JP2004326732A - Method for identifying error parameter of simulation robot - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain an optimum error parameter used for a simulation robot and to identify its value . <P>SOLUTION: A tuning path candidate consisting of a plurality of representative attitude of the simulation robot is created, the tuning path candidate is applied to each robot mechanism model for a standard robot and the simulation robot, and an error between the two models is operated. An error parameter necessary so as to reduce the error and its value are operated and, while both of them are being considered, an evaluation path consisting of a plurality of attitudes different from the tuning path candidate for the simulation robot is applied to the respective robot mechanism models of the standard robot and the simulation robot, and an error between the two models is evaluated. When the error parameter is minutely fluctuated to the evaluation path attitude whose error is evaluated to be large, the error parameter and the tuning path attitude which give a great influence on movement of the robot mechanism model for the simulation robot are selected, a selected error parameter is added, and a processing from a stage at which the error is operated to the selected tuning path attitude to a stage at which a tuning path attitude is selected is repeated. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method for identifying an error parameter used for a simulation robot in order to correct an error between a real robot and a simulation robot.

  In recent years, the operation confirmation of a robot has been often performed by simulation. The simulation displays a three-dimensional model of the robot and peripheral equipment including workpieces and jigs on a graphic display, examines their arrangement, and examines the workability, reachability, existence of interference, and cycle time of the robot. It is performed to examine the feasibility of the equipment from the viewpoint of and to create a robot operation program.

  In the simulation, since the robots and peripheral equipment arranged on the display are all created based on CAD data which is design data, they do not always match the robots and peripheral equipment actually arranged on the line. For example, the installation position of the actual robot may not match the installation position of the robot on the display, and similarly, the tool shape of the robot and the installation position and shape of the peripheral equipment may not match. Further, in a small portion, an error may occur in a teaching position at the time of simulation due to a shift of a zero point position of an encoder for detecting a position of a robot arm or an effect of bending of an arm at the time of gripping a work.

From the above, the motion of the robot operated by the simulation and the motion of the actual robot slightly differ. For this reason, conventionally, as shown in Patent Literature 1 below, a robot on a display and a robot of an actual machine are moved based on teaching data created by simulation to measure an error in the movement of both robots. The movement of the robot on the display is corrected so that the error is eliminated, so that the robot on the display performs the same movement as the robot of the actual machine.
JP-A-2002-36156

  However, in the conventional method as described above, in order to reduce the hand position error occurring between the robot at the time of the simulation and the actual robot, an error parameter that is significant in reducing the error is required. It is necessary to prepare for many robot mechanism models.

  However, since error parameters are selected based on human intuition and experience, an insignificant error parameter may be selected, and the error cannot be easily reduced. In addition, when the number of parameters and the robot mechanism model increase, the work time required to verify the reduction of the error increases exponentially.

  Further, the identification of the error parameter is performed so that the total sum of the hand position errors with respect to the input posture data group is minimized. However, when the hand position is evaluated in a posture far from the identification input posture, good accuracy is not necessarily obtained. I can't get it. Also, if more input postures are given to improve the accuracy, it takes too much time for the measurement in advance.

  The present invention has been made to solve the above-described problems of the conventional technology, and is used for a simulation robot to correct an error between a real robot and a simulation robot. An object of the present invention is to provide a simulation robot error parameter identification method capable of identifying an optimal error parameter.

  In order to solve the above problems and achieve the object, a simulation robot error parameter identification method according to the present invention creates a tuning path candidate including a plurality of representative postures of a simulation robot, and generates the created tuning path. The path candidates are applied to the robot mechanism model of the reference robot and the robot mechanism model of the simulation robot, an error between the two robot mechanism models is calculated, and an error parameter necessary for reducing the error and an error parameter of the error parameter are calculated. Calculate the evaluation path consisting of a plurality of postures different from the tuning path candidate of the simulation robot, and calculate the calculated error parameter and the robot mechanism model of the reference robot in consideration of the value of the error parameter. The simulation robot When the error parameter is slightly fluctuated with respect to the evaluation path posture evaluated as having a large error, the error of the simulation robot is evaluated. An error parameter having a large influence on the movement of the robot mechanism model is selected, and a tuning path posture that has a large effect on the movement of the robot mechanism model of the simulation robot is set to a tuning path candidate having a large influence degree of the selected error parameter. It is characterized in that a process is selected from among them, a selected error parameter is added, and a process from the error calculation stage to the tuning pass posture selection stage is repeated for the selected tuning pass posture.

