JP2016039737A - Motor controller and motor control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To achieve a flexible and highly accurate joint control.SOLUTION: A motor controller (1) includes: a dynamic friction compensator (6) for generating a dynamic friction compensatory signal and provide it to a motor (2) on the basis of a rotational speed signal detected with a rotational speed sensor (4) and a predetermined dynamic friction model of a reduction gear (3); and a static friction compensator (7) for generating a static friction compensatory signal on the basis of a torque signal detected with a torque sensor (5) and providing it to the motor (2).SELECTED DRAWING: Figure 1

Description

  The present invention relates to a motor control device and a motor control method for controlling a motor having a reduction gear.

  Many conventional robots represented by industrial applications have been designed for the purpose of performing work that requires precise positioning. For this reason, the joints of these robots are rigidly controlled by high gain feedback compensation, and it is necessary to place them in a space separated from humans in order to avoid accidents.

  However, in recent years, there has been a demand for robots that share the environment with people, interact with each other, and support human work. Such a robot joint needs to have the contradictory properties of rigidity for ensuring sufficient positioning accuracy for performing work and flexibility at the time of contact.

  As a method of ensuring the above flexibility, the use of mechanical flexible elements such as springs and dampers immediately comes to mind (Non-Patent Document 1). However, such a mechanical flexible element causes a significant decrease in control accuracy. In order to achieve both rigidity and flexibility, the actuator and its force transmission mechanism must have back drivability.

  Here, back drivability means that when a force is applied to an output node, the output node is movable, and the force applied to the output node is transmitted to the input node side, so that the drive machine can be driven in reverse. Shall mean.

  Conventional robots have high torque and low rotational speed characteristics with a compact configuration combining a rotary electromagnetic motor with a high reduction ratio reduction gear. However, since a high reduction ratio also amplifies the friction inside the reduction gear, it is a direct cause of impairing back drivability. In order to avoid the occurrence of friction, there is a research using a direct drive motor (Non-Patent Document 2), but this has caused an excessive increase in size of the joint. There is also research using a linear drive actuator by electromagnetic drive (Non-patent Document 3), but sufficient thrust cannot be obtained. A hydraulic drive linear drive actuator (Non-Patent Document 4) has also been developed, but is expensive and not necessarily widely available.

  Thus, at the present time, it is difficult to achieve the required characteristics of having both rigidity and flexibility only with hardware. On the other hand, in a combination of a conventional rotary electromagnetic motor and a high reduction ratio reducer, a method of solving the problem of software that compensates for the friction force generated inside the reducer and achieves both rigidity and flexibility, Several reports have been made in recent years (Non-patent document 5, Non-patent document 6, Non-patent document 7).

  Moreover, the structure which compensates the dynamic friction force which arises inside a reduction gear by feedforward control is disclosed (nonpatent literature 8, patent documents 1, patent documents 2).

Patent No.4985659 specification (registered on May 11, 2012) Patent No. 3752758 (Registered on December 22, 2005)

G. Pratt, M. Williamson, P. Dillworth, J. Pratt, K. Ullandand A. Wright, “Stiffness Is n’t Everything,” Preprintsof the Fourth International Symposium on Experimental Robotics, 1995. H. Asada, T. Kanade, I. Takeyama, “Control of a Direct-Drive Arm,” Journal of Dynamic Systems, Measurement, and Control, Vol.105, 1983, pp.136-242. Yoshihiro NAKATA, Tianyi YU, Hiroshi ISHIGURO, Katsuhiro HIRATA “Small Size Linear Vernier Motor for Artificial Muscle Applications” The Japan Society of Applied Electromagnetics and Mechanics, Vol. 20, No.1, 2012, pp.113-118. Hiroshi Kaminaga, Tomoya Amari, Yukihiro Katayama, Junya Ono, Yuto Shimoyama, and Yoshihiro Nakayama, “Backdrivability Analysis of Electro-Hydrostatic Actuator and Series Dissipative Actuation Model” 2010 IEEE International Conference on Robotics and Automation, Anchorage, Alaska, USA, May 3 -8, 2010, pp.4204-4211. G. Hirzinger, N. Sporner, A. Albu-Schaffer, M. Hahnle, R. Krenn, A. Pascucci, M. Schedl “DLR's torque-controlled light weight robot III: Are we reaching the technological limits now?” Proceedings of the 2002 IEEE International Conference on Robotics and Automation Washington, DC, USA, May 2002, pp. 1710-1716. T. Kawakami, K. Ayusawa, H. Kaminaga and Yoshihiko Nakamura, “High-Fidelity Joint Drive System by Torque Feedback Control Using High Precision Linear Encoder,” Proceedings of 2010 IEEE International Conference on Robotics and Automation, Anchorage, Alaska, USA, May 3-8, 2010, pp.3904-3909. Kenichiro Nagasaki, Yasunori Kawanami, Satoru Shimizu, Takashi Kito, Toshimitsu Tsuboi, Atsushi Miyamoto, Tetsuharu Fukushima and Hideki Shimomura (Sony Corporation) “Whole-body Cooperative Force Control for a Two-Wheeled Mobile Robot Using Generalized Inverse Dynamics and Idealized Joint Units” 2010 IEEE International Conference on Robotics and Automation, Anchorage, Alaska, USA, May 3-8, 2010, pp.3377-3383. Oussama Khatib, Peter Thaulad, Taizo Yoshizawa, and Jachcung Park (Stanford University, Artificial Intelligence Laboratory) (Honda Research Institute USA, Inc.) “Torque-Position Transformer for Task Control of Position Controlled Robots” 2008 IEEE International Conference on Robotics and Automation Pasadena, CA, USA, May 19-23, 2008, pp.1729-1734.

