JP2006146572A - Servo control apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To more accurately determine a frictional force that varies with time and temperature, thus making it possible to keep the control performance of position control or flexible control high at all times without depending on time and temperature variations. <P>SOLUTION: A servo control apparatus includes a simulation position controller 15 and a simulation speed controller 16, both of which have the same configurations as actual control; a mechanical model computing part 18 for simulating a control target; a comparison part 19 for comparing the actual control with simulations; a friction model creating part 20 for creating a friction model based on the output of the comparison part 19; and a compensating part 21 for compensating a friction compensation value based on the output of the friction model creating part 20. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a servo control device that controls a mechanical device such as a robot.

  Conventional servo control devices are designed to improve the positioning accuracy and trajectory accuracy due to servo delay and overshoot when performing high-speed and high-accuracy position control and flexible control. Compensation is performed in consideration of the frictional force that is one of the causes. For example, with respect to flexible control in a robot, in order to increase flexibility, a constant friction force is compensated at the time of transition to flexible control (see Patent Document 1). In FIG. 17, reference numeral 101 denotes a position command generation unit that outputs a position command. 102 is a position controller, 103 is a speed controller, 105 is an amplifier, 106 is a motor, 107 is a rotation detector, and 108 is a load machine. In 102 to 103, a position command is issued from the position command creating unit 101. The torque is input and output to the amplifier 105 to drive the motor 106 and the load machine 108. Further, feedback control is performed by the rotation detector 107. In such a configuration, when shifting from position control to flexible control, the integrated value accumulated in the integrator of the speed controller 103 is stored in the memory 104, and this value is assumed assuming that there is no influence of gravity. Is compensated for in the torque command as a friction compensation value during flexible control. At the same time, the integrator of the speed controller 103 is turned off. Thus, a constant frictional force is compensated at the time of transition to flexible control.

  In addition, in order to perform position control that does not cause overshoot, in the two-degree-of-freedom control that performs feedforward including the model to be controlled, the Coulomb friction force is estimated and compensated from the torque command output to the control target. (See Patent Document 2). In FIG. 18, reference numeral 201 denotes a command generation circuit that outputs a position command. Reference numeral 202 denotes a first position control circuit, 203 denotes a first speed control circuit, 204 denotes a motor simulation circuit, and the position command from the command generation circuit 201 is input by the 202 to 204, and the first speed is controlled. A feedforward calculation is performed with the torque command from the control circuit 203 and the simulated position response and simulated speed response from the motor simulation circuit 204 as outputs, and a feedforward command is given to the feedback portion. 207 is a second position control circuit, 208 is a second speed control circuit, 209 is a torque control circuit, 210 is a motor, 211 is a rotation detector, 212 is a load machine, and feedforward commands from the above 202 to 204 And the motor 210 and the load machine 212 are driven. Further, feedback control is performed by the rotation detector 211. In such a two-degree-of-freedom control configuration, the friction force is compensated using the friction model circuit 205 and the friction correction circuit 206. The friction correction circuit 206 extracts the Coulomb friction Tf from the torque command input to the torque control circuit 209 as shown in the following equation.

Tf = ((T2 + T3) / 2) -T1 (1)
However, T1 is a torque command immediately before starting, T2 is a torque command immediately after acceleration, and T3 is a torque command immediately before stopping.
At the same time, a reset value for resetting the integrator in the second speed control circuit 208 is also obtained. The friction model circuit 205 uses the output from the friction correction circuit 206 as a limit value, and outputs a compensation torque in response to a speed input from the rotation detector 211. When the compensation torque is output, the integrator in the second speed control circuit 208 is reset by the integral reset value from the friction correction circuit 206. In this way, the frictional force corresponding to the Coulomb friction accumulated in the integrator in the second speed control circuit 208 is extracted and compensated.

  In addition, there is a means for performing observer control that generates and compensates for a frictional force based on a friction coefficient obtained in advance (see Patent Document 3).

As described above, the conventional servo control device, when performing position control or flexible control, extracts the frictional force accumulated in the integrator, or obtains the frictional force based on the coefficient obtained in advance and compensates the friction. is there.
JP-A-9-76184 (page 2-3, FIG. 1) JP-A-8-331881 (page 6-10, FIGS. 1 to 12) JP 2002-178281 A (page 4-5, FIG. 3)

Patent Document 1 is designed to compensate for a certain frictional force at the time of transition to flexible control, and when the control target includes a speed reducer, the grease accumulated unevenly in the speed reducer is evenly adapted and the temperature rises. As a result, the viscous friction (viscosity coefficient of friction) changes with time and temperature, and the torque command is not output in proportion to the friction, so that there is a problem that the flexibility cannot be sufficiently exhibited.
Further, in the case of Patent Document 2, even when Coulomb friction changes with time, compensation is made for viscous friction with a large time change. There was a problem that the frictional force changing with time and temperature was not accurately determined.
In the case of Patent Document 3, a frictional force is generated and compensated based on a friction coefficient obtained in advance, and a temporal change and a temperature change of the friction coefficient are not taken into consideration. In addition, there is a problem that the friction coefficient cannot be switched dynamically with respect to time change and temperature change.

