TW396096B - System and method for calibrating a hexapod positioning device - Google Patents

Abstract

A system and method for calibrating a positioning device, such as a hexapod machining center, is disclosed. The system comprises a gage nest, having a plurality of sensors and being connected to the positioning device; an artifact, disposed in a work area defined by the positioning device; a first computer connected to said positioning device, for controlling and recording precision movements of the positioning device with respect to the artifact; and a second computer for using the recorded precision movement as input data in order to simulate the operation of the positioning device, thereby iteractively determining the actual geometry of the positioning device.

Description

V. Description of the invention (1) Background of the invention L. Invention Fan Lu The present invention is generally related to machines, such as units, coordinated measuring machines, and positioner machines: workers; Talking Robot Center "2. These directions and the three rotation directions are centered. / 、 工 / 、 Can be explained in 3 linear L · previous cases ㈣ ^^^ U.S. Patent 5,538,37 " granted to ^ Cao, * 1 996/0 9/1 7 U.S. Patent 5 issued to Sheldon „R 9 / 19 土 id and ^ iddin.s & Lewis ^ ^ 4 ^ ； ^ ^ Patent unveiled a six-legged robot tool, package, and tool platform ΆΓ :; 置 ”楼 —Worker's good job 1 罝 fork Paul Ya drive-tools. By manipulating the actuator # 定 #, along each linear and rotational axis relative to the work piece. Six-axis positioning provides the power of the operator's precision tools == the function of the machine tool. ,, silly π & small:: ϊ = ϊ 的 正! The degree is determined by the machine at the time of calibration and setting. Each d = = ί ® Take 1 foot robot tool as an example. The 12 pivot points must be correct and long sound. The position of the actuator can be defined relative to the position of the broken tool. The total positive reading for the work piece's positive and mechanical benefits. A standard method for determining the precision of a machine is to measure its individual plates. E.g

i Page 5 _ 丨 5. Description of the invention (2) In conventional presses, the guidance combined with machine movement must be accurately measured, machined or scraped, monitored, and maintained within known tolerance limits. In order to reduce the increase in shame (ie, the larger total error obtained by adding the errors of the components), the tolerance value of each component must be minimized. In the above example, the error in the linearity of the guide directly affects 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 will lead to extremely high-diameter manufacturing costs and / or structural complexity. For example, each component must be manufactured and maintained with extremely accurate standards compared with other requesters, and no increase in tolerance will occur. In addition, determine how the error of each component makes the total The process of increasing machine errors is extremely difficult for a six-legged robot center. In the following works: JA Soons, " ERROR ANALYSIS OF A HEXAPOD MACHINE TOOL, " Lambda, _Map 1.97, Thjrd International Conference and Exhibi t i.on__on, Laser Metrology, Machine PerXormance '1 9 97/07/1 5-1 In 7, the author analyzes the possible measurement errors produced by the six-leg press of the National Standards and Technology Center, and the difficulty in determining these errors. In a six-legged robot tool, the visibility of each actuator must also support changes when the tool platform moves. The acceleration when the platform moves will also exert a variable force on the different components of the machine. As a result, these forces are subject to elastic deflections at any known time, so the geometry of the machine will always slightly modify these elastic deflections and reduce the correctness of the center of the six-legged robot. The test method is to compensate for the error caused by the changing position and acceleration of the machine. The error is the output position error (such as the six-foot robot tool in the tool itself, page 6 V. Description of the invention (3) --- 3: f) £. . These errors are stored in the machine control system and used to adjust by measuring the machine wheel out position to _ A. .., which needs to be measured per-total approximate f: J ΐ is known as an orphan, especially when other axes ^ Setting error. Limited number of measurements required in a six-legged robotic tool. When there is a strong influence on positional errors, if so there is a need to provide an improved device or other machine with a method to easily correct the positioning of the loading center. The degree of freedom of the eight-dimensional festival, especially with a six-legged machine. Therefore, the principal of the present invention is to provide an improved system and method for correcting multi-dimensional freedom, using another device of the present invention. The method and device for judging a seed under the smart component do not need to be precise in the precise connecting rod size of another object of the present invention. Device input for a method and system to move the machined product through the wheel out to the machine size, and the machine executes the related invention of the present invention-the eyepiece stop robot size of the initial shake f to provide a method and device, in the present invention The other purpose is to calibrate the center of a six-legged robot. : Used to determine machine geometry, methods and devices, continue to settle with a minimum amount of settings. 、 It is easy to mention the need of a special machine. In order to accomplish this, the method is to provide a corrective positioning device provided by Jiaming.

°° Precise movement defined by the characteristics of desire, J. V. Description of the invention (4) ίΚί A moving platform on the base by many actuators, the I > has many pivot positions. The method includes the following steps: Actuator's actuator length change data is used to respond to the positioning device; 、, move, and (b) Use the actuator length change data to execute Choi 1 set the curtain in the tolerance limit, the multi-frame axis is straight 0,

A method for calibrating the center of a six-legged robot is provided. The center and the platform each have a plurality of pivots. In the six-legged robot, there are a plurality of extendable feet, each of which is connected to one of the many pivots of the platform. Xi pivot between one. The method includes the following steps: (a) the estimated value of the fixed earth-Zhongxi holding axis position; (b) when the platform is in a reference position, the estimated values of the positions of many pivot vehicles in the mobile platform are defined; (c) when the platform is in At this reference position, calculate the different actuator lengths from the estimated values described in steps (a) and (b); (d) place a precision ball near the platform; (e) install a subtractor on the platform ' It is designed to indicate the distance between the unique point on the platform and the center of the precision ball; (f) Get the required platform instruction data to move the platform to a position, and the sensor starts measuring the distance described in step (e); ( g) moving the platform according to the platform instruction data; (h) obtaining the distance described in step (e) from the sensor; (i) obtaining the determined platform instruction data to move the platform to another location, and reducing the step (e The distance described in '); (J ·) moves the platform according to the platform instruction data; (k) repeats step (h) two (j) until the distance described in step (e) falls within a preset range. Around +; (1) when the platform is in the test position described in step (c), Output the change in actuator length to when the platform meets the conditions of step (k); (m) when the platform is not precise

I. Explanation of the invention (5) When the last position on the dense ball is obtained, the required platform instruction data is obtained to move the platform. At different positions, the sensor starts measuring the distance described in step (e); (η) Repeat steps (g)-(m) until the final position of the platform on the precision ball; (〇) When the platform is not on the last precision ball, place a precision ball in a new position near the platform; (P) Repeat steps (F)-(0) until the platform is on the last precision sphere; (Q) Judge enough accurate estimates of the actual positions of the multiple axes. These many pivots are determined by (q 丨) estimating the precise ball position; (q 2) using the platform instruction data of step (1) to simulate the movement of the platform to generate a simulated platform position, and the sensor is centered in precision ' On the ball; calculate the error between the simulated platform position and the estimated position of the precision ball; (q4) adjust the precision ball position and actuator pivot position to reduce the error described in step (q3); (q5) repeat Steps (q2)-(q4) until the error described in step (q3) falls within the preset range; (q6) Outputs the last Z axis position generated in step (q5), where the last pole draw position represents the actual pivot A sufficiently accurate estimate of the position. It is human-oriented to provide a system for calibrating a positioning device. The system includes a 5 measurement kit 'with a plurality of sensors and connected to the positioning device. A first computer connected to the positioning device, used to control and record the relative movement of the positioning device as 作为 ί ,,, .. The second computer is used to use the recorded precision device to simulate the positioning device. Overlay determination to locate the date and month + # # Step: (al) Fork—a method for calibrating the positioning device, including the following steps] Estimating the geometry of the positioning device at the reference position; (a2) Near the platform; (a 3) Install the sensor on the platform, which is designed to mean

Page 9 V. Description of the invention (6) Show platform movement and processing σ _ platform instruction data to shift = offset between fine f features; (a4) get the precise features of the required work; f platform to a position,俾 The sensor starts measuring plus U6) Move the platform from the sensor according to the platform instruction data, · The platform instruction data is two; the offset described in (a3); U7) It is determined that the platform is used to operate the platform Go to another position and reduce the offset; (a8) until the sensor drifts down / moves flat D, (a 9) repeat steps (a 6)-(a 8) change the data and set step (7 to the preset range) (A〗 〇) According to the actuator position step U9), the change in the position geometry is output to the last position of the satisfying step Htr (all) When the platform is not in the same position with respect to the processed product, the r The instruction data is to move the platform to-without repeating the steps to process the precision feature of 0 mouth; (al2) the heavy product is not on the platform to the platform in the final position; (a) when processing, approaching the position, place the processed product on the flat surface Near △,-U12) The measurer is not π., Thousand. Move it to the previous step (a2) I. (ai4) Repeat steps (a3)-(al3) until the processing gate is in the last one approaching the platform. And "plus" invention also provides a kind of correction positioning device = = table and mounted on: the wheel of the table, and more package △. (A2) t it ball plate 'supports many precision balls, in work: 中 if I Γ kit, supporting many sensors, the only point and precision ball on the Spindle®, 20% pre-measurement measurement kit = = among the many precision balls; (a4) the required number of records = J length X change The instruction is to complete step (a3) as the actuator length change data.