  However, the configurations disclosed in Non-Patent Documents 3 to 5 above all measure torque and control at the acceleration level, and are based on the use of a highly accurate (and therefore expensive) torque sensor. It is not widely available.

  Further, the configurations taught in Non-Patent Document 6 and Patent Documents 1 and 2 cannot compensate for the static friction force generated inside the reduction gear.

  An object of the present invention is to realize flexible and highly accurate joint control with an inexpensive hardware configuration and software in a combination of a rotary electromagnetic motor and a high reduction ratio reduction gear.

  In order to solve the above problems, a motor control device according to the present invention is a motor control device that controls a motor having a speed reducer, and a rotational speed sensor that detects a rotational speed of an output shaft of the speed reducer; Based on a torque sensor that detects torque of the output shaft of the speed reducer, a rotational speed signal that represents the rotational speed detected by the rotational speed sensor, and a predetermined dynamic friction model of the speed reducer Corresponding to a dynamic friction compensator that generates a dynamic friction compensation signal that compensates the dynamic friction torque of the machine and supplies it to the motor, a torque signal that represents the torque detected by the torque sensor, and a rotational speed of the output shaft of the speed reducer A static friction compensation signal for compensating for the static friction torque of the reduction gear based on a motor inertia torque signal generated based on an angular acceleration signal representing the angular acceleration Generates and characterized in that a static friction compensator supplied to the motor.

  According to this feature, the dynamic friction torque of the speed reducer is compensated by the dynamic friction compensator. For this reason, an inexpensive torque sensor with a small bandwidth can be adopted as the torque sensor for the static friction compensator that compensates for the static friction torque of the reduction gear. As a result, flexible and highly accurate joint control can be realized with an inexpensive hardware configuration and software.

  In the motor control device according to the present invention, it is preferable that the dynamic friction model is provided with a voltage applied to the motor as a function of a rotation speed of an output shaft of the speed reducer.

  According to the above configuration, the dynamic friction compensation signal for compensating the dynamic friction torque of the speed reducer is supplied to the motor based on the rotational speed of the output shaft of the speed reducer based on the rotational speed sensor and the predetermined dynamic friction model of the speed reducer. It can be applied in advance to the input signal.

  In the motor control device according to the present invention, the static friction compensation signal includes the motor inertia torque signal, a torque estimation signal for estimating the torque based on the torque signal, and an applied voltage signal representing an applied voltage to the motor. It is preferable to be generated based on the above.

  According to the above configuration, the static friction compensation signal for compensating the static friction torque of the reduction gear can be fed back to the input signal to the motor.

  In the motor control device according to the present invention, it is preferable that the speed reducer is a gear speed reducer.

  According to the above configuration, the rotation of the motor can be decelerated at a high reduction ratio with a simple configuration.

  In the motor control device according to the present invention, the gear reducer is preferably a wave gear reducer.

  According to the above configuration, the rotation of the motor can be decelerated at a high reduction ratio with a simple configuration.

  A motor control method according to the present invention is a motor control method for controlling a motor having a speed reducer, the rotational speed detecting step for detecting the rotational speed of the output shaft of the speed reducer, and the torque of the output shaft of the speed reducer. Compensation of the dynamic friction torque of the speed reducer based on a torque detection step for detecting the rotational speed, a rotational speed signal representing the rotational speed detected by the rotational speed detection step, and a predetermined dynamic friction model of the speed reducer A dynamic friction compensation step that generates and supplies the dynamic friction compensation signal to the motor, a torque signal that represents the torque detected by the torque detection step, and an angular acceleration that corresponds to the rotational speed of the output shaft of the speed reducer Based on the motor inertia torque signal generated based on the angular acceleration signal, a static friction compensation signal for compensating the static friction torque of the reduction gear is generated to generate the mode. Characterized in that it comprises a static friction compensation step of supplying a.