The present invention has been made in view of such problems, and by using a simulation, the frictional force that changes with time and temperature can be obtained more accurately, and the time change and without depending on the change of time and temperature. It is an object of the present invention to provide a servo control device and method that can dynamically switch the friction coefficient in response to a temperature change and always maintain high control performance of position control and flexible control.
It is another object of the present invention to perform friction compensation corresponding to a time change with a simple configuration without generating a friction model, and always maintain high control performance of position control and flexible control.
Furthermore, the object is to more accurately determine and compensate for the frictional force even when an unknown external force is applied due to contact work or collision.

  In order to solve the above problem, the present invention is configured as follows.

The invention according to claim 1 is a servo control device that includes a position controller and a speed controller, and controls a control object including a motor.
A simulation position controller having the same configuration as the position controller and inputting the same command;
A simulation speed controller having the same configuration as the speed controller;
A mechanical model calculation unit that simulates the control object and feeds back to the simulation position controller and the simulation speed controller;
A comparator for comparing the output of the speed controller and the output of the simulation speed controller;
A friction model generation unit that generates a friction model from the output of the comparison unit;
A compensation unit for performing compensation based on the output of the friction model generation unit;
It is characterized by comprising.

  According to a second aspect of the present invention, the comparison unit obtains a friction force estimation error from the output of the speed controller and the output of the simulation speed controller, and the friction force estimation error is determined based on a predetermined value. It is characterized by determining to update the friction model when it becomes larger.

  According to a third aspect of the present invention, the friction model generation unit is constant from a speed when a constant speed operation is performed at a plurality of speeds at regular time intervals and an average value of the constant speed output of the comparison unit. A friction model constant for each time is obtained and stored, and the compensation unit interpolates the stored friction model constant to obtain a friction model constant for each operation time and perform friction compensation. It is.

  According to a fourth aspect of the present invention, the friction model generation unit obtains and stores a friction model constant for each temperature using information from a temperature sensor provided in a control target, and the compensation unit stores the friction model constant. Further, by interpolating the friction model constant with respect to temperature, the friction model constant at the temperature detected by the temperature sensor is obtained to perform friction compensation.

The invention according to claim 5 is a servo control device that includes a position controller and a speed controller, and controls a control object including a motor.
A simulation position controller having the same configuration as the position controller and inputting the same command;
A simulation speed controller having the same configuration as the speed controller;
A mechanical model calculation unit that simulates the control object and feeds back to the simulation position controller and the simulation speed controller;
A comparator for comparing the output of the speed controller and the output of the simulation speed controller;
A friction separating unit that subtracts and separates the output of the comparison unit from the output of the speed controller;
A compensation unit for compensating the output of the comparison unit;
It is characterized by comprising.

