 1,=下平台慣性矩?皁（假設在小負度下是常數） rpu,=來自Cg的矩ν吊於上平台上的功率辟1 來自Cg的矩臂另於T竿台上的功率拜I r^:=來自Cu的经¥吊於上平台上的氣動？單晉ί r^=來自Cg的廷背罔於下平台上的氣動译晉1 r:;=來自Cg的矩臂污於安装辟I 下平台的Cg的ϋ令直線加速 c二下平台的機器弍荇令角加速·其t手與整'登系托The action formula of the if calculator in the idle record B is fy = power. ¾7 force factor ep = power unit force vector F; = aerodynamic force vector M., = platform purchase amount under platform quality F:-' ,, install our 7 force vector (from the platform under numerical integration) 1 .. = up + platform moment of inertia continuity = when abc 2 0, Min Yu X, Y, Z upper 5 moments of inertia moment R 2 rotation moment Zhang uses the rhetoric upper platform 1, = moment of inertia of the lower platform? (Assuming constant under a small negative degree) rpu, = moment ν from Cg hangs from the power on the upper platform 辟 1 moment arm from Cg other than T Power on the pole platform I r ^: = Pneumatic from Cu via hanging from the upper platform? Shan Jin r r == Pneumatic translation from Cg's court on the lower platform 1 r:; = The moment arm from Cg is dirty to install the Cg of the platform I to order the linear acceleration of the machine on the second platform Commanding Angle Acceleration · Its t-hands and the whole '

第47頁 五、發明說明（44) 3一: a t r 10 : £1一- Π : ar 彐1. ίF1 1-—-·-丨日 L^l/iel ml-ac ^.-+-31^-=^Page 47 V. Description of the invention (44) 31: atr 10: £ 1 1-Π: ar 彐 1. ίF1 1 ----... 丨 L ^ l / iel ml-ac ^ .- +-31 ^ -= ^

3I f _ ίΤΓ + wi3I f _ ίΤΓ + wi

PI I I .wu ml 一: 31 U"丨 κΐ - C 2:--1 * f - § 3:3colvc eac-^OWM d. ... 丨 〃--:-;fy; :y 卡 ii;j.-PI II .wu ml 1: 31 U " 丨 κΐ-C 2:-1 * f-§ 3: 3colvc eac- ^ OWM d. ... 丨 〃-:-; fy;: y card ii; j .-

// 卩 fvi 一 W 二 d-— I一’一综 c:n-s.-,:,~;-s 耷->0< 二 s) -S-K:----- -''-SI-'--· Ac 一- 頁 五、發明說明（4〇) -μ <η il "Ί?κ—.- ,^ν 4/+，c. V. μ ml i.? ml+/i// 卩 fvi one W two d-— I one 'one comprehensive c: ns .-,:, ~; -s 耷-> 0 < two s) -SK: ----- -''- SI- '-· Ac 1-Page 5. Explanation of the invention (4〇) -μ < η il " Ί? Κ —.-, ^ ν 4 / +, c. V. μ ml i.? Ml + / i

IS; HI I 13: fl ffIS; HI I 13: fl ff

$r0 -&(-PJr /x - :^-+^9)-+^^1-11- fg - mi g) ml 1¾ 3-^-;T:5.-P·^ b' "-p ¾(芄a-tTliflt n?) t-*R{f-rh-flQ-^) -.-1:(名 pa--fff - fg - wl) - ml {fla- ym), /i+ml 12” .fl * -· .. ml -fyi. TL-U · —L^ru^r$ r0-& (-PJr / x-: ^-+ ^ 9)-+ ^^ 1-11- fg-mi g) ml 1¾ 3-^-; T: 5.-P · ^ b '" -p ¾ (芄 a-tTliflt n?) t- * R (f-rh-flQ- ^) -.- 1: (name pa--fff-fg-wl)-ml (fla- ym), / i + ml 12 ”.fl *-· .. ml -fyi. TL-U · —L ^ ru ^ r

第49頁Page 49

Claims (1)