  The present invention produces an effect that flexible and highly accurate joint control can be realized by an inexpensive hardware configuration and software in a combination of a rotary electromagnetic motor and a high reduction ratio reduction gear.

It is a block diagram which shows the structure of the motor control apparatus which concerns on embodiment. It is a block diagram for demonstrating the friction compensation by the dynamic friction compensator provided in the said motor control apparatus. It is a block diagram for demonstrating the friction compensation by the static friction compensator provided in the said motor control apparatus, and the friction compensation by the said dynamic friction compensator. (A) (b) is a perspective view which shows the structure of the motor and speed reducer which were provided in the experimental apparatus of the said motor control apparatus. It is a perspective view which shows the structure of the torque sensor provided in the experimental apparatus of the said motor control apparatus. (A) (b) is a figure for demonstrating operation | movement of the said torque sensor. It is a graph which shows the relationship between the angular velocity of the output shaft of the reduction gear provided in the said motor control apparatus, and the PWM duty ratio showing the torque of the said output shaft. It is a graph which shows the relationship between the output voltage of the said torque sensor, and the torque of the said output shaft. It is a block diagram of the control system which consists of the feedback (PID) controller and feedforward controller which were provided in the experimental apparatus of the said motor control apparatus. (A)-(d) is a wave form diagram which shows the torque value measured by the said control system, and the value of the output rotation angle of a reduction gear. (A) is a wave form diagram which shows the simulation result corresponding to FIG.10 (d), (b) is a wave form diagram which shows the simulation result supposing the precision improvement of torque estimation in FIG.10 (d).

  Hereinafter, embodiments of the present invention will be described in detail.

(Overview)
First, the dynamic friction torque of a wave gear reducer (Harmonic Drive (registered trademark)) is identified, and the torque for compensating this dynamic friction torque is given by feedforward, so that most of the friction torque is compensated. As a result, the required bandwidth of the torque sensor required to compensate the residual friction can be reduced. Then, an inexpensive torque sensor using a photo interrupter is configured and calibrated. For compensation of static friction torque and the like, a disturbance observer using this torque sensor is configured. As a result of applying the two-degree-of-freedom control to the friction-compensated motor and applying a disturbance to the position-controlled motor, it was confirmed that good characteristics were obtained.

(Embodiment)
(Configuration of motor control device 1)
FIG. 1 is a block diagram illustrating a configuration of a motor control device 1 according to the embodiment. The motor control device 1 controls a motor 2 having a speed reducer 3. The motor 2 is, for example, a rotary electromagnetic motor. The reduction gear 3 is, for example, a gear reduction gear that reduces the speed with a high reduction ratio using mechanical meshing. In the gear reducer, when the gear rotates at a predetermined rotational speed, a frictional force is generated by contact between the tooth tips of the gear. By adding a voltage signal that preliminarily compensates for the frictional force due to the contact of the gear to the voltage signal supplied to the motor 2, the frictional force due to the gear of the reduction gear 3 can be canceled.

  The motor control device 1 includes a drive circuit 8, a rotation speed sensor 4, and a torque sensor 5. The drive circuit 8 supplies a voltage signal for driving the motor 2 to the motor. The rotational speed sensor 4 detects the rotational speed of the output shaft of the speed reducer 3. The torque sensor 5 detects the torque of the output shaft of the speed reducer 3.

  The motor control device 1 is provided with a dynamic friction compensator 6 and a static friction compensator 7. The dynamic friction compensator 6 is based on a rotational speed signal that is detected by the rotational speed sensor 4 and that represents the rotational speed of the output shaft of the speed reducer 3 and a predetermined dynamic friction model of the speed reducer 3. A dynamic friction compensation signal for compensating the torque is generated and supplied to the motor 2. The static friction compensator 7 is generated based on the torque signal representing the torque of the output shaft detected by the torque sensor 5 and the angular acceleration signal representing the angular acceleration corresponding to the rotational speed of the output shaft of the speed reducer 3. Based on the motor inertia torque signal, a static friction compensation signal for compensating the static friction torque of the speed reducer 3 is generated and supplied to the motor 2.

(Feed forward friction compensation control)
FIG. 2 is a block diagram for explaining friction compensation by the dynamic friction compensator 6 provided in the motor control device 1.

