JP6502634B2 - Control device - Google Patents

Description

  The present invention relates to a technique for controlling a control target that can be modeled as a so-called two-inertia system.

  In the following, (n) on the right shoulder indicates a reference number at the end of the specification.

In recent years, high resolution and low cost encoders have been developed, and inexpensive high resolution encoders have come to be used in various industrial fields. Conventionally, in the field of machine tools, industrial robots, welfare robots, etc., semi-closed control is generally performed in which drive side information is fed back using a drive side encoder ^{(1) (2)} . However, in the semi-closed control, the axial torsion between the motor and the load degrades the positioning accuracy on the load side for final positioning. Since the final positioning accuracy can be improved by performing full-closed control by using an encoder for feedback not only on the drive side but also on the load side, encoders are also used on the load side in the field of machine tools etc. ing. In the industrial robot, it has been considered difficult to attach an encoder on the load side, but in recent years, industrial robot modules having an encoder on the load side have also been proposed. Thus, it is thought that the use of the apparatus provided with the encoder on the load side will be expanded from now on aiming at the improvement of the final positioning accuracy.

Despite the industrial flow like this, it is difficult to say that research using the load side encoder has been sufficiently performed. Therefore, the inventors have proposed a control method using a load side encoder ^{(3) to (5)} . In the future, the cost reduction of encoders is expected to further expand the use of encoders on the load side, and a new control method in a system that can obtain encoder information on both the drive side and the load side is required.

Therefore, the present inventors are studying an axial torque control method using a load side encoder. By controlling the axial torque in the two-inertia system, it is considered that in the industrial robot, high-level operations such as assembly operations become possible. In addition, in the robot for welfare or EV (Electric Vehicle), it is possible to improve the convenience and the safety, and it is considered that the back drivability can be realized ^{(6, 7)} .

In the field of wearable robots and the like where back drivability is required, Serial Elastic Actuators (SEA) are often used ^{(8) (9)} . Generally, when the control target has an element with low rigidity, the control performance is degraded, so it is desirable to make the rigidity of the control target as high as possible ⁽¹⁰⁾ . The SEA can be said to be an actuator with back drivability secured at the expense of control performance by having a flexible spring element.

  Also, as a conventional shaft torque control method, there is a method of controlling shaft torque using a torque sensor. However, if a torque sensor is used, the rigidity of the entire system is reduced because the sensor portion is deformed.

  Therefore, there is a need for a method of controlling the axial torque without using a torque sensor that causes a reduction in rigidity and without changing the hardware as in the SEA. According to the present invention, it is considered possible to realize the back drivability of a servomotor having a reduction gear.

  The present invention has been made in view of the above situation. The main object of the present invention is to provide a technique for controlling the power transmitted by the power transmission unit using position information on the load side.

  The means for solving the above-mentioned problems can be described as the following items.

(Item 1)
A drive unit, a power transmission unit, a load unit, a drive-side position detection unit, a load-side position detection unit, and a control unit;
The power transmission unit is configured to transmit power from the drive unit to the load unit.
The drive side position detection unit is configured to obtain drive side position information in the drive unit,
The load-side position detection unit is configured to acquire load-side position information in the load unit,
The control unit is configured to control the power transmitted by the power transmission unit by calculating a command value to the drive unit using the load-side position information and the drive-side position information. Control unit.

(Item 2)
The control unit includes a non-linear element compensation FF unit,
The control device according to Item 1, wherein the non-linear element compensation FF unit is configured to positively feed forward compensate the non-linear element in the power transmission unit.

(Item 3)
The control unit includes a drive side position FB unit,
The control device according to Item 1 or 2, wherein the drive-side position FB unit performs feedback control of the drive-side position using the drive-side position information.

(Item 4)
The control unit includes a power FB unit,
The power FB unit is configured to perform feedback control of an estimated value of power in the power transmission unit using the estimated power estimated using the drive-side position information. Control device as described.

(Item 5)
The control unit includes a drive side position FF unit,
The drive-side position FF unit raises the control band at the drive-side position by performing feed-forward compensation on the command value to the drive unit such that the actual position matches the position command value. The control device according to any one of items 1 to 4, which is configured.

(Item 6)
The drive unit is a rotary motor,
The power transmission unit is configured to transmit the power using a rotary shaft,
The control device according to any one of Items 1 to 5, wherein the power is an axial torque at the rotating shaft.

(Item 7)
The control device according to any one of Items 1 to 6, wherein the position is an angle or an angular velocity.

(Item 8)
The drive unit is a linear motor,
The power transmission unit is configured to transmit a translational force from the drive unit to the load side.
The control device according to any one of Items 1 to 5, wherein the power is a translational force transmitted by the power transmission unit.