  According to the simulation robot error parameter identification method of the present invention, the error parameter and its value are automatically selected based on the error generated between the robot mechanism model of the reference robot and the robot mechanism model of the simulation robot. Therefore, an insignificant error parameter is not selected, and the motion of the simulation robot can be easily brought close to the motion of the robot on site.

  Exemplary embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

  FIG. 1 is a block diagram showing a schematic configuration of a robot facility for executing the method according to the present invention.

  The robot equipment includes a robot 10, a controller 20, a terminal device 30, a display 40, and a CAD device 50. The robot equipment performs a simulation to set a number of error parameters for the simulation robot, identifies error parameter values, creates teaching data, and causes the robot 10 to perform a playback operation based on the teaching data. I do.

  The robot 10 is, for example, a six-axis real machine robot actually arranged at a production site. The controller 20 sets error parameters of the simulation robot corresponding to the robot 10, identifies the values of the error parameters, creates teaching data of the robot 10 by simulation, and performs a playback operation of the robot 10 using the created teaching data. I do. The controller 20 stores a robot mechanism model of the reference robot, a robot mechanism model of the simulation robot, an error parameter unique to each robot mechanism model, an error parameter value, and an evaluation path.

  Here, the reference robot is a virtual robot that moves on the display, and is a robot that has been calibrated so as to move as close as possible to the actual robot that is actually operating at the site. The simulation robot is a virtual robot that moves on the display like the reference robot, and is a robot that moves close to a real robot arranged at the site. When performing the simulation, the model of the reference robot and the model of the simulation robot are the same model. In the method of the present invention, the error parameter to be used and the value of the error parameter are determined so that the movement of the simulation robot is the same as the movement of the reference robot. However, as described above, the reference robot actually operates at the site. Since the robot moves as close as possible to the actual robot, if the motion of the simulation robot is matched with the motion of the reference robot, the motion of the simulation robot will eventually move It becomes something close to.

  The terminal device 30 performs an operation necessary for processing of the method of the present invention, and instructs a playback operation of the robot 10. The display 40 displays a reference robot and a simulation robot, and displays instructions required for the playback operation of the robot 10. The CAD device 50 stores the shape data and arrangement data of the robot 10 and the constituent elements (arms and hands) of the robot 10, and the shape data and arrangement data of peripheral equipment on the site where the robot 10 is arranged. Therefore, these data of the CAD device 50 are used by the controller 20 when simulating the simulation robot.

  FIG. 2 is a main flowchart showing the procedure of the error parameter identification method for the simulation robot according to the present invention. The method according to the present invention is performed by the controller 20. First, a schematic processing procedure of the method according to the present invention will be described with reference to FIG.

  First, a tuning path candidate including a plurality of representative postures of the simulation robot is created. This tuning path candidate is performed by operating the terminal device 30, and for example, postures of m simulation robots are created. The simulation robot can take various postures, and a representative posture (a posture likely to be used in the work performed by the robot 10) among the postures is created (S10).

  Next, for each posture of the created tuning path candidate, how much influence (the degree of influence) on the hand position of the simulation robot when each arm constituting the simulation robot is moved by a small amount. Is calculated. That is, the degree of influence of the minute fluctuation of each error parameter on the movement of each tuning path posture is calculated (S20). Then, a tuning path posture to be used to identify the error parameter of the simulation robot is created (initialization is initially performed with the number of postures being zero) (S25).

  Next, the created tuning path candidates are applied to the robot mechanism model of the reference robot having representative characteristics and the robot mechanism model of the simulation robot, and the error between the two robot mechanism models is calculated. An evaluation path, which is different from the tuning path candidate and is composed of a plurality of postures of the simulation robot, is applied to the robot mechanism model of the reference robot and the robot mechanism model of the simulation robot in consideration of the obtained error. Then, the error between the two robot mechanism models is evaluated (S30).

  It is determined whether or not the evaluation value calculated in the evaluation of the tuning path is within a predetermined maximum error (S40). If the evaluation value is within the maximum target error (S40: YES), the robot calibration is performed. If the processing shifts to the processing to be executed and is not within the target maximum error (S40: NO), an error parameter having a large degree of influence is selected for a posture having a large evaluation value that is greatly affected by the error parameter, and the selected error parameter is selected. A tuning path posture having a large influence is added to the tuning path (S50).