(Identification of dynamic friction characteristics)
Viscous friction and dynamic friction can be identified relatively easily as a function of speed. Friction torque generated inside the speed reducer 3 by identifying the dynamic friction torque inside the speed reducer 3 as a function of the rotational speed of the output shaft provided in the speed reducer 3 and superimposing it on the input voltage signal to the motor 2 It can be assumed that most of this can be compensated.

  Consider a drive unit comprising a motor 2 and a speed reducer 3. The circuit equation of the motor 2 is

It is represented by However, the rotation angle of θ motor 2, e is the voltage applied to the motor 2 (applied voltage signal), i is the current flowing through the motor 2, K _E is the back electromotive force constant of the motor 2, R is between the motor 2 terminal Resistance, L is the self-inductance of the motor 2. The equation of motion of motor 2 is

It is. Where J is inertia from the shaft of the motor 2 to the output shaft of the speed reducer 3, γ is a reduction ratio of the speed reducer 3, τ _m is torque generated by the motor 2, and τ _f is dynamic friction generated inside the speed reducer 3. Torque, τ _e is an external load torque. Further, when the motor 2 is a DC brush motor, tau _m and i by using the torque constant _{K T,}

Are associated with each other. Assuming that the self-inductance L is sufficiently small, and further considering no-load and steady state, equations (1) to (3) are

Can be organized like Assuming the frictional torque tau _f and impact of dry and viscosity is dominant, combined with the back EMF of the motor 2, the applied voltage e of the motor 2 is given as a function of the rotation speed d [theta] / dt. The relationship between the applied voltage e to the motor 2 and the rotational speed dθ / dt is obtained experimentally, and the dynamic friction model of the speed reducer 3 is identified.

(Feed forward compensation of dynamic friction torque)
Using the identified dynamic friction model, feedforward compensation as shown in FIG. 2 is attempted. The rotational speed of the output shaft of the speed reducer 3 is obtained from the value of the encoder constituting the rotational speed sensor 4, and the voltage _eff (dynamic friction compensation signal) necessary for compensating the dynamic friction torque is calculated from the above-mentioned dynamic friction model. By superimposing this on the control voltage e _c, it can be canceled by combining the dynamic friction torque and the motor back electromotive force of the reduction gear 3.

(Residual friction compensation by disturbance observer 9)
FIG. 3 is a block diagram for explaining friction compensation by the static friction compensator 7 and friction compensation by the dynamic friction compensator 6 provided in the motor control device 1.

(Design of disturbance observer 9)
Friction torque near zero speed is distinguished as static friction torque. Unlike dynamic friction torque, it is difficult to assume a friction model as a function of some value. Therefore, it is considered that a disturbance observer 9 using the torque sensor 5 is configured to compensate for the static friction torque.

  If all the friction torques are compensated by the dynamic friction compensator 6 described with reference to FIGS. 1 and 2, the behavior of the motor 2 with the speed reducer 3 compensated for the viscous resistance is a theoretical model.

Should follow. Here, we have a e _u the input voltage.

  However, in practice, nonlinear friction torque and modeling error remain, and the motor 2 with the speed reducer 3 does not behave as theoretically.

  These nonlinear friction torques, disturbance voltage conversion values including modeling errors,

The equation of motion of the motor 2 with the speed reducer 3 is

It becomes. Therefore, a disturbance observer 9 as shown in FIG. 3 is used to attempt an ideal response. However, the output voltage characteristic of the torque sensor 5 is h (τ), its inverse function is h ⁻¹ (e), and each subscript n represents a nominal value.

The disturbance observer 9 includes a load torque estimated value (torque estimation signal) τ _ee given by a disturbance torque estimator 10 to which a torque signal corresponding to the disturbance is input, and an output shaft of the speed reducer 3 detected by the rotational speed sensor 4. The output torque of the motor 2 with the speed reducer 3 is calculated from the motor inertia torque signal generated based on the rotational angular acceleration dθ _M / dt (angular acceleration signal) calculated from the rotational speed. Then, the output torque is converted to a converted element 16 the voltage of the disturbance observer 9, by taking the difference between the applied voltage signal corresponding to the input voltage e _u, the signal representative of the estimated voltage conversion value of the residual friction torque

Get. Next, by feeding back the residual friction torque estimated voltage conversion value signal representative of the .DELTA.e _{ff (static} friction compensation signal) is stabilized through the secondary delay element 17, the input voltage e _u is added to the control voltage e _c ,

And a response closer to the model of Equation (5) can be realized.

