US20080065240A1 - Position control method, position control device, and medium storage device having disturbance suppression function - Google Patents

Abstract

In a position control device having a disturbance suppression function, loop gain is calibrated without stopping disturbance suppression control. The position control device has, for disturbance suppression control, a feedback controller for changing the loop characteristic, a table for storing a target gain according to a disturbance frequency, and a gain calibration section for calibrating open loop gain. The gain is calibrated using a target gain in the table according to the change of the loop characteristic of the feedback controller. Open loop gain can be calibrated without interrupting the disturbance suppression control, so the open loop gain can be accurately calibrated without being affected by disturbance, and accurate position control is possible.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
• This application is based upon and claims the benefit of priority from the prior Japanese Patent Application No. 2006-246451, filed on Sep. 12, 2006, the entire contents of which are incorporated herein by reference.
BACKGROUND OF THE INVENTION
• 1. Field of the Invention
• The present invention relates to a position control method, position control device and medium storage device for controlling the position of an object by suppressing disturbance, and more particularly to a position control method, position control device, and medium storage device for suppressing the position deviation of an object due to disturbance.
• 2. Description of the Related Art
• A position control device is for controlling the position of an object to a specified position and is widely used in various fields. For example, in a medium storage device, such as a magnetic disk device and an optical disk device, it is extremely critical to accurately position the head to a target track to improve recording density.
• In this position control, it is known that disturbance affects positioning accuracy. In order to suppress this disturbance by a control system, a method of installing a filter for suppressing specific frequencies (e.g. see U.S. Pat. No. 6,487,028 B1, R. J. Bickel and M. Tomizuka: “Disturbance observer based hybrid impedance control”, (Proceedings of the American Control Conference, 1995, pp. 729-733)), and a method of suppressing such disturbance as eccentricity by observer control (e.g. see Japanese Patent Application Laid-Open No. H7-50075 and Japanese Patent Application Laid-Open No. 2000-21104) have been proposed.
• On the other hand, in order for the positioning control system to operate accurately, calibration of gain of open loop characteristics (open loop gain) is indispensable. The open loop gain changes if the characteristics of elements constituting a feedback loop change (e.g. temperature, age-based deterioration). Always maintaining this open loop gain at an optimum contributes to the performance of the feedback loop. This calibration is performed when the power of a device is turned ON, when temperature is changed, or when a predetermined time elapsed, for example, so as to calibrate the open loop gain at an optimum.
• As a method for calibrating a gain, applying a sine wave disturbance to a position or current and acquiring and comparing waveforms before and after applying the sine wave disturbance so as to measure the gain of the open loop characteristic is known (e.g. see Japanese Patent Application Laid-Open No. H11-328891, Japanese Patent Application Laid-Open No. H8-167160, and Japanese Patent Application Laid-Open No. H11-96704).
• In the case of this conventional gain calibration method, a target gain of the loop characteristic for adjusting the gain and a sine wave disturbance frequency for adjusting the gain are fixed values. In other words, in the case of the conventional gain calibration method, gain is calibrated assuming that the position control system has only one type of characteristic during gain adjustment.
• However in the case of a position control system with a disturbance suppression function, the characteristics of the position control system change according to the disturbance frequency to be suppressed. For example, in a control system using adaptive control, the loop characteristic of the control system is different from that before disturbance is applied, when the position control system is following up disturbance based on the adaptive control in a state of applied disturbance vibration.
• In order to execute gain adjustment while this adaptive control is following up disturbance, it is necessary to stop the adaptive control and switch to the control system for gain adjustment. In this case, the external vibrations cannot be sufficiently suppressed and positioning accuracy drops since the adaptive control is stopped. As a result, accuracy of gain adjustment drops.
• In other words, in the case of such a position control system as an adaptive control system, disturbance is suppressed by changing the loop characteristic according to disturbance, so it is difficult to accurately calibrate the open loop gain of the control system using prior art. Particularly if suppression width is taken wide or if disturbance in a high frequency area is suppressed to meet the recent demand for adapting a wide range of disturbance frequencies, the original characteristics of the controller are influenced, so it is more difficult to calibrate gain accurately.