Translated fromChinese

六、申請專利範s 1. 一種在定位裝置用以補償扭西之方法，定位裝置具有 一裝在眾多枢軸位置之眾多致動器上之可移動平台，該方 法包含以下步驟·· 收集定位裝置之順應性資訊； 根據收集之順應性資訊而補償移動指令；以及 根據補償移動指令而移動定位裝置致動器。 2. 如申請專利範圍第1項之方法，其中收集步驟更包 含： 發展定位裝置之有限元件模型；以及 使用得自至少一施力及一加速應用之有限元件模型， 以計算扭西而收集靈敏度資料。 3. 如申請專利範圍第2項之方法，其中有限元件模型包 含眾多節點，而收集靈敏度資料之步驟更包含： 固定有限元件模型之輸出；以及 對於眾多節點之每一者， (1)在一卡氏方向施加單位力及單位加速之一至有 限元件模型中眾多節點之一； (2 )在各卡氏方向於有限元件模型中眾多節點之至 少一者記錄產生之扭西； (3)在各卡氏方向重覆步驟（1 )至（2);以及（4)在眾 多節點之每一者重復步-驟（3 );( 5 )在一卡氏方向施加一單 位加速至有限元件模型；（6 )在各卡氏方向於眾多節點之 至少一者記錄產生之扭西；以及（7 )在加速之各卡氏方向 重覆步肆（5 )至（6 ) 36. Application for patents 1. A method for compensating for twisting in a positioning device, the positioning device has a movable platform mounted on a plurality of actuators at a plurality of pivot positions, and the method includes the following steps: Compliance information; compensating a movement instruction according to the collected compliance information; and moving the positioning device actuator according to the compensating movement instruction. 2. The method of item 1 in the scope of patent application, wherein the collection step further includes: developing a finite element model of the positioning device; and using a finite element model obtained from at least one force application and an accelerated application to calculate distortion and collect sensitivity data. 3. For the method in the second item of the patent application, where the finite element model includes many nodes, and the step of collecting sensitivity data further includes: fixing the output of the finite element model; and for each of the many nodes, (1) a Apply one of the unit force and the unit acceleration in the direction of the Kelvin to one of the many nodes in the finite element model; (2) record the twists generated by at least one of the many nodes in the finite element model in each of the Kelvin directions; (3) in each Repeat the steps (1) to (2) in the Karst direction; and (4) repeat the steps-steps (3) at each of the many nodes; (5) apply a unit acceleration to the finite element model in a Karst direction; (6) Twist generated by recording at least one of a plurality of nodes in each Karst direction; and (7) Repeat steps (5) to (6) 3 in each Karst direction in acceleration第50頁 六、申請專利範圍 4. 如申請專利範圍第1項之方法，其中補償步驟更包 含： 將框轴位置製模； 計算致動器力； 計算平台加速； 將致動器施力及加速施加至收集之順應性資訊以判定 極轴位置中之扭曲； 根據扭西而調整樞轴位置；以及 根據樞軸位置之調整而補償移動指令至定位裝置。 5. 如申請專利範圍第1項之方法，其中定位裝置係一六 鞞機器人中心。 6. —種補償順應性系統，包含： 一定位裝置；以及 一微電腦被配置以控制定位裝置，並使用關於定位裝 置之順應性資訊，並根據順應性資訊而調整定位裝置之控 制。 7. 如申請專利範圍第6項之順應性補償系統，其中定位 裝置更包含眾多致動器，連接在眾多框轴位置中之上平台 與下平台之間；以及 其中微電蹈更包含： 一機器控制配置成控制定位裝置： 機器樞軸模型配置成將定位裝置之眾多樞軲位置製 * 4 . ， 一致動器力計算器配置成使闬定位裝置之眾多特性以Page 50 VI. Patent Application Range 4. The method of the first patent application range, wherein the compensation step further includes: molding the position of the frame axis; calculating the actuator force; calculating the acceleration of the platform; applying force to the actuator and Accelerate the compliance information collected to determine the distortion in the position of the polar axis; adjust the pivot position according to the twist; and compensate the movement instruction to the positioning device according to the adjustment of the pivot position. 5. For the method in the first scope of the patent application, the positioning device is a sixteen robot center. 