(Item 9)
A control device including a drive unit, a power transmission unit, a load unit, a drive-side position detection unit, a load-side position detection unit, and a control unit is used.
The power transmission unit is configured to transmit power from the drive unit to the load unit.
The drive side position detection unit is configured to obtain drive side position information in the drive unit,
The load-side position detection unit is configured to acquire load-side position information in the load unit,
The control unit controls the power transmitted by the power transmission unit by calculating a command value to the drive unit using the load-side position information and the drive-side position information. Control method.

(Item 10)
A computer program for causing a computer to execute the steps according to item 9.

  This computer program can be used as an appropriate recording medium (eg, an optical recording medium such as a CD-ROM or a DVD disk, a magnetic recording medium such as a hard disk or a flexible disk, or a magneto-optical recording medium such as an MO disk). It can be stored. This computer program can be transmitted via a communication line such as the Internet. It is possible to realize the configuration of the control unit using a computer program.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to provide the technique which controls the motive power transmitted with a power transmission part using the positional information on the load side.

FIG. 1 is a block diagram conceptually showing a control device according to an embodiment of the present invention. It is explanatory drawing which represented 2 inertia systems notionally. It is a block diagram at the time of modeling 2 inertia systems. FIG. 2 is a block diagram of Example 1; It is a graph which shows the target value response frequency characteristic of axial torque in Example 1. In Example 1, it is a graph which shows the comparison by the presence or absence of axial torque FF control when there is no angular velocity FF of the step response of axial torque. In Example 1, it is a graph which shows the comparison by the presence or absence of angular velocity FF control when it is set as axial torque FF of step response of axial torque. In Example 1, it is a graph which shows the step response of axial torque. This graph represents a comparison with and without nonlinear compensation in the axial torque FF. In Example 1, it is a graph which shows the sine wave response of axial torque. This graph represents a comparison with and without nonlinear compensation in the axial torque FF. 5 is a graph showing the feasibility of back drivability in the first embodiment. FIG. 10 (a) shows a comparison of shaft torque responses depending on the presence or absence of the proposed control when load side disturbance is applied. FIG. 10 (b) shows a comparison of the load side angle according to the presence or absence of the proposed control when load side disturbance is introduced. It is a graph which shows the target value response frequency characteristic of axial torque in an example of an experiment. In an experimental example, it is a graph which shows the comparison by axial torque FF presence or absence about the step response of axial torque. In an experimental example, it is a graph which shows the comparison by angular velocity FF presence or absence about the step response of axial torque. FIG. 7 is a block diagram of Example 2; FIG. 7 is a block diagram of Example 3; It is a table | surface for demonstrating Example 1-3 mutual relationship.

(Embodiment)
A control device according to an embodiment of the present invention will be described with reference to FIG. The control device includes a drive unit 1, a power transmission unit 2, a load unit 3, a drive-side position detection unit 4, a load-side position detection unit 5, and a control unit 6.

  The power transmission unit 2 is configured to transmit power from the drive unit 1 to the load unit 3. For example, various mechanisms such as a gear, a belt, and a ball screw can be used as the power transmission unit 2. Moreover, in this embodiment, although the power transmission part 2 with flexibility is assumed, since it is difficult to assume a physically perfect rigid body, all the power transmission mechanisms that exist in reality have flexibility. Can.

  The drive side position detection unit 4 is configured to acquire drive side position information in the drive unit 1. The drive side position detection unit 4 can be realized, for example, using an encoder.

  The load-side position detection unit 5 is configured to acquire load-side position information in the load unit 3. The load-side position detection unit 5 can also be realized using, for example, an encoder.

  Here, in this embodiment, the position is not limited to the position or velocity in the translational direction, but is used in the meaning including the angle or angular velocity of the rotating member.

  The drive unit 1 is, for example, a rotary motor. In this case, the power transmission unit 2 can be configured to transmit power using the rotation shaft. Further, in this case, power can be grasped as shaft torque in the rotating shaft.

  The drive unit 1 may be, for example, a linear motor. In this case, the power transmission unit 2 can be configured to transmit a translational force from the drive unit 1 to the load side. Also, in this case, the power can be grasped as a translational force transmitted by the power transmission unit 2.

  Further, while the drive unit 1 is a rotary motor, the power transmission unit 2 can be configured using a ball screw mechanism that generates a translational force using a rotation shaft. In this case, the power can be grasped as an axial torque on the rotary shaft or a translational force generated by the ball screw mechanism.

(Control unit)
The control unit 6 is configured to control the power transmitted by the power transmission unit by calculating a command value to the drive unit using the load side position information and the drive side position information.

  The control unit 6 includes a non-linear element compensation FF unit 61, a drive-side position FB unit 62, a power FB unit 63, and a drive-side position FF unit 64.