The invention described in claim 6 is characterized in that an unknown external force estimation unit for estimating an unknown external force applied to the control target is provided.
According to the seventh aspect of the present invention, after setting a default value in the constant portion of the friction model so that the default friction model can be generated, the step 1 of turning on the servo power supply;
Step 2 for determining whether a predetermined time has passed at a predetermined time interval, measuring a time for determining friction model update, and compensating for friction without updating the friction model if the predetermined time has not passed ,
Based on the currently stored friction model, the comparison unit calculates the difference between the torque command Tref of the speed controller and the torque command Tsim of the simulation speed controller to obtain a torque command difference value ΔT, and the torque command difference value ΔT and Step 3 for storing time series data of motor speed v;
The torque command difference value ΔT is compared with a predetermined value ΔTmax determined in advance. When | ΔT | <ΔTmax, the friction model is updated without updating, and when | ΔT |> ΔTmax, the friction model is determined to be updated. Step 4 to
When updating the friction model, the torque command difference value ΔT is added to the friction force f ′ by the friction model to set the friction force f to f = f ′ + ΔT, and the friction force f and the motor speed v at that time are stored. Step 5 and
Obtaining a friction model equation by linear approximation based on the friction force f and the motor speed v;
Step 7 for obtaining the frictional force f corresponding to the motor speed v from the friction model equation as f = cv + f0 and compensating the obtained frictional force f as a compensation value;
If the servo power is not off, the process returns to step 2 and the above is repeated, and if the servo power is off, the process ends.
Further, the invention according to claim 8 is a step 1 of performing a familiar operation by repeated operation,
It is determined whether a predetermined time has passed at a predetermined time interval, the time for generating the friction model is measured, and if the predetermined time has not passed, the process returns to (Step 1) and the familiar operation is performed. ,
A predetermined constant speed operation is performed with the friction compensation off, and the constant speed output average value Tmean (v at the motor speed v from the constant speed output ΔT _i (i = 1, 2,..., N) from the comparison unit 19. ) As Tmean (v) = (ΔT ₁ + ΔT ₂ +... + ΔT _n ) / n;
Obtaining a friction model by linear approximation based on the motor speed v and the constant speed output average value Tmean (v);
Storing the friction model, and storing the viscous friction coefficient c and the Coulomb friction force f0 in the database as functions c (t _i ), f0 (t _i ) at time t _i (i = 1, 2,...); ,
It is determined that the friction model has converged when the friction model constants c (t _i ) and f0 (t _i ) have converged to a predetermined constant value. Step 6 returning to 1;
A servo control method comprising a step 7 in which a friction force f corresponding to a motor speed v is obtained from a friction model equation, and the actual operation is performed by compensating the obtained friction force f.
According to a ninth aspect of the present invention, in the step 7, the friction model formula is obtained by the following formula based on the data stored in accordance with the operation time t '.
f = c (t ') v + f0 (t') (8)
c (t ') = ((c (t _{i + 1} ) -c (t _i )) / (t _{i + 1} -t _i )} (t'-t _i ) + c (t _i )
(t _i ≦ t ′ <t _{i + 1} ), c (t ′) = c (t _e ) (t _e ≦ t ′) (9)
f0 (t ') = {(f0 (t _{i + 1} ) -f0 (t _i )) / (t _{i + 1} -t _i )} (t'-t _i ) + f0 (t _i )
(t _i ≦ t ′ <t _{i + 1} ), f0 (t ′) = f0 (t _e ) (t _e ≦ t ′) (10)
However, i = 1, 2, 3,..., And t _e is the time when the friction model converges. Further, in the invention according to claim 10, in the step 7, the friction model equation is expressed as temperature instead of time. Based on the data stored according to tp _i (i = 1, 2,...), the following equation is used.
f = c (tp ') v + f0 (tp')
c (tp ') = ((c (tp _{i + 1} ) -c (tp _i )) / (tp _{i + 1} -tp _i )} (tp'-tp _i ) + c (tp _i )
(tp _i ≤tp '<tp _{i + 1} ), c (tp') = c (tp _e ) (tp _e ≤tp ')
f0 (tp ') = ((f0 (tp _{i + 1} ) -f0 (tp _i )) / (tp _{i + 1} -tp _i )} (tp'-tp _i ) + f0 (tp _i )
(tp _i ≤tp '<tp _{i + 1} ), f0 (tp') = f0 (tp _e ) (tp _e ≤tp ')
However i = 1,2,3, is a ..., tp _e is the temperature at which the friction model has converged

According to the first and second aspects of the present invention, the necessity of the friction model correction and the friction model correction are performed by using a simulation and comparing with the actual control, so that the frictional force changing with time is optimal. Compensation can be realized, and the control performance of the position control and the flexible control can always be kept high without depending on the change of time.
According to the third aspect of the present invention, by generating the friction model by the specific operation in advance, it is possible to execute the more accurate friction compensation according to the time change.
Further, according to the invention described in claim 4, since the frictional force changing with temperature is obtained and compensated by using the temperature sensor, the control performance of the position control and the flexible control is always improved without depending on the temperature change. Can be maintained.
Further, according to the invention described in claim 5, it is possible to perform friction compensation corresponding to a time change with a simple configuration without generating a friction model, and always maintain high control performance of position control and flexible control. Can do.
According to the sixth aspect of the present invention, it is possible to more accurately determine and compensate for the frictional force even when an unknown external force is applied due to contact work or collision.
Further, according to the invention described in claims 7 to 9, it is possible to realize the optimum compensation for the frictional force changing with time, and the control performance of the position control and the flexible control always without depending on the change of time. Can be kept high.
According to the invention described in claim 10, since the frictional force changing with temperature is obtained and compensated by using the temperature sensor, the control performance of the position control and the flexible control is always improved without depending on the temperature change. Can be maintained.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a configuration diagram of a servo control apparatus according to a first embodiment of the present invention. 2 and 3 show examples of objects to which the present invention is applied. FIG. 2 shows an example of an XY table, and FIG. 3 shows an example of application to an industrial robot. For example, in the case of a control target as shown in FIGS. 2 and 3, these calculations are executed by the X-axis amplifier 302 and the Y-axis amplifier 303 which are module type servo amplifiers in the XY table. In the case of an industrial robot, the servo motor control system is generally executed by a multi-axis amplifier, so the following calculation is executed by the robot multi-axis amplifier 404 shown in the figure.