Page 10

I. Description of the invention (7), ^: ί Record the actuator length change data at a platform reference position. U5.) Rotate the platform to another position on the ball, and maintain the center of the measurement kit. 2 =, on, (a 6) Record the second actuator length required for s (if used as the actuation H length change data, where the actuator length change data is recorded relative to the horizontal placement; (a7) repeat step (a5): ( =） 位 密 ίΐΠ ;: and U8) Repeat steps "3)-" 6) for many fine settings = Also Ϊ "kind of correction positioning device such as the first item in the scope of patent application, the test position defines the second pivot position Estimated value; (b2) The first estimated value of the position of the fixed work port; (b 3) Simulating the trajectory of the processed product by activating a moving platform; (b4) Judgment step m: the error between the precise features of the mouth; " 5) Value according to the error; ⑽) Repeat steps mb3) _ (b5), the error in the statistically set tenth quarter has been reduced to a minimum, so that the value of many drag shafts is toward an acceptable error. Actual pivot position, r X ^ positioning device.罝 and 歜, to correct this simple description, the drawings are part of the description, which can explain the present example, and explain the present invention better than the above general description and the following best practices. Since, in each figure, the same number ": The detailed description comes from:" Substantiates the same elements' "Figure 1 is a six-footed robot center correction system according to a preferred embodiment of the present invention

Page 11 V. Description of the invention (8) ___ The perspective view of the system; = is a side view of the center of the six-legged robot in Fig. 1; Figs. 3A and 3β are respectively-an unbearable view of a profiler arm extending the actuator foot and mounted on a tool The length of the extendable actuator in the arm; Figure, which can be used to measure the six-footed robot Yin Xin of Figure 2; Figure 4 is a diagram according to the present invention; Figure 1 is a general block diagram of the control system of the correction system of Figure 1; A six-pi side perspective view; an enlarged view of the robot's heart measurement kit and ball board. Figure 6 is a perspective view connected to Figure 2_ ^; a bottom view of the robot's center pivot box's measurement kit 7 is connected to Figure 2 and six views; The top view of the ball board of the center table. The flowchart of FIG. 8 and the flowchart of FIG. 9 are the ball rotation procedures of the people who need them. Detailed block diagram; Figure 1 of a boat 泣 i 旰 湖 力 块 圚, the reaction of the previous figure steps to perform Figure 12 inverse to Figure 14 detailed actuator force gauge head. Vstep is used to execute the figure 1 3 Feedforward calculator for reaction feedforward calculation

Page 12 V. Description of the invention (9) '' ------ Figure 15 shows the signal flow diagram of Figure 5 and the operation of the motion calculator; Figure 16 shows the flow chart steps to collect the machine sensitivity model of Figure 15 Compliance information; 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 forklift system 100 includes a six-legged robot center 11 and is connected to a row of equipment cabinets 2 through power and control cable cables 1 25. This row of cabinets 120 preferably includes: a microcomputer cabinet 丨 30 ′, a servo power control system 丨, and a tool power control system 150. The microcomputer body 130 and the servo power control system 4 40 will be described in detail with reference to FIG. 4. Tool power control system 丨 50 operation = The motor shaft is centered on the rotating machine and is manufactured by Weifang Fangla ^^ hddings & Uwis. A pneumatic spring 1 60 fixed at the center of a six-legged robot " foot 1 70 contains an embodiment of a vibration isolation system, the details of which are described below. It should be noted, however, that the technique of the present invention can be used in a variety of other positioning devices other than the center of a two-legged robot. The preferred embodiment of the robot center 110. Hexapod 2′〇: bottom or lower platform 200, which is driven by an actuator ϊ = Λ Λ tool platform 205. The actuator mechanism 210 preferably includes an extended cup or actuator 215, which is connected to the lower platform 200 at a plurality of low cabinet shaft positions 215a, and is at the armor plate platform 20 5. For example, you can use the heart to set 215b to connect to the κ% silk mechanism to construct a unique cry9. However, the preferred embodiment includes ball snails and ballads, ^ α Although it is a private mechanism, the present invention is not limited to This type of actuator. … 215 can also include pneumatic cylinders, hydraulic red bodies, or racks

Page 13 V. Description of the invention (10) and gear device. 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 a work piece 230. When the actuator mechanism 210 moves the upper platform 2005 along a preset path relative to the lower platform 2000, the tool 225 interacts with the work piece 230. The six-legged robotic center 110 may also contain numerous balances 235 to counteract new gravity. The balance 235 is preferably a nitrogen-filled spring, which is connected at the pivot position 25 ° between the bottom 戋 platform 200 and the upper or movable platform 205. The foot robot center 110 also includes different kinds of vibration isolation elements.

For example, as shown in Fig. 1, the center of the six-legged robot is isolated from the ground by two pneumatic bombs, and each spring is positioned on each side of the three feet of the machine. The pads 245, which extend from three of the lower platform 2000: the female to the ground, all perform the same function, that is, isolate the solid connection between the upper flat σ 205 and the lower platform 20 and the external support structure. The separation element can include the current surplus vibration cry & μ vibration isolation, elasticity;:, damping. ° ㈣’combination of different springs and shock absorbers ^ Ϊ J5A5t38 " 13173 f ^ ^ ^ ^ ^ ^ ^ ^ ^ Note that in the picture: No. 6, 2 and 42 get 'the square that has been incorporated for the reference center *, 7 flat two = heart U °, which is similar to the vertical machine system, then the next flat a combined $$: 205 under. Use without vibration isolation. In this example, the following 夂 交 二 ® the fixed base of this Tiancheng Jiyi., The ancestral temples of the following are not different, that is, it is Miao * Zeng bottom, or relative to the lower platform of the upper platform △, the ceremony a π P bases his dagger on different mobile platforms as described above + 〇 '疋 0. However, π Bayer and in U.S. Patent No. 5,538,373, and

Ⅴ V. Description of the Invention (I "-U2: In various other models of the Robot Center, the lower platform is separated from the seasonal cake and the lower isolation is separated by the vibration isolation system. It is not necessary to treat the vibration partition as a machine. The fixed base is used, and Gushi 'σ% is different for the base. Moreover, at the horizontal robot center ^, the lower platform 200 is not even located below the upper platform 205, but at = branch building !. Therefore, It should be remembered that the name 5§J used in this article not only refers to the specific type of machine center in Figure 2, but ί 疋 means many other positioning equipment and robot center configurations. 位 i Γ 3 ί A t ~ 1 ^ Medium '^ Many measuring transducer monitoring machines such as a transducer are combined with each actuator 215. US special: 5: 38,373 discloses the use of a "solid tool arm, separate from the actuator 215, specifically including a length Measuring transducer. A possible length measurement is shown in Figure ⑽ and ⑽. Laser interferometers are shown in detail later. 测量 The measuring transducer can be configured as a separate tool arm, which is connected to its position. The pivot on the upper / lower platform The shaft or the buildable interior is as follows. ^ 15 In Figure 3A, a tool arm 300 is formed by concentric outer and inner tubes 31 and 315, respectively, so as to be embedded in a beam 325. A laser beam of a laser light source 330 Enter the sealed hollow interior of the tool arm through the window 3 35 and then reflect from the mirror 3 40 to the stem / stepper device 3.45, which is split into 2 beam components. One beam component proceeds to retroreflection over the entire length of the concentric tube It is mounted on the sealed end of the outer tube ^ 〇. Light is reflected downward from the tube toward the interferometer device 345. Then in the interferometer device 345, the two beam components are recombined, and the combined beam components are determined according to Its phase interferes constructively or destructively with each other.

Page 15

I. V. Description of the invention (13) ~~-Therefore, the laser interferometer measuring transducer 3 6 0 can accurately measure the distance between two points, that is, the interferometer device 345 and the retroreflector 3 5 0 are in the ball screw driver 2 1 on 5 positions. This distance can be mathematically converted to determine the positions of the pivots 2 1 5 a and 2 1 5 b. The combined interferometer and actuator are therefore preferred embodiments, so the remainder of the invention will be described using a pivot position, which is a reference to the pivot of the actuator 2 1 5 which is connected to the stick that separates the tool arm The shafts are opposite. However, the present invention is suitable for any kind of measurement configuration, because it is the correction ball operating distance of the actual positioning device to determine the estimated pivot position of the tool arm, and whether it is related to the tool arm and the actuator 2 as described below 5 intersect 'has nothing to do. Reference is now made to FIG. 4 ′, which shows a microcomputer cabinet 13 and a power supply and control system 4O of the calibration system of FIG. The control system 40 includes a microcomputer 41 and includes an input and output device 412, at least one central processing unit 414, a random access memory 416, and a read-only memory 418. The microcomputer 4 1 0 contains one more. For example, the control body 4 2 0 is used to store the control software of the calibration system 4 〇 and the data about the calibration and control of the six-footed robot center Π 0. An operation interface 4 2 5 is connected to the microcomputer 4 1 〇 The operation interface 4 2 5 can include, for example, a display screen,

Keyboard ' printer or other output device, and / or input pointing device. The microcomputer 410 is connected to the processor 430, and the process 1 430 provides a mathematical conversion of the Karst coordinate system instructions to the actuator length. The processor 4 30 is connected to a pair of digital signal processors (D SP s) 4 4 0, and each is connected to a plurality of servo control stages 443 of each actuator 2 1 5. Each control stage 443 includes a servo amplifier 445, a motor 4 5 0 'encoder 4 5 5, an actuator driver 4 6 0 (as shown in the ball screw mechanism of the actuator 2 1 5), and a measuring transducer 4 6 5 (as shown in Figure 3 B laser interferometer measurement