  The non-linear element compensation FF unit 61 is configured to positively feed forward compensate the non-linear element in the power transmission unit 2. Further, the non-linear element compensation FF unit 61 of the present embodiment explicitly performs feed-forward compensation not only on non-linear elements but also on linear elements.

  The drive side position FB unit 62 performs feedback control of the drive side position using the drive side position information.

  The power FB unit 63 is configured to perform feedback control of the estimated value of the power in the power transmission unit using the estimated power estimated using the drive side position information.

  The drive-side position FF unit 64 is configured to raise the control band at the drive-side position by performing feed-forward compensation so that the actual position matches the position command value with respect to the command value to the drive unit. There is.

  The specific configuration and operation of the control unit 6 will be described in more detail in the following embodiment.

Example 1
Hereafter, the Example which actualized the control apparatus of above-mentioned embodiment is described. In the following description of the embodiment, the same reference numerals are used for elements that are basically in common with the reference numerals used in the above-described embodiment.

(About 2 inertia system)
First, a plant to which the present embodiment is applied is a so-called two-inertia system. Here, the concept of the two-inertia system will be described with reference to FIG. In this plant, a rotary motor is used as the drive unit 1. The drive unit 1 is configured to rotationally drive the load unit 3 via the power transmission unit 2 having rigidity (shaft torsional rigidity in the case of a rotation shaft) K.

A block diagram obtained by modeling the plant of FIG. 2 is shown in FIG. In FIG. 3, the inertia moment, viscous friction coefficient, and input torque of the drive motor (drive unit) are J _M , D _M , and T _M, and the inertia moment, viscous friction coefficient, disturbance of the load side (load portion) is J _L , D _L , d _L , the axial torsional rigidity of the rotary shaft is K, and the reduction gear ratio of the power transmission unit 2 is R. Further, the axial torque Ts, the driving side angle is theta _M, the drive-side angular velocity omega _M, load angle theta _L, the load-side angular velocity omega _L, torsional angle delta _theta.

The plant contains various non-linear elements ⁽¹¹⁾ . In FIG. 3, the backlash which is one of the non-linear elements included in the power transmission unit 2 of the plant is modeled and expressed as a dead zone.

  In addition, an example of the parameter value which the plant of FIG. 3 can take is shown in the following Table 1.

(Preliminary research on shaft torque control)
The axial torque control can be roughly classified into two types, control methods ^{(12) and (13)} using a torque sensor, and control methods ^{(14) and (15)} using a reaction force observer without using a torque sensor. In the case of a method using a torque sensor, there are problems such as high cost, a decrease in rigidity, and an influence on plant characteristics, a low sensor bandwidth, and an influence of sensor noise ⁽¹⁶⁾ .

  On the other hand, in the case of using a reaction force observer, it is possible to avoid the demerit like a torque sensor such as lowering the rigidity of the mechanism.

However, since forces including friction and modeling errors are estimated, it is important to estimate the axial torque by removing these components well, and many studies have been conducted ⁽¹⁷⁾ . In this embodiment, the axial torque is estimated by a reaction force observer (an example of a control unit) using the drive side position information detected by the drive side encoder (drive side position detection unit). Furthermore, in the present embodiment, by combining the load side encoder information (load side position information), more accurate shaft torque control is made possible.

(Control method in this embodiment)
The control method of the present embodiment uses the load side encoder (load side position detection unit) in view of the widespread use of encoders on the load side in recent years as demand for final positioning accuracy increases in various fields. It is the axial torque control method in 2 inertia system. Since information on both the drive side and the load side can be obtained, in this control method, it is possible to control the axial twist angular velocity of the rotating shaft in the power transmission unit 2 and to control the axial torque precisely. In addition, since the axial twist angular velocity can be controlled, it is possible to design a controller of an FF (feed forward) that explicitly takes into account the non-linear element compensation of the axial twist portion such as backlash.

A block diagram including the control unit 6 used in this control method is shown in FIG. In this figure, C _P is a P controller for driving side angular velocity, C _PI is a PI controller for axial torque FB (feedback) control, and Q _FF is a primary for realizing FF control in the inner loop of angular velocity control. A low-pass filter, T _s hat indicates an axial torque estimated by the disturbance observer, Q indicates a low-pass filter of the disturbance observer, and τ p indicates a time constant of pseudo differentiation. Also, the subscript n indicates a nominal value, and * indicates a command value to the control unit. A portion of the block diagram of FIG. 4 includes a block diagram substantially equivalent to the plant (control target) shown in FIG.

  The control system shown in FIG. 4 can be roughly divided into three parts. The first is a drive-side angular velocity control unit for controlling the axial twist angular velocity, the second is an axial torque FF control unit that generates an axial twist angular velocity command value from the axial torque command value, and the third is a drive-side reaction force It is an FB control unit of the shaft torque estimated by the observer.

