JP3081518B2 - Robot rigidity identification method and device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for identifying the rigidity of a robot, and more particularly to a method and an apparatus for identifying the rigidity of a low rigidity portion in a robot link system.

[0002]

2. Description of the Related Art Until now, robot rigidity (spring constant)
Although various methods have been developed to identify the stiffness, all of them address the lack of rigidity of the reduction gears and the like that exist in the power transmission system. For example, a technique disclosed in Japanese Patent Application Laid-Open No. 03-117580 (hereinafter referred to as Conventional Technique 1) relates to a robot having a rotating joint as shown in FIG. Here, the coefficient b ₀ ~b ₂ of the transfer function G from the input torque of the robot until the motor angular velocity (s), a ₀ ~a
₃ is identified first.

(Equation 1)Then, from the identified coefficients, the inertia M of the motor is calculated._M,viscosity
Coefficient D_M, Drive unit (load side) inertia M_A, Viscosity coefficient
D_A, Joint spring constant K_G, The viscosity coefficient D of the spring element _GTo
Calculate back using the following equation.

(Equation 2)

[0003] In the identification of the transfer function G (s), a discrete ARMA (Amto Regressive Moving Averr) is used.
After identification by age) Model ^{D (z -1) / C (} z -1), the resulting discrete-system model ^{D (z -1) / C (} z -1) impulse model of a continuous-time system and Response approximation and transfer function G
(S). Here, since the system is treated in the form of a transfer function, it is assumed that the system is linear. However, in general, a robot has nonlinearity and cannot be described in a linear manner.
For this reason, an acceleration sensor is mounted on the robot, the joint angular acceleration of the arm (load side) is measured, and in the case of a two-link robot, the transfer function is expressed as two inputs and one output as follows.

[0004] In order to achieve the above object, an invention according to claim 1 is a method for identifying the rigidity of the robot based at least on output data from a sensor attached to the robot. After separating each parameter representing the characteristics of the robot into a rigidity parameter or a parameter containing rigidity inside, and other non-rigidity parameters,
For output data from external and / or internal sensors
Non-rigidity parameters identified by frequency analysis
In the rigidity identification method for a robot for identifying a rigidity parameter based on the frequency analysis result and the non-rigidity parameter,
When identifying the parameters, the robot
From the operation command data, a robot
After removing high-frequency components higher than the natural frequency in the neutral mode
Is input to the robot as a rigidity identification method for the robot. The invention according to claim 2 is based on at least output data from a sensor attached to the robot.
A method of identifying a rigidity of the robot, the parameters representing the characteristics of the robot, the parameters including the stiffness parameter or rigid internally, after separated into a non-rigid parameters of otherwise elastic deformation of the robot Out of the department
Equivalent to the Jacobi matrix up to the position of the field sensor or its inverse matrix
The matrix to be calculated is calculated based on the non-rigid parameters identified separately.
Calculate the moment applied to the elastic deformation part
Based on the output data of the external sensor, moment and matrix
In the rigidity identification method for a robot that identifies the rigidity parameters,
From the robot operation command data used for identification, the robot
Above the natural frequency in the rigid mode operating as a rigid body
The operation command data after removing high frequency components
This is a method for identifying the rigidity of a robot, characterized by being input to a robot.

The independent transfer function G for each axis
A discrete system model of ₁ (s) and G ₂ (s) is identified, and a physical parameter is determined from a coefficient obtained by approximating a continuous time system. Also, Zhou and Maeda, “Measurement of joint stiffness of a vertical articulated robot manipulator” (The Robotics Society of Japan,
Vol. 13, No. 3, pp. 390/396 (19
95)) (hereinafter referred to as “prior art 2”).
Then, a laser displacement meter and a six-axis force sensor are attached to the hand of the robot, and an external force is applied to the hand. Then, the external force F is measured by the force sensor, and the displacement ΔX of the hand position is measured by the displacement meter. Then, torque τ applied to each joint by external force
And the deformation ΔΘ of each joint generated at that time, the following relationship is established. ^{τ = J (Θ) T F} ΔΘ = J (Θ) -1 ΔX However, J (Θ) is the Jacobian matrix of the joint angle theta to the hand position X. The joint stiffness matrix K = diag (k) having the spring constant k of each joint as a diagonal element can be estimated from the least square method or the like by the following equation. ^{ΔΘ = K -1 τ J (Θ} ) where ^{-1 Δ X = K -1 J (} Θ) T F is the simultaneous identification of measuring and joint stiffness and servo stiffness simultaneously, independent identification to identify only the joints stiffness Two techniques are described. What is a problem now is joint stiffness. If simultaneous identification is performed, it is necessary to obtain joint stiffness by subtracting deformation due to servo stiffness from the spring constant describing joint stiffness and servo stiffness as one spring element. There is. In the case of independent identification, as shown in FIG. 27, joint rigidity is identified in a state where the motor shaft is locked and the posture is fixed using a jig or the like.

[0006]

The above-mentioned conventional method for identifying the rigidity of a robot has the following problems. (1) In the prior art 1, each coefficient of the transfer function is identified, and a physical parameter is calculated from a nonlinear equation including the obtained coefficient. However, this nonlinear equation includes a product or division of the physical parameter, and the value of the physical parameter calculated by a subtle change in the coefficient of the transfer function greatly changes. Therefore, the value may be a physically impossible value (eg, a negative value). Therefore, the response characteristics of the identified transfer function may match those of the actual device, but the physical parameters obtained from the coefficients may be far apart from the actual device, making it difficult to control using the physical parameters. It is. That is, the rigidity is not identified here, but only the coefficient of the transfer function.

Further, in this case, it is possible to identify a one-link manipulator, but it is considered difficult to identify a manipulator having two or more links for the following reasons. In the multi-link manipulator, since the dynamic characteristics change depending on the joint angle θ and its differential value, there is no linear transfer function, and the transfer function also changes depending on the posture or motion. That is, the transfer function is identified here on the assumption that the transfer function does not change. However, since the transfer function itself changes in the case of a robot, the transfer function cannot be accurately identified. The only way to identify the transfer function somewhat accurately is to limit the attitude and motion at the time of parameter identification. However, in this case, the fixed signal is also limited, so that the reliability of the identified transfer function is low. Furthermore, in order to obtain a model in all regions in order to limit the posture and movement, it is necessary to identify the transfer function for each posture and movement, and the physical parameters obtained from them also depend on the posture and movement. This is because an irrational state of changing the state may occur.

(2) In the prior art 2, an external force is applied to the hand, and the elastic deformation and the external force generated at that time are measured to identify the rigidity of the joint. However, the rigidity of the wrist axis (4 to 6 axes) is lower than that of the main axis (1 to 3 axes in the case of the vertical articulated robot), and changes greatly due to an external force applied to the hand.
Therefore, when simultaneously measuring the rigidity of both the main shaft and the wrist shaft, the elastic deformation of the main shaft becomes small, and the rigidity of the main shaft cannot be accurately identified. For this purpose, the posture to apply the external force and the direction to apply the external force are determined so that the elastic deformation of one axis appears remarkably. In the simultaneous identification, the posture is fixed in the servo state. The axis is fixed. However, simultaneous identification can only provide the combined stiffness of the joint stiffness and the servo stiffness, and the servo stiffness is greatly deviated due to the effects of friction and so on. Have difficulty. That is, the joint stiffness can be identified only by the independent identification. In this case, a large jig as shown in FIG. 27 is required, and the joint stiffness cannot be easily identified. Further, in this case, an external force is applied from the outside of the robot (hand position). At this time, since the robot is stopped, the influence of friction is large. Therefore,
In an industrial robot or the like, the external force is not transmitted to the elastically deformable portion, and the required rigidity does not always coincide with the rigidity during the operation of the robot. SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a method and an apparatus for identifying rigidity of a robot with high identification accuracy.