6. A compensation compliance system including: a positioning device; and a microcomputer configured to control the positioning device and use compliance information about the positioning device, and adjust the control of the positioning device based on the compliance information. 7. For example, the compliance compensation system of the 6th scope of the patent application, wherein the positioning device further includes a plurality of actuators connected between the upper platform and the lower platform in a plurality of frame axis positions; and the micro-electric track further includes: The machine control is configured to control the positioning device: The machine pivot model is configured to control the many pivot positions of the positioning device * 4. The actuator force calculator is configured to make the many characteristics of the positioning device to第51 I 六、申請專利範圍 判定眾多致動器在定位裝置上之施力； 一平台加速計算器；以及 一機器靈敏度模型配置成將關於定位裝置之收集順應 性資訊製模，以判定因施力與加速而導致眾多致動器之扭 西。 8. 如申請專利範圍第6項之順應性補償系統，其中定位 裝置係一六腳機器人中心。 9. 一種補償六腳機器人中心扭面之方法，在眾多框軸位 置，該中心具有一上平台藉由眾多致動器及眾多平衡而接 至一下平台，該方法包含： (a )發展六腳機器人中心之獨立有限元件模型； (b )固定有限元件模型之輸出； (c )選擇上平台上眾多樞轴位置之一； (d)在一卡氏方向施加一單位力至上平台上取多樞轴 位置之一，並且在各卡氏方向記録上平台上眾多框韩位置 之位移； (e )在各卡氏方向重覆步驟（d ); (f )在上平台上眾多樞轴位置之每一者重覆步驟（c )至 (e ); (g) 在下平台上選擇眾多樞軸位置之一； (h) 在一卡氏方向施加一單位力至下平台上幕多桓袖 位置之一，並且在各卡氏方向記録下平台上眾·多檀軸位置 之位移： (i) 在各卡氏方向重覆步琢（h);No. 51 I 6. The scope of patent application determines the force exerted by many actuators on the positioning device; a platform acceleration calculator; and a machine sensitivity model configured to model the collected compliance information about the positioning device to determine the cause Force and acceleration cause twisting of many actuators. 8. If the compliance compensation system of item 6 of the patent application, the positioning device is a six-leg robot center. 9. A method for compensating the torsional plane of the center of a six-legged robot. At the position of many frame axes, the center has an upper platform connected to the lower platform by numerous actuators and many balances. The method includes: (a) developing a six-legged robot Independent finite element model of the robot center; (b) Output of the fixed finite element model; (c) Select one of the many pivot positions on the upper platform; (d) Apply a unit force to the upper platform in the direction of Karst to take multiple pivots One of the axis positions, and record the displacements of the positions of the many frames on the platform in the directions of each K.K .; (e) repeat the steps (d) in each direction of K.K .; (f) each of the many pivot positions on the platform. One repeats steps (c) to (e); (g) selects one of a number of pivot positions on the lower platform; (h) applies a unit of force in one Karst direction to one of the positions on the upper platform of the lower platform , And record the displacement of the position of the multi-dot axis on the platform in each of the directions of Karst: (i) repeated steps in each direction of Karst (h);六、申請專利範圍 (J)在下平台上眾多樞軸位置之每一者重覆步驟（g)至 ⑴； (k) 在平衡上選擇眾多樞軸位置之一； (l) 在--^氏方向施加一單位力至平衡上眾多樞軸位 置之一，並且在各卡氏方向記錄上平台上眾多樞軸位置之 位移； (m) 在各卡氏方向重覆步驟（1); (η )在平衡上眾多樞軸位置之每一者重覆步驟（k)至 (m)； (〇)重覆步驟（c)至（f)及步驟（k)至（η)以一施加之單 位加速來取代施加之單位力； (ρ )在一機器控制中將眾多框軸位置製模； (q )計算致動器力； (r )施加致動器力至步驟（d )，（ h )，（ e )及（〇 )中記錄之 資料； (s) 根據致動器力施加至記錄資料所導致之扭曲而調 整眾多揠軸位置；以及 (t) 极據植轴位置調整而補償六腳機器人中心致動器 之移動指令。 