Page 17 V. Description of the invention (14) Converter 360). The feeding amplifier 44 5 supplies the correct power to the motor 45 0, which contains an encoder 4 5 5. The encoder 4 5 5 Saki motor speed signal is sent back to provide a closed speed loop in the actuator control system. The feedback signal from the encoder 4 5 5 is connected to a suitable digital signal processor 440. When the servo amplifier 445 receives a signal to start a new motor 4 50, the motor moves the actuator driver 4 6 0, and extends or retracts the actuator 2 15, so it provides an electrical method to move on the upper platform 2 0. 5. The stiffness converter 4 6 5 (operation as described above) sends the precise change in length of the actuator 2 1 5 back to the digital signal processor 4 4 0, which then transmits this information to the processor 4 3 0 Microcomputer 41 0. In a preferred embodiment, the microcomputer 4 1 0 is a 0 丄 (1 (^ 11 忌 5 & 1 ^ \ ^ 3 company manufactured 0 ^ 8 0 0 0 controller; processor 43 0) —Generally Intel PENTIUM (trademark of Intei) processor; digital signal processor 440 is a 21 81 DSPs type manufactured by Ana log Devices, Nowa, Mass .; servo amplifier 4 4 5 is Kol, Raf, Victoria The BDS type 4 amplifier manufactured by lmorgen Industrial Dr. Ves; the motor 450 is a B604 brushless DC motor of Kollmorgen; and the actuator driver 460 and the laser measurement converter 465 are as described previously in FIG. 3B The details of the software executed on the above hardware will be described with reference to Figs. 12-17. In the normal operation of the six-legged robot center 110, the servo system includes DSPs 4 40 and servo control stages 443, which provide applications Closed feedback loop to position the machine. However, in the corrective rotation (such as the ball rotation described below) order, the information from the correction device 470 (such as the measurement kit described below) is included in

Page 18 V. Description of the invention (15) ------- In the larger feedback loop, the loop contains a digital signal processor 44. Because in the calibration mode, the servo system further includes a processor 43 and a calibration device. Although the above device and computer include the preferred embodiments of the present invention, many skilled alternatives can be used by those skilled in the art. For example, if the dagger function of the processor 4 ^ 〇 'is incorporated into a more powerful microcomputer 4 1 0, the processing of the cry 430 can be omitted. In fact, Giddings & Lewis' CMC 800 0 controller & included an IBM-compatible personal computer to perform many of its functions, especially using the new interface. Although the work performed by each computer or processor shown in FIG. 4 is explained with reference to a special hardware device, the technology of the present invention should not be interpreted as using only these devices. FIG. 5 provides an enlarged side perspective view of a measurement kit 505 ball plate 510 connected to the center of a six-legged robot 110. The measurement kit 505 is connected to the shaft box 2 2 0 ′ on the upper platform 2 05 and the ball plate 5 10 is connected to the work platform 2 4 0 of the lower platform 200. By controlling the movement of the upper platform 2 0 ', the calibration system moves the measuring unit 5 0 5 with respect to the ball 5 1 0 to record the precision actuator information used to calibrate the machine. This calibration procedure is explained in detail below. Reference is made to FIG. 6 ′, which shows a bottom perspective view of a measurement kit 505 in a preferred embodiment of the calibration system. The measurement kit 505 contains many sensors 605 located near a single point in the figure. The sensor 6 0 5 is configured to detect the displacement of the upper flat table 205 relative to the ball on the ball 5 10. FIG. 7 shows a top perspective view of the ball plate 5 1 0. In a preferred embodiment, the ball plate 5 1 0 includes six precision balls 7 0 5 arranged near the center precision ball and located on the hexagonal base 710. Jiao Luo. Although the combination of Figures 6 and 7 shows the

Page 19 V. Description of the invention (16) The sensor 605 can measure seven precision balls 705 arranged in a six-pin pattern, but the present invention is not limited to such a ball plate and sensor configuration. Conversely, the purpose of the ball plate 5 i 〇 is any kind of processed product to be measured, such as a ball plate, a cylinder, a panel, a step meter, a movable ball and a laser interferometer beam in any configuration. In the preferred embodiment, the sensors 605 are conventional linear variable differential converters (LVDTs). Each LVDT produces an electrical output signal that is proportional to the displacement of its movable core. Each sensor 605 has an end portion 61 which makes mechanical contact with a precision ball 705 during calibration. The end portion 61 is mechanically connected to the LVDT movable core and is made of magnetic material. The movable core is located between the primary and secondary coils so that the primary coil is located between the secondary coils. Therefore, the LVDTs convert the body input, such as the movement of the end portion 61, into an output voltage, which corresponds to the displacement starting from the 0 position. For further explanation of LVDTs, please refer to Schaevitz Engineering of Pennsauken, Handbook (Measurement and Control, ΗΒ-84Γ, Congress Book Card No. 76-24971) published by NJ Company in October, 1993. ° θ is mounted on the shaft E 22 0 The sensor 6 0 5 can monitor or measure each process by determining (a) the precise nature of the processed product such as surface characteristics and (b) the offset between the position shift of the stage 2 0 5 Different types of processed products can use different types of sensors 6 0 5. For example, LVDTs can be used to measure the described cylinders, panels, and tool balls, and laser interferometers can be used to monitor the upper platform in a certain section. The linear movement in distance, the distance is equal to the integral number of the wavelength of the light. Moreover, the type of the processed product can determine the type of precision movement.

Page 20 5. Description of the invention (17) Examples include: moving along a line, parallel to a plane ^, moving a precise distance, parallel to a line # point rotation. The movement of the surface, rotation around a line, and around ί The flow chart will now be explained, and the system will be calibrated to find out six two: Ming two how to use it by computer simulation, completed in the present invention: L, precision size. Briefly (1) Performing the ball-to-human positive technology on the machine using a two-stage process: dense movement to make the measurement suite 5.5 Sa: J: 2: ί: human center 1 10 for precision ball 705 movement, Take “3 μ” to control the actuating mechanism 21 to accurately measure the data on a special positioning device around the I, and execute on the computer. The preferred reality in the center of a six-legged robot is: the positions of several pivot positions ma and 215b surrounding the actuator are refined ... as shown in Figure 8 to provide a flowchart; ^ to explain the first part of a pure touchdown process The detailed explanation must first collect measurement data from a series of precision machine movements, saying ^ 'moving to: determine the precise geometry or link size of the six-footed robot center i! 〇'. In the transfer: 2 beast example, the parameter data called the change in actuator length is collected as described below. Starting from step 800, the approximate estimated values of the pivot positions of the actuators 205 are determined at step 805 based on the known values of the machine manufacturing dimensions. In the preferred embodiment, when the machine is the center of a six-legged robot, each of the six actuators 2 and 5 has two pivot positions 215a & 215b, that is, a total of 12 pivot positions. Some manufacturing dimensions are preferably accurate to within ± 1/2 inches of the actual value,

V. Description of the invention (18)-The actual value is final. Defined by the Cartesian pedestal. These measurement pivot positions are measured from absolute or fixed coordinate systems associated with the lower platform 200. An example of the estimated frame axis position of the actuator 21 5 is in the lower pivot position 2] The order of 5a is [17. 7,-38.2'-23.6], and in the upper pivot position 215b 3 [_40 · 9 ' 10 · 2 3, 2 3 · 6]. The unit for this example is inches, but two for any other unit. Note that because the absolute coordinate / system is fixed at the lower |, whether the position of the lower axis is actually moved relative to the ground does not matter. There are 6 points defining the pivot fixed at the base or lower platform 200. The points define fixed at Pivot sleeve on the platform or movable platform 205. And when there are 6 platforms moving, a numerical specification must be assigned to each platform bit. Is there 36: =? Degrees of freedom, so specify 6 unique-coordinates to define the dead position. 3 moving coordinates and 3 rotating coordinates. For the superposition, there is a group of 6 coordinates [^ 2, ^ (:] (where ^: mother one, linear coordinates, and A, B are rotation coordinates) to define the flat / and ^ 疋 6 锢 actuation on Position of the device pivot and 6 upper platform pivot positions' position: J time is also referred to as the platform pivot below, and so there are various displacements relative to the platform with the coordinates [〇, 〇〇- platform. Position in the reference position To define the 6 points of the upper platform pivot, the displacement from the fixed pivot to the above defined upper level 3,5,2,4,6] can be completed as follows: L, Α · Define the upper platform belongs to and The mobile coordinate system and system connected to it quickly so that only on page 22 V. Description of the invention (19) When the upper platform is located at the reference position, the mobile coordinate system will overlap with the fixed absolute coordinate system of the lower platform; B. From the reference Start position 'rotate the upper platform by 6 units around the Z axis of the mobile coordinate system; C. Start from the reference position' rotate the upper platform by 4 units around the gamma axis of the mobile coordinate system; D. Start from the reference position, Rotate the upper platform 2 units around the X axis of the mobile coordinate system; E. Move in the Z direction 5 units on the platform; F. 3 units on the platform in the Y direction; and G. 1 unit on the platform in the X direction. Note 'The above rotation and movement sequence should not be changed in the preferred embodiment. Therefore, the platform The application of such displacement will cause the position of the upper platform pivot. Recall that the coordinates [X, γ, Z, A, B, C] define the upper platform position, and the pivot position is at the reference position [0, 〇, 〇, 〇 , 0, 〇], and the position coordinates [χ, γ, z, A, B, C] are only equal to [0, 〇, 〇, 〇, 〇, 〇] at the reference position. Note that it is important to This reference position can be generated again when the machine has been turned off. Because the actuator 215 is connected between the pivot positions 215a and 215b, the length of each actuator can be approximately estimated through the use of Bishop's theorem. 810, calculate these estimated values of the actuator length at the reference position of the platform, and call them the actuator length of the reference position. From the size of the machine, an approximate estimate of the center of the sensor 605 in the mobile coordinate system can be obtained. An example of such a sensor position estimate is the shift of the platform 2 0 5 [〇, 〇, 〇] in the moving coordinate system. In addition, also in absolute coordinates