(Inexpensive torque measurement system)
Many systems that have been developed in the past achieve an ideal response by acceleration-based control using the disturbance observer 9. However, if all disturbances are to be compensated by the disturbance observer 9, a torque sensor with a large bandwidth is required, and the system becomes expensive. Therefore, FIG. 2 describes the feedforward compensation of the dynamic friction torque that can be modeled by the speed reference. As shown in FIGS. 1 and 3, by using the feedforward compensation of the dynamic friction torque by the dynamic friction compensator 6 together with the feedback compensation of the static friction torque by the static friction compensator 7, even an inexpensive torque sensor 5 having a small bandwidth can be compared. It is possible to perform disturbance compensation with high accuracy.

  4A and 4B are perspective views showing configurations of the motor 2 and the speed reducer 3 provided in the experimental device of the motor control device 1. FIG. FIG. 5 is a perspective view showing the configuration of the torque sensor 5 provided in the experimental device of the motor control device 1. 6A and 6B are diagrams for explaining the operation of the torque sensor 5.

  The motor 2 and the speed reducer 3 are provided in parallel on the substrate. The shaft of the motor 2 and the input shaft of the speed reducer 3 are coupled by a timing belt 11. An arm 12 is attached to the output shaft of the speed reducer 3. A ring 14 is fixed between the arm 12 and the speed reducer 3. A slit 15 is formed in the ring 14. A photo interrupter 13 constituting the torque sensor 5 is arranged at a position corresponding to the slit 15 of the arm 12.

  In this embodiment, the torque sensor 5 provided with the photo interrupter 13 is employed. The photo interrupter 13 detects a change in light amount caused by an object blocking between the light emitting diode and the phototransistor as a change in output voltage.

  As shown in FIGS. 5 and 6, a photo interrupter 13 is provided on the arm 12, and a slit 15 is formed on the ring 14. When a torque is applied to the arm 12, a relative displacement occurs between the photo interrupter 13 and the slit 15 due to bending deformation, and the amount of light passing through the slit 15 changes. Therefore, the principle of the photo interrupter 13 can be applied to torque measurement.

  The torque sensor 5 using the photo interrupter 13 has the merit of being small and light and inexpensive, except that the output is nonlinear and the temperature is highly sensitive. In order to compensate for the disturbance with the above-described disturbance observer 9 accurately, the characteristics of the torque sensor 5 are identified by experiments, and the actually applied torque is correctly estimated based on the voltage output from the torque sensor 5. Calibration is important.

(Experiment)
(Identification experiment of dynamic friction torque and viscous resistance)
4 (a) and 4 (b) comprising a motor 2 (maxon RE-max 29) and a reduction gear 3 (harmonic drive (registered trademark) reduction gear (SHG-14-80-2UH-SP), wave gear reduction gear). Using the drive unit shown, an experiment was performed to measure the dynamic friction torque generated in the speed reducer 3 when the motor 2 rotates based on the equation (4).

  Since the motor 2 is controlled from the drive circuit 8 including a microcomputer, the voltage is given by PWM (Pulse Width Modulation) and adjusted by the duty ratio. The experiment was performed by measuring the change in the rotation speed due to the increase / decrease of the duty ratio, and the duty ratio was increased / decreased by 1.0% in the stationary state and 0.2% in the rotational state, up to ± 64.0%. However, the sign of the duty ratio corresponds to the sign of the voltage applied to the motor 2. The rotational speed was measured by taking 1000 samples at a period of 2 [ms] every time the duty ratio was increased or decreased by 1% to 0.2%, and taking the average. The result at this time is shown in FIG.

    FIG. 7 is a graph showing the relationship between the angular speed of the output shaft of the speed reducer 3 provided in the motor control device 1 and the PWM duty ratio representing the torque of the output shaft. Temperature drift occurs due to the on-resistance temperature characteristics of the MOS FET (Metal-Oxide-Semiconductor Field-Effect Transistor), which is a switching element, but the rotational speed is the sum of the viscosity characteristics and the dryness characteristics that are sign functions. Can be read from the figure.

(Characteristic confirmation of torque sensor 5)
FIG. 8 is a graph showing the relationship between the output voltage of the torque sensor 5 and the torque of the output shaft of the speed reducer 3. A removable arm 12 is attached to the experimental machine shown in FIG. 4A as shown in FIG. 4B, and the relationship between the torque applied to the arm 12 and the output voltage of the torque sensor 5 is measured by experiment. This result is the graph shown in FIG. 8, and a curve C having a hysteresis that cannot be ignored was confirmed. This hysteresis is considered to be largely caused by the viscosity and plastic deformation of the material of the arm 12. A large drift due to temperature change was also confirmed. Originally, it is desirable that these hysteresis and drift are compensated by appropriate model and parameter identification. However, in this embodiment, it is assumed that a proportional relationship such as a straight line L shown in FIG. 8 is established between the output voltage of the photo interrupter 13 constituting the torque sensor 5 and the torque of the output shaft of the speed reducer 3. The control performance under low torque estimation is evaluated.