  In the above procedure, the creation of the tuning path candidates (S10) and the calculation of the influence of the error parameter candidate posture in each tuning path (S20) are performed before the evaluation of the tuning paths (S30). However, the processing of S10 and S20 may be performed in parallel with the processing of S30.

  FIG. 3 is a flowchart showing a detailed processing procedure of S10 shown in FIG. This flowchart constitutes a tuning path candidate creation stage. The contents of this procedure will be described in detail with reference to FIGS.

  The operator inputs the number of divisions m of the operating range from the terminal device 30 (S11). Further, the operation range (min, max) of each axis is input (S12). For example, as shown in FIG. 4, when the axis JV1 to the axis JV6 can move in the operation range shown in the figure, the minimum and maximum values of the operation range of each axis are input.

  The controller 20 calculates the designated angle by dividing the operation range of each axis by the input division number m as shown in FIG. 4 (S13). The controller 20 creates an initial posture matrix that is a candidate for a tuning path by combining the designated angles of the respective axes (S14). For example, as shown in FIG. 4, the axis JV1 is in the position of posture 3, the axis JV2 is in the position of posture 4, the axis JV3 is the position of posture 2, the axis JV4 is the position of posture 7, the axis JV5 is the position of posture 6, and the axis is JV5. Assuming that the case where JV6 is set to the position of posture 8 is one candidate for a tuning path, the initial posture matrix is a posture matrix as shown in FIG.

  The controller 20 extracts and creates the shape data and arrangement data of the robot 10 and the constituent elements (arms and hands) of the robot 10 and the shape data and arrangement data of the peripheral equipment on the site where the robot 10 is arranged from the CAD device 50. It is checked whether or not the simulation robot interferes with peripheral equipment based on the posture of the posture matrix (S15). If the initial posture does not cause interference with peripheral equipment (S16: NO), the combination of the designated angles of each axis is changed to create the next posture matrix. Note that the posture matrix creates a plurality of postures of the simulation robot without using the designated angle used for each axis of 1 degree (S17). For example, excluding the motion range used in the initial posture, as shown in FIG. 5, the following posture matrix (the axis JV1 is the position at the designated angle 4, the axis JV2 is the position at the designated angle 3, and the axis JV3 is the designated angle 1 is set, the axis JV4 is set to the position of the specified angle 6, the axis JV5 is set to the position of the specified angle 5, and the axis JV6 is set to the position of the specified angle 6. The created posture is also checked for interference with peripheral equipment in the process of S15.

  Since the operation range of each axis is divided by m, a very large number of combinations of designated angles can be created. In the case of the present embodiment, the same designated angle is not used for each axis. In addition, except for a posture that is expected to interfere with peripheral equipment, a representative posture is selected from postures that may be actually used for work, and the number of postures finally obtained is m ( (Specific numbers are 10 to 20). As an algorithm for selecting a representative posture, a generally used genetic algorithm and optimization algorithm are used. That is, by executing the flowchart of FIG. 3, finally, m tuning path candidates that do not interfere with peripheral equipment are selected.

  FIG. 6 is a flowchart showing a detailed processing procedure of S20 shown in FIG. This flowchart constitutes the influence calculation stage. The contents of this procedure will be described in detail with reference to FIGS.

  When m tuning path candidates are created, the controller 20 then applies the postures of the tuning path candidates to the robot mechanism model of the simulation robot, and the error parameters (q1 to qn) of the simulation robot change slightly. Then, the amount of displacement of the hand position is calculated (S21). Optimal values are selected for error parameters (q1 to qn) of the simulation robot, and are prepared in advance as a table. To minutely displace the error parameter of the simulation robot means that the simulation robot is in a certain posture and only one axis is moved by a minute distance in the operable direction of the axis. As shown in FIG. 7, it is assumed that the postures of the m tuning path candidates are X1 to Xm, and the error parameters q1 to qn of the simulation robot are slightly changed for each of the tuning path positions. On the other hand, the hand position changes. For example, when the error parameter q1 of the simulation robot is slightly changed, the tuning path attitude X1 is 0.53, the tuning path attitude X2 is 2.56, the tuning path attitude X3 is 3.52, and the tuning path attitude Xm is 1. 01. When the error parameter qn of the simulation robot is slightly changed, the tuning path attitude X1 is 1.23, the tuning path attitude X2 is 3.01, the tuning path attitude X3 is 3.67, and the tuning path attitude Xm is 1. It will be 25.