Page 23

I. Description of the invention (20) An approximate estimate of the ball position can be obtained. An example of the ball position in the target system is [5, 0 ^] in the absolute platform of the lower platform 2 0 0 in step 820, the microcomputer 41 0 calculates and private information 'to move the upper platform 2 5 to the new position', the required The platform command data 605 made mechanical contact with a precision ball 705. 'In order for the actuator of the measurement kit sensor to feed back data to control the servo step 825, the processor 43 0 moves 2 5 to the corresponding platform command data, so that the actuator includes the upper platform system as part of the control system. , Basin position. As described above, the servo will be further described with reference to FIG. 9 to control the servo 8 for positioning the machine. The following steps refer to 8 3 0, Servo system control output. The output of the ball rotation correction procedure at step is referred to as the data file. This output indicates that, in the preferred embodiment, let s = equal to & device length change data. The center of one of them makes each precision movement of 50 precision spinners and seven precision balls 705. It should be noted that: the formula, that is, the cue ball rotates a total of 350 special applications, the desired precision, the number of force movements is based on the type of machine, although it is selected as% in the preferred embodiment; the type and the type of movement, statistics A sufficient number of moves is sufficient.佾 益 栘 动 仁 疋 八 要 疋 The servo system controller is in step. After the step M data, the microcomputer 410 is released at two degrees, and the current position of the recorded actuator length change 205 is consistent with step 835 in the embodiment. On the platform, the most H position is shown. The precision 50 m sub-bit water platform instruction J material is used. At step 840, the required position of the platform by the microcomputer 410 has the -angle direction. The angle directions of the same-balls are different. Then in step 825, repeat

Page 24 ί The data collection process. If the microcomputer 4 judges the current position table _705 last position of the upper platform 205, then the microcomputer 4 丨 0 then proceeds to step 845 " ^^ $ 密 球 70 5 is the last of the seven balls on the ball 5 1 0 . ^ 疋 9 疋;: month The computer 41 0 transmits the next required right at step 8 50 = 疋 'breaks the upper platform 205 to the next time-precision ball' and repeats the above process at step. If the microcomputer 41 〇 judges that the head ball is on & one ball 51 0, the last of the seven balls, then in step 815, the ball is the ball order. When completed, the actuator length change data, 1 table _ jade, turn, weight positive distance reference position and actuated H length is better than 35 Q positions ^ length is different, that is stored in memory 420. Figure 9 in between provides a detailed description of step 825 of Figure 8, which * is performed and operated in the ball rotation correction routine to be green on the processor. Figures 8 of steps 8 0, 840, and δ50 provide this information. Don't come out. The processor 43 0 is based on the instruction [1,3,5,2,4,6] of the platform instruction information at step 91 and the reference position \ [: 立 〇 置 如 乂 'platform position platform pivot The position (as in step 805 in Fig. 8) of Cheng Jin ,, 0, 0] is estimated as Niu Q1c, and the new Huai axis position. At step 915, the processor 43 uses the command from the platform 205 to the platform to move the length to move the upper 430 to actuate the servo tether. As j asks, location. At step 920, the processor of FIG. 4 digitally sends a new actuator length to 44 °. Each servo amplifier ... drives its "number" to cause each actuator drive shaft 460 to adjust each actuator 215 to the new V. Description of the invention (22) ------ Actuator length. Measuring transducer 4 β 5 Θ Crane · = Capacity I, μ η ,, & Α ^ Feed the poor signal to the digital signal processor 4 4 0 to ensure that the dispenser 2 1 5 is precisely extended by seven degrees. Therefore, the platform 2 5 : Move to flat? / Eight ends return to the new actuator long ϊ: Γ-'Λ Note: Yes Note that the six-footed robot center 110 is still used == ίϊ Therefore, the movement of step 9 0 will not completely change with the processing π ^ test fit, even if it exactly matches the actual length of the actuator 605 calculated by the reference device 430 in the reference position. Therefore the sensor actuates / applies 10 and precision ball 70 5 as a mechanical Contact, when the correct position of the ancient precision ball 705 on the platform 205 and the device 605 is displayed, at least one sensor is displayed for the end part 610. In step 925, this displacement feedback is used as the sensor snowball and I processing. The voltage difference of the processor 430. At step 930, the processor 430 is limited to q whether it is within a predetermined limit. In a preferred embodiment, this is predetermined. ^ qmV 'If the electric difference is not within the predetermined limit', the processor 430 senses that the voltage difference is converted into new platform instruction data, which is designed to minimize the resulting displacement. For example, the platform position [1, 3, 5 ， 2,4,6] By using the new platform instruction of 〇1_12.99,5.01,2,4.6]. This conversion is based on the use of a series of known coordinate systems to move along the sensor axis. Similar conversions of * and 6 Shan Juan Du, Cheng Cheng. This similar conversion is required in the preferred embodiment 'because the sensor's metric axis at 2.5 is not orthogonal. · * Integral = step 940' by The integral and proportional feedback from LVDTs, that is, the use of benefits, Γ, drives Lvs to become completely empty and uses a stable proportional increase = step to rectify the new platform instruction data. Then, the processing type 430 will control the return journey (at 0 'and use the adjusted The platform instruction data is used to repeat the above steps. After each step of steps 91 0 to 94 0, the level is continuously adjusted.

Page 26 V. Description of the invention (23) instruction data until the output voltage of the sensor 605 decreases to the above-mentioned displacement limit. When this occurs in step 930 ', in step 950 and the reference position (such as step 810 in Fig. 8), the current actuator length is removed from the actuator length. Finally, at step 9 5 5, these differences are outputted as actuator length data changes and the control of the server system is completed at step 9 60. f When a statistically sufficient number of actuator length change data is collected, it is best to use the above-mentioned ball rotation correction procedure, that is, to find out the size of the machine and any unknown characteristics of the product. Examples of the latter include the laser interferometer measuring the direction and coordinates of a point on the line, the position and direction of the line defined by the center of the cylinder; the position and direction of the plane defined by the panel; and the center of rotation defined by the tool ball. "Shi Li" ^ "Wang Si" In many robot center applications, it can be expected to use laser-generated lines instead of ball plates as processed products to perform correction rotation. In this example + +, + „Ύ 疋 does not determine the position of the center of the precision ball, but instead builds the laser > 1- 1 λ beam direction, and the laser interferometer measures in several wavelength ranges Moving up = the coordinates of the point. In this example, the corrective rotation may include a laser and a ball rotation. In fact, other types of processed products can be measured during the corrective rotation. The ball 5 is described as soon as the actuation is collected The change data of the device Changxiu, preferably from the calibration program, can be performed by performing simulation analysis on the computer. ^ X. JJL Τγ_ 〇ι + *, six-legged machine. The proposed analysis uses the actuator length change data to repeat the analysis. Override determination 1 0 »1 η Many of the polar axis positions 21 5a and 2151 of the robot center 1 10). Refer to the details of this process in Figures 10 and 11. Analog points 7 Figure 10 provides a flowchart with multiple steps These steps can be performed on the microcomputer 410 or on the microcomputer 410

Page 27 V. Description of the invention (24) '---- Implemented in completely separate and different computer systems. Simulate and analyze a computer's actuator length change data file, by using any large storage device (such as a floppy disk 'CD-R⑽) on the Internet or through any conventional connection to the Internet ( Such as LAN or to transmit. Therefore, the computer system's micro, brain 41 or the above + different computer systems) will be referred to as the platform simulator. In the preferred embodiment, the platform simulator is a personal computer included in the aforementioned CNC 800. Starting from step 1000 in FIG. 10, the platform simulator receives the actuator length change data at step 105. In step 1010, the platform simulator makes an initial estimation of the position of the processed product, such as seven precision balls. These estimates are based on previous displacement knowledge of the precision sphere on the lower platform 200. Refer to the above-mentioned absolute coordinate system fixed to the lower platform to define the ball position. Examples of the estimated values of the seven ball positions in the absolute χ_γ-ζ coordinate system are: [〇, 〇, 〇]; [_8 5 〇〇]. Bu 4.3, -7.5, 0]; [4.3, a 7.5, 〇]; [8 5, 〇., 〇,], [4.3, 7. 5, 0]; [-4. 3, 7. 5, 〇]. In step 1015, the platform simulator makes initial estimates of the pivot positions of the lower platform 215 & and the upper platform 21 5b, respectively. These estimates are derived from knowledge of machine dimensions in manufacturing tolerance limits. An example of the pivot position estimation is described above. > J step 1020, the platform simulator uses the initial position of the ball position and pivot position and the actuator length change data to simulate the platform movement in each precise movement of the ball rotation correction program. Therefore, the simulation was performed 350 times. Figure u shows detailed steps for performing this platform movement simulation and will be explained below. Simulation of platform movement provides 350 positions on platform 205 to microcomputer