[0009]

According to a first aspect of the present invention, there is provided a method for identifying the rigidity of a robot based on at least output data from a sensor attached to the robot. In the robot rigidity identification method for separately identifying each parameter representing the characteristics of the robot into a rigidity parameter or a parameter including rigidity inside and a non-rigidity parameter other than the parameters, an external sensor and / or an internal sensor are used. It is configured as a robot rigidity identification method that performs frequency analysis on output data from a field sensor and identifies rigidity parameters based on separately prepared non-rigidity parameters and the above frequency analysis results. . The second invention is a method for identifying the rigidity of the robot based on at least output data from a sensor attached to the robot, wherein each parameter representing the characteristics of the robot is stored with a rigidity parameter or rigidity inside. Separate the parameters to be included and the other non-rigid parameters and then individually identify the rigidity of the robot. In the method of identifying the rigidity of the robot, a matrix corresponding to the Jacobian matrix from the robot's elastic deformation part to the position of the external sensor or its inverse Is calculated, and the moment applied to the elastic deformation portion is calculated based on the separately prepared non-rigidity parameter,
A stiffness identification method for a robot, wherein a stiffness parameter is identified based on output data, a moment, and a matrix of the external sensor.

[0010] The invention according to claim 3 is at least as follows.
Based on output data from sensors attached to the robot
A method for identifying the stiffness of the robot,
Each parameter that represents the characteristics of the above robot is
Parameters that include internal data or rigidity, and other non-
Separated from the rigidity parameter, the external sensor and / or
Performs frequency analysis on output data from the internal sensor.
And the frequency analysis results
Stiffness of a robot that identifies stiffness parameters based on results
Identify the above non-rigid parameters in the method
At this time, the operation command data of the robot used for the identification and
From the output data from the above sensors, make the robot a rigid body
Frequency component higher than natural frequency in rigid mode
The operation command data and output data after removing
Robot that identifies the above non-rigid parameters based on the
This is a method for identifying the rigidity of a vehicle. The invention according to claim 4
Is at least from the sensor attached to the robot
Identify the stiffness of the robot based on the output data
A method comprising the steps of:
Is a stiffness parameter or a parameter that contains stiffness inside
And the other non-rigid parameters,
Jacobi from the elastically deformed part of the bot to the position of the external sensor
Calculate matrix corresponding to matrix or its inverse, and identify separately
Based on the determined non-rigidity parameter
Is calculated, and the output data of the external
Identify stiffness parameters based on moments and matrices
In the method for identifying the rigidity of a robot,
When identifying the data, the operation of the robot used for the identification
From the command data and the output data from the above sensor,
Below the natural frequency in the rigid mode in which the unit operates as a rigid body
The operation command data after removing the high frequency components
Of the above non-rigid parameters based on the
This is a method for identifying the rigidity of the robot that performs the determination.

[0011] The invention according to claim 5 has at least
Based on output data from sensors attached to the robot
A method for identifying the stiffness of the robot,
Each parameter that represents the characteristics of the above robot is
Parameters that include internal data or rigidity, and other non-
Separated from the rigidity parameter, the external sensor and / or
Performs frequency analysis on output data from the internal sensor.
And the frequency analysis results
Stiffness of a robot that identifies stiffness parameters based on results
Identify the above non-rigid parameters in the method
At this time, the operation command data of the robot used for the identification and
The output data from the above sensors is
Only the natural frequency component in the rigid mode operating as a body
Based on the operation command data and output data after
Identifying the non-rigid parameter.
This is a bot rigidity identification method. Further, according to claim 6
Akira is based on at least the sensor attached to the robot.
The rigidity of the robot based on the output data of
Method that includes parameters that represent the characteristics of the robot.
Parameter, or a parameter containing rigidity inside
And the other non-rigid parameters,
Jacobi from the elastically deformed part of the bot to the position of the external sensor
Calculate matrix corresponding to matrix or its inverse, and identify separately
Based on the determined non-rigidity parameter
Is calculated, and the output data of the external
Identify stiffness parameters based on moments and matrices
When identifying the above-mentioned non-rigidity parameter in the robot rigidity identification method, the operation command data of the robot used for the identification and the output data from the sensor are used to calculate the natural frequency component in the rigid mode in which the robot operates as a rigid body. carrying out the identification of the non-rigid parameter based on said operating command data and the output data after none only
This is a method for identifying rigidity of a robot characterized by the following. further,
The invention according to claim 7 is the invention according to any one of claims 1 to 6 above.
In the rigidity identification method for a robot described above, when the rigidity parameter is identified, not only the elastic deformation of the movable part of the robot in the movable direction but also the elasticity of the movable part of the robot in the predetermined direction other than the movable direction and the elasticity of the support part in the predetermined direction. the use of model considering also the direction shall be the gist thereof.

The invention according to claim 8 is an apparatus for identifying the rigidity of the robot based on at least output data from a sensor attached to the robot,
First stiffness parameter identification means for separately identifying each parameter representing the characteristics of the robot into a stiffness parameter or a parameter including stiffness therein, and other non-rigid parameters, and individually identifying each parameter. Non-rigid
Parameter identification means, the first rigid parameter
Means for identifying data from an external sensor and / or an internal sensor.
Frequency analysis means for performing frequency analysis on output data
The first non-rigid parameter identifying means is separately identified
Based on the obtained non-rigidity parameters and the above frequency analysis results
In the rigidity identification device for a robot, wherein the rigidity parameter is identified by the first method, the first non-rigidity parameter identification means includes:
Operating finger of robot used for identification of the above non-rigid parameter
From the command data, a rigid mode that operates the robot as a rigid body
After removing high-frequency components higher than the natural frequency
Operation data is input to the robot.
A device for identifying the rigidity of a robot.