  Description will be made by paying attention to one axis among the servo amplifiers of the module type and the multi-axis mechanism as described above. In FIG. 1, 11 is a position controller, which inputs a position command and a position response of a controlled object, and outputs a speed command. FIG. 4 shows a configuration example of the position controller 11. Kp in the figure is a position loop gain. A speed controller 12 receives the speed command output from the position controller 11 and the position response of the control target, and outputs a torque command. FIG. 5 shows a configuration example of the speed controller 12. In the figure, Kv is a speed loop gain, T is an integration time constant, Jm is a moment of inertia of the motor, and s and 1 / s represent differentiation and integration in Laplace transform, respectively. Reference numeral 14 denotes a control object, which includes a motor and a load machine coupled to the motor, and is driven according to a torque command output from the speed controller 12.

  The present invention is characterized by having 15 to 22, which will be described below. A simulation position controller 15 has the same configuration as the position controller 11 and inputs the same command. A simulation speed controller 16 has the same configuration as the speed controller 12 and outputs a torque command. Reference numeral 18 denotes a mechanical model calculation unit which simulates the control object 14 and feeds back to the simulation position controller 15 and the simulation speed controller 16 in response to a torque command input from the simulation speed controller 16. FIG. 6 shows an example in which the control object 14 is approximated by a two-inertia system in the mechanical model calculation unit 18. Here, Jm is the inertia moment of the motor, N is the reduction ratio, Kc is the spring constant of the decelerator, Jl is the inertia moment on the load side, and 1 / s represents the integral in the Laplace transform. The mechanical model calculation unit 18 models the mechanism to be controlled as a two-inertia mechanism, but when gravity is applied, the gravity part is also modeled, except for the frictional force. To.

  Thus, the simulation can be executed by configuring the simulation position controller 15, the simulation speed controller 16, and the mechanical model calculation unit 18, and an ideal state quantity is obtained. The simulation is performed in parallel with the actual control processing of the position controller 11 and the speed controller 12, and as a result, the state quantity in the actual control and the ideal state quantity can be compared. Reference numeral 19 denotes a comparison unit, which detects a difference between actual control and simulation command torque by calculating a difference between the torque command of the speed controller 12 and the torque command of the simulation speed controller 16. A friction model generation unit 20 generates a friction model by obtaining the relationship between the output of the comparison unit 19 and the motor speed. A compensation unit 21 obtains a friction compensation value corresponding to the motor speed based on the friction model generated by the friction model generation unit 20, and adds the friction compensation value to the output of the speed controller 12 to obtain a friction force. To compensate. Then, a torque command in which the frictional force is compensated is output, and the control object 14 is controlled through a torque control circuit (not shown).

Next, the procedure for carrying out the present invention will be described with reference to the flowchart of FIG.
Turn on the servo power. At this time, a default value is already set in the constant part of the friction model, and a default friction model can be generated. (Step 1)
It is determined whether a predetermined time has passed at a predetermined time interval, and the time for determining the friction model update is measured. If the fixed time has not passed, friction compensation is performed without updating the friction model. (Step 2)
Based on the currently stored friction model, actual control processing of the position controller 11 and speed controller 12 and simulation of the simulation position controller 15, simulation speed controller 16, and mechanical model computing unit 18 when friction compensation is performed. Compare by. That is, the comparison unit 19 calculates the difference between the torque command Tref of the speed controller 12 and the torque command Tsim of the simulation speed controller 16 as follows, and outputs a torque command difference value ΔT. At this time, time series data of the torque command difference value ΔT and the motor speed v is stored.
ΔT = Tref-Tsim (2)