SUMMARY OF THE INVENTION
• With the foregoing in view, it is an object of the present invention to provide a position control method, position control device, and medium storage device for accurately calibrating the open loop gain in a position control system using the disturbance adaptive control.
• It is another object of the present invention to provide a position control method, position control device, and medium storage device for accurately calibrating the open loop gain even while disturbance is being applied.
• It is still another object of the present invention to provide a position control method, position control device, and medium storage device for accurately calibrating the open loop gain, even if the disturbance suppression frequency is set from the outside.
• It is still another object of the present invention to provide a position control method, position control device, and medium storage device for accurately calibrating the open loop gain, even if the position control system adapts to disturbance suppression frequencies in a wide range.
• A position control method according to the present invention is a position control method for controlling a position of an object to a predetermined position by an actuator, having: a step of computing a position error based on a target position of the object and a current position of the object; a step of computing a control value in which a disturbance frequency component is suppressed, using a predetermined feedback loop based on the position error, and multiplying the result by a loop gain, so as to compute a drive value of the actuator; a step of fetching a target loop gain according to the disturbance frequency from a table; a step of adding the disturbance of a measurement frequency to the feedback loop and measuring a loop gain of the feedback loop; and a step of calibrating the loop gain of the control value computing step based on the measured loop gain and the target loop gain.
• A position control device for controlling a position of an object to a predetermined position by an actuator according to the present invention, has: a control section for computing a position error based on a target position of the object and a current position of the object, computing a control value in which a disturbance frequency component is suppressed, using a predetermined feedback loop based on the position error, and multiplying the result by a loop gain, so as to compute a drive value of the actuator; and a table for storing a target loop gain according to the disturbance frequency, wherein the control section fetches the target loop gain according to the disturbance frequency from the table, adds the disturbance of a measurement frequency to the feedback loop, measures a loop gain of the feedback loop, and calibrates the loop gain in the control value computing step based on the measured loop gain and the target loop gain.
• A medium storage device according to the present invention, has: a head for at least reading data on a storage medium; an actuator for positioning the head to a predetermined position on the storage medium; a control section for computing a position error based on a target position of the head and a current position acquired from the head, computing a control value in which a disturbance frequency component is suppressed, using a predetermined feedback loop based on the position error, and multiplying the result by a loop gain, so as to compute a drive value of the actuator; and a table for storing a target loop gain according to the disturbance frequency, wherein the control section fetches the target loop gain according to the disturbance frequency from the table, adds the disturbance of a measurement frequency to the feedback loop, measures a loop gain of the feedback loop, and calibrates the loop gain in the control value computing step based on the measured loop gain and the target loop gain.
• In the present invention, it is preferable that the step of fetching the target loop gain further has a step of fetching a measurement frequency according to the disturbance frequency, and the measurement step further has a step of adding the disturbance of the fetched measurement frequency to the feedback loop and measuring a loop gain of the feedback loop.
• Also in the present invention, it is preferable that the step of fetching the measurement frequency further has a step of fetching a measurement frequency which does not overlap with the disturbance frequency.
• Also in the present invention, it is preferable that the drive value computing step further has a step of estimating the disturbance frequency according to the position error based on an adaptive control, computing a control value in which the disturbance frequency component is suppressed, according to the estimated disturbance frequency, and multiplying the result by a loop gain, so as to compute a drive value of the actuator.
• It is also preferable that the present invention further has a step of interrupting the estimation of the disturbance frequency according to the position error during calibration of the loop gain.
• Also in the present invention, it is preferable that the drive value computing step further has a step of estimating the disturbance frequency according to the position error based on an adaptive control, and changing a constant of a controller according to the estimated disturbance frequency, and a step of computing a control value in which the disturbance frequency component is suppressed, using the changed controller according to the position error, and multiplying the result by a loop gain so as to compute a drive value of the actuator.