According to a ninth aspect of the present invention, there is provided an apparatus for identifying the rigidity of the robot based at least on output data from a sensor attached to the robot,
A second stiffness parameter identification means for separating each parameter representing the characteristics of the robot into a stiffness parameter or a parameter containing stiffness therein and a non-rigidity parameter other than the stiffness parameter, and individually identifying each parameter. Non-rigid
Parameter identification means, and the second rigid parameter
The data identification means detects the external sensor from the elastic deformation part of the robot.
A matrix corresponding to the Jacobi matrix up to the position or its inverse
Matrix operation means for performing the operation, and the second non-rigid parameter separately
Based on the non-rigid parameters identified by the
To calculate the moment applied to the elastic deformation part
And output data of the external sensor.
The stiffness parameters based on the moment, the matrix and
In the rigidity identification device for a robot ,
The non-rigid parameter identification means is configured to determine the non-rigid parameter
From the robot operation command data used for identification, the robot
Above the natural frequency in the rigid mode operating as a rigid body
The operation command data after removing high frequency components
A rigidity identification device for a robot, characterized in that it is input to a robot. The invention according to claim 10 has a small
At least the output data from the sensor attached to the robot
Is a device that identifies the rigidity of the robot based on the
Therefore, the parameters representing the above robot characteristics are
Parameters or parameters that contain stiffness internally,
After separating them into external non-rigid parameters,
A third stiffness parameter identifying means for separately identifying and a third non-stiffness parameter identifying means;
And rigidity parameter identification means.
The parameter identification means is an external sensor and / or an internal sensor;
Analysis that performs frequency analysis on output data from
And the third non-rigid parameter identification means is
Between the identified non-rigid parameter and the frequency analysis result
Stiffness of robot based on stiffness parameter identification
In the identification device, the third non-rigid parameter identification method
The step is the position of the robot used to identify the non-rigid parameter.
From the operation command data and the output data from the above sensor,
Natural vibration in rigid mode in which the robot operates as a rigid body
The operation command after removing more than a few high-frequency components respectively
The non-rigidity parameter based on the data and output data
Robot rigidity characterized by identifying
Sex identification device. In addition, the invention according to claim 11
Output data from sensors attached to the robot.
Is a device that identifies the stiffness of the robot based on data
The parameters representing the characteristics of the robot
Parameters that include the properties parameter or stiffness inside,
After separating them into non-rigid parameters other than
A fourth stiffness parameter identifying means for individually identifying and a fourth stiffness parameter identifying means;
Non-rigid parameter identification means,
The parameter identification means can detect the external
It corresponds to the Jacobi matrix up to the sensor position or its inverse matrix.
Matrix calculating means for calculating a matrix, and a fourth non-rigid
Non-rigid parameters identified by the sex parameter identification means
Calculate the moment applied to the elastic deformation part based on the data
And a moment calculating means for calculating the output of the external sensor.
Stiffness parameters based on force data, moments and matrices
Robot rigidity identification device
The fourth non-rigid parameter identification means may be a non-rigid parameter identifying means.
Robot operation command data used for meter identification
Robot as a rigid body from the output data from the sensor
High frequency components higher than the natural frequency in the stiffness mode that operates
The operation command data and output data after each removal
The above-mentioned non-rigidity parameter is identified based on
A rigidity identification device for a robot. Ma
In addition, the invention according to claim 12 is at least applied to a robot.
Based on the output data from the attached sensor,
An apparatus for identifying the rigidity of a robot, comprising:
Parameters representing the characteristics of
Parameters that contain
The fifth method is to separate each
Stiffness parameter identification means and fifth non-rigid parameter identification
Means for identifying the fifth rigidity parameter.
Is the output data from the external and / or internal sensors
Frequency analysis means for performing frequency analysis on
Non-rigidity separately identified by the fifth non-rigidity parameter identification means
Parameters based on the parameters and the above frequency analysis results.
In the robot rigidity identification device that identifies the meter
The fifth non-rigid parameter identifying means may
Command data of the robot used for identification of performance parameters
And the output data from the above sensors
Of the natural frequency component in the rigid mode that operates as a rigid body
Based on the operation command data and output data
And to identify the above non-rigid parameters.
A rigidity identification device for a robot, characterized in that: Ma
In addition, the invention according to claim 13 applies at least to the robot.
Based on the output data from the attached sensor,
An apparatus for identifying the rigidity of a robot, comprising:
Parameters representing the characteristics of
Parameters that contain
6
Stiffness parameter identification means and sixth non-rigid parameter identification
Means for identifying the sixth stiffness parameter.
Between the elastic deformation part of the robot and the position of the external sensor
Row for calculating the matrix corresponding to the Jacobi matrix or its inverse
A column calculating means, and a sixth non-rigid parameter identifying means
Based on the non-rigid parameter identified by the step
Calculation to calculate the moment applied to the deformable part
Means and the output data of the external sensor and the momentum.
Stiffness parameters are identified based on
In the bot rigidity identification device, the sixth non-rigid parameter
Meter identification means used to identify the above non-rigid parameters
Robot operation command data and output from the above sensor
The data is used to calculate the rigidity of the robot as a rigid body.
The operation command after only the natural frequency component in the mode
The non-rigidity parameter based on the data and output data
Robot rigidity characterized by identifying
Sex identification device. The invention according to claim 14 is:
The rigidity of the robot according to any one of claims 8 to 13.
In the identification device, when identifying the above stiffness parameters,
If only the elastic deformation of the movable part of the robot in the movable direction
And the predetermined direction other than the movable direction of the movable part and the predetermined
The use of a model that also considers the elastic direction in
The summary of the The robot according to the first, third, and fifth aspects of the present invention.
The method for identifying the rigidity of
Based on the output data from the sensor
This is a method to identify the rigidity of the robot.
Each parameter that represents stiffness parameter or stiffness
And other non-rigid parameters
Separated from external and / or internal sensors
Frequency analysis was performed on the output data, and the non-
Stiffness based on the stiffness parameter and the above frequency analysis results
Basics related to rigidity identification "identify parameters"
Configuration and has in common, engaging below claims 1, 3, 5
The above basic configuration of the robot rigidity identification method
The first basic configuration related to identification ". In addition,
The method for identifying the rigidity of a robot according to claims 2, 4, and 6
Output data from sensors attached to the robot.
Method to identify the rigidity of the robot based on the data
The parameters representing the characteristics of the robot
Parameters that include the properties parameter or stiffness inside,
After separating into non-rigid parameters other than
Jacobi matrix from the elastic deformation part to the position of the external sensor or
A matrix corresponding to the inverse matrix is calculated, and the non-rigid
Based on the elasticity parameter
The output data of the external sensor and the moment
Stiffness parameters are identified based on
The basic configuration related to rigidity identification is common,
The robot rigidity identification method according to claims 2, 4, and 6 is described below.
The above basic configuration is referred to as “the second basic configuration related to rigidity identification”.
"Mr." In addition, the robot according to claim 8, 10, or 12
The rigidity identification device of the
Based on the output data from the sensor
A device for identifying the rigidity of the robot
Each parameter that represents stiffness parameter or stiffness
And other non-rigid parameters
After separation, they are individually identified (first and
(3,5) stiffness parameter identification means and (1,3,3)
(Fifth) non-rigid parameter identification means;
If the means for identifying the sexual parameter is an external sensor and / or an internal sensor,
Frequency to perform frequency analysis on output data from sensor
Includes analysis means and separates the above non-rigid parameter identification means
Based on the identified non-rigid parameters and the above frequency analysis results
Stiffness parameters based on
The alternative basic configuration is common.
The above basics of the rigidity identification device of the robot according to 10, 12
Configuration is called "third basic configuration related to rigidity identification"
U. Further, the rigidity of the robot according to claim 9, 11, or 13 is provided.
The identification device says, “At least the security
Based on the output data from the sensor
Device that identifies the robot
Parameter is a stiffness parameter or a parameter that contains stiffness inside.
And parameters were separated into a non-rigid parameters otherwise
In the above, each is individually identified (second, fourth, sixth
Stiffness parameter identification means and (second, fourth, sixth)
Non-rigid parameter identification means,
The data identification means uses an external sensor from the elastic deformation part of the robot.
Matrix corresponding to the Jacobi matrix up to the position or its inverse
Matrix operation means for calculating the non-rigidity parameter
Based on the non-rigid parameters identified by the identification means
Moment for calculating the moment applied to the elastic deformation part
And the output data of the external sensor and the
Stiffness parameters based on the
Common to the basic configuration related to rigidity identification
The rigidity of the robot according to claims 9, 11, and 13
The above basic configuration of the gender identification device
"Fourth basic configuration".

[0014]

DETAILED DESCRIPTION OF THE INVENTION AND

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments and examples of the present invention will be described below with reference to the accompanying drawings to provide an understanding of the present invention.
The following embodiments and examples are mere examples embodying the present invention, and do not limit the technical scope of the present invention. Here, Fig. 1 relates to the above rigidity identification .
Schematic diagram showing a schematic configuration of a first basic configuration applicable first device A1, 2, 3 is an explanatory view showing an operation example of the apparatus A1, the first base 4 is involved identifying the stiffness Constitution
Is a schematic diagram showing a schematic configuration of a second device A2 to which can be applied ,
FIGS. 5 to 7 show the non-rigidity parameter identification of the present invention which is used together with the first basic configuration relating to the rigidity identification .
Schematic diagram showing a schematic configuration of applicable third to fifth device A3~A5 how alternative, FIG. 8 according to the stiffness identification
FIG. 9 to FIG. 23 are schematic diagrams showing a strict model or simplified models M1 to M10 and the like, and FIG. 24 is a diagram showing the rigidity of the sixth device A6 to which the second basic configuration can be applied.
FIG. 25 is a schematic diagram showing a schematic configuration of a seventh device A7 according to a modification of the first basic configuration relating to the rigidity , and FIG.
Eighth device A8 according to a modification of the second basic configuration involved
It is a schematic diagram which shows the schematic structure of.

[ According to Claims 1, 3, 5, 8, 10, 12 ]
That INVENTION stiffness identification method of a robot according to claim 1, 3 and 5, the upper
With the first basic configuration related to the rigidity identification
And the rigidity identification of the robot according to claim 1, 3, or 5.
Claims 8, 10, and 12 wherein the method is applicable respectively.
Such a rigidity identification device for a robot is, as shown in FIG.
A third basic configuration related to rigidity identification is provided. Below
Is a robot rigidity identification device according to claims 8, 10, and 12.
The embodiment of the installation is the robot according to the first, third and fifth aspects.
The rigidity identification method will be described as being included in an embodiment.

First, the first basics relating to the above rigidity identification
Embodiment of the configuration (and the third basics related to the above rigidity identification)
Embodiments of the configuration, the same applies hereinafter) will be described focusing on the basic principle. Here, an example is shown for a 6-axis vertical articulated robot, but the identification of the stiffness of the arm section (main axes 1 to 3 in terms of joints) will be mainly described, and the wrist section (joint joints) will be described. Speaking of the wrist axis 4-6
The description of (axis) is omitted for the following reason. That is,
This is because the effect of friction on the wrist axis is greater than that of elastic deformation or inertia deformation, and stiffness is hardly used for control or the like even if stiffness is identified, so there is little need to identify stiffness. However, the rigidity of the wrist shaft can be identified in the same manner as the method of identifying the rigidity of the main shaft, and the only thing to pay attention to here is to sufficiently consider the influence of friction. One method of removing the influence of friction is to superimpose a dither signal consisting of a relatively high frequency on the identification signal input to the wrist axis at the time of identification, and remove the dither signal component from the identification data during identification calculation to perform identification. ， Can cancel the effect of friction.