FIG. 8A shows a graph of the command Tref and the torque command Tsim, and FIG. 8B shows a graph of the torque command difference value ΔT. When the torque command difference value ΔT, that is, the frictional force estimation error is larger than a predetermined value, it indicates that the frictional force has changed and a deviation from the friction model has occurred. (Step 3)
It is determined whether to update the friction model by detecting a deviation from the friction model due to a change in the friction force. That is, as shown in FIG. 8B, it is compared whether the torque command difference value ΔT is larger than a predetermined value ΔTmax. In the mathematical expression, the determination is made by the following expression in consideration of the sign of the torque command difference value ΔT.
| ΔT |> ΔTmax (3)
If equation (3) is not satisfied, friction compensation is performed without updating the friction model. (Step 4)
When it is necessary to update the friction model, it is necessary to obtain the relationship between the friction force and the motor speed and update the friction model formula. For that purpose, the torque command difference value ΔT obtained in (Step 3) is added to the frictional force f ′ by the friction model, and the frictional force f is obtained again by the following equation.
f = f '+ ΔT (4)
The frictional force f and the motor speed v at that time are stored. (Step 5)
The friction model generation unit 20 obtains a friction model formula from the friction force f and the motor speed v. FIG. 9 is a graph showing the relationship between the speed and the friction force when obtaining the friction model equation. The friction force f and the motor speed v obtained in (Step 5) are plotted, and the least square method or the like is plotted. A friction model equation is obtained by linear approximation using. That is, the present viscous friction coefficient c and Coulomb friction force f0 are obtained. (Step 6)
The compensation unit 21 obtains the friction force f corresponding to the motor speed v from the friction model equation generated and updated by the friction model generation unit 20 as follows, and compensates the obtained friction force f as a compensation value. . (Step 7)
f = cv + f0 (5)
If the servo power is not off, return to (Step 2) and repeat the above. If the servo power is off, finish. (Step 8)
In the above example, ΔT is obtained every certain time and the frictional force estimation error is monitored. However, the frictional force estimation error is always monitored, and when the frictional force estimation error exceeds a predetermined value, the friction model is monitored. May be updated.

  In this way, it is possible to monitor the change in the friction force at regular intervals and update the friction model to perform the friction compensation according to the change in time, so that the control performance can always be kept high.

FIG. 10 is a block diagram of the servo control apparatus according to the second embodiment of the present invention. The torque limiter 13, the model torque limiter 17, and the unknown external force estimating unit 22 are added to the first embodiment. This is particularly effective during flexible control where a torque command is restricted or an unknown external force is applied by contact or collision.
The torque limiter 13 limits the torque command, which is the output of the speed controller 12, with a predetermined value. The model torque limiter 17 has the same configuration as the torque limiter, and limits the torque command from the simulation speed controller 16 by a predetermined value. The unknown external force estimation unit 22 estimates an unknown external force applied to the control target 14 from the torque command output from the compensation unit 21 and the position response of the control target 14. FIG. 11 shows a configuration example of the unknown external force estimation unit 22. In the figure, Jm is the moment of inertia of the motor, and s represents the differential in Laplace transform. The dead zone is for separating the unknown external force and the change in the frictional force. When the absolute value of the input exceeds a certain set value, the input value is output as an unknown external force.
By inputting the estimated unknown external force to the mechanical model calculation unit 18, when an unknown external force is received, such as when it collides with a contact operation or an obstacle, the simulation is reflected, and the simulation excluding the friction force is performed. Make it accurate. Then, as in the first embodiment, the frictional force is obtained and compensated in comparison with the actual control.

  In this way, even when an unknown external force is received due to a contact operation or a collision, the simulation excluding the frictional force can be performed accurately, so that the frictional force can be determined and compensated more accurately.

  Since the configuration of the third embodiment is the same as that of the first embodiment, description thereof is omitted.

In the third embodiment, in order to generate a more accurate friction model, the friction model is generated off-line by a specific operation in advance, instead of generating the friction model online as in the first embodiment. The implementation procedure will be described with reference to the flowchart of FIG.
Familiar operation is performed by a specific repetitive action. (Step 1)
It is determined whether a predetermined time has passed at a predetermined time interval, and the time for generating the friction model is measured. If the predetermined time has not elapsed, the process returns to (Step 1) and the familiar operation is performed. (Step 2)
A constant speed operation is performed at a predetermined number of patterns, and a comparison is made between actual control and simulation. At this time, friction compensation is turned off. Then, the constant speed output average value Tmean (v) at the motor speed v is obtained from the output ΔT _i (i = 1, 2,..., N) at the constant speed from the comparison unit 19 as follows. (Step 3)
Tmean (v) = (ΔT ₁ + ΔT ₂ + ... + ΔT _n ) / n (6)
A friction model at that time is determined from the motor speed v and the constant speed output average value Tmean (v) by linear approximation as in the first embodiment. (Step 4)
The friction model obtained in (Step 4) is stored. That is, the viscous friction coefficient c and the Coulomb friction force f0 in the equation (5) are stored in the database as functions c (t _i ), f0 (t _i ) at time t _i (i = 1, 2,...). (Step 5)
Whether the friction model constants c (t _i ) and f 0 (t _i ) have converged to a constant value is determined as follows, and it is determined whether the friction model has converged.
| c (t _i ) -c (t _i-1 ) | <e _c and | f0 (t _i ) -f0 (t _i-1 ) | <e _f0 (7)
However, i = 2, 3,..., And e _c and e _f0 are predetermined convergence determination values. If not converged, return to (Step 1). (Step 6)
The compensation unit 21 obtains the friction force f corresponding to the motor speed v from the friction model equation generated by the friction model generation unit 20 as in the first embodiment, and performs the operation by compensating the obtained friction force f. At this time, the friction model formula is changed to the following formula based on the data stored according to the operation time t ′.
f = c (t ') v + f0 (t') (8)
c (t ') = ((c (t _{i + 1} ) -c (t _i )) / (t _{i + 1} -t _i )} (t'-t _i ) + c (t _i )
(t _i ≦ t ′ <t _{i + 1} ), c (t ′) = c (t _e ) (t _e ≦ t ′) (9)
f0 (t ') = {(f0 (t _{i + 1} ) -f0 (t _i )) / (t _{i + 1} -t _i )} (t'-t _i ) + f0 (t _i )
(t _i ≦ t ′ <t _{i + 1} ), f0 (t ′) = f0 (t _e ) (t _e ≦ t ′) (10)
However, i = 1, 2, 3,..., And t _e is the time when the friction model converges. (Step 7)