• Also in the present invention, it is preferable that the drive value computing step further has a step of estimating the disturbance frequency according to the position error based on an adaptive control, and changing a constant of a controller constructed with an observer according to the estimated disturbance frequency, and a computing step of computing a control value in which the disturbance frequency component is suppressed, according to the position error using the changed observer, and multiplying the result by a loop gain, so as to compute a drive value of the actuator.
• Because of disturbance suppression control, open loop gain is calibrated using a target gain according to a disturbance frequency, even if a loop characteristic of a feedback controller changes, so the open loop gain can be calibrated without interrupting the disturbance suppression control. Therefore the open loop gain can be accurately calibrated without being affected by disturbance, and accurate position control is possible.
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1 is a block diagram depicting a medium storage device according to an embodiment of the present invention;
• FIG. 2 is a diagram depicting the position signals of the medium storage device inFIG. 1;
• FIG. 3 is a diagram depicting details of the position signals inFIG. 2;
• FIG. 4 is a block diagram depicting a position control system according to the first embodiment of the present invention;
• FIG. 5 shows a target gain table inFIG. 4;
• FIG. 6 is a characteristic diagram of the sensitivity function of the feedback loop inFIG. 4;
• FIG. 7 is an open loop characteristic diagram of the feedback loop inFIG. 4;
• FIG. 8 is a flow chart depicting the gain calibration processing of the gain calibration block inFIG. 4;
• FIG. 9 is a diagram depicting the sine wave disturbance of the gain calibration processing inFIG. 8;
• FIG. 10 is a block diagram depicting a position control system according to the second embodiment of the present invention;
• FIG. 11 shows a target gain table of the embodiment inFIG. 10;
• FIG. 12 is a block diagram depicting a position control system according to the third embodiment of the present invention;
• FIG. 13 is a block diagram depicting an embodiment where the controller inFIG. 12 is constructed with a current observer;
• FIG. 14 shows a parameter table of the embodiment inFIG. 13;
• FIG. 15 shows another parameter table of the embodiment inFIG. 13;
• FIG. 16 is a diagram depicting a measurement frequency in the parameter table inFIG. 15;
• FIG. 17 shows still another parameter table of the embodiment inFIG. 13;
• FIG. 18 is a diagram depicting a measurement frequency of the parameter table inFIG. 17;
• FIG. 19 is a block diagram depicting a position control system according to the fourth embodiment of the present invention; and
• FIG. 20 is a block diagram depicting a position control system according to the fifth embodiment of the present invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
• Embodiments of the present invention will now be described in the sequence of the position control device, first embodiment of the position control device, loop gain calibration processing, second embodiment, third embodiment, design of disturbance observer, fourth embodiment, fifth embodiment, and other embodiments, but the present invention is not limited to these embodiments.
Position Control Device
• FIG. 1 is a block diagram depicting the position control device according to an embodiment of the present invention,FIG. 2 is a diagram depicting the arrangement of the position signals of the magnetic disk inFIG. 1, andFIG. 3 is a diagram depicting the position signals of the magnetic disk inFIG. 1 andFIG. 2.
• FIG. 1 shows a magnetic disk device, which is a type of disk device, as a position control device. AsFIG. 1 shows, in a magnetic disk device, amagnetic disk4, which is a magnetic storage medium, is installed at arotation axis2 of aspindle motor5. Thespindle motor5 rotates themagnetic disk4. An actuator (VCM)1 has amagnetic head3 at the tip, and moves themagnetic head3 in the radius direction of themagnetic disk4.
• Theactuator1 is comprised of a voice coil motor (VCM) which rotates with the rotation axis as the center. InFIG. 1, twomagnetic disks4 are mounted on the magnetic disk device, and fourmagnetic heads3 are simultaneously driven by thesame actuator1.
• Themagnetic head3 is a separation type magnetic head which has a read element and a write element. Themagnetic head3 is comprised of a read element, including a magneto-resistance (MR) element, stacked on a slider, and a write element, including a write coil, stacked thereon.