FIG. 1 shows a case where the present invention is applied to a 6-axis vertical articulated parallel link robot. Here, the robot 1 is fixed in an arbitrary posture, vibration is generated by hammering, and the vibration is captured by the acceleration sensor 2 arranged at the hand. The vibration signal captured by the acceleration sensor 2 is guided to a frequency analysis unit 3 (corresponding to a frequency analysis unit), where a frequency analysis such as FFT is performed, and the analysis result is separately obtained and stored in the memory 4. The stiffness parameter is identified by the identification unit 5 (corresponding to the first , third, or fifth stiffness parameter identification means) using a stiffness parameter or a parameter (a non-stiffness parameter) other than a parameter including stiffness therein. Now, it is assumed that the inertia parameter and the viscosity parameter of the arm of the robot are given as non-rigid parameters. At this time, if the joint angle θ in a certain posture is given, the arm inertia matrix J _A (θ) is calculated from the arm inertia parameter. For example,
If the wrist axis is ignored and the three-link parallel link manipulator is used, the inertia matrix J _A (θ) is given by the following equation.

[0018]

(Equation 4) Here, the joint angle vector is θ = [θ ₁ θ ₂ θ ₃ ] ^T , J ₂ , J ₃ , and J ₂₃ are a part of the arm inertia parameter, and J _A1 (θ) is also the arm inertia parameter. It is a value that can be calculated from θ.

If the data of the acceleration sensor is subjected to frequency analysis with the above preparation, data such as the natural frequency can be obtained. Now, focusing on the vibration on the XZ plane in FIG. 1, the natural frequencies ω ₂ and ω _{3 in} the W ₂ and W ₃ directions match the natural frequencies of the second and third links, respectively.
If ω ₂ ≧ ω ₃ , the following equation holds.

(Equation 5) Here, K ₂ and K ₃ are the respective spring constants of the second and third joints.

[0020]

(Equation 6)The spring constants K of the second and third joints_Two, K_ThreeIs decided
Is done. Also ω_Two≤ω _ThreeThen, in equation (3) above,
Only +-of the root is inverted. On the other hand, the Y-axis direction
Frequency ω₁Notice that the spring constant of the first joint
K₁Is given by the following equation.

(Equation 7)

That is, when obtaining the spring constant of the first joint, the spring constant K ₁ is determined by the inertia parameter J ₁ and the natural frequency ω
It is given by a simple formula expressed as the product of ₁ . Therefore, if the inertia parameter is known, as a result of performing the frequency analysis, the spring constant can be easily obtained as compared with the related art 1. Also, the spring constants K _{2 of} the second and third joints,
K ₃ is expressed by an equation including a square root, etc.
This is due to interference between the second link and the third link, and is the same as in the case of the two-link manipulator of the first related art. In the prior art 1, interference is separately approximated by a transfer function as a disturbance, and the information is not used although the transfer function of the disturbance is identified. That is, since the information is handled in the form of a transfer function, disturbance information cannot be extracted, and physical parameters are derived only from transfer functions other than disturbance. On the other hand, the first
In the embodiment of the basic configuration, useful information such as interference and attitude change expressed as a disturbance in the prior art 1 is also incorporated, and the expression of the spring constant is slightly complicated in order to identify a more accurate value. I have. However, this complexity is the same as in the prior art 1
It is several steps simpler than obtaining physical parameters from the nonlinear equation in

As described above, the spring constant of each joint can be obtained simply by performing hammering in a certain posture and performing frequency analysis of the acceleration in the X, Y, and Z directions.
In order to further improve the accuracy of the identified spring constants, hammering is performed in various postures 1, 2,... And the respective spring constants may be obtained by the least square method or the like based on the following equation.

(Equation 8) In the above equation (5), it is described that the least square method or the like is performed only with the expression of each spring constant alone. However, not only the expression of each spring constant but also the following relationship can be used simultaneously. Good.

[0023]

(Equation 9) Here, the acceleration sensor is disposed at the hand, but may be disposed at an arbitrary position, and may be hit at any position by hammering. Further, in the prior art 2, it is necessary to use a large jig or to maintain a specific posture and apply an external force to the hand portion in a specific direction. On the other hand, the first
In the embodiment of this configuration, no jig or the like is required at all, and it is possible to perform hammering at an arbitrary position in an arbitrary direction in an arbitrary direction. Therefore, it is possible to prevent loss of information due to performing only a specific posture and increase in rigidity identification errors of the main shaft (1 to 3 axes) due to flexibility of the wrist (4 to 6 axes) due to application of an external force to the hand. Can be.

Further, the first basics relating to the above rigidity identification
In the embodiment of the configuration, the acceleration sensor is used as the external sensor, but the point is that the natural frequency and the like can be specified by frequency analysis. Therefore, the external sensor is not limited to the acceleration sensor, but may be a position sensor such as a speed sensor, a gyro, or a displacement sensor. Furthermore, in the embodiment of the first basic configuration relating to the rigidity identification, the spring constant was identified from the natural frequency obtained by the frequency analysis, but not only at the natural frequency but also at an arbitrary frequency. Similarly, the relationship between the spring constant and the gain and phase can be derived from the gain and phase information, and the spring constant can be similarly identified based on the relationship. However, the method using the natural frequency shown this time can derive the spring constant more accurately than the identification based on the gain and the phase.

In the above description, the identification method is described assuming an arbitrary posture. However, as shown in FIG. 2, when the position of the center of gravity of the third link of the robot 1 is perpendicular to the second link, the interference term Disappears. Therefore, J ₂₃ cos (θ ₃
−θ ₂ ) becomes 0, and the equation including the square root used when identifying the spring constants of the second and third joints is further simplified as follows.

(Equation 10) Thus, the spring constant can be easily obtained. However, the interference term disappears only in the case of the parallel link manipulator, and in the case of the serial link manipulator, the interference term becomes small but does not disappear. Therefore, to be precise, the spring constants of the second and third joints are expressed by the above-described equation (3). Further, since the posture is limited here, information in other postures cannot be obtained, and the accuracy of the identified spring constant may be degraded.

In FIGS. 1 and 2, the vibration constant is excited by hammering, and the vibration is analyzed to identify the spring constant. However, merely applying an external force to the hammering or the hand does not apply a force to a portion of the speed reducer or the like that is elastically deformed during the operation of the robot. For this reason,
In some cases, the stiffness of the elastically deformed portion during these operations cannot be measured. Therefore, as shown in FIG. 3, the robot 1 is moved from the position indicated by the dotted line to the position indicated by the solid line, is rapidly decelerated, and the vibration at that time is measured. In this way, the portion that is actually elastically deformed during the operation of the robot is deformed, and the rigidity to be obtained can be identified. Alternatively, as shown in FIG. 4, the robot 1 is arbitrarily operated, and when the robot is a rigid link, the position or speed or acceleration at which the hand will move is converted into the output from the motor of each joint. It is derived by performing the forward conversion in 6. Then, by calculating the difference from the output of the external sensor 2, only the external sensor component generated by the elastic deformation is extracted. By performing frequency analysis on the extracted data by the frequency analysis unit 3 and performing the same processing as described above, it is possible to identify the spring constant of a portion that is elastically deformed when the robot actually operates.