  In this way, by generating a friction model by a specific operation in advance, it is possible to execute more accurate friction compensation according to a change in time, and it is possible to always maintain high control performance.

FIG. 13 is a configuration diagram of Embodiment 4 of the servo control device of the present invention. The difference from the third embodiment is that a temperature sensor 23 is attached to the controlled object 14, and the friction model is determined based on the temperature information of the temperature sensor 23. Instead of the time in the third embodiment, the constants c (tp _i ), f0 (tp _i ) (i = 1, 2,...) Of the friction model equation based on the temperature tp _i (i = 1, 2,...) Are obtained and stored. Keep it. However, the temperatures tp _i , c (tp _i ), and f0 (tp _i ) are stored in ascending order of the temperature tp _i . Similarly to the third embodiment, the frictional force f corresponding to the motor speed v is obtained, and the operation is performed by compensating the obtained frictional force f. At this time, the friction model equation is changed as the following equation based on the data stored in accordance with the temperature tp ′ of the temperature sensor 23.
f = c (tp ') v + f0 (tp') (11)
c (tp ') = ((c (tp _{i + 1} ) -c (tp _i )) / (tp _{i + 1} -tp _i )} (tp'-tp _i ) + c (tp _i )
(tp _i ≤tp '<tp _{i + 1} ), c (tp') = c (tp _e ) (tp _e ≤tp ') (12)
f0 (tp ') = ((f0 (tp _{i + 1} ) -f0 (tp _i )) / (tp _{i + 1} -tp _i )} (tp'-tp _i ) + f0 (tp _i )
(tp _i ≤tp '<tp _{i + 1} ), f0 (tp') = f0 (tp _e ) (tp _e ≤tp ') (13)
However i = 1,2,3, is ..., is tp _e is the temperature at which the friction model has converged.

  In this way, by generating a frictional force that changes with temperature, friction compensation can be performed according to the temperature change, and the control performance of the position control and flexible control is always kept high regardless of the temperature change. can do.

  FIG. 14 is a configuration diagram of Embodiment 5 of the servo control apparatus of the present invention. The difference from the first embodiment is that the friction model generation unit 20 is not provided, and the friction separation unit 24 is provided in the preceding stage of the compensation unit 21, and the output of the comparison unit 19 is output from the torque command that is the output of the speed controller 12. The frictional force is separated by subtracting, and the compensation unit 21 adds the output of the comparison unit 19 to the output of the frictional separation unit 24 to compensate the frictional force. That is, by obtaining the torque command for acceleration for separating the friction force from the torque command using the output of the comparison unit 19 as a friction force and following the position command, and compensating the friction force in the torque command for acceleration, This means that the friction force is always compensated without generating a friction model. This is effective when the torque limiter is used only for the torque command for acceleration as described below.

  As in the second embodiment, a torque limiter 13, a model torque limiter 17, and an unknown external force estimation unit 22 can be added as shown in FIG. In the fifth embodiment, this configuration including a torque limiter is more general. Particularly in the case of flexible control, since the torque limit value of the torque limiter 13 is small, it is effective to limit the torque only to the acceleration component by separating the torque command into the acceleration component and the friction component. This will be described with reference to the graph of FIG. FIG. 16 shows a torque waveform in the configuration of FIG. The torque command difference shown in FIG. 16B from the torque command Tref which is the output of the speed controller 12 shown in FIG. 16A and the torque command Tsim which is the output of the simulation speed controller 16 shown in FIG. A value ΔT is determined. Since the torque command difference value ΔT corresponds to a frictional force, the friction separation unit 24 takes the difference between the torque command Tref and the torque command difference value ΔT to obtain an acceleration amount as shown in FIG. The extracted torque command can be obtained. By applying a torque limiter to this and adding the torque command difference value ΔT, a friction-compensated torque command can be generated as shown in FIG. Even when the torque limit such as the flexible control is small, the torque command is separated into the acceleration component and the friction component, and the desired torque command can be output by compensating the friction component after applying the torque limitation to the acceleration component. .