(Control experiment)
A position control experiment was performed on the motor control device 1 shown in FIG. The actual parameters are
R _n = 2.36 [Ω],
K _Tn = 2.58 × 10-2 [N · m / A],
J _n = 7.01 × 10 −6 [kg · m ² ],
γ = 120,
ω _L = 250 [rad / s],
It is.

  When feedforward friction compensation is applied to the dynamic friction torque of the harmonic drive (registered trademark) type reduction gear 3 identified as shown in FIG. 7, the dynamic friction torque is controlled to prevent the control from becoming unstable. 5-10% smaller compensation torque was converted into voltage and superimposed. The torque difference between the identified dynamic friction torque and the smaller compensation torque is compensated by feedback control by the static friction compensator 7.

Friction torque that could not be compensated by this feedforward friction compensation is compensated by the disturbance observer 9. The transfer function from the ideal motor input voltage _eu to the output rotation angle θ with the friction torque fully compensated is

It is represented by In contrast, PID controller

And a feedforward controller to improve the target value response

A two-degree-of-freedom controller consisting of the above was designed to constitute a control system as shown in FIG.

However, _{K P} is a proportional gain of PID control, _{K I} is the integral gain, _{K D} is its differential gain. ω _n and αω _n are the natural frequencies of the feedback system, ζ is the damping coefficient,

There is a relationship. ω ′ _n and αω ′ _n are natural frequencies related to the target value response, and by taking a large value, it is possible to achieve a target value response with high responsiveness.

A small PID gain considering the disturbance response was set for the above system, and a two-degree-of-freedom control experiment with a control period of 2 [ms] was performed. Each parameter is
ω _n = 4.50 [rad / s],
ω ′ _n = 50.0 [rad / s],
ζ = 0.60,
α = 0.80,
It was.

The experiment
Experiment (a) No compensation,
Experiment (b) Feedforward compensation only,
Experiment (c) Disturbance observer only
Experiment (d) Four combinations of feed forward compensation and disturbance observer were used. In each experiment, the following four operations were given.
1. A target value is set so that the angle of the output shaft of the speed reducer 3 is 10.0 °.
2. Slowly apply disturbance torque.
3. Unload disturbance torque.
4). Load disturbance torque quickly and instantaneously.

  10A to 10D are graphs showing the measured torque value and the output rotation angle (joint rotation angle) value of the speed reducer 3. The measured torque contains a lot of noise derived from the PWM frequency and the commercial frequency of the power supply, and hysteresis and drift are also seen. Compare response of output angle to target value and disturbance torque.

  Referring to FIG. 10A, the friction torque is very large without compensation, and the joint rotation angle cannot respond at all to the target value and the disturbance torque. That the PID gain set in the present embodiment is very small well represents the target value response under this condition.

  Referring to FIG. 10 (b), in the case of only feedforward compensation (dynamic friction compensator 6), the response is improved to some extent, but it stops due to residual friction, and due to the divergence of the integrator of feedback control due to this, The behavior of the joint rotation angle is oscillatory.

  Referring to FIG. 10C, in the case of only the disturbance observer 9 (static friction compensator 7), both the target value response and the disturbance response are greatly improved. However, it can be seen that it is oscillating near the target value. This is presumably because a delay occurs in compensation due to the second-order delay element 17 inserted for stabilizing the disturbance observer 9.

  Referring to FIG. 10D, the combined use of the feedforward compensation (dynamic friction compensator 6) and the disturbance observer 9 (static friction compensator 7) can reduce the compensation amount of the disturbance observer 9, and FIG. The delay that appeared in c) can be eliminated. However, even in this case, vibration occurs in the vicinity of the target value, and complete convergence cannot be achieved.