As described above, the stiffness identification
In the embodiment of the first basic configuration, a stiffness parameter whose natural frequency and the like match well by frequency analysis can be obtained. In addition, by separately obtaining the non-rigidity parameter, the problem due to linearization as in the related art 1 does not occur, and the non-rigidity parameter can be easily applied to a multiple link manipulator. Furthermore, by obtaining data corresponding to the joint angle, angular acceleration, or angular acceleration of the robot that cannot be measured by the internal sensor from the external sensor, the rigidity parameter can be more accurately identified. Subsequently, the rigidity
Used together with the first basic configuration
Therefore, non-rigidity which is a feature of the invention according to claims 1, 3, and 5
Configurations related to parameter identification (claims 8, 10, 12
To identify the non-rigidity parameter that is the feature of the invention according to
An alternative configuration, hereinafter the same) will be described. Claims above
Claim 1 to which the method for identifying rigidity of a robot according to Claim 1 is applicable.
Item 8. Non-rigid parameter in the robot rigidity identification device according to item 8.
FIG. 5 shows a configuration related to the identification of the meter . Usually, as described in the related art 1, as an identification signal (a signal input to a control target to obtain identification data),
A signal including a wide band from a low frequency to a high frequency such as an M-sequence signal or white noise is used. However, when identifying non-rigid parameters, the identification model is usually described by a rigid body model, and the manipulator needs to operate as a rigid body. For this reason, when an identification signal including a high frequency is input, natural vibrations are excited, and the robot 1 makes a motion that cannot be described by a rigid body model, so that a non-rigid parameter cannot be accurately identified. Therefore, the identification signal generating unit 7 of FIG. 5, for example, generates an identification signal u represented by the following formula, taking the angular frequency omega _n sufficiently smaller than the natural frequency.

[0028]

[Equation 11] Here, _An is the amplitude, and φ _n is the phase shift. The non-rigidity parameter is identified by the identification unit 8 based on the data obtained at this time, and stored in the memory 9. This identification result is used for rigidity identification of the robot. Alternatively, FIG.
As shown in (1), the output from the identification signal generation unit 7 may be filtered by a filter 10 to remove the high frequency that excites the natural vibration, and then commanded to each motor (the device A3 or A3). Identification in device A4
Signal generation unit 7 and identification unit 8 serve as first non-rigid identification means
Equivalent) . This makes it possible to easily obtain the above-mentioned identification signal u even in an actual robot or the like. For example, when the motor driver of each joint of the robot 1 has an analog input, it is only necessary to insert an analog filter in front of the input terminal of the motor driver. However, recently, digital input motor drivers are becoming widespread, and there are cases where the driver input cannot be filtered by an analog filter. Further, there is a case where the filtered identification signal or the identification operation itself cannot be performed due to a limitation on the operation of the robot or a limitation of the robot controller. In these cases, it is necessary to perform identification from data on the operation of the robot in accordance with data on the operation of the robot as usual and instructions to the robot including high-frequency components. As shown in FIG.
This is to remove the high frequency by filtering the input / output data (identification data) of the robot including the high frequency with the filter 10, respectively.
The robot rigidity identification method according to claim 3 is applicable.
The non-rigid part of the robot rigidity identification device according to claim 10.
This is a configuration related to parameter identification. In device A5
Identification unit 8 and filter 10 have the same third non-rigidity parameter.
It corresponds to the setting means. In this case, the same effect as when the filtered identification signal is input can be easily realized. This procedure is shown by the following formula. First, in the three-link manipulator, the equation of motion of the second link when the first link is fixed is as follows.

[0029]

(Equation 12) Based on this equation of motion, the data can be described in a matrix expression using data obtained in various postures as follows.

(Equation 13) In an ordinary identification method, the inertia parameters J ₂ and J ₂₃ are identified based on the matrix expression by a least square method or the like. However, here, both sides of the above equation (10) are filtered. That is, the matrix and vector elements other than the inertia parameters J ₂ and J ₂₃ in the above equation (10) are each filtered, and the parameters may be identified using the least square method or the like by the matrix expression obtained by the following equation.

[0030]

[Equation 14] Here, the pre-processing unit 11 in FIG. 7 is for compensating for the non-linearity of the robot 1, and calculates each non-linear element of each matrix and vector before filtering. In the above, identification of inertia in the rigid body mode in which the robot operates as a rigid body has been mainly described. Although the above-described method for identifying non-rigid parameters is particularly effective for identifying inertia in the rigid mode, similarly, if viscosity and gravity parameters are incorporated as parameters, those parameters can also be accurately identified.

By the way, the first related to the rigidity identification is as follows.
In the embodiment of the basic configuration , instead of the inertia in the rigid body mode,
Inertia on the load side (arm) is required. In order to obtain the inertia on the load side, the inertia on the input side of the speed reducer may be subtracted from the inertia in the rigid body mode. The inertia on the input side of the speed reducer includes the inertia of the rotor of the motor and the input gear of the speed reducer. These values can be easily obtained from a catalog or the like. Also, the accuracy of the catalog value is usually very high. Therefore, there is no problem even if the input side inertia is calculated from the catalog, and an accurate value can be easily obtained. Further, in the above description, the natural vibration is not excited by the filtering or the like. On the contrary, the natural vibration can be excited and the inertia parameter on the load side can be obtained from the data at the time of the natural vibration. The configuration at this time is shown in FIG.
However, what is different from the above is the filtering characteristics and the like (the rigidity of the robot according to claim 5).
An identification method, and a method for identifying rigidity of a robot according to claim 12.
The filtering characteristics are different in FIG.
The filtered filter 10 and the identification unit 8 form a fifth non-rigid
Parameter identification means). Speaking of how to obtain load side inertia and friction, the motor hardly moves at the time of natural vibration, and only the load (arm) vibrates. At this time, since the state relating to the motor can be deleted from the motion equation of the robot, the motion equation is as follows.

[0032]

(Equation 15) Equation (12) is the equation of motion of the second link when the first link is fixed in the three-link manipulator, similarly to the above equation (9). The suffix A indicates the state of the arm, and _{JA2 and the} like are the inertia parameters of the arm.
Based on the equation of motion expressed by the above equation (12), parameter identification is possible as in the case of the above equation (9). However, here, it is necessary to select or filter the input so as to excite the natural vibration. If the friction parameter is added to the equation of motion expressed by the above equation (12), the friction parameter of the arm can be identified. This makes it possible to directly identify the inertia and friction of the arm. However, in this method, the influence of friction is great, and it is necessary to pay attention to the identification of friction in order to accurately identify the inertia on the load side. For example, it is necessary to add conditions such as the magnitude of the frictional force and the inertia force at the time of the natural vibration to be the same as the constraint conditions at the time of the previous identification, and to ensure the identification accuracy. Therefore, if the inertia on the input side can be calculated from catalog values, etc., it is necessary to identify the inertia with high accuracy.
As mentioned earlier, it is advisable to use a method of identifying non-rigid parameters that does not excite natural vibrations.

The aboveFirst basic configuration related to rigidity identification
Is an identification method based on frequency analysis.
A model with a good number of matches is obtained. However, elastic deformation
When compensating the amount, the amount of deformation, not the natural frequency
The actual size of the model must match the actual model
You. For this reason, in the domain of deformation, not in the frequency domain,
Parameter identification better matches the amount of deformation
Is obtained.The second involved in rigidity identification (and
4) Basic configurationFocuses on this point.
This will be described below. [Claims 2, 4, 6, 9, 11, and 13invention〕The robot rigidity identification method according to claims 2, 4, and 6 is
With the second basic configuration related to the rigidity identification
And the rigidity identification of the robot according to claims 2, 4, and 6.
Claims 9, 11, and 13 wherein the method is applicable respectively.
As shown in FIG.
A fourth basic configuration related to rigidity identification is provided. Below
Is a rigidity identification device for a robot according to claims 9, 11, and 13.
The embodiment of the device is the robot according to the second, fourth and sixth aspects.
Stiffness identification method This will be described in connection with the embodiment. As shown in FIG.
,The second basic configuration related to the rigidity identification (and the fourth basic configuration)
Basic configuration ofIs the elastic deformation part of the robot 1
Jacobi matrix from the position of the external sensor 2 to its position or its inverse matrix
Calculate matrices Jθ and Jε corresponding toIdentifyNon-rigid
Based on the elasticity parameter
Of the external sensor 2 and the moment M
Identify stiffness parameters based on matrices Jθ, Jε
To. NonIdentification of stiffness parametersConfigurations involvedIsAbove
For the first basic configuration related to gender identificationStatedeachMethod
UsingTo.That is, the second (and the second)
Combination of the basic configuration of 4) and the above devices A3 to A5
The identification signal generator in the device A3 or A4
7 and the identification unit 8 serve as second non-rigid parameter identification means.
The identification unit 8 and the filter 10 in the device A5
4 corresponds to the non-rigid parameter identification means,
The filtering characteristics of the filter 10
The filter 10 and the identification unit 8 in the case are the sixth non-rigid parameter.
Data identification means.