• Aposition detection circuit7 converts the position signals (analog signals) read by themagnetic head3 into digital signals. A read/write (R/W)circuit10 controls the reading and writing of themagnetic head3. A spindle motor (SPM)device circuit8 drives thespindle motor5. A voice coil motor (VCM)drive circuit6 supplies drive current to the voice coil motor (VCM)1, and drives theVCM1.
• A microcontroller (MCU)14 detects (demodulates) the current position from the digital position signals from theposition detection circuit7, and computes a VCM drive command value according to an error between the detected current position and a target position. In other words, themicrocontroller14 performs position demodulation and servo control, including the disturbance suppression described inFIG. 4 and later. A read only memory (ROM)13 stores a control program of theMCU14. A random access memory (RAM)12 stores data for processing of theMCU14.
• A hard disk controller (HDC)11 judges a position in one track based on a sector number of a servo signal, and records/reproduces the data. A random access memory (RAM) forbuffer15 temporarily stores the read data or write data. TheHDC11 communicates with a host via an interface IF, such as USB, ATA and SCSI. Abus9 connects these composing elements.
• AsFIG. 2 shows, on themagnetic disk4, servo signals (position signals)16 are arrayed in each track in the circumferential direction from the outer circumference to the inner circumference with an equal interval. Each track has a plurality of sectors, and the solid lines inFIG. 2 indicate positions where the servo signals16 are recorded. AsFIG. 3 shows, the position signal is comprised of a servo mark ServoMark, track number GrayCode, index Index, and offset information (servo burst) PosA, PosB, PosC and PosD. The dotted line inFIG. 3 shows the track center.
• The position signals inFIG. 3 are read by thehead3, and the position of the magnetic head in the radius direction is detected using the track number GrayCode and offset information PosA, PosB, PosC and PosD. Also the position of the magnetic head in the circumference direction is acquired based on the index signal Index.
• For example, the sector number when the index signal is detected is set to No. 0, which is counted up every time the servo signal is detected, so as to acquire the sector number of each sector of the track. The sector number of the servo signals is used as a reference when data is recorded/reproduced. There is one index signal in one track. The sector number may be set instead of the index signal.
• TheMCU14 inFIG. 1 confirms the position of theactuator1 through theposition detection circuit7, performs servo computation, and supplies appropriate current to theVCM1. In other words, in seek control, the head is moved to the target position through the transition from coarse control, settling control and following control. For all these controls, the current position of the head must be detected.
• To confirm the position like this, the servo signals are recorded on the magnetic disk in advance, as mentioned inFIG. 2. In other words, asFIG. 3 shows, servo marks which indicate the start position of the servo signal, gray code which indicates the track number, index signal, and signals PosA to PosD which indicate the offset are recorded on the magnetic disk in advance. These signals are read by the magnetic head, and these servo signals are converted into digital values by theposition detection circuit7.
First Embodiment of Position Control System
• FIG. 4 is a block diagram depicting a first embodiment of the position control system of the present invention, and is a block diagram of a position control system for suppressing disturbance which theMCU14 inFIG. 1 executes.FIG. 5 shows the target gain table inFIG. 4, andFIG. 6 andFIG. 7 are characteristic diagrams of the disturbance adaptive control inFIG. 4.
• The position control system inFIG. 4 controls a disturbance suppression compensation function of a controller which is set from the outside, or according to a detected disturbance frequency Fdist. A gain adjustment function is added to this position control system. Anerror computing block24 subtracts an observation position (current position) ‘y’ from a target position ‘r’ to compute a position error ‘e’.
• According to the position error ‘e’, acontroller20 computes a drive instruction value ‘u’ of a plant22 (1,3) to make the position error ‘e’. Thecontroller20 computes the drive instruction value ‘u’ by a known PID control, PI control+LeadLag, and observer control, for example.