  Next, the controller 20 obtains the maximum value of the displacement amount for each error parameter (S22). As described above, the numerical values obtained by slightly changing the error parameters q1 to qn of the simulation robot and calculating the amount of displacement of the hand position of the simulation robot with respect to the posture of each tuning path candidate were as shown in FIG. Then, the maximum value of the displacement of the hand position with respect to the error parameter q1 of the simulation robot is 3.52, that of q2 is 3.21, that of q3 is 2.41, and that of qn is 3.67.

  Then, the controller 20 normalizes each displacement amount based on the maximum displacement amount of each error parameter (S23). This normalization is performed as follows. For example, if the maximum value of the displacement of the hand position with respect to the error parameter q1 of the simulation robot is 5.00, the displacement of the hand position of each tuning path posture X1 to Xm with respect to the error parameter q1 of the simulation robot is calculated as Divide by 5.00. If the maximum value of the displacement of the hand position with respect to the error parameter q2 of the simulation robot is 4.00, the displacement of the hand position of each tuning path posture X1 to Xm with respect to the error parameter q2 of the simulation robot is calculated as Divide by 4.00. Similarly, after the error parameter q3 of the simulation robot, the displacement of the hand position of each error parameter is divided by the maximum value of the displacement of the hand position of each error parameter.

  When this normalization is performed, for example, as shown in FIG. 9, the magnitude of the influence of each error parameter on the displacement of the hand position is quantified for each tuning pass posture. The normalized numerical value indicates the degree of influence. For example, it can be seen that the error parameter q1 of the simulation robot has a large effect on the displacement of the hand position of the simulation robot with respect to the tuning path posture X3, but has little effect on the tuning path posture X1.

  FIG. 10 is a flowchart showing a detailed processing procedure of S30 shown in FIG. This flowchart constitutes an error parameter calculation stage and an error evaluation stage. The contents of this procedure will be described in detail with reference to FIG.

  First, the controller 20 initializes a tuning path of the simulation robot. That is, the values of all error parameters are set to 0 (S31). Next, the controller 20 applies this tuning path to the reference robot and the simulation robot to operate both robots (S32). As shown in FIG. 11, a robot mechanism model of the reference robot exists in the controller 20. Here, the reference robot is a robot that has been calibrated so as to move as close as possible to the actual robot actually operating at the site. Further, the simulation robot is a robot of the A-type machine for making the robot move close to the robot 10 of the real machine arranged at the site. The model of the reference robot and the model of the robot of Unit A are the same model.

  The controller 20 obtains a hand position error of the robot of the No. A robot with respect to the reference robot from the result of operating using the tuning path (S33). That is, the error between the hand position of the reference robot operated by the simulation and the hand position of the robot of Unit A, which is the simulation robot, is obtained. Then, the controller 20 calculates an error parameter necessary for reducing the error and the value of the error parameter, and stores them in its own storage device (S34).

  With the above processing, the error between the hand position when the tuning path is applied to the reference robot, the hand position when the tuning path is applied to the simulation robot A, and the hand position between the two robots is obtained. Briefly, when a certain posture is taken, it is possible to know how much error the simulation robot has with respect to the actual robot arranged at the site. The processing of steps S32 and S33 constitutes an error calculation step, and the processing of step S34 constitutes an error parameter calculation step.

  Further, as shown in FIG. 11, the controller 20 calculates an evaluation path having a plurality of postures different from the tuning path of the simulation robot and a simulation of the robot mechanism model of the reference robot in consideration of the calculated error parameter values. Both robots are operated by applying the robot mechanism model to the robot mechanism model (S35). The evaluation path is either the attitude used for past production, the attitude used for current production, or the attitude that will be used for future production. The controller 20 obtains an error of the hand position of the A-th robot with respect to the reference robot from the result of the operation using the evaluation path (S36). Originally, the error in which the reference robot and the robot of Unit A, which is the simulation robot, are at exactly the same hand position is taken into account in the tuning pass. Should be exactly the same as the hand position. However, when the posture of the simulation robot is different, the hand position is slightly different. The controller 20 sets the maximum value of the hand position error with respect to the evaluation path as the evaluation value (S37).

  FIG. 12 is a flowchart showing a detailed processing procedure of S50 shown in FIG. This flowchart constitutes an error parameter selection step and a tuning path attitude selection step. The contents of this procedure will be described in detail with reference to FIGS.