Page 28

V. Description of the invention (25) --- 4 1 〇 The main thing is; in the embodiment of Che Fujia, this one corresponds to the position of the precision ball on the seven-down, because the rear position is straight ^ „U two on the ball Rotate the correction process itchy love Η * sensor 6 0 5 is exactly centered in one of the seven precision ball positions, 'a, each of the 350 actuator length change data points. Recorded in step 1 02 5, platform The simulator subtracts the seven estimates in step 1010 from the last 350 ball positions in step 1 020 to get the difference /, and sets the square of these differences to produce a cost, function. In step 丨1 Adjust the ball position and pivot position to reduce the cost function. In the better 30, adjust the estimated ball position and pivot position and use 蚵

Davidon-Fletcher-PowelUD-F-P) non-threaded algorithm error. Although the conventional D_F_p method is used in the preferred embodiment, its speed can be used less, and other minimization techniques such as the steepest descent or slope are used. You can refer to Stanford (3181 ^ 〇1 '(1) University's 1). (; .1 ^ 6111 ^^ 1 / the book Introduction to m Nonlinear Programming, especially the further information on the DFP method in Chapter 9.3. Early In step 1 03 5, the platform simulator uses the adjusted pivot position and ball position and the original actuator length change data to simulate the platform movement again. As mentioned above, the master makes one time at a time, and the ball rotation correction program In each of the 350 actuator data points collected during the period, the platform simulator calculates the movement of the platform. In step 1040, the newly adjusted seven in step 1030 are used. The cost function described in step 1025 is recalculated based on the number of ball positions and the 350 ball positions in step 1035. In step, the current cost function value is used and the previous cost function value is subtracted (that is, at the first time). When the flowchart is executed, the previous cost function is the value calculated in step 105, otherwise it is 5. The invention description (26) is the previous cost function value in step recommendation) to calculate the cost change and divide the current cost function value by The difference is 1. 5 ... The cost change calculated in the device judgment step 1045: 乂: = In the preferred embodiment, the preset limit is 1, 9. If the difference is greater than the pre-right limit, the platform simulator returns to steps 1 0 30 to 105 0. ^ ινόϋ ϊ ϊ The coverage is limited to two: 1 water ° 5J eight: the stage handle simulator determines whether the change in cost 'is less than the last estimated value of the preset polar pivot position and estimates the position of the interpolation after the step, and in the step 1 〇 6 〇 End the simulation analysis. The actual size of these machines is in order to perform the ball rotation correction procedure within an acceptable error range. The interval is to be described in more detail in steps 1 0 2 0 and 1 0 3 5 platform moving wedge die: the device does not know! The platform simulator of the ball position 55 M /-. Does not care about the input pivot position. The platform simulator assumes that it represents a correct machine with the new Π simulator and then uses the actuator length change data to Find: Set [〇, 〇〇〇〇〇〇 For example, the position of the upper platform 2 0 5 before the simulation is the reference position change data for library 2 and ^ after the simulation has occurred 'If the actuator is long [1, 3, 9 4 R Then the upper platform \ 05 will be at an offset position such as completely different and 猸, so Gu Zhan ^ In order to complete the simulation of the platform movement in any computer with the correction system according to the present invention, the simulation step of the platform movement can be performed in step 11 (Γ〇Η). The axis position and estimation 'platform simulator First receive a set of polar pivot positions to calculate the position in step 1105. In step six, the robot center 1100 uses a different-area car from the platform position (ie, the reference position at [0, 0, 0, 0, 0, 0, 0]).

Page 30 V. Description of Invention (27)

Position) and add the actuator length # X ^,, ^, ”X data to the desired special platform position. The result is the actual actuator length of the displaced platform position. Step 6: The center of the six-footed robot is 1115, and the estimated value of the new platform position is determined or received at 1 hour. For example, the reference position ^, 办, and an acceptable initial estimate of the platform position. At step 11 2 0, ^ may be updated ^ SL ·. Js,, «f ^ Thousands of simulators convert thousands of reference positions (and upper platform reference pivot positions) into estimated table positions. In step 1125, Bishop's theorem is used to 'calculate each pair of lower levels △ The distance between the frame axis of the actuator and the pivot of the upper platform actuator. The result is the estimated actuator length of the new platform position oblique = ^ =., Center °: Calculate the actuator length error at step 11 3 0 ' The sum of the squared differences, which is the difference between the estimated actuator length in step 1125 and the actual actuator length in step 111. At step ^ 35, the platform simulator adjusts the estimated $ platform position to reduce steps j 丨 3〇 Actuator length as defined in this adjustment. The known Newtonian Lassen method is performed in the preferred embodiment, although $ can use any algorithm designed to minimize the value of the multivariate function. See Johan Wiley, P. Henrici, 1964 Basics of Numerical Analysis by Author (Congress Book Card No. 64_2384〇) 'Especially Chapter 5, further information on Newton's Lasheng method. In step 11 4 0, the platform simulator will adjust the adjustment amount required in step 1 1 3 5 with Compared with the preset limit. If the adjustment is greater than the preset limit, the platform simulator returns to step 1 11 5 and repeats the steps of the six-footed robot center 1 11 5 to I 1 4 0. Through the upper platform position or the estimated new position, Repeat the adjustment, and the adjustment required in step II 40 will eventually become smaller and smaller. In theory, it can be as small as the largest number of known computers that perform simulations, because when estimating

Page 31

I. Explanation of the invention (28) When the actuator length approaches the actual actuator length found in the six-footed robot center 1 1 0, the change in the -command position will eventually approach 0. In a preferred embodiment, the preset adjustment limit used in step 1 1 40 is defined as the absolute value of the change in any / coordinate X, Y, Z, A, B, c must be less than 1.0 〇 E ~ 6. If the adjustment meets the limit in step 1140, the platform simulator outputs the command position as the position of the upper platform in step 11 4 5. For example, if the command position is [〇. 01, -0.2, 0, 2, 〇: 〇] when continuously passing through the flowchart, the platform simulator will output the position as the position of the upper platform 2 05, or as above The position of one of the precision balls 705 is described. The simulation of the platform movement is completed at step 1105. It is important to note that the present invention is not limited to this preferred embodiment of simulation analysis. Instead, any simulation can be used that uses similar measurements to actuate and estimate pivot positions to find new machine positions. . J 的 1's upper J4 program 'When the actual machine executes-special processing σ 疋 疋, ore, column precision fork movement, you can correct the six-legged robot by collecting data on actuator length changes, ... Snap buckle has not been repeated on the simulation computer to find out "this Beaiya is used in the out-of-the-way position of a positioning device provided by Ming. Therefore, the position of the pivot of this hairpin. Zheng Zhengyao: surprise: movable The platform's extremely precise connection can be performed on the right cup. You can make a charge-set correction movement. The receiver can then control the correction field periodically on any remote control computer. . At this point, you can now call the Jiren Center to confirm i. Therefore, the calibration technology of the present invention can be accurate without long length. At high cost, it can ensure the extremely high diameter and precise positioning device. He knows the calibration system as above. As shown in FIG. 4, the microcomputer 41 controls the human center so that

P. 32 j

Five 'Description of the Invention (29) The platform 2 0 5 moves along a preset path. The microcomputer 4 10 moves the platform by transmitting platform instruction data in the form of a new desired position. For example, if Shangping & is currently at position [0, 0, 0, 0, 0, 0], the preset path will require moving to the new position [1, 2, 3, 5, 5, 5] of Q. The platform command data can also be [!, 2 3 5, 5, 5] * ′, where the asterisk * indicates that the position is used as a command or command to locate the dream to move it to that position, and is opposite to the current position itself. This flat two: The data must be converted into "to cause the actuator to move" by referring to the hardware described in Figs. 4 and 12 to 17. In the following Figures 12 to 17, how to translate the plain command data into a precise actuator movement is explained in detail. Referring to FIG. 12, the microcomputer 410 transmits platform instruction data [also, ya, 2, 8, 8, (:] * to the trajectory generator 1205, which is preferably inside the microcomputer 41 {). It should be noted that the instructions [X, γ, ζ, Α, β, c] are relative to the lower platform execution line generator 1 205. This Karst machine platform instruction data is converted into command bits f, speed and acceleration (p, v , A) a continuous string and output all three. In the preferred embodiment, this is done using an S curve:

v = dp / dt = 3.An2, 9 D

Where p (t) is the temporal velocity of the time function, a (t) is the position of the time function ^ 蚪, and the instantaneous description of V (t) 述 the time function is calculated for each P, v & a table number of instants Acceleration, and t is time. Above Z) and an angle u, B, C) singular consists of 2 vectors, a straight line α, γ, song. The coefficients Λ, B, c, and β are produced by the trajectory

Page 33 V. Description of the invention (30) The generator calculates ’in a certain way, that is, it does not exceed the limits of jitter, acceleration, and speed. -The line generator generates at least one S-curve, and its continuous performance image is used for the position [fork, y, 2, 8, 1, (^] * input generator 1205. Another method for calculating instantaneous speed and acceleration Is the difference between the first and second values of the position command string p (t).