• Again multiplication block26 multiplies the drive instruction value ‘u’ from thecontroller20 by a gain which is set (open loop gain), and outputs the result. A power amplifier, which is not illustrated, converts this output to a drive current I of the plant22 (1,3), and drives the plant22 (1,3).
• A disturbancesuppression compensation block30 changes a characteristic (e.g. constant) of thecontroller20 according to a disturbance suppression frequency which is set from the outside, or an estimated disturbance frequency Fdist, and adds a disturbance frequency suppression characteristic by thecontroller20. This suppression characteristic will be described using the sensitivity function (1/(1+CP)) and open loop characteristic (CP). C is a characteristic of a controller, and P is a characteristic of a plant.
• FIG. 6 shows an example of the frequency characteristic of a sensitivity function of the position control system, where a frequency (Hz) vs. gain characteristic (dB) is shown on the top, and a frequency (Hz) vs. phase characteristic (deg) is shown at the bottom ofFIG. 6. AsFIG. 6 shows, the sensitivity function is changed as shown in the thin line from the original characteristic of the controller which is shown in the thick line, according to the frequency to be suppressed. Here the sensitivity function in the case of suppressing a 500 Hz disturbance frequency is shown.
• FIG. 7 shows an example of a frequency characteristic of the open loop characteristic of the position control system corresponding toFIG. 6, where a frequency (Hz) vs. gain characteristic (dB) is shown on the top, and a frequency (Hz) vs. phase characteristic (deg) is shown at the bottom ofFIG. 7. AsFIG. 7 shows, the open loop characteristic is changed as shown in the thin line from the original characteristic of the controller which is shown in the thick line, according to the frequency to be suppressed. Here the open loop characteristic in the case of suppressing a 500 Hz disturbance frequency is shown.
• Acontroller20 which implements this sensitivity function or open loop characteristic for suppressing a specific disturbance frequency will be concretely described inFIG. 12 or later.
• Back inFIG. 4, again calibration block34 applies a sine wave disturbance SD having a predetermined frequency according to a gain calibration instruction, detects signals in a loop before and after applying the sine wave disturbance, and calibrates a gain in thegain multiplication block26. InFIG. 4, a sine wave disturbance SD for measurement is applied from anaddition block28 to a feedback loop according to a position (position error) which is an input of thecontroller20, and position errors before and after applying the sine wave disturbance are observed. A target gain table32 stores a target gain G corresponding to a disturbance frequency Fdist, supplies the target gain G corresponding to the disturbance frequency Fdist to thegain calibration block34, and uses it as a reference of the gain calibration in thegain calibration block34.
• This target gain table34 stores target gains TG1, TG2, . . . , TGnwo for each disturbance frequency Fdist, as shown inFIG. 5. This target gain is decided corresponding to the loop characteristic, which changes by the disturbance suppression control, as described inFIG. 6 andFIG. 7.
Gain Calibration Processing
• Now the calibration processing of thegain calibration block34 will be described.FIG. 8 is a flow chart depicting the gain calibration processing which thegain calibration block34 executes, andFIG. 9 is a diagram depicting the sine wave disturbance thereof.
• (S10) Thegain calibration block34 acquires a target gain TG corresponding to a disturbance frequency from the target gain table34.
• (S12) Thegain calibration block34 applies a sine wave disturbance SD to theaddition block28.FIG. 9 shows an example of a waveform of a sine wave disturbance SD to be applied, and a sine wave at 800 Hz, for example, is used.
• (S14) Thegain calibration block34 observes signals S1 and S2 before and after the disturbance is applied. InFIG. 4, the input stage of theaddition block28 is observed (acquired) as a signal (position error) S1 before the disturbance is applied, and the output stage of theaddition block28 as a signal (position error) S2 after the disturbance is applied.
• (S16) Thegain calibration block34 performs DFT (Digital Fourier Transfer) operation on each of the observed signals S1 and S2, so as to determine the level (amplitude) of each signal S1 and S2. This is repeated for N number of turns of the disk (e.g. 10 turns), and the sums thereof ΣS1 and ΣS2 are calculated.