Hereinafter, the second basic principle relating to the rigidity identification will be described.
The configuration will be described focusing on its basic principle. the above
In the second basic configuration relating to the rigidity identification, as shown in FIG. 8, information on the amount of deformation that cannot be measured by the inner field sensor is measured by the outer field sensor 2, and the calculation unit 12 (the matrix calculation means and the moment calculation means) The amount of deformation is obtained by using the moment M calculated from the motor rotation angle θM, the Jacobian matrices Jθ, Jε, and the like. Then, the identification unit 5 (corresponding to the second , fourth, or sixth stiffness parameter identification means) performs parameter identification in the region of the deformation amount, thereby obtaining a model in which the deformation amounts match well. Teeth or to <br/>, in the conventional model which allows elastic deformation only in the joint portion can not sufficiently describe the behavior of the robot actual, which leads to the deterioration of the identification accuracy. Therefore, based on a stricter model, it is possible to improve the identification accuracy and the model accuracy. Also, when identification is performed by frequency analysis, only one natural mode appears in the conventional model, but in the strict model, it is possible to handle multiple modes, and furthermore, natural mode due to interference of each axis. It is also possible to handle the deviation. Therefore, a very high-precision model can be obtained.

The following description is made by taking a three-axis vertical articulated robot as an example. In each of the prior arts 1 and 2, the spring elements k ₁ θ to k ₃ θ exist only in the rotary joint, and the spring elements also have elastic deformation ε ₁ θ to ε in the rotation axis direction of the rotary joint.
_It was based on a model that allowed only 3θ and did not allow any bending deformation in the other direction or elastic deformation in the extension direction. here,
As shown in FIG. 9, in addition to the elastic deformation of the movable parts such as the rotary joints, the elastic deformation of the support part supporting the movable parts is also considered. Further, by adopting the model M1 that allows elastic deformation in an arbitrary direction, a more accurate modeling can be performed. However, in FIG.
k _{iR, iS} represents a spring driven by the i-th joint and a spring supporting the same, and X, Y, Z fixed to the link.
The elastic deformation in the direction is ε _{iRX, Y, Z, iSX, Y, Z} , and the spring constant is k
_{iRX, Y, Z, iSX, Y, Z} and bending moment are M
_{Let iRX, Y, Z, iSX, Y, Z.}

It is assumed that the inertia and viscosity parameters have already been identified by the above-described non-rigid parameter identification method and the like. Here, the spring constant K = diag (1 /
( _{i.sub.iR *, iS *} ) (subscript * indicates omission of symbol). First, the hand position x of the robot and the rotation angle of the motor θ _M = [θ _M1 θ _M2 θ _M3 ] ^T are measured, and the displacement e of the hand caused by the elastic deformation is calculated by the following equation. e = x−L (θ _M ) (L (θ _M ): forward conversion from θ _M to x) (13) Next, a moment M (a vertical vector having elements of M _{iR * and iS *} ) is _expressed as It is calculated from the inertia / viscosity / gravity parameters already identified and the inertia parameters obtained from the geometric conditions. Then, using the least squares method, the spring constant K
Can be obtained. Here, ε is a vertical vector having ε _{iR *, iS *} as elements, and Jε is a Jacobian matrix from ε to e.

(Equation 16)

As described above, parameter identification can be performed. In addition, the above method can be applied to a conventional model that does not allow elastic deformation only in a joint. FIG. 10 shows a model M2 including elastic deformation near the fourth joint, but the model can be identified in the same manner. 9 and 10
The rigorous model of
When identifying a robot,
Not only elastic deformation but also a predetermined direction other than the movable direction of the movable part
And a model that also considers the elastic direction of the support in the predetermined direction
The invention according to claim 7 or claim 14 may be used.
This is an embodiment of the present invention. FIG. 11 shows the result of parameter identification based on both the model M2 in FIG. 10 and the conventional model. In FIG. 11, the theoretical value and the actual measurement of the hand position displacement of the robot by the model M2 (this model) obtained by the parameter identification by the method of the second embodiment and the conventional model in the displacement of the hand position caused by elastic deformation. Values. In the conventional model, since only the elastic deformation in the joint rotation direction is handled, it is impossible to describe the elastic deformation. Therefore, the theoretical values slightly vary, but generally agree with the actually measured values. Furthermore, when parameter identification is performed based on this model, the theoretical value and the measured value of the model agree very well, and it can be seen that the rigidity parameter and other parameters are accurately identified.

However, the so-called strict elastic deformation models M1 and M2 in FIGS. 9 and 10 have redundant elastic deformation, and the models M3 to M10 are simplified as shown in FIGS. This method is also effective. FIG. 12 shows a model M3 in which a plurality of integrated elastic deformations are represented by one spring element among the elastic deformations. Further, FIG. 13 shows a model M in which the center of the elastic deformation of a certain spring element is provided in an area including each center of a plurality of elastic deformations represented by the spring element as shown in FIG.
4. Further, FIGS. 15 to 18 show that there are a plurality of one spring elements representing the above-mentioned plurality of elastic deformations, and when the values are close to each other, it is assumed that the spring coefficients of the elastic deformation in a predetermined direction are the same. Are models M5 to M8 in which a plurality of spring elements are further integrated. 19 and FIG.
Are models M9 and M10 that are separated into components corresponding to the respective elastic deformation directions, and among components in the same deformation direction, components higher than a predetermined stiffness component are included in low components.

For example, according to the model M10 of FIG.
FIG. 21 shows the result of parameter identification by this method (see the Δ mark in the figure). Compared to the case without simplification (see the circle in the figure), the simplification has a slight variation, but it is better than the conventional model (see the circle in the figure). Match. These differences depend on the model. The model obtained by parameter identification in this method is useful for exact models, simplified models, and even conventional models. You can see that there is. FIG. 22 shows a case where joint angles are measured based on the conventional model without using a Jacobian matrix or the like, and rigidity is identified from the joint angles and the motor rotation angles (× in the figure).
), A case where parameters are identified by the present method (（in the figure), and a case where parameters are identified by the present method in accordance with a stricter model (○ in the figure). From the figure, the models identified based on this method (○ and ■ in the figure)
(Marks) agree well with the actual values, and in particular, the accuracy is better when the model is based on a strict model (marked with a circle in the figure). On the other hand, in the case of the conventional technology based on the conventional model that does not use the Jacobi matrix or the like (marked by x in the figure), the displacement due to the elastic deformation of the hand position cannot be represented at all. Further, for reference, FIG. 23 shows a compensation result of the amount of elastic deformation using a model whose parameters are identified by the present method based on the model M10 of FIG. As a result of performing compensation based on the model obtained by this method, elastic deformation is compensated and the amplitude is kept constant as shown in FIG. 23 (b). As shown in FIG. 23 (a), the amplitude increases as the frequency increases due to elastic deformation. From the above, it can be seen that the rigidity parameters and the like are accurately identified by the parameter identification according to the present method.

Since the amount of elastic deformation is generally very small, high measurement accuracy is required when position information is measured by an external sensor and used for identification. However, a sensor having high measurement accuracy generally has a small measurement range. Therefore, it is possible to ensure high measurement accuracy and improve identification accuracy by causing the robot to perform a small motion within the measurement range. For example, if the minute motion of the hand position measured by the external sensor with high precision is Δx, and the elastic deformation, hand position displacement, moment, and motor angle caused by the minute motion are Δε, Δe, ΔM, Δθ _M , respectively,
Equations (13) and (14) are

[Equation 17]

(Equation 18) And the parameter can be identified with high accuracy. Further, if the same is performed in each posture, it is possible to absorb a change due to the posture.

The absorption of the change in the posture here means that identification is performed based on a model that takes into account non-linearity such as a change in the posture, instead of having a model for each posture as in the prior art 1. Therefore, according to the method of performing the micro motion, non-linearity such as a change in posture can be strictly absorbed. This method uses the first and second methods related to the rigidity identification.
It is needless to say that the present invention can be applied to any of the basic configurations . For example, FIG. 24 shows a first example of the rigidity identification .
A case of applying the basic configuration, Figure 25 is the stiffness identification
This is a case where the present invention is applied to the second basic configuration . FIG.
In 5, the arithmetic unit 12 calculates the moment M and each Jacobi matrix J
Although θ, Jε, and the like are calculated from the rotation angle θM of the motor, a joint angle, a joint angle command value, and a joint angle calculated back from the hand position x may be used instead of the motor rotation angle. This is the same for the device A6.