  This cause was considered using simulation. Possible causes are that a controller that ignores the viscoelasticity of the timing belt 11 is configured, the torque estimation accuracy of the torque sensor 5 is low, the control cycle is not sufficient, and the residual friction compensation of the disturbance observer 9 is It can be delayed. In order to verify these, a two-inertia system considering the viscoelasticity of the timing belt 11

As an actual model. However,
J _m is the inertia of the output shaft of the motor 2,
θ _m is the rotation angle of the motor 2,
_Jg is the inertia of the input shaft of the reducer 3,
θ _g is the rotation angle of the input shaft of the speed reducer 3,
It is.
Also,
τ _g is a torque acting on the input shaft of the speed reducer 3,
When the elastic coefficient of the timing belt 11 is K _b and the viscosity coefficient is C _b ,

It is expressed. From preliminary experiments, it is known that K _b ≈10 [N · m / rad] and C _b ≈0.01 [N · m · s / rad]. The simulation results did not change even when the system could be approximated. Therefore, it can be said that the vibration is not caused by the timing belt 11.

On the other hand, in the present embodiment, the control cycle is 2 [ms], but by setting this control cycle smaller, the cutoff frequency ω _L of the LPF (Low Pass Filter) for stabilizing the disturbance observer 9 is set. I found that I could take it big. That is, it is considered that smooth control can be realized by reducing the delay of compensation and reducing the influence of residual friction. Furthermore, in order to remove minute vibrations, it was found that it is necessary to improve the accuracy of torque estimation including countermeasures for hysteresis and drift.

FIG. 11 shows the above. FIG. 11A is a waveform diagram showing a simulation result corresponding to FIG. 10D, and FIG. 11B is a waveform diagram showing a simulation result assuming an improvement in torque estimation accuracy in FIG. 10D. FIG. 11A shows a waveform diagram showing the simulation of the experiment (d). In the experiment (d), the control period is 500 [μs], and the cutoff frequency of the LPF is ω _L = 500 [rad / s]. FIG. 11B shows a waveform diagram representing a simulation result assuming an improvement in accuracy of torque estimation (here, reduction of noise).

(Modification)
In the above-described embodiment, the example in which the rotation of the motor 2 is decelerated by the gear reduction device that decelerates at a high reduction ratio using mechanical meshing has been described. However, the present invention is not limited to this. For example, the rotation of the motor 2 may be decelerated by a hydraulic mechanism. In this case, the dynamic friction compensator 6 compensates for the viscous friction generated between the wall surface of the pipe and the oil.

  In the embodiment, the motor 2 is a rotary electromagnetic motor. However, the present invention is not limited to this. For example, the motor 2 may be a linear motor.

  Furthermore, in the embodiment, an example in which the speed reducer 3 is a wave gear speed reducer has been described, but the present invention is not limited to this. The reducer 3 may be a gear reducer other than a wave gear reducer, such as a planetary gear reducer.

(Summary of embodiment)
In the present embodiment, in the combination of the rotary electromagnetic motor 2 and the reduction gear 3 with a high reduction ratio, soft and highly accurate joint control in software is realized with an inexpensive hardware configuration. In addition, a method for reducing the required bandwidth of the torque sensor 5 by using both the disturbance observer 9 and feedforward friction compensation and performing disturbance compensation at a low cost and with relatively high accuracy has been proposed. In addition, by conducting a position control experiment in the presence of disturbances, it was shown that the proposed method can improve the response even in a drive system with large friction.

  The cause of hindering the soft control of the robot is mainly the friction torque generated inside the speed reducer 3. Measuring and feeding back the disturbance in the conventional configuration is equivalent to estimating and compensating for the friction torque generated inside the speed reducer 3 after all. In the present embodiment, the friction torque generated inside the reduction gear 3 is identified in advance, and the friction torque that must be compensated by feedback is reduced by compensating the friction torque in feed forward. As a result, the accuracy required for the torque sensor 5 that measures the disturbance is reduced.

  Specifically, the unit price of the torque sensor required in the conventional configuration is about several hundred thousand yen to about one million yen, but this embodiment reduces it to about several hundred yen. Further, according to the present embodiment, it is not necessary to use a flexible mechanical element such as a spring or a pneumatic actuator that significantly reduces the control accuracy, and the degree of compensation of the friction torque can be adjusted by software. Therefore, the positioning accuracy can be freely recovered during the work.

  The internal friction of a rotary speed reducer such as the gear type speed reducer 3 is a function of only the rotational speed and does not depend on the load inertia and the rotation angle. Therefore, it is possible to identify the characteristics of the friction torque generated inside the speed reducer 3 at the stage of a single unit including the motor 2 and the speed reducer 3 before being incorporated into the robot. The rotational speed of the speed reducer 3 can be measured relatively easily by the rotational speed sensor 4 using an encoder. For this reason, it is possible to superimpose a feedforward signal for canceling out the internal friction of the speed reducer 3 on the input voltage signal to the motor 2 in advance based on the measured rotational speed.