• (S18) Thegain calibration block34 divides the sum of the amplitudes of the signals S1 before the disturbance is applied, that is ΣS1, by the sum of the amplitudes of the signals S2 after the disturbance is applied, that is ΣS2, so as to calculate an open loop gain Tm.
• (S20) Then thegain calibration block34 calculates a correction gain Tc by dividing the target gain TG corresponding to the disturbance frequency, which is acquired in step S10, by the measured open loop gain Tm.
• (S22) Thegain calibration block34 sets this correction gain Tc in thegain calibration block26, and ends the calibration.
• It is desirable to set this target gain TG for each disturbance frequency, but this increases the memory capacity of thetarget gain memory32. Therefore asFIG. 5 shows, the table32 may store the target gain TG at a multiple of Fr (e.g. rotation frequency), so that for disturbance frequencies between these multiple frequencies, a target gain corresponding to the disturbance frequency Fr is calculated by interpolation.
• By suppressing disturbance like this, the open loop gain is calibrated using a target gain according to the disturbance frequency, so the open loop gain can be calibrated without interrupting disturbance suppression control, even if the loop characteristic of thecontroller20 is changed. Therefore the open loop gain can be accurately calibrated without being affected by the disturbance.
Second Embodiment of Position Control System
• FIG. 10 is a block diagram depicting a second embodiment of the position control system of the present invention. InFIG. 10, composing elements the same as those inFIG. 4 are denoted with the same reference symbols.
• Just likeFIG. 4, anerror computing block24 subtracts an observation position (current position) y from the target position r to compute a position error e. Acontroller20 computes, according to position the error e, a drive instruction value u of a plant22 (1,3) to make the position error e zero. Thecontroller20 computes the drive instruction value u by a known PID control, PI control+LeadLag, and observer control, for example.
• Again multiplication block26 multiplies the drive instruction value u from thecontroller20 by a gain which is set (open loop gain), and outputs the result. A power amplifier, which is not illustrated, converts this output into a drive current I of the plant22 (1,3), and drives the plant22 (1,3).
• A disturbancesuppression compensation block30 changes the characteristic (e.g. constant) of thecontroller20 according to a disturbance suppression frequency which is set from the outside, or an estimated disturbance frequency Fdist, and adds a disturbance frequency suppression characteristic by thecontroller20.
• Again calibration block34 applies a sine wave disturbance SD having a predetermined frequency according to a gain calibration instruction, detects signals in a loop before and after applying the sine wave disturbance, and calibrates a gain in thegain multiplication block26. InFIG. 11, a sine wave disturbance SD for measurement is applied from an addition block28-1 to a feedback loop according to a current level (drive instruction value) which is an output of the controller, and currents before and after applying the sine wave disturbance are observed. A target gain table32 stores a target gain G corresponding to a disturbance frequency Fdist, supplies the target gain G corresponding to the disturbance frequency Fdist to thegain calibration block34, and uses it as a reference of the gain calibration in thegain calibration block34.
• This target gain table34 stores target gains TG1, TG2, . . . , TGn for each disturbance frequency Fdist, as shown inFIG. 5 orFIG. 10. This target gain is decided corresponding to the loop characteristic, which changes by the disturbance suppression control, as described inFIG. 6 andFIG. 7.
• This gain calibration processing is the same asFIG. 8, except that the observation target is the current level of the output stage. In this way, gain can be calibrated also by observing the current level.
• FIG. 11 shows another target gain table. Compared with the table32 inFIG. 5, this table32-1 has a measurement frequency column in addition to the target gain, for each disturbance frequency Fdist. In other words, the measurement frequency can also be changed according to the disturbance frequency Fdist.
• Here if the disturbance frequency Fdist is a predetermined measurement frequency fsd, the measurement frequency is changed to fsd+α (α≠0). In other words, applying a measurement frequency the same as a disturbance frequency to the control system as a disturbance while the position control system is controlling disturbance suppression means that the disturbance frequency is supplied into the loop while suppressing the same disturbance frequency.