Further, the strict model or the simplified model can be applied to the first basic configuration relating to the rigidity identification . For example, the simplified model M1 shown in FIG.
For 0, it is as follows. That is, ε ″ _2SY , ε ″
The stiffness in the _2RY direction is the spring constant K ₂ ,
In the equations of κ ₂ and κ ₃ such as the equation (3) used for identifying K ₃ , the inertia parameters J ₂ , J ₃ and J ₂₃ are simply changed to the inertia parameters around the elastic deformation shown in FIG. I get it. The inertia parameter around the center of elastic deformation is
It can be calculated from known information such as the inertia parameters J ₂ , J ₃ , J ₂₃ , gravity parameters, the weight of each arm, and the position of the center of elastic deformation. Also, it is related to the above rigidity identification.
In the first basic configuration , the spring constant K is calculated based on the simple equation (4) by focusing on the natural vibration in the Y-axis direction.
Had identified _1, model focuses M10 in the Y-axis direction ε _"2SX, ε" of Figure 20 _{for 2SZ} direction of the elastic deformation interfere, respective configurations of simple expressions as described above Cannot be identified. However, the spring constant K
As if the _2, K ₃ have been interference, ε _"2SX, ε"
The rigidity in the _2SZ direction is expressed by the same formula as κ ₂ and κ ₃ and can be identified.

Further, in order to easily identify the rigidity in the ε ″ _2SX and ε ″ _2SZ directions, it is necessary to eliminate mutual interference. However, the easiest method is the position of an external sensor for measuring vibration. Is set in the X-axis direction or the Z-axis direction from the center of elastic deformation and identified. In this case, the stiffness can be individually identified, but the posture is limited, which may cause deterioration of the identification accuracy. As described above, in any case, the rigidity can be accurately identified.
As a result, accurate robot control can be performed based on the obtained parameters. Although the method for identifying the rigidity of a robot has been described above, the same can be applied to a link system other than a robot in actual use.

[0044]

Since the robot rigidity identification method and apparatus according to the present invention are configured as described above, it is possible to accurately identify rigidity. As a result, accurate robot control can be performed based on the obtained parameters.

[Brief description of the drawings]

FIG. 1 is a first basic configuration related to rigidity identification of the present invention .
And a schematic diagram for explaining the device A1 .

FIG. 2 is an explanatory diagram showing an operation example of the device A1.

Figure 3 is an explanatory diagram showing an operation example of the device A1.

FIG. 4 is a first basic configuration relating to rigidity identification of the present invention .
The schematic diagram which shows the modified example of FIG.

FIG. 5 shows a configuration related to identification of non-rigidity parameters , and
For explaining a schematic configuration of a Bisoreni applicable device A3
Schematic diagram of.

FIG. 6 shows a configuration related to identification of non-rigidity parameters , and
And a schematic configuration of an apparatus A4 applicable thereto.
Schematic diagram of.

FIG. 7 shows a configuration related to identification of non-rigidity parameters , and
And to explain the schematic configuration of the device A5 applicable thereto.
Schematic diagram of.

FIG. 8 shows a second basic structure related to rigidity identification of the present invention .
Formed, and schematic diagram for explaining a schematic structure of the device A6.

FIG. 9 is an explanatory diagram showing a strict model M1.

FIG. 10 is an explanatory diagram showing a strict model M2.

FIG. 11 is an explanatory diagram showing actual measured values and theoretical values of hand position displacement amounts.

FIG. 12 is an explanatory diagram showing a simplified model M3.

FIG. 13 is an explanatory diagram showing a simplified model M4.

FIG. 14 is an explanatory diagram showing a method for creating a simplified model M4.

FIG. 15 is an explanatory diagram showing a simplified model M5.

FIG. 16 is an explanatory diagram showing a simplified model M6.

FIG. 17 is an explanatory diagram showing a simplified model M7.

FIG. 18 is an explanatory diagram showing a simplified model M8.

FIG. 19 is an explanatory diagram showing a simplified model M9.

FIG. 20 is an explanatory diagram showing a simplified model M10.

FIG. 21 is an explanatory diagram showing measured values and theoretical values of the amount of deformation.

FIG. 22 is an explanatory diagram showing actual measured values and theoretical values of the amount of deformation.

FIG. 23 is an explanatory diagram showing a vibration response of a hand position amplitude.

FIG. 24 shows a first basic configuration related to the rigidity identification .
FIG. 3 is a schematic diagram showing a schematic configuration of such an apparatus A7.

FIG. 25 shows a second basic configuration related to the rigidity identification .
Schematic diagram showing a schematic configuration of a device A8 according.

FIG. 26 is a schematic diagram showing a schematic configuration of a device A0 to which a conventional robot rigidity identification method can be applied, and a block diagram of a control system thereof.

FIG. 27 is a schematic diagram showing a schematic configuration of a device A0 ′ to which a conventional robot rigidity identification method can be applied.

[Explanation of symbols]

A1 to A8 robot rigidity identification device 1 robot 2 sensor 3 frequency analysis unit (corresponding to frequency analysis means) 4 memory 5 identification unit (corresponding to first and second rigidity parameter identification means) 6 ... Forward conversion unit 7 Identification signal generation unit 8 Non-rigid parameter identification unit 9 Memory 10 Filter 11 Preprocessing unit 12 Operation unit (corresponding to matrix and moment operation means)

Continuation of front page (56) References JP-A-7-152429 (JP, A) JP-A-6-99380 (JP, A) JP-A-3-117580 (JP, A) JP-A-7-168605 (JP) JP-A-3-282718 (JP, A) JP-A-2-224991 (JP, A) JP-A-64-71681 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB Name) B24J 13/08 G05B 11/36 G05B 13/02

Claims (14)