(Drive-side angular velocity control)
First, the first drive-side angular velocity control unit (corresponding to the drive-side position FB and the drive-side position FF) will be described. This control method performs axial torque control by controlling axial torsional angular velocity. After controlling the drive-side angular velocity, which is a co-located system, the axial twist angular velocity is controlled by combining the load-side angular velocity. The drive-side angular velocity control has two degrees of freedom control, and performs feedforward control and feedback control. Here, from FIG. 4, the axial twist angular velocity Δω is given by equation (1).

  Therefore, the command value of the drive side angular velocity can be generated as shown in equation (2) using the axial twist angular velocity command value and the load side angular velocity.

The driving side angular velocity is controlled by using a disturbance observer using (J _Mn s + D _Mn ) as an inverse model and performing P control using Cp (driving side position FB unit). That is, the driving-side position FB unit 62 of the present embodiment, by using the drive-side position information omega _M, it has a configuration in which feedback control of the drive-side position. In addition, drive-side angular velocity control is control with two degrees of freedom, and the control band is improved. By the improvement of the band of the inner loop, the improvement of the response of the shaft torque control (described later) which is the outer loop can be expected. FF control (drive-side position FF portion) of angular velocity control in the inner loop is realized using a first-order low-pass filter Q _FF for making proper (J _Mn s + D _Mn ). That is, in the drive-side position FF unit 64 of this example, the actual position (for example, the actual drive-side angular velocity or angle) of the drive unit 1 matches the position command value ω ^* _M with respect to the command value for the drive unit 1 As described above, the control band at the drive side position is raised by performing feed forward compensation.

(Axle torque FF control)
Next, in the second axial torque FF control unit (corresponding to the non-linear element compensation FF unit), using the axial torque command value T ^* _s , the drive-side angular velocity command value ω ^* represented by the above equation (2) Make _M Considering the inverse model from Δω to Ts shown in FIG. 4, ω ^* _M is generated using the inverse of axial torsional stiffness (linear element), the inverse function of nonlinear element, and differentiation. Differentiation is realized by pseudo-differentiation of time constant τp. In this embodiment, since backlash is assumed as a non-linear element and modeling is performed with a dead zone, an inverse function model of the dead zone is used as non-linear element compensation. That is, the non-linear element compensation FF unit 61 in this example generates Δω ^* using the above-described elements, and uses this to positively feed-forward compensate the non-linear elements and linear elements in the power transmission unit 2. There is.

(Axis torque FB control)
In the third shaft torque FB control unit (power FB unit), the shaft torque Ts hat estimated by the drive side reaction force observer is fed back, and the pole arrangement is performed for Ts = (k / s) Δω by PI control. Do. By performing PI control, it is possible to follow the command value without steady-state deviation. Considering the delay of the low-pass filter Q of the drive side reaction force observer, Q is also added to the axial torque command value T ^* s. That is, the power FB unit 63 of this example is configured to feedback-control the estimated value of the power in the power transmission unit 2 using the estimated power Ts hat estimated using the drive-side position information ω _M.

By providing the above configuration, the control unit 6 of this example uses the load-side position information ω _L and the drive-side position information ω _M as a whole to input an input value T _M to the drive unit 1 (that is, control by calculating a command value) to the plant from section has a configuration for controlling power T _s (axial torque) transmitted by the power transmission unit 2.

(simulation)
The effectiveness of the control system proposed above was verified by simulation. The model used for the simulation is a two-inertia system model shown in FIG. 3, and the parameters use the values of Table 1 described above. Also for the sake of simplicity, non-linear elements and non-linear compensation elements are not included in the simulation unless otherwise specified.

(1) Design of Drive-side Angular Velocity Controller First, the P controller for the drive-side angular velocity was designed to have a band of 180 Hz. The cutoff frequency of the disturbance observer is 30 Hz. It is considered that the influence of the reaction force of the shaft torque is not interfered by the disturbance observer, and (J _Mn s + D _Mn ) is used as a model of the FF controller to realize a low pass of 1 kHz cutoff frequency for realization Use a filter.

(2) Design and Response of Shaft Torque Controller The PI controller for shaft torque performs pole arrangement with respect to Ts = (k / s) Δω, and performs multiple root arrangement at 30 Hz. The cutoff frequency of the pseudo differential is 1 kHz, and the cutoff frequency of the driving reaction force observer is 30 Hz. Since it is not assumed that a torque sensor is used, the shaft torque to be fed back and the shaft torque which can be detected by the experiment become the shaft torque estimated by the drive side reaction force observer, so the characteristics of the estimated shaft torque are examined. Note that no modeling error is given in the observer or FF control.

  At this time, the frequency characteristic of the target value response of the shaft torque is as shown in FIG. The solid line indicates the case where FF control of angular velocity is performed, and the dashed-dotted line indicates the case where FF control of angular velocity is not performed. By putting the FF control in the inner loop, the band of the target value response of the shaft torque which is the outer loop is improved from 20 Hz to 27 Hz.