• Therefore the sine wave disturbance of the measurement frequency is also suppressed by the disturbance suppression function, and accurate open loop gain cannot be measured. So the measurement frequency is shifted so as not to overlap with the disturbance frequency, then the open loop gain can be accurately measured.
• The table32-1 inFIG. 11 can also be applied to the first embodiment inFIG. 4.
Third Embodiment
• FIG. 12 is a block diagram depicting a third embodiment of the position control system of the present invention,FIG. 13 is a block diagram when the control system inFIG. 12 is constructed with a current observer, andFIG. 14 is the parameter table inFIG. 13.
• FIG. 12 shows a position control system for detecting a disturbance frequency and suppressing the disturbance by adaptive control, and inFIG. 12, composing elements the same as those inFIG. 4 are denoted with the same reference symbols.
• In other words, anerror computing block24 subtracts an observation position (current position) ‘y’ from a target position ‘r’ to compute a position error ‘e’. According to the position error ‘e’, acontroller20 computes a drive instruction value ‘u’ of a plant22 (1,3) to make the position error e zero. Thecontroller20 computes the drive instruction value ‘u’ by an observer control to be described later inFIG. 13, for example.
• Again multiplication block26 multiplies the drive instruction value ‘u’ from thecontroller20 by a gain which is set (open loop gain), and outputs the result. A power amplifier, which is not illustrated, converts this output into a drive current I of the plant22 (1,3), and drives the plant22 (1,3).
• A disturbancesuppression compensation block30 is comprised of an external suppression adaptive control system. This adaptive control system has a ω estimation section30-1 for estimating a disturbance frequency Fdist (ω) according to an adaptive rule based on the position error of thecontroller20, and a table30-2 for storing estimated gains L and A of thecontroller20 according to an estimated frequency (angular frequency ω in this case).
• Again calibration block34 applies a sine wave disturbance SD having a predetermined frequency according to a gain calibration instruction, detects signals in a loop before and after applying the sine wave disturbance, and calibrates a gain in thegain multiplication block26. InFIG. 12, a sine wave disturbance SD for measurement is applied from theaddition block28 to a feedback loop according to a position which is an input of the controller, and positions before and after applying the sine wave disturbance are observed.
• A target gain table32 stores a target gain G corresponding to a disturbance frequency Fdist, supplies the target gain G corresponding to the disturbance frequency Fdist from the ω estimation section30-1 to thegain calibration block34, and uses it as a reference of the gain calibration in thegain calibration block34.
• The present embodiment also has a switch SW for stopping the input of a position error of thecontroller20 to the ω estimation section30-1 during gain calibration. By this, the ω estimation section30-1 maintains the estimated disturbance frequency before starting gain calibration during gain calibration. Therefore during gain calibration, the disturbance suppression control is executed, but adaptive control is interrupted so that unnecessary adaptive control is not performed by a sine wave disturbance for measuring the gain.
• In the present embodiment, the characteristic (e.g. constant) of thecontroller20 is changed according to an estimated disturbance frequency Fdist, and the disturbance frequency suppression characteristic is added by thecontroller20.
• The present embodiment will be described in more detail using thecontroller20 based on a current observer inFIG. 13. InFIG. 13, composing elements the same as those inFIG. 12 are shown with the same reference symbols. The current observer shown inFIG. 13 is a current observer which includes bias compensation shown in the following Expressions (1), (2), (3), (4) and (5).