  The controller 20 takes out the evaluation path posture evaluated in step S40 of FIG. 2 when the evaluation value exceeds the target maximum error (S51). If the number of the evaluation path postures is three, for example, as shown in FIG. 13, they are XT1, XT2, and XT3. Next, the controller calculates the amount of displacement of the hand position when the error parameters (q1 to qn) of the simulation robot slightly fluctuate in the extracted evaluation path posture. This process is the same as the process of S21 in FIG. 6 (S52). Then, as shown in FIG. 14, the degree of influence is obtained by normalizing the amount of change of each error parameter. This processing is substantially the same as the processing of S23 in FIG. 6 (S53). Further, as shown in FIG. 15, the controller 20 calculates an average value of the degree of influence for each error parameter. This average value is an average value for all evaluation path postures (X1 to XT3) (S54).

  Next, the controller 20 selects three error parameters of the simulation robot having a large average value of the degree of influence. For example, as shown in FIG. 16, since the average value of the degree of influence is 0.74, 0.37, and 0.2, q3, q1, and qn are selected as error parameters of the simulation robot. Is performed (S55).

  The controller 20 selects a tuning path posture suitable for identifying the selected error parameter from the m tuning path candidates obtained in the process of S10 in FIG. In this selection, a search is started from a tuning path candidate having a high degree of influence of an error parameter in a tuning path candidate in the workspace, and a joint angle distribution (distribution with respect to the posture in the tuning path) related to the error parameter is classified into a small number of categories. This is done by selecting the tuning pass attitude that enters. For example, since the error parameter of the simulation robot having the largest average value of the influence is q3, if the tuning path posture that has the most influence on this error parameter q3 is selected, it is Xm as shown in FIG. This tuning path posture Xm is added to the tuning path as shown in FIG. Further, since the error parameter of the simulation robot having the second largest average value of the influence degree is q1, if the tuning path posture that most influences the error parameter q1 is X3, the posture X3 is represented by As shown in FIG. 18, it is added to the tuning path.

  In this way, an evaluation path having a large error is determined, an error parameter having a large effect on the movement of the evaluation path is obtained, and a tuning path attitude having a large influence on the error parameter is added to the tuning path again. When an error parameter that has a large effect on the evaluation path is added, no matter what posture the simulation robot takes, an error parameter value is created to reduce the error in the hand position in that posture. Therefore, it is possible to reproduce an error-free movement with respect to the reference robot, in other words, the robot of the actual machine, and it is possible to minimize the error of the hand position in that posture during playback.

  As described above, it is possible to evaluate the simulation robot model identified in a certain tuning path, determine a posture with poor accuracy, and select an error parameter of the simulation robot that most affects the accuracy of the hand position in that posture. Thus, by setting the optimum value of the error parameter, the hand position error of the simulation robot can be reliably reduced. In addition, since a minimum number of error parameters required to achieve the target accuracy are selected, the number of postures for identifying the error parameters may be reduced, and the measurement work on site is facilitated.

  In addition, various postures that can be taken in the workspace of the simulation robot are created in advance, and the posture suitable for identifying the simulation robot error parameters is selected from these. The resulting error can be reliably reduced. Since the tuning path created by the method of the present invention has the minimum number of postures required for identifying the error parameter of the simulation robot, the measurement time on site can be reduced.

  Furthermore, the degree of influence of the error parameter on the hand position in various postures is calculated, and the relationship between the error parameter and those postures is created as a determinant, so the effect of the error parameter on the hand position in each posture is calculated. The degree can be quantified, and the optimum error parameter for correcting the error can be selected in the posture. Further, the posture for identifying the error parameter of the simulation robot can be quickly obtained.

  Further, since the simulation robot is within the movable range of each axis and can generate an arbitrary posture matrix that does not interfere with peripheral equipment, human work for confirming interference is not required.

  It can be used as a technology to easily bring the movement of the simulation robot closer to the movement of the robot on site.