Further, FIG. 6 shows a step response of the axial torque, and shows a comparison between the case where axial torque FF control is present and the case where axial torque FF control is not present in a state where angular velocity FF control is not present. At 0.050 s, input a command value passed through a 1st order low-pass filter with a cutoff frequency of 30 Hz to a step command of 8.0 Nm, and then enter -10 Nm of step disturbance as the load side disturbance d _L at 0.15 s There is. When a dotted line indicates a step command value through a low pass filter, and a solid line indicates axial torque FF control, a dashed dotted line indicates a case where axial torque FF control is not input. It can be seen that the start of the response is quickened by inserting the FF control. When there is no FF control of the shaft torque, the command value is reached quickly but with a somewhat oscillatory response. The disturbance response does not differ depending on the presence or absence of the shaft torque FF control.

  Next, FIG. 7 shows a step response of the shaft torque, and shows a comparison with and without the angular velocity FF control in a state where the shaft torque FF control is present. When a dotted line indicates a step command value through a low pass filter, and a solid line indicates angular velocity FF control, a dashed-dotted line indicates a case where angular velocity FF control is not inserted. Although FF is overshot by the FF control, quick response and load side disturbance suppression characteristics are improved.

Further, FIG. 8 shows the step response of the axial torque, and in the case where the solid line compensates for the non-linear element in the part of the FF control, the dashed-dotted line does not compensate for the non-linear element. In this simulation, the non-linear element is modeled as a dead zone, and the width of the dead zone is the assumed non-linear element, the maximum backlash width (5.3e ⁻⁵ rad) of the catalog value of the harmonic gear. The initial position is at the center of the backlash. It can be seen that, by FF compensating the non-linear element, the oscillation of the response can be suppressed and the control performance can be improved.

  In addition, a sine wave response is shown in FIG. 9 for examination at the time of speed reversal. The command value has an amplitude of 1 Nm and an angular frequency of 50 rad / s. When the dotted line indicates the sine wave command value and the solid line indicates the non-linear element compensation, the dashed-dotted line indicates the case where the non-linear element compensation is not inserted. In the case where the non-linear element compensation is not included, the amplitude of the initial response is reduced only because the initial position is located at the center of the backlash and the influence of the non-linear portion is half of the speed reversal It is from. It can be seen that the control performance is improved by performing the non-linear element compensation.

  In this simulation, the non-linear component compensation model is a perfect inverse model of non-linear components included in the plant, but it is actually difficult to accurately model the non-linear components of the plant. Therefore, in order to improve the control performance of the proposed method, modeling that reproduces non-linear elements well and has an inverse function is important.

Finally, simulations relating to realization of back drivability are shown in FIGS. 10 (a) and 10 (b). With T ^* s set to 0 Nm, a step disturbance of -10 Nm is input when the load side disturbance d _L is 0.050 s. In the case where the solid line is the control of this example, the dashed-dotted line is the case where nothing is controlled. In the control of this embodiment, the shaft torque can be reduced to 0 Nm in about 0.050 s. Further, as shown in FIG. 10 (b), since the amount of movement of the load side angle is larger with respect to the same load side disturbance as compared with the case where nothing is controlled, the back drivability is not affected by the friction on the drive side. It is thought that it has been realized.

(Experimental example)
An experimental example is shown below. The parameters of the controller in the experimental example are the same as in the simulation. In this experimental example, the controller is implemented by discretizing at 5 kHz by Tustin conversion. The angular velocity was obtained by applying a 1 kHz first-order low-pass filter to the 5 kHz backward difference of the angle information obtained from each encoder. Furthermore, the axial torque was estimated by a drive side reaction force observer with a cutoff frequency of 30 Hz.

  The frequency characteristic of the target value response of the axial torque by the experiment is as shown in FIG. Since this is the case without the angular velocity FF control, the simulation corresponds to the dashed dotted line in FIG. As shown in FIG. 11, the band was 18 Hz.

  The experimental results of the axial torque step response of this control method are shown in FIG. 12 and FIG. In FIG. 12, the dotted line indicates the step command, and the solid line indicates the comparison with the case where axial torque FF control is not performed and the alternate long and short dashed line indicates the case where axial torque FF control is not performed. This corresponds to FIG. 6 in the simulation. From FIG. 12, it can be seen that the response rise is better in the case of the axial torque FF control, as in the simulation.

  In FIG. 13, the dotted line indicates the step command, and the solid line indicates the comparison in the case where the angular velocity FF control is performed, and the dashed-dotted line indicates the case in which the angular velocity FF control is not performed. This corresponds to FIG. 7 in the simulation. In the experiment, although the improvement of the response of the shaft torque was not seen so much, it can be considered as a modeling error or sensor noise. When angular velocity FF control is applied, more noise is added to the estimated shaft torque. This is considered to be because the addition of the FF control in the inner loop increases the influence of the noise of the axial torque estimated value caused by the modeling error and the sensor noise.