  In this way, friction compensation corresponding to time changes can be performed with a simple configuration without generating a friction model, and control performance can always be maintained high.

  The present invention is an apparatus for performing position control and flexible control of industrial robots and consumer robots having a friction element in a mechanical part in order to more accurately determine and compensate for friction force that changes with time and temperature using simulation. Widely applicable with respect to.

1 is a configuration diagram of a servo control device showing Embodiment 1 of the present invention. The image figure (XY stage) showing the 1st application example of this invention Image diagram showing a second application example of the present invention (robot) Block diagram of the position controller of the present invention Block diagram of the speed controller of the present invention Block diagram of mechanical model calculation unit of the present invention The flowchart which shows the implementation procedure of Example 1 of this invention. Explanatory drawing which shows the torque command comparison method in Example 1 of this invention. Explanatory drawing which shows the friction model production | generation of this invention Configuration diagram of servo control apparatus showing Embodiment 2 of the present invention Block diagram of unknown external force estimation unit of the present invention The flowchart which shows the implementation procedure of Example 3 of this invention. Configuration diagram of servo control apparatus showing Embodiment 4 of the present invention Basic configuration diagram of servo control apparatus showing Embodiment 5 of the present invention Application configuration diagram of servo control apparatus showing Embodiment 5 of the present invention Explanatory drawing which shows the torque command production | generation method in Example 5 of this invention. Configuration diagram of a conventional servo control device (Patent Document 1) Configuration diagram of conventional servo control device (Patent Document 2)

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Servo control apparatus 11, 102 Position controller 12, 103 Speed controller 13 Torque limiter 14 Control object 15 Simulation position controller 16 Simulation speed controller 17 Model torque limiter 18 Mechanical model calculating part 19 Comparison part 20 Friction model production | generation part 21 Compensation Unit 22 unknown external force estimation unit 23 temperature sensor 24 friction separation unit 101 position command generation unit 104 memory 105 amplifier 106, 210 motor 107, 211 rotation detector 108, 212 load machine 201 position generation circuit 202 first position control circuit 203 first 1 speed control circuit 204 motor simulation circuit 205 friction model circuit 206 friction correction circuit 207 second position control circuit 208 second speed control circuit 209 torque control circuit 301 control arithmetic unit 302 X-axis amplifier 303 Y-axis amplifier 304 X Y-table device 305 X-axis motor 306 Y-axis motor 307 Table 401 Articulated robot 402 Robot control panel 403 Robot multi-axis control arithmetic unit 404 Robot multi-axis amplifier

Claims (10)