(57) [Claims] 1. A method for identifying rigidity of a robot based on at least output data from a sensor attached to the robot, wherein each parameter representing a characteristic of the robot includes a rigidity parameter or a rigidity inside. Parameters and other non-rigid parameters, and output data from the external sensor and / or internal sensor.
Frequency analysis of the non-rigid parameter
Rigidity parameters based on the meter and the above frequency analysis results
When identifying the above non-rigidity parameter in the robot rigidity identification method for identifying
From the robot's operation command data,
Frequency component higher than natural frequency in rigid mode
Enter the operation command data after removing
Stiffness identification method of a robot characterized by comprising. 2. A method for identifying the rigidity of the robot based on at least output data from a sensor attached to the robot, wherein each parameter representing a characteristic of the robot includes a rigidity parameter or a rigidity inside. After separating the parameters from the other non-rigid parameters, from the elastic deformation part of the robot to the position of the external sensor
Computes the matrix corresponding to the Jacobi matrix or its inverse,
The above elastic deformation based on the non-rigid parameter identified separately
Calculates the moment applied to the part and outputs the output of the external sensor
Stiffness parameters based on data, moments and matrices
When identifying the above non-rigidity parameter in the robot rigidity identification method for identifying
From the robot's operation command data,
Frequency component higher than natural frequency in rigid mode
Enter the operation command data after removing
Stiffness identification method of a robot characterized by comprising. 3. At least a cell attached to the robot.
Based on the output data from the sensor
Is a method of identifying
Parameter is a stiffness parameter or a parameter that contains stiffness inside.
Parameters and other non-rigid parameters
The output data from the external and / or internal sensors
Frequency analysis of the non-rigid parameter
Rigidity parameters based on the meter and the above frequency analysis results
When identifying the above non-rigidity parameter in the robot rigidity identification method for identifying
Robot operation command data and output from the above sensor
Rigid mode that operates the robot as a rigid body from the data
After removing high frequency components above the natural frequency at
On the basis of the operation command data and the output data of
A robot rigidity identification method that identifies gender parameters. 4. At least a cell attached to the robot.
Based on the output data from the sensor
Is a method of identifying
Parameter is a stiffness parameter or a parameter that contains stiffness inside.
Parameters and other non-rigid parameters
Above, from the robot's elastic deformation to the position of the external sensor
Computes the matrix corresponding to the Jacobi matrix or its inverse,
The above elastic deformation based on the non-rigid parameter identified separately
Calculates the moment applied to the part and outputs the output of the external sensor
Stiffness parameters based on data, moments and matrices
When identifying the above non-rigidity parameter in the robot rigidity identification method for identifying
Robot operation command data and output from the above sensor
Rigid mode that operates the robot as a rigid body from the data
After removing high frequency components above the natural frequency at
On the basis of the operation command data and the output data of
A robot rigidity identification method that identifies gender parameters. 5. At least a cell attached to a robot.
Based on the output data from the sensor
Is a method of identifying
Parameter is a stiffness parameter or a parameter that contains stiffness inside.
Parameters and other non-rigid parameters
The output data from the external and / or internal sensors
Frequency analysis of the non-rigid parameter
Rigidity parameters based on the meter and the above frequency analysis results
When identifying the above non-rigidity parameter in the robot rigidity identification method for identifying
Robot operation command data and output from the above sensor
The data is used to calculate the rigidity of the robot as a rigid body.
The operation command after only the natural frequency component in the mode
The non-rigidity parameter based on the data and output data
Method of stiffness identification of robot characterized by identification of robot
Law. 6. At least a cell attached to a robot.
Based on the output data from the sensor
Is a method of identifying
Parameter is a stiffness parameter or a parameter that contains stiffness inside.
Parameters and other non-rigid parameters
Above, from the robot's elastic deformation to the position of the external sensor
Computes the matrix corresponding to the Jacobi matrix or its inverse,
The above elastic deformation based on the non-rigid parameter identified separately
Calculates the moment applied to the part and outputs the output of the external sensor
Stiffness parameters based on data, moments and matrices
When identifying the above non-rigid parameters in the robot rigidity identification method, the robot's operation command data and the output data from the sensor are used to identify the natural vibration in the rigid mode in which the robot operates as a rigid body. A rigidity identification method for a robot, characterized in that the non-rigidity parameter is identified based on the operation command data and the output data after only a few components. 7. When identifying the rigidity parameter, not only elastic deformation of the movable part of the robot in the movable direction but also a predetermined direction other than the movable direction of the movable part and an elastic direction of the support part in the predetermined direction are considered. The rigidity identification method for a robot according to any one of claims 1 to 6 , wherein the model is a model obtained by using the model. 8. An apparatus for identifying the rigidity of the robot based on at least output data from a sensor attached to the robot, wherein each parameter representing a characteristic of the robot includes a rigidity parameter or a rigidity inside. parameters and, after separated into a non-rigid parameters otherwise, the first stiffness parameter identifying each individual
Identification means and first non-rigid parameter identification means
The first stiffness parameter identifying means is an external sensor.
And / or frequency for output data from the internal sensor
A frequency analysis means for performing an analysis;
Non-rigid parameters separately identified by parameter identification means
Stiffness parameters are identified based on the frequency analysis results
In the rigidity identification device for a robot, the first non-rigidity parameter identifying means includes:
Robot operation command data used for parameter identification
The rigid mode in which the robot operates as a rigid body
The operation command data after removing high-frequency components higher than the frequency
Data to be input to the robot.
Bot rigidity identification device. 9. An apparatus for identifying the rigidity of the robot based on at least output data from a sensor attached to the robot, wherein each parameter representing the characteristic of the robot includes a rigidity parameter or a rigidity inside. parameters and, after separated into a non-rigid parameters otherwise, the second stiffness parameter identifying each individual
Identification means and second non-rigid parameter identification means
And the second stiffness parameter identifying means is provided for the robot.
Jacobi matrix from the elastic deformation part to the position of the external sensor or
Matrix calculation means for calculating a matrix corresponding to the inverse matrix;
Separately identified by the second non-rigid parameter identification means.
Based on the determined non-rigidity parameter
Moment calculating means for calculating the moment
Based on the output data, moment and matrix of the external sensor
A rigidity parameter identifying means for identifying a rigidity parameter of the robot based on the rigidity parameter.
Robot operation command data used for parameter identification
The rigid mode in which the robot operates as a rigid body
The operation command data after removing high-frequency components higher than the frequency
Data to be input to the robot.
Bot rigidity identification device. 10. At least attached to the robot
Based on the output data from the sensor,
This is a device that identifies the characteristics of a robot and expresses the characteristics of the robot.
Each parameter contains rigidity parameter or rigidity inside
Parameters and other non-rigid parameters
And a third stiffness parameter to identify each individually.
Data identification means and third non-rigidity parameter identification means
The third stiffness parameter identifying means is an external sensor.
And / or frequency for output data from the internal sensor
A frequency analysis means for performing an analysis, wherein the third non-rigid
Non-rigid parameters separately identified by parameter identification means
Stiffness parameters are identified based on the frequency analysis results
Robot rigidity identification device The third non-rigid parameter identification means is configured to detect the non-rigid parameter.
Used for parameter identification Robot operation command data and
From the output data from the above sensors, make the robot a rigid body
Frequency component higher than natural frequency in rigid mode
The operation command data and output data after removing
The above non-rigid parameters are identified based on the
An apparatus for identifying rigidity of a robot. 11. At least attached to the robot
Based on the output data from the sensor,
This is a device that identifies the characteristics of a robot and expresses the characteristics of the robot.
Each parameter contains rigidity parameter or rigidity inside
Parameters and other non-rigid parameters
And a fourth stiffness parameter to identify each individually.
Data identification means and fourth non-rigid parameter identification means
And the fourth stiffness parameter identifying means is provided for the robot.
Jacobi matrix from the elastic deformation part to the position of the external sensor or
Matrix calculation means for calculating a matrix corresponding to the inverse matrix;
Separately identified by the fourth non-rigid parameter identification means.
Based on the determined non-rigidity parameter
Moment calculating means for calculating the moment
Based on the output data, moment and matrix of the external sensor
Stiffness parameter of the robot
In the fixed device, The fourth non-rigid parameter identification means is configured to detect the non-rigid parameter.
Robot operation command data used for parameter identification and
From the output data from the above sensors, make the robot a rigid body
Frequency component higher than natural frequency in rigid mode
The operation command data and output data after removing
The above non-rigid parameters are identified based on the
An apparatus for identifying rigidity of a robot. 12. At least attached to the robot
Based on the output data from the sensor,
This is a device that identifies the characteristics of a robot and expresses the characteristics of the robot.
Each parameter contains rigidity parameter or rigidity inside
Parameters and other non-rigid parameters
And a fifth stiffness parameter to identify each
Data identification means and fifth non-rigid parameter identification means
The fifth stiffness parameter identifying means may be an external sensor.
And / or frequency for output data from the internal sensor
A frequency analysis means for performing an analysis, wherein the fifth non-rigid
Non-rigid parameters separately identified by parameter identification means
Stiffness parameters are identified based on the frequency analysis results
Robot rigidity identification device The fifth non-rigid parameter identification means may include a non-rigid parameter.
Robot operation command data used for parameter identification and
The output data from the above sensors is
Only the natural frequency component in the rigid mode operating as a body
Based on the operation command data and output data after
Identification of the above non-rigid parameters
A rigidity identification device for a robot, comprising: Claim 13 At least attached to the robot
Based on the output data from the sensor,
This is a device that identifies the characteristics of a robot and expresses the characteristics of the robot.
Each parameter contains rigidity parameter or rigidity inside
Parameters and other non-rigid parameters
And the sixth rigidity parameter to identify each individually.
Data identification means and sixth non-rigid parameter identification means
And the sixth stiffness parameter identifying means can
Jacobi matrix from the elastic deformation part to the position of the external sensor or
Matrix calculation means for calculating a matrix corresponding to the inverse matrix;
Separately identified by the sixth non-rigid parameter identification means.
Based on the determined non-rigidity parameter
Moment calculating means for calculating the moment
Based on the output data, moment and matrix of the external sensor
Stiffness parameter of the robot
In the fixed device, The sixth non-rigid parameter identification means may include the non-rigid parameter.
Robot operation command data used for parameter identification and
The output data from the above sensors is
Only the natural frequency component in the rigid mode operating as a body
Based on the operation command data and output data after
Note that the above non-rigidity parameters are identified.
A device for identifying the rigidity of a robot. 14. In identifying the above stiffness parameters,
Not only elastic deformation of the movable part of the robot in the movable direction,
A predetermined direction other than the movable direction of the movable part and a predetermined direction of the support part
8. A model in which an elastic direction toward the surface is also taken into consideration.
4. The robot rigidity identification device according to any one of 3.