(Advantages of this embodiment)
In the present embodiment, in consideration of the current situation where the use of an encoder on the load side has been increasing in recent years, an axial torque control method using a load-side encoder has been proposed. By controlling the axial torque in the two-inertia system and controlling the torque transmitted to the load, more sophisticated work is possible in a robot or a machine tool. And the controller of the proposed method was designed, the control performance was evaluated by simulation and experiment, and it was shown that the back drivability can be realized.

Although this embodiment is a simple one using the inverse model of backlash as nonlinear element compensation, it is considered that nonlinear element compensation considering the initial position obtained by both the drive side and load side encoders is also possible. . Further, in the present embodiment, the step command or 0 is given as the command value of the shaft torque control, but more sophisticated control such as impedance control or collision detection can be considered ^{(14) to (16)} . Although the proposed control method is applied to a robot in this embodiment, application to other fields such as a machine tool capable of modeling in a two-inertia system is also possible.

(Example 2)
Next, an example in which the above-described first embodiment is modified is shown as a second embodiment in FIG. Detailed descriptions of parts similar to or substantially equivalent to those of the first embodiment will be omitted. In addition, although the notation (the position of the integrator, etc.) of the block diagram of the plant is slightly different in the first embodiment and the second embodiment, they are theoretically equivalent. In each control method, it is corrected suitably and written in order to be easy to read.

  In the second embodiment, the inner loop is used for angle control. In this case, unlike the case of angular velocity control, the inner loop controls not the axial twist angular velocity but the axial twist angle. Therefore, no pseudo differentiation is required in the FF control part. Also, in place of Ts = (k / s) Δω, it is assumed that pole arrangement is performed for Ts = kΔθ, so I control is used. However, it is a drawback of the second embodiment that the band of the inner loop can not be increased as compared with the first embodiment in which the inner loop is subjected to angular velocity control. The drive side angle control is PD control. The FF in the inner loop of the second embodiment is not the drive-side angular velocity as in the first embodiment, but the FF control of the drive-side angle control.

Drive-side position FB unit 62 of the present embodiment, by using the drive-side position information theta _M, it has a configuration in which feedback control of the drive-side position.

Further, in the drive-side position FF unit 64 of this example, the actual position (for example, the actual drive-side angular velocity or angle) of the drive unit 1 matches the position command value θ ^* _M with respect to the command value for the drive unit 1 As described above, the control band at the drive side position is raised by performing feed forward compensation.

Further, the non-linear element compensation FF unit 61 in this example is configured to generate Δθ ^* and use it to positively feed-forward compensate non-linear elements and linear elements in the power transmission unit 2.

Furthermore, the power FB unit 6 of this embodiment, using the estimated power Ts hat estimated using the drive-side position information theta _M, has a configuration in which feedback control of the estimated value of the power in the power transmission unit 2.

Furthermore, the control unit 6 of this example uses the load-side position information θ _L and the drive-side position information θ _M as a whole to input an input value T _M to the drive unit 1 (that is, a command from the control unit to the plant By calculating the value), the power T _s (shaft torque) transmitted by the power transmission unit 2 is controlled.

(Example 3)
Next, another example in which the above-described first embodiment is modified is shown as a third embodiment in FIG. Also in FIG. 15, the plant is equivalent to the first embodiment.

  In the third embodiment, by performing the axial torque FF control with respect to Δθ, the pseudo differentiation of the axial torque FF control unit, which is the defect of the first embodiment, becomes unnecessary. In addition, it is considered that by making the inner loop angular velocity control, the band can be made higher than in the case where angle control is performed as in the second embodiment.

  This is made possible by assembling in parallel two outer loops of drive side angle control and shaft torque control in an inner loop of drive side angular velocity control. However, in the third embodiment, since there is a possibility that the axial torque FB control and the drive-side angular control FB interfere with each other, a design that avoids this is desirable.

  The drive-side angle control in the third embodiment is P control.

Drive-side position FB unit 62 of the present embodiment, by using the drive-side position information theta _M, it has a configuration in which feedback control of the drive-side position.

Further, in the drive-side position FF unit 64 of this example, the actual position (for example, the actual drive-side angular velocity or angle) of the drive unit 1 matches the position command value ω ^* _M with respect to the command value for the drive unit 1 As described above, the control band at the drive side position is raised by performing feed forward compensation.

Further, the non-linear element compensation FF unit 61 in this example is configured to generate Δθ ^* and use it to positively feed-forward compensate non-linear elements and linear elements in the power transmission unit 2.