Referring again to FIG. 12, the trajectory generator 1205 outputs the commanded relative position, speed, and acceleration to the reaction feedforward calculator 12 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 215 to generate the trajectory generator 1205. Relative acceleration. The steps required to perform this converter will be further explained with reference to FIG. 13. The reaction feedforward calculator 1210 outputs the actuator force F * to the motion calculator 1215 and the servo feedback system 1225. The motion calculator 1215 uses the commanded position and the actuator force F * to determine the commanded actuator length, which includes an error correction caused by the elastic distortion of the six-footed robot center 110, which is caused by the actuator force F * and Other solvency is generated. The routine signal flow diagram required to execute the command actuator length will be further explained with reference to FIG. 15. The commanded actuator length L * is output to the servo return system 1 225 and to the differentiator 1220. The differentiator 1220 provides the rate of change of the actuator length L (that is, the differential of dL / dt length L with respect to time) to the servo feedback system 2 2 5. The servo feedback system 1 225 provides the command torque to the control stage 443 of each actuator 215 as shown in FIG. 4. The control stage 443 includes a servo amplifier 445, a motor 4 50, an encoder 4 55, an actuator drive shaft 4 6.0, and a measurement transducer 4 65. Each control stage 443 provides feedback from encoder 4 55 and measuring transducer 465 to servo feedback

Page 34 V. Description of the invention (31) 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 * . Preferably, the reaction feedforward calculator 1210, the motion calculator 1215, and the differentiator 1220 are mounted on an Intel PENTIUM CPU 430. The servo feedback system 1 225 is mounted on the digital signal processor 440 as shown in FIG. 4. Referring now to FIG. 13, there is shown a general flowchart step for performing a reaction feedforward calculation performed by the reaction feedforward calculator 1210. After step 1300, the command relative position, speed, and acceleration of the wire execution generator 1205 are received in step 1305. At step 1310, the commanded relative position velocity, acceleration, and lower platform force are used to calculate the required actuator force to provide the desired relative acceleration. This calculation can be done by solving the system with 6 formulas for 6 actuator forces F at the same time. For example, in a preferred embodiment, the formula system is expressed in a matrix form: ^ ^ CPmat · F = B (1) where CPmat is a coupling matrix with: above) 6 rows of columns are filled with actuators 2 1 5 The unit vector and the 6 rows in the 3 columns below are filled with the relative unit torque to inertia ratio, where the unit torque is the moment of the actuator unit vector relative to the gravity of the platform :. Calculating the product of the foot unit vector and the center of gravity from the center of the actuator to the actuator's pivot can be done. In the formula (1), F is a vector containing 6 actuator forces.篁 (3 columns above) in the formula for the linear system is the force. The quantities in the angle system formula (3 columns below) are

Phase angle plus confusion. Therefore, the first three elements of the 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, γ, and z, which is calculated as the following formula:

Mu / (Mu + Μ1) • (Ml .a + sum — foot) (2) where Mu is the mass of the upper platform, M1 is the mass of the lower platform, and & is the upper platform relative to the lower platform in step 1 305 Commanded acceleration. Sum_f07 is the sum of all the linear forces applied to the lower platform by the 'vibration isolation component' in step 1325. The next 3 elements of the B vector are the relative angular acceleration of the instruction from step 1 305. By first determining the value of the facet matrix CPmat '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) at step 1310 to find the commanded actuator force F *, the actuator force is output from the reaction feedforward calculator 1210. 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 force equals mass times acceleration,

Page 36 V. Description of the invention (33) —-—-- All the forces on the lower platform are divided by the mass of the following platform to obtain the linear acceleration. 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 was integrated to obtain a speed of 200 on the lower platform. The speed can be re-integrated to obtain the position of the lower platform 200. The new position, speed, and acceleration of the platform 20 at the step 1 32 5 1 are used to calculate the position and distance of the vibration isolation component 2 45, which can be used when combined with the hard and vibration isolation characteristics of the isolation component. Calculate the force exerted by the isolating component on the lower platform. The force exerted by the vibration isolation component on the lower platform 200, and the new position, speed, and acceleration of the lower platform 200 are used again in step 1310 to determine the B vector 'and repeat the above process. Referring now to FIG. 14, the calculation of the actuator force in step 1310 is described starting with step 1400 in the required steps. As described above, the coupling matrix CPmat and the B vector must be established to obtain the actuator force vector F in the formula. In step 1405, the three PENTIUM processors 430 calculate the moment arms of the upper platform 205 and the lower platform 200. The moment arm is a vector by subtracting the center of the gravitational position from the position of the polar axis on the actuator connected to the Z axis position. In step 1410, the PENTIUM processor 430 calculates a unit torque vector of the upper platform 200 and the lower platform 200. The unit torque vector is equal to the cross product of the unit vector of the actuator 215 and the relevant moment arm. At step 1415, the PENTIUM processor 440 obtains the relevant angular acceleration from the thread generator 1 2 05. At step 1420, the coupling matrix is populated 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:

Page 37 V. Description of the invention (34) (iu [. (RT.Tugn))-RT. (Rlc. (Lli. (RlcT * Tlgn))), ⑷ where this vector is used 6 times (one for each actuator ) In order to fill the rows in the lower 3 columns of the transformation matrix CPmat with 6 vectors. The term lui is the inverse of the inertia matrix of the upper platform, replacing the input-matrix in the denominator of Equation 4. A similar Π i is the inverse of the 2 0 0 inertia matrix of the lower platform. The term Tugn represents the 6 vectors of the upper platform 205 to which each of the 6 actuators belongs. A similar 'term Tlgn represents 6 unit torque vectors of the lower platform 2000 to which each of the 6 actuators belongs. The rotation matrices r and Me of the upper and lower platforms respectively describe the rotation of the platform relative to the overall coordinate system. An example of the rotation matrix R is: sb] -sa * cb] ca * cb] [cb * cc -cb »sc (ca * sc + sa * sb * cc ca * cc-saesb * sc [sa * sc-ca ^ sb * cc sa «cc + caesb * sc where: sa = sin (A) ca = cos (A) sb = sin (B) cb = cos (B) sc = sin (C) cc = cos (C) appears in The R and Rlc transpose matrices of formula (4) can be used to express the quantities in the same coordinate system. At step 1425 'PENTIUM processing 1 §430 calculates the three values above the g vector according to formula (2) . Use the following formula to fill the bottom 3 values of the B vector: (5) maccc_ang + R 丁 • Rlc.lli.RlcT.foot—ang—tot

Page 38 V. Invention description (35) where maccc_ang is the relative angular acceleration of the instruction from the generator 1205, R and Rlc are 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. At step 1435, the digital signal processor 440 establishes a b-vector based on the above value. At step 1 440, the PENTIUM processor 430 inverts the coupling matrix CPmat and multiplies it by the B vector to find the actuator force vector F. Then, the command actuator force F * (or the force vector ί0) of each actuator 215 to the motion calculator 1215 ′ of FIG. 12 is ended at step 1 445. Referring to FIG. 15 ′, the motion calculator of FIG. 12 is explained. The signal flow chart of 1215. What is input to the motion calculator 1215 is the command position vector P (which contains 6 position commands of X ', Z, A, B, and C) from the trajectory generator and the actuation. The force F *. The motion calculator 1215 uses the rotation command position data [A, B, C] in the rotation matrix generator 1505 to generate a rotation matrix r. The position constant of the upper axis of the light axis (which is shown in step 1 of FIG. 10) The pivot position vector determined in 05 5) is added to the adder 1 52 5 and input to the multiplier 1510 to rotate the pivot vector by R. The motion calculator 1215 then performs the linear command position of the line generator 12 05 Data [X, Y, Z] is added to the rotary pivot position movement calculator 1215 in the adder 1515 using the actuator force and pivot red am two tassel kung fu knife r winter criminal shaft position calculator 1520 actuator The early vector is used to calculate the upper platform 205 and the lower displacement. These pivot displacements are due to the machine platform α is the actuator force "and inertial force. These distortions are referred to as self-deflection. This is due to the consideration of Fig. 16 to illustrate the collection of this compliance asset: 妩 compliance. Steps taken by the householder. Will be on the platform Pivot compliance model The sub-table pivot position constant. Similar to the upper foot upper ten mouth processing in the above-mentioned upper flat controller 15 25, the lower platform pivot