10. —種在定位裝置用以補償扭ώ之裝置，定位裝置具 有一裝在眾多樞軸位置之眾多致動器上之可移動平台，該 裝置包含： 用以改集定位裝置之順應性資訊之裝置； 根據收集之順應性資訊而用以補償移動指令之裝置；Sixth, the scope of patent application (J) Repeat steps (g) to ⑴ for each of the many pivot positions on the lower platform; (k) Select one of the many pivot positions on the balance; (l) At-^ 's Apply a unit of force in one direction to one of the many pivot positions on the balance, and record the displacements of the many pivot positions on the platform in the directions of each Kelvin; (m) Repeat step (1) in each direction of Kelvin; (η) Repeat steps (k) to (m) for each of the numerous pivot positions on the balance; (〇) Repeat steps (c) to (f) and steps (k) to (η) to accelerate in an applied unit To replace the applied unit force; (ρ) mold a number of frame axis positions in a machine control; (q) calculate the actuator force; (r) apply the actuator force to steps (d), (h), (E) and the information recorded in (0); (s) adjusting the position of many axes according to the distortion caused by the actuator force applied to the recorded data; and (t) compensating the six-legged robot according to the axis adjustment Move command of center actuator. 10. —A device for compensating for twisting in a positioning device, the positioning device has a movable platform mounted on a plurality of actuators at a plurality of pivot positions, and the device includes: used to collect compliance information of the positioning device Devices; devices used to compensate for movement instructions based on compliance information collected;第53頁 六、申猜專利範圍 以及 根據補償移動指令而用以移動定位裝置之裝置3 1 1.如申請專利範圍第1 〇項之裝置，其中闬以收集順應 性資訊之裝置更包含： 定位裝置之有限元件模型；以及 .藉由記錄施力及加速所導致有限元件模型之扭西而用 以收集靈敏度資料之之裝置。 1 2.如申請專利範圍第1 0項之裝置，其中有限元件模型 包含眾多節點及用以收集靈敏度資料之裝置，更包含： 有限元件模型之固定輸出； 在一卡氏方向用以施加任一單位'力至有限元件模型中 眾多節點之任一者之裝置；以及 在各卡氏方向於有限元件模型中眾多節點之至少一者 周以記錄產生之扭面之裝置。 1 3.如申請專利範圍第1 0項之裝置，其中闬以補償之裝 置更包含： 用以将樞軸位置製模之裝置； 闬以計算致動器力之裝置； 闬以計算平台加速之裝置； 罔以施加致動器力及加速至改集之順應性資訊，以判 定棍軸位置中之担曲之裝置； 根據扭西而罔以調整樞韩位置之裝置；以及 根據框轴位置之調整而闱以補償移動指令至定位裝置 之裝置。Page 53 6. The scope of the patent application and the device used to move the positioning device according to the compensation movement instruction 3 1 1. As for the device in the scope of patent application No. 10, the device that collects compliance information further includes: positioning A finite element model of the device; and a device for collecting sensitivity data by recording distortions of the finite element model caused by force and acceleration. 1 2. The device according to item 10 of the scope of patent application, wherein the finite element model includes a plurality of nodes and a device for collecting sensitivity data, and further includes: a fixed output of the finite element model; used to apply any A device that applies a unit of force to any of a number of nodes in the finite element model; and a device that records at least one of the nodes in the finite element model in each Kelvin direction to record the resulting twisted surface. 1 3. The device according to item 10 of the scope of patent application, among which the device for compensation further includes: a device for molding a pivot position; 轴 a device for calculating an actuator force; 闬 a device for accelerating by a computing platform Device; 施加 a device that applies actuator force and accelerates to the compliance information for the compilation to determine the tension in the position of the stick shaft; a device that adjusts the position of the pivot shaft according to the twist; and a device based on the position of the frame shaft The device is adjusted to compensate for the movement instruction to the positioning device.第54頁Page 54