Furthermore, the power FB unit 63 of this example is configured to feedback-control the estimated value of the power in the power transmission unit 2 using the estimated power Ts hat estimated using the drive side position information ω _M.

Furthermore, the control unit 6 of this example uses the load-side position information θ _L and the drive-side position information θ _M as a whole to input an input value T _M to the drive unit 1 (that is, a command from the control unit to the plant By calculating the value), the power T _s (shaft torque) transmitted by the power transmission unit 2 is controlled.

  The other configurations and operations in the second and third embodiments are the same as in the first embodiment, and thus the detailed description will be omitted.

  A comparison of Examples 1, 2 and 3 is shown in FIG. Classification can be made as to whether the axis twist angle is controlled or the axis twist angular velocity is controlled. Furthermore, the control of the inner loop can be classified into drive angle control or drive angle control.

  The contents of the present invention are not limited to the above embodiments. The present invention can add various changes to a specific configuration within the scope described in the claims.

(List of references cited in this specification)
(1) J. Hu and F. Hu: "Analysis and Design to the Asymptotic Behavior of the Semi / Full / Dual-closed Loop Feedback Position Control System", Proc. Of the SICE Annual Conference 2012, pp. 1096-1101, (2012).
(2) F. Sakai, M. Hikizu, Y. Kamiya and H. Seki: "Precision Positioning Which Contains Flexible Mechanical System: Fusion of Semi-closed and Full-closed Loop Control", J. JSPE, vol. 66, no 9, (2000) (in Japanese).
(3) M. Aoki, H. Fujimoto, Y. Hori and T. Takahashi: "Robust Resonance Suppression based on Self Resonance Cancellation Control and Self Resonance Cancellation Disturbance Observer for Application to Humanoid Robot", IEEJ Trans. On Industry Applications, vol. 134, no. 4, pp. 376-383, (2014).
(4) K. Sakata, H. Asaumi, K. Hirachi, K. Saiki and H. Fujimoto: "Frequency Separation Self Resonance Cancellation for Vibration Suppression Control of a Large-Scale Stage Using Multiple Position Sensors", Proc. Annual Conference of IEEE Ind. Electron Society (IECON-2013) pp. 6478-6483, (2013).
(5) S. Yamada, H. Fujimoto and Y. Hori: "Vibration Suppression Control of Two-Inertia System Using Load-Side State Variables with High-Resolution Encoder", Proc. Of IEE of Japan Technical Meeting Record, IIC- 14 -140, MEC-14-128, pp. 107-112, (2014).
(6) B. Na, J. Bae and K. Kong: "Back-drivability recovery of a full lower extremity assistive robot", Proc. Of the 12th International Conference on Control, Automation and Systems (ICCAS), pp. 1030-1034, (2012).
(7) P. Weiss, P. Zenker and E. Maehle: "Feed-forward Friction and Inertia Compensation for Improving Back-drivability of Motors", Proc. Of the 12th International Conference on Control, Automation, Robotics and Vision ICARCV-2012), pp. 288-293, (2013).
(8) N. Paine, S. Oh and L. Sentis: "Design and Control Considerations for High-Performance Series Elastic Actuators", IEEE Trans. On Mechatronics, vol. 19, no. 3, (2014).
(9) K. Kong, J. Bae and M. Tomizuka: "A Compact Rotary Series Elastic Act for Human Assistive Systems", IEEE Trans. Mechatronics, vol. 17, no. 2, (2012).
(10) Y. Kuroki, Y. Kosaka, T. Takahashi, E. Niwa, H. Kaminaga and Y. Nakamura: "Cr-N Alloy Thin Film Based Torque Sensors and Joint Torque Servo Systems for Compliant Robot Control", IEEE International Conference on Robotics and Automation (ICRA-2013), pp. 4954-4959, (2013).
(11) M. Iwasaki, M. Kainuma, M. Yamamoto and Y. Okitsu: "Compensation by Exact Linearization Method for Nonlinear Components in Positioning Device with Harmonic Drive Gearings", Journal of JSPE, vol. 78, no. 7, pp. 624-630, (2012).
(12) T. Kawakami, K. Ayusawa, H. Kaminaga and Y. Nakamura: "High-Fidelity Joint Drive System by Torque Feedback Control Using High Precision Linear Encoder", IEEE International Conference on Robotics and Automation (ICRA-2010), pp. 3904-3909, (2010).
(13) T. Kawakami, K. Ayusawa, H. Kaminaga, C. Ott and Y. Nakamura: "Development of Joint Driving Mechanism with Rigid Torque Sensor and its Torque Control Systems", Proc. , no. 08-4, pp. 1-4, 2008, (in Japanese).
(14) S. Oh and Y. Hori: "Experimental Study on Impedance and Impact Force Control by Position Feedback Control", Proc. Of IEE of Japan Technical Meeting Record, IIC-09-117, pp. 31-36, (2009) ), (in Japanese).
(15) N. Shibata, T. Yoshioka, K. Ohishi and T. Miyazaki: "High Reliable Force Sensorless Contact Detection of Contact Motion", Proc. Of IEE of Japan Technical Meeting Record, IIC-12-151, pp. 107 -112, (2012), (in Japanese).
(16) S. Katsura, Y. Matsumoto and K. Ohnishi: "Modeling of Force Sensing and Validation of Disturbance Observer for Force Control", IEEE Trans. On Industrial Electronics, vol. 54, no. 1, (2007).
(17) C. Mitsantisuk, M. Nandapaya, K. Ohishi and S. Katsura: "Design for Sensorless Force Control of Flexible Robot by Using Resonance Ration Control Based on Coefficient Diagram Method", Automatika, vol. 54, no. pp. 62-73, (2013).