Page 39 5. Description of the invention (36) The position constant, which means that the pivot position vector from 1 05 5 is added to the lower platform pivot displacement in the adder 1530. Now send the upper and lower platform pivot position vectors to the actuator length calculator 1 540, which uses Pyth's theorem to calculate the actuation: vector = small. This size is the actuator length. The motion calculator η "also uses the pure universal joint compliance calculator 1 545 to calculate the scalar component represented by the rigidity of the actuator joint and the balance ring compliance. This joint is shown as the universal joints 380 and 382 of FIG. Type 11 connector & actuator i 550's actuator length Λ head / fruit 'into order ㈣ n long ㈣. Fruit ^ motion calculator 1215 output information ί i: Figure 1 ΐ [Beishu pivot deflection calculation Flow chart of the steps taken by the device 1 520 to collect the compliance model. For example, the accuracy of the two-axis human center 110 is determined by the microcomputer 41 ° and the accuracy of 215 '. It is also slightly moved as described above. The frequency of the six actuator forces (extended tension package inflammation and white: "extensible actuators (addition of compression and dispersion mass =) degrees") is also frequently reduced. The three balancing forces from the balance 235 start at step 16 00 in Figure 16 and are executed on the software software and constructed on the flat A work station. CAEDS finite element software is a structural dynamics research company; Step 16 0 5), the shaft E 220 of 205 is fixed to the product 7 in the ground. In the model, the upper platform is reduced in curvature It is very small. This is limited because the purpose is to make the deflection force or record displacement of this component have :: ;; ~ type = many nodes, so that you can apply force here. In the definition of several force nodes, the force is applied at 6 pivot positions 21 5 b in σ; page 5. Fifth, the description of the invention (37) and the force can be applied at the balance yoke position 250 of the upper platform 205. The 6 displacement nodes in the 3 Karst curves include the actuator pivot position 21 5b in the upper platform 2 05. The finite element model also provides deflections in the displacement nodes in response to X, Y, and Z The acceleration of the upper platform in the direction. In step 1610 'apply a unit acceleration to the upper platform 205 in one of the three Karz directions X, γ, z. The deflection response of the actuator pivot 215b on the upper platform 205 continues at step Recorded in 1 6 1 5 where each Kal's component is recorded in each of the 6 pivots 215b. Step 1620 causes the process to repeat at steps 615 and 1615 to accelerate independently in each Kal's direction. Record After the accelerated response from the upper platform component, a force application point is selected in step 1625 and a The potential force is applied at the selected application point in the single-direction X, Y, and Z (step 1630). The deflection response of the actuator polar axis 215b on the upper platform or the tug movement is recorded in step 1 63 5 , Where each Kelvin component is recorded in each of the 6 actuator pivots in the upper platform 205. The steps make the process repeat steps 1630 and 1635 by applying unit forces independently in each Kelvin direction and still at selected nodes Step 1 6 4 5 causes all nodes to perform steps 1 62 5 to 1 6 40, where a unit force is applied in the node ... 6 'actuator pivots 215b and 3 balance pivots 2 50. The analysis of the lower platform 200 starts at step 1650 '. At step 1655, the above-mentioned IBM workstation is executed on the CAEDS finite element software to construct a finite element model of the lower platform 200. The model of the worktable 240 of the lower platform 200 is fixed, because the purpose is to minimize the deflection of this component. A finite element model like a platform above and below contains numerous nodes to force or record displacements here. In the preferred embodiment, several force nodes are defined so that

Page 41 V. Description of the invention (38) The force is applied here, and the force can be applied at 6 pivot positions 2 丨 5a in the lower platform. The displacement nodes in the three Kago components of the deflection were recorded, including the actuator pivot position 2 15 a in the lower platform 2000. After sighing the finite element model of the platform 200, the process proceeds to steps 1 625 to 1 645, that is, apply a unit force at a point of force in the single-Karst direction, and record all actuators in the platform 2000 The pivotal element 21 5a finally flexes the Karst component. Finally, when all the pivot positions / movements have been recorded, the process ends at step 1660. Note that the response deflection of the force is calculated using the finite element model. When implemented in a finite element model, the actual σ, σ, and 〇 can be consolidated. The force component can be applied to the force node physically, and the displacement response of the displacement node can be measured physically. Axis = ί d / T! News can be used at the center of the six-footed robot center 11 〇 multiply leaves i out of the child. The actual Karst component type of the force applied to the force node is in response to the corresponding finite element model deflection in response to the actual Karst component of the speed of the unit piece applied to the force node multiplied by the calculated axis of the displacement node. , Accelerate with the adder pivot unit. The principle of overlap allows the forces or accelerations to be added to the in-situ platform 205 to add displacement nodes (ie, the actuator pivots) to add, which is very important in the appendix "The mathematical formation of the platform The stacking f is here for reference. C The matrix contains the calculated finite plant wall and the unit force of Appendix A is best determined according to Figure 6. Respond to ^ ^ * t ^ ^ # ^ ^ Block 13 Page 42 ί 5. Description of the system (39). The generated P vector contains the output of the desired axis deflection calculator * 1 5 20. In the phase, between " displacement ', it is pivoted in Figure 15 to the reference position [ 0, 0, 0, 0, 〇, 〇] This displacement is provided in the calibration / crypton system. The actuator drag axis coordinates are defined. The actuator pivot coordinates in the τ position will be suppressed. The > main idea The above is the number of nodes and the number of $ 4. The number of fields is only used for the preferred embodiment. The present invention is not limited to this purpose. On the contrary, Xi Xilin Congshi agent soil-r, on the other hand, is hot and limited The component model can understand that the variations of Figure 16 can also be used to provide any statistically sufficient number of elastic distortions to accurately construct a six-legged robot center11. Compliance model. 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 graph of the servo feedback system of FIG. 12 2 servo. The input of the feedback system 1225 is the command actuator force, the command actuator length L * and the differential dL * / dt of the command actuator length. The digital signal processor 440 sends the command actuator force F * to the unit converter. 1 7 05 in order to convert the force into torque. This can be done by multiplying the commanded actuator force F * by the lead of the above platform 205 and then dividing by 2 * pi. This is the first of the three torques added in the adder 1710 One to find the output command torque T *. The digital signal processor 4 40 sends the commanded actuator length L * to the adder 1715 to subtract the actuator length L, which is fed back from the measuring transducer 4 6 5 The length error of the actuator is fed through a proportional-integral-derivative (PID) 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. Then 3 torques are added Adder 1720, and the result is 3 added to adder 1710 Torque second false torque integral gain Ki is used as a low-frequency enhancement, or hard to strengthen, while the differential gain Kd is used to increase the damping stable

Page 43 Explanation (40) V. Invention; Workstation. It also sends the commanded actuator length to the converter lug, which converts the commanded actuator length L * into the commanded angular position θ * of the actuator screw. The commanded angular position β * is then subtracted from the angular position 0 of the encoder 4 4 5 feedback signal by a subtractor 1 73 0 to calculate an angular error. A proportional gain Kρ and differential gains s and Kd are then given an angular error for a reason similar to that described above. The generated torque is then added to the adder 7 3 5 and the result is that the angle is the sum of the differential torques. The sum of the angular error moments is sent to another adder 1740. The digital signal processor 440 also inputs the differential actuator length dL " d1T to the conversion piping 1 74 5 to convert it to an angular velocity. Then multiply the angular velocity by the gain to overcome the viscous friction between the motor 450 and the ball screw, and add the resulting error to the adder 174 0. The angular velocity is also differentiated by 3 to obtain the angular force of 0, which is multiplied by the inertia of the motor and the actuator screw, and the torque generated will be

Moment addition adder 1 740 〇 Adder 1 740's last known extension a 丄 AO + set the torque 疋 3 added: to complete the command torque ”to the feeding amplifier 445. Front Ϊ: Wide " monthly compliance compensation The technology uses the reaction to feed the clearing force data generated by 2 10. However, the data uses the same compliance and sexual compensation steps. ° Heart /, other materials can be used for this purpose, but it is generally a torque back : It is difficult to use it. "隹 ###, and will make the reference to 'the following Appendix B description-1210 detailed operation formula' and Appendix C description ^ The two appendixes attached to the calculator are in This is for reference. Differential in δ direction. Although the specific < money.% Example of the present invention has been shown and explained here, it is familiar with this technology

Page 44

J V. Invention Description (41) The surgeon can still make further amendments 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.

Page 45 V. Description of the Invention (42)

Appendix A Pivot Displacement Compensation of Various Upper Platforms The pivot displacement vector can be calculated by the following formula: P5-=. Cl- (lT-Fu) a where

Cl = matrix of 27 compliance matrices of the upper space box R = rotation matrix of the upper platform

Fu = 6 vectors of foot force vectors, and 3 equilibrium force vectors a, b, c P = 6 pivot displacement vectors C1 are of the form: cltf cu qc ^ qs q, ^ ^ q .a ^-C5 * ° 36 C34. Sa ^ C46 ^ -4 * CJ1 Ca. C23 C54 CJ5 C5tf CJ * C5> Cjfe. Jiang Ca C "Ca •.

Each element in Cl is singular from the hardness matrix. For example, element C23 is the compliance of pivot 2 in response to the force on foot 3. These compliance matrices are of the form:

Page 46

I V. Description of the invention (43) Its ζζ. ^ Ά3Τ- I15 5L5, IA ·-Ύ C ,, 丨 " 素 · 素 AxMfx is a deflection in the 乂 direction in response to a change in the force in the X direction Ή4- > τ >. Each element is a three-dimensional foot impulse vector. > JZ. The form of the foot force vector is, for example, F1 is a foot.] The force is determined by the position W1, such as ψΐ F2 ^ The form of the shift vector ^ is: (? «? 3?, P /? 5),

F

FJ

T a F〇 = force vector

Pi only area car. • Weidan music vector ί 1 '.;-..