In a servo control device that includes a position controller and a speed controller and controls a control target including a motor,
A simulation position controller having the same configuration as the position controller and inputting the same command;
A simulation speed controller having the same configuration as the speed controller;
A mechanical model calculation unit that simulates the control object and feeds back to the simulation position controller and the simulation speed controller;
A comparator for comparing the output of the speed controller and the output of the simulation speed controller;
A friction model generation unit that generates a friction model from the output of the comparison unit;
A compensation unit for performing compensation based on the output of the friction model generation unit;
A servo control device comprising:   The comparison unit obtains a friction force estimation error from the output of the speed controller and the output of the simulation speed controller, and updates the friction model when the friction force estimation error becomes larger than a predetermined value. 2. The servo control device according to claim 1, wherein the servo control device is determined as follows.   The friction model generation unit obtains and stores a friction model constant at a fixed time from a speed when a constant speed operation is performed at a plurality of speeds at a fixed time and an average value of the constant speed output of the comparison unit. 2. The servo control apparatus according to claim 1, wherein the compensation unit interpolates the stored friction model constant to obtain a friction model constant for each operation time and perform friction compensation.   The friction model generation unit obtains and stores a friction model constant for each temperature using information of a temperature sensor provided in a control target, and the compensation unit interpolates the stored friction model constant with respect to temperature. 4. The servo control device according to claim 3, wherein friction compensation is performed by obtaining a friction model constant at the temperature detected by the temperature sensor. In a servo control device that includes a position controller and a speed controller and controls a control target including a motor,
A simulation position controller having the same configuration as the position controller and inputting the same command;
A simulation speed controller having the same configuration as the speed controller;
A mechanical model calculation unit that simulates the control object and feeds back to the simulation position controller and the simulation speed controller;
A comparator for comparing the output of the speed controller and the output of the simulation speed controller;
A friction separating unit that subtracts and separates the output of the comparison unit from the output of the speed controller;
A compensation unit for compensating the output of the comparison unit;
A servo control device comprising:   6. The servo control apparatus according to claim 1, further comprising an unknown external force estimation unit that estimates an unknown external force applied to the control target. After setting a default value in the constant part of the friction model so that a default friction model can be generated, step 1 of turning on the servo power supply,
Step 2 for determining whether a predetermined time has passed at a predetermined time interval, measuring a time for determining friction model update, and compensating for friction without updating the friction model if the predetermined time has not passed ,
Based on the currently stored friction model, the comparison unit calculates the difference between the torque command Tref of the speed controller and the torque command Tsim of the simulation speed controller to obtain a torque command difference value ΔT, and the torque command difference value ΔT and Step 3 for storing time series data of motor speed v;
The torque command difference value ΔT is compared with a predetermined value ΔTmax determined in advance. When | ΔT | <ΔTmax, the friction model is updated without updating, and when | ΔT |> ΔTmax, the friction model is determined to be updated. Step 4 to
When updating the friction model, the torque command difference value ΔT is added to the friction force f ′ by the friction model to set the friction force f to f = f ′ + ΔT, and the friction force f and the motor speed v at that time are stored. Step 5 and
Obtaining a friction model equation by linear approximation based on the friction force f and the motor speed v;
Step 7 for obtaining the frictional force f corresponding to the motor speed v from the friction model equation as f = cv + f0 and compensating the obtained frictional force f as a compensation value;
When the servo power is not off, the servo control method includes step 8 which returns to step 2 and repeats the above, and ends when the servo power is off. Step 1 for running-in by repeated operation,
It is determined whether a predetermined time has elapsed at a predetermined time interval, the time for generating the friction model is measured, and if the predetermined time has not elapsed, the process returns to step 1 to perform the familiar operation;
A predetermined constant speed operation is performed with the friction compensation off, and the constant speed output average value Tmean (v at the motor speed v from the constant speed output ΔT _i (i = 1, 2,..., N) from the comparison unit 19. ) As Tmean (v) = (ΔT ₁ + ΔT ₂ +... + ΔT _n ) / n;
Obtaining a friction model by linear approximation based on the motor speed v and the constant speed output average value Tmean (v);
Storing the friction model, and storing the viscous friction coefficient c and the Coulomb friction force f0 in the database as functions c (t _i ), f0 (t _i ) at time t _i (i = 1, 2,...); ,
It is determined that the friction model has converged when the friction model constants c (t _i ) and f0 (t _i ) have converged to a predetermined constant value. Step 6 returning to 1;
A servo control method comprising a step 7 in which a friction force f corresponding to a motor speed v is obtained from a friction model equation, and the actual operation is performed by compensating the obtained friction force f. 9. The servo control method according to claim 8, wherein in step 7, the friction model formula is obtained by the following formula based on data stored in accordance with the operation time t '.
f = c (t ') v + f0 (t')
c (t ') = ((c (t _{i + 1} ) -c (t _i )) / (t _{i + 1} -t _i )} (t'-t _i ) + c (t _i )
(t _i ≤t '<t _{i + 1} ), c (t') = c (t _e ) (t _e ≤t ')
f0 (t ') = {(f0 (t _{i + 1} ) -f0 (t _i )) / (t _{i + 1} -t _i )} (t'-t _i ) + f0 (t _i )
(t _i ≤t '<t _{i + 1} ), f0 (t') = f0 (t _e ) (t _e ≤t ')
Where i = 1,2,3, ... and t _e is the time when the friction model converges 9. The servo control according to claim 8, wherein in step 7, the friction model equation is obtained by the following equation based on data stored in accordance with the temperature tp _i (i = 1, 2,...) Instead of time. Method.
f = c (tp ') v + f0 (tp')
c (tp ') = ((c (tp _{i + 1} ) -c (tp _i )) / (tp _{i + 1} -tp _i )} (tp'-tp _i ) + c (tp _i )
(tp _i ≤tp '<tp _{i + 1} ), c (tp') = c (tp _e ) (tp _e ≤tp ')
f0 (tp ') = ((f0 (tp _{i + 1} ) -f0 (tp _i )) / (tp _{i + 1} -tp _i )} (tp'-tp _i ) + f0 (tp _i )
(tp _i ≤tp '<tp _{i + 1} ), f0 (tp') = f0 (tp _e ) (tp _e ≤tp ')
However i = 1,2,3, is a ..., tp _e is the temperature at which the friction model has converged

Images