1 is a block diagram illustrating a schematic configuration of a robot control device that performs a method according to the present invention. 6 is a main flowchart showing a processing procedure of a method for creating an error parameter of a robot according to the present invention. 3 is a flowchart showing a detailed processing procedure of S10 shown in FIG. FIG. 4 is a diagram provided for explanation of a processing procedure in the flowchart of FIG. 3. FIG. 4 is a diagram provided for explanation of a processing procedure in the flowchart of FIG. 3. 3 is a flowchart showing a detailed processing procedure of S20 shown in FIG. FIG. 7 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 6. FIG. 7 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 6. FIG. 7 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 6. 3 is a flowchart illustrating a detailed processing procedure of S30 illustrated in FIG. 2. FIG. 11 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 10. FIG. 12 is a flowchart showing a detailed processing procedure of S50 shown in FIG. FIG. 13 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 12. FIG. 13 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 12. FIG. 13 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 12. FIG. 13 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 12. FIG. 13 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 12. FIG. 13 is a diagram provided for explanation of the processing procedure of the flowchart in FIG. 12.

Explanation of reference numerals

10 robots,
20 controllers,
30 terminal devices,
40 displays,
50 CAD device.

Claims (10)

A tuning path candidate creating step of creating a tuning path candidate including a plurality of representative postures of the simulation robot;
Applying the created tuning path candidate to the robot mechanism model of the reference robot and the robot mechanism model of the simulation robot, and calculating an error between the two robot mechanism models;
An error parameter calculation step of calculating an error parameter required to reduce the error and a value of the error parameter,
An evaluation path consisting of a plurality of postures different from the tuning path candidates of the simulation robot is calculated by taking into account the calculated error parameter and the value of the error parameter, and a robot mechanism model of the reference robot and a robot of the simulation robot. An error evaluation step of applying to the mechanism model and evaluating an error between the two robot mechanism models;
An error parameter selecting step of selecting an error parameter having a large degree of influence on the movement of the robot mechanism model of the simulation robot when the error parameter is slightly changed with respect to the evaluation path posture evaluated as having a large error. When,
A tuning path attitude that greatly affects the movement of the robot mechanism model of the simulation robot, and a tuning path attitude selecting step of selecting from among tuning path candidates having a large degree of influence of the selected error parameter;
Adding the selected error parameter, an error parameter identification step of repeating the processing from the error calculation step to the tuning path attitude selection step for the selected tuning path attitude,
A method for identifying an error parameter of a simulation robot, comprising: The tuning path candidate creation step includes:
Inputting an operation range of each axis constituting the robot and the number of divisions of the operation range;
Calculating the designated angle by dividing the operating range of each axis by the number of divisions,
Creating a candidate for the tuning path by combining the designated angles of each axis;
The method for identifying error parameters of a simulation robot according to claim 1, comprising: The step of creating a candidate for the tuning path includes:
Creating a plurality of poses without using the designated angles used for each axis once,
Excluding the posture from which interference with the peripheral equipment is expected from the plurality of created postures,
The error parameter identification method for a simulation robot according to claim 2, comprising:   2. The method according to claim 1, wherein all error parameters are initially set to 0 in the error calculation step.   2. The error parameter identification method for a simulation robot according to claim 1, wherein the error between the two robot mechanism models is calculated based on the hand positions of the two robot mechanism models. The motion of the robot mechanism model of the simulation robot in each tuning path posture and each evaluation path posture when each error parameter is slightly changed between the tuning path candidate creation step and the error parameter calculation step. The method further includes an influence degree calculation step of calculating the degree of influence,
2. The error parameter identification method for a simulation robot according to claim 1, wherein in the error parameter calculation step, a necessary error parameter and a value of the error parameter are calculated based on the degree of influence. The influence degree calculating step includes:
Calculating the amount of change in the motion of the robot mechanism model of the simulation robot for each of the tuning path posture and the evaluation path posture when the value of each error parameter is slightly changed;
Extracting a maximum value of the variation for each error parameter;
Calculating the degree of influence by dividing the variation of each error parameter by the maximum value for each extracted error parameter;
7. The error parameter identification method for a simulation robot according to claim 6, comprising:   The method according to claim 1, wherein the evaluation path is one of a posture used for past production, a posture used for current production, and a posture that will be used for future production. The error parameter identification method of the simulation robot described in the above. The error parameter selecting step includes:
Calculating a degree of influence on the movement of the robot mechanism model of the simulation robot when the value of each error parameter is slightly changed with respect to the evaluation path posture evaluated as having a large error;
Calculating the average value of the degrees of influence for each error parameter;
Extracting the largest error parameter of the average value;
The method for identifying error parameters of a simulation robot according to claim 1, comprising: The error parameter identification step includes:
2. The error of the simulation robot according to claim 1, wherein a process from the error calculation step to the tuning path posture selection step is repeated until an error between the two robot mechanism models reaches a target value in the error evaluation step. Parameter identification method.

Images