Rotating system: ..L ·.. ·. ^^ 离 leak "

^ The action formula of the reaction feedforward calculator fp scalar power output foot force ep = power foot force unit g amount:

Fs = Pneumatic bomb meal: Force vector Mass of upper platform Μι = Mass of lower platform! FF1 = Force of mounting foot | f date

Iv = Integration of the upper platform moment of inertia self-breaking value and lower platform) IUi ~ € abc = 0, wider than γ v 7 R = rotation matrix to spin ^ up ^, z's upper platform inertia matrix ίui: 来 t ΐ (: j , The angle is constant) V, the power foot from ^ on Cg],; == the power foot from the CU's torque port I reu = upper from cg = aerodynamic spring on the stage ln = from Cg The pneumatic spring in the H platform is straight: the line is far away ". 3 V acceleration, which is almost parallel to the overall system V. Description of the invention (45) 11: pr sr; ^ fJLiLLi: " t ^ w — ΞΜ •-* · 1- /, -1, ml + fc 10 ″ rn'-lfl- + f both a:; ttfvll r * l {mll * cl-· Η1Η1. · ×: 丨 lnl ^- -*-* c ^ E · V s-^^ vf 5 4 fe ^-zi = cl ml wr * f 丨 6''ΡΓ Ϊ au -tol S: au 厶: ol * 丨

T? FH wu 31 1: ml al "-丨 κΐ 丨 Zuo 2: / is X f-wu 3: solveroach for aT '%

Ze CB3ht s Xue Jin (KHS) wlllJ ^ + ab, sll-twrni ^^ s ^ N® / i = Mlil_L-^^ s1f4 Ac Page 49 V. Description of the invention (46) -M σ \ ί l. F ^-p 「-t μ 15; fl 丨 14: fl n 13: fl ff l:-p» r IPI / fllt-tlf ±-fg \ (ml tlt ft ../ 0 ml + μ ml:- & (-tuM.T: -H) af t M3) tT: (HIl-tl- nlQ- 芑 3) 罘 百 a-^ 1¾-fg-s ·· ηα

Jnl-lHtt-h- fg-wl)-ml {fg-yiu), 12; .fl-^ e · ^ '-. Ml + / i. C 3— n-rf] Two 7.5:? /

* hl-rA τμ-t--tn · — M. jlr-5-:

Page 50

Claims (1)

6. Scope of Patent Application 1. A method for calibrating a positioning device, the device has a movable platform connected to a substrate through a plurality of actuators in a plurality of pivot positions, and the method includes the following steps: (a) obtaining calibration data To instruct the movable platform to move relative to a series of processed products; and (b) by simulating the movement of the movable platform relative to the processed products, using the calibration data to perform a computer simulation analysis to determine the Numerous box axis positions. 2. The method for calibrating a positioning device as described in the first item of the patent application, wherein step (a) further includes the following steps: (a 1) estimating the geometric position of the positioning device at a reference position; (a 2) placing a processed product Near the platform; (a 3) Install a sensor on the platform, which is designed to indicate the offset between the movement of the platform and the precision features of the processed products; U 4) Obtain the required platform instruction data to move the platform to a position, feel The detector starts to measure the precise characteristics of the processed product; (a 5) Move the platform according to the platform instruction data;, (a 6) Obtain the offset described in step (a 3) from the sensor; (a 7) get a judgment (A 8) Move the platform according to the platform instruction data; U9) Repeat steps (a6)-(a8) until the sensor offset falls within a pre-set Set the range; (a 1 0) Set the geometric position of step (a 1) according to the actuator length change data. I VI. The changes in the scope of the patent application are output to the geometric position of the positioning device that satisfies step (a 9); (a 1 1) When the platform is not in the final position on the processed product, obtain the required platform instruction data to move the platform To a different position, the sensor starts measuring the precise characteristics of the processed product; (al2) Repeat steps (a6)-(all) until the platform is in the final position; (a 1 3) when the processed product is not close to the last of the platform In the position, the processed product is placed near the platform, and the precise characteristics of the processed product define the movement of the platform, which is different from the measurer in the previous steps (a 2)-(a 1 2); (al4) Repeat step (a3) -(al3) until the processed product is near the last of the platform. 3. The method for calibrating the positioning device as described in the second item of the patent application, wherein the offset is a position offset with a zero error defined by the sensor. 4. The method for calibrating a positioning device according to item 1 of the scope of patent application, wherein the positioning device further includes a workbench mounted on the base and a shaft box mounted on the platform, and step (a) further includes the following steps: (a 1) Install a ball board to support many precision balls on the workbench; (a 2) Install a measurement kit to support many sensors. In the shaft box, many sensors are designed to detect the unique points on the measurement kit. The coincidence of the center of the precision ball is called centering; (a3) the mobile platform 俾 measurement kit is centered on one of the many precision balls; (a 4) record the required first actuator length change instruction to complete the step ( a3) Actuator length change data, which records actuator length change data relative to a platform reference position; (a 5) Rotate the platform to another position on the ball, maintain the measurement kit, and center it Page 52 6. The scope of the patent application is on the ball; (a 6) Record the required second actuator length change instruction to complete step (a5) as the actuator length change data, where relative to the platform reference position Record actuator length change data; (a 7) Repeat steps (a 5)-(a 6) multiple times; and (a8) Repeat steps (a3)-(a6) for the other of many precision balls . 5. The method for calibrating the positioning device as described in the first item of the patent application scope, wherein the precise movements are defined by the platform tracking the precise characteristics of the processed product. 6. The method for calibrating the positioning device according to the first patent application range, wherein step (b) further includes the following steps: (b 1) defines a second estimated value of a plurality of pivot positions at the reference position of the platform; (b2) defines the processing The first estimated value of the product position; (b 3) Use the actuator length change data to simulate the trajectory of the processed product by moving the platform; (b 4) Simulate the movement and processed product precision characteristics performed in the determination step (b 3) (B5) update the second estimated value of the position of the processed product and the first estimated values of many polar axis positions according to the error; and (b6) repeat steps (b3)-(b5) multiple times to make a Let the error in the range be minimized, so that the second estimated values of the plurality of Z axis positions converge towards the actual Z axis position within an acceptable error, thereby correcting the positioning device. 7. For the method in the first scope of the patent application, where the correction data is the department P.53 6. Scope of patent application List the change in length of each actuator during movement. 8. The method of claim 7 in which the positioning device is a six-legged robot center. 9. A method for calibrating a six-foot robot center, the center having a base and a platform each having a plurality of pivots, and the six-foot robot center also having a plurality of extendable feet, each foot being connected to one of the many pivots of the platform Between one and one of the many pivots of the base, the method includes the following steps: (a) define and define estimates of the positions of many polar axes in the base; V (b) define the mobile platform when the platform is in a reference position Estimates of the multiple Z axis positions; (c) When the platform is at the reference position, calculate the actuator length from the estimates described in steps (a) and (b); (d) Place a precision ball near the platform; (e) Install a sensor on the platform, which is designed to indicate the distance between the unique point on the platform and the center of the precision ball; (f) Get the required platform instruction data to move The platform reaches a position, the sensor starts measuring the distance described in step (e);. (G) moves the platform according to the platform instruction data; (h) obtains the distance described in step (e) from the sensor (I) get the determined platform instruction data to move the platform to another location, and reduce the distance described in step (e); (j) move the platform according to the platform instruction data; (k) repeat step (h) -(j) until the distance described in step (e) falls within a preset range; Page 54 6. Scope of patent application (1) When the platform is in the reference position described in step (C), the change in actuator length is output according to the actuator length change data to when the platform meets step (k) (M) when the platform is not in the final position on the precision ball, obtain the required platform instruction data to move the platform to a different position, and the sensor starts measuring the distance described in step (e); ( η) Repeat steps (g)-(m) until the final position of the platform on the precision ball; (〇) When the platform is not on the last precision ball, place a precision ball in a new position near the platform; (P) Repeat Repeat steps (f)-(o) until the platform is on the final precision ball; (q) Determine a sufficiently accurate estimate of the actual positions of the numerous pivots, which are determined by: (Q 1) estimating the position of the precision ball; (q 2) Use the platform instruction data of step (1) to simulate the platform movement to generate the simulated platform position with the sensor centered on the precision ball; (q 3) Calculate the distance between the simulated platform position and the estimated position of the precision ball Error, (q 4) adjust the estimated position of the precision ball With the actuator pivot position to reduce the error described in step (q3); (q5) repeat steps (q2)-(q4) until the error described in step (q3) falls within a preset range; ( q6) Output the last pivot position generated in step (q5), where the last pivot position represents a sufficiently accurate estimate of the actual Z axis position. Page 55 6. Scope of patent application 1 0. A system for calibrating a positioning device, the system includes: a measurement kit with a plurality of sensors connected to the positioning device; a processed product defined on the positioning device Excavation in the working area; a first computer connected to the positioning device for controlling and recording the precise movement of the positioning device relative to the processed product; and a second computer for using the recorded precise movement as input data, The simulation positioning device is used to repeatedly determine the actual geometric position of the positioning device. . '. 1 1. The system for calibrating the positioning device according to item 10 of the patent application, wherein the first and second computers are a computer. Page 56