DESCRIPTION OF SYMBOLS 1 drive part 2 power transmission part 3 load part 4 drive side position detection part 5 load side position detection part 6 control part 61 nonlinear element compensation FF part 62 drive side position FB part 63 power FB part 64 drive side position FF part

Claims (10)

A drive unit, a power transmission unit, a load unit, a drive-side position detection unit, a load-side position detection unit, and a control unit;
The power transmission unit is configured to transmit power from the drive unit to the load unit.
The drive side position detection unit is configured to obtain drive side position information in the drive unit,
The load-side position detection unit is configured to acquire load-side position information in the load unit,
The control unit uses the load-side position information (ω _L or θ _L ) and the drive-side position information (ω _M or θ _M ) to generate a command value (ω ^* _M or θ ^* ) to the drive unit ^. By calculating _{M 2} ) , the power (Ts) transmitted by the power transmission unit is controlled ,
The control unit includes a non-linear element compensation FF unit,
The non-linear element compensation FF unit generates a position output value (Δω ^* or Δθ ^* ) for feed-forward compensating a non-linear element in the power transmission unit using an inverse function of the non-linear element Has become
Furthermore, the input value to the non-linear compensation FF unit is the command value (T ^* s) of the power in the power transmission unit ,
The control unit uses the result of the addition of the output value (Δω ^* or Δθ ^* ) for the position from the non-linear compensation FF unit and the load-side position information (ω _L or θ _L ) to perform the driving. A control device configured to generate a command value (ω ^* _M or θ ^* _M ) to a unit. The control device according to claim 1 , wherein an inverse function model of a dead zone in the power transmission unit is used as the inverse function of the non-linear element . The control unit includes a drive side position FB unit,
The control device according to claim 1, wherein the drive-side position FB unit performs feedback control of the drive-side position using the drive-side position information. The control unit includes a power FB unit,
The power FB unit is configured to perform feedback control of an estimated value of the power in the power transmission unit using the estimated power estimated using the drive-side position information. Control device described in. The control unit includes a drive side position FF unit,
The drive-side position FF unit raises the control band at the drive-side position by performing feed-forward compensation on the command value to the drive unit such that the actual position matches the position command value. The control device according to any one of claims 1 to 4. The drive unit is a rotary motor,
The power transmission unit is configured to transmit the power using a rotary shaft,
The control device according to any one of claims 1 to 5, wherein the power is an axial torque at the rotating shaft. The control device according to any one of claims 1 to 6, wherein the position is an angle or an angular velocity. The drive unit is a linear motor,
The power transmission unit is configured to transmit a translational force from the drive unit to the load side.
The control device according to any one of claims 1 to 5, wherein the power is a translational force transmitted by the power transmission unit. A control device including a drive unit, a power transmission unit, a load unit, a drive-side position detection unit, a load-side position detection unit, and a control unit is used.
The power transmission unit is configured to transmit power from the drive unit to the load unit.
The drive side position detection unit is configured to obtain drive side position information in the drive unit,
The load-side position detection unit is configured to acquire load-side position information in the load unit,
The control unit controls the power transmitted by the power transmission unit by calculating a command value to the drive unit using the load-side position information and the drive-side position information. Yes,
Here, the control unit includes a non-linear element compensation FF unit,
The step by the control unit is:
The non-linear element compensation FF unit generating an output value (Δω ^* or Δθ ^* ) for position using feedforward compensation of the non-linear element in the power transmission unit , using an inverse function of the non-linear element; ,
The control unit uses the result of the addition of the output value (Δω ^* or Δθ ^* ) for the position from the non-linear compensation FF unit and the load-side position information (ω _L or θ _L ) to perform the driving. Generating a command value (ω ^* _M or θ ^* _M ) to the unit;
Here, a control method in which an input value to the non-linear compensation FF unit is a command value (T ^* s) of power in the power transmission unit .   A computer program for causing a computer to execute the steps according to claim 9.

Images