  Since the static friction torque generated when the stationary torque acts on the reduction gear 3 cannot be estimated, in order to compensate for the static friction torque, it is necessary to measure this with the torque sensor 5, but this is required. Since the measurement accuracy is not so high, the torque sensor 5 is an inexpensive sensor and is sufficiently practical.

  As described above, according to the present embodiment, motor control having both softness and hardness can be realized at a very low cost by combining feedback control and feedforward control. This embodiment greatly facilitates the automation of work that conventional robots have not been good at, that is, work that involves frequent and irregular collisions and contact with objects and people, and is likely to lead to the creation of new industries. .

[Example of software implementation]
The control blocks (particularly the dynamic friction compensator 6 and the static friction compensator 7) of the motor control device 1 may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or a CPU (Central It may be realized by software using a Processing Unit.

  In the latter case, the dynamic friction compensator 6 and the static friction compensator 7 are recorded with a CPU that executes instructions of a program that is software that realizes each function, and the program and various data that can be read by a computer (or CPU). A ROM (Read Only Memory) or a storage device (these are referred to as “recording media”), a RAM (Random Access Memory) for expanding the program, and the like are provided. And the objective of this invention is achieved when a computer (or CPU) reads the said program from the said recording medium and runs it. As the recording medium, a “non-temporary tangible medium” such as a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. The program may be supplied to the computer via an arbitrary transmission medium (such as a communication network or a broadcast wave) that can transmit the program. The present invention can also be realized in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

  The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

  The present invention can be used for a motor control device and a motor control method for controlling a motor having a reduction gear.

  The motor control device and motor control method according to the present invention can be used for machine tools and robots, and in particular, robots that perform tasks for the first time when they collide with objects, such as nursing robots, disasters, etc. It can be used for local reconstruction support robots and disaster rescue robots.

  Furthermore, the present invention can be used for assembly work by a robot of various complex shaped parts in a cell-based production system.

  Furthermore, the present invention can be used for life support and welfare care support for the elderly and persons with disabilities.

  Furthermore, the present invention can be used for remote robot work including advanced equipment operation in a super-complex environment such as a nuclear power plant.

DESCRIPTION OF SYMBOLS 1 Motor controller 2 Motor 3 Reducer 4 Rotation speed sensor 5 Torque sensor 6 Dynamic friction compensator 7 Static friction compensator 8 Drive circuit 9 Disturbance observer 10 Torque estimator 11 Timing belt 12 Arm 13 Photo interrupter 14 Ring 15 Slit

Claims (6)

A motor control device for controlling a motor having a speed reducer,
A rotational speed sensor for detecting the rotational speed of the output shaft of the speed reducer;
A torque sensor for detecting the torque of the output shaft of the speed reducer;
Based on a rotational speed signal representing the rotational speed detected by the rotational speed sensor and a predetermined dynamic friction model of the speed reducer, a dynamic friction compensation signal for compensating the dynamic friction torque of the speed reducer is generated, and A dynamic friction compensator to be supplied to the motor;
Based on a torque signal representing the torque detected by the torque sensor and a motor inertia torque signal generated based on an angular acceleration signal representing an angular acceleration corresponding to the rotational speed of the output shaft of the speed reducer, A motor control device comprising: a static friction compensator that generates a static friction compensation signal that compensates for the static friction torque of the speed reducer and supplies the static friction compensation signal to the motor.   The motor control device according to claim 1, wherein the dynamic friction model is provided with a voltage applied to the motor as a function of a rotational speed of an output shaft of the speed reducer.   The static friction compensation signal is generated based on the motor inertia torque signal, a torque estimation signal for estimating the torque based on the torque signal, and an applied voltage signal representing an applied voltage to the motor. The motor control device according to 1.   The motor control device according to claim 1, wherein the speed reducer is a gear speed reducer.   The motor control device according to claim 4, wherein the gear reducer is a wave gear reducer. A motor control method for controlling a motor having a reduction gear,
A rotational speed detection step of detecting the rotational speed of the output shaft of the speed reducer;
A torque detection step of detecting torque of the output shaft of the speed reducer;
Based on a rotational speed signal representing the rotational speed detected by the rotational speed detecting step and a predetermined dynamic friction model of the speed reducer, a dynamic friction compensation signal for compensating the dynamic friction torque of the speed reducer is generated. A dynamic friction compensation step for supplying the motor;
Based on the torque signal representing the torque detected by the torque detection step and the motor inertia torque signal generated based on the angular acceleration signal representing the angular acceleration corresponding to the rotational speed of the output shaft of the speed reducer, And a static friction compensation step of generating a static friction compensation signal for compensating the static friction torque of the speed reducer and supplying the static friction compensation signal to the motor.