• $\begin{array}{cc}\left(\begin{array}{c}x\left(k\right)\\ v\left(k\right)\\ b\left(k\right)\\ z1\left(k\right)\\ z2\left(k\right)\end{array}\right)=\left(\begin{array}{c}x\left(k\right)\\ v\left(k\right)\\ b\left(k\right)\\ z1\left(k\right)\\ z2\left(k\right)\end{array}\right)+\left(\begin{array}{c}L1\\ L2\\ L3\\ L4\\ L5\end{array}\right)\left(y\left(k\right)-x\left(k\right)\right)& \left(1\right)\\ u\left(k\right)=-\left(\begin{array}{cc}F1& F2\end{array}\right)\left(\begin{array}{c}x\left(k\right)\\ v\left(k\right)\end{array}\right)& \left(2\right)\\ \mathrm{uout}\left(k\right)=u\left(k\right)-\left(\begin{array}{ccc}F3& F4& F5\end{array}\right)\left(\begin{array}{c}b\left(k\right)\\ z1\left(k\right)\\ z2\left(k\right)\end{array}\right)& \left(3\right)\\ \left(\begin{array}{c}x\left(k+1\right)\\ v\left(k+1\right)\end{array}\right)=\left(\begin{array}{cc}1& 1\\ 0& 1\end{array}\right)\left(\begin{array}{c}x\left(k\right)\\ v\left(k\right)\end{array}\right)+\frac{\mathrm{Bl}}{m}\frac{1}{\mathrm{Lp}}T^2\left(\begin{array}{c}1/2\\ 1\end{array}\right)u\left(k\right)& \left(4\right)\\ \begin{array}{c}b\left(k+1\right)=b\left(k\right)\\ \left(\begin{array}{c}z1\left(k+1\right)\\ z2\left(k+1\right)\end{array}\right)=\left(\begin{array}{cc}a11& a12\\ a21& a22\end{array}\right)\left(\begin{array}{c}z1\left(k\right)\\ z2\left(k\right)\end{array}\right)\end{array}\}& \left(5\right)\end{array}$
• In other words, this embodiment is an example of an adaptive control system where thecontroller20 is separated into a model of the controller and a disturbance model.
• InFIG. 13, afirst computing block24 computes an actual position error er[k] by subtracting a target position ‘r’ from an observation position y[k] which is acquired by demodulating the servo information read by thehead3. Asecond computing block40 computes an estimated position error e[k] by subtracting an estimated position x[k] of the observer from the actual position error er[k].
• In the controller model, this estimated position error e[k] is input to astate estimation block42, and an estimated correction value (right hand side of Expression (1)) is computed using an estimated gain La (L1, L2) of the controller. And in theaddition block44, the result is added with state quantities (left hand side of Expression (1)) x[k] and v[k] from thedelay block46, and estimated position x[k] and estimated velocity v[k] are acquired, as shown in Expression (1). In Expression (1), the estimated position error e[k] is indicated by (y[k]−x[k]).
• The estimated values x[k] and v[k] are multiplied by a state feedback gain (−Fa=F1, F2) in afourth computing block48, and a first drive value u[k] for theactuator1 is acquired, as shown in Expression (2). On the other hand, the estimated values x[k] and v[k] of Expression (1) from anaddition block44 are multiplied by an estimated gain Aa (2×2 matrix (1,0) in Expression (4)) in afifth computing block52.
• The drive value u[k] in thefourth computing block48 is multiplied by an estimated gain Ba (a value by which u[k] in Expression (4) is multiplied) in asixth computing block50. Both of the multiplication results are added in anaddition block54, and estimated state quantities x[k+1] and v[k+1] of the next sample in Expression (4) are acquired.
• The estimated state quantity of the next sample is input to adelay block46, and is corrected with the estimated correction value in thestate estimation block42. And for the estimated value of Expression (1) from theaddition block44, the estimated position x[k] is acquired in aseventh computing block56, and is input to thesecond computing block40.
• In the disturbance model, on the other hand, the estimated position error e[k] is input to astate estimation block60 of the disturbance, and an estimated correction value (right hand side of Expression (1)) is computed using estimated gains Ld1 (L3, L4, L5). And the result is added with a state quantity (left hand side of Expression (1)) from adelay block62 in anaddition block66, and an estimated bias value b[k] and estimated disturbance suppression values z1[k] and z2[k] are acquired, as shown in Expression (1).
• The estimated values b[k], z1[k] and z2[k] are multiplied by a state feedback gain (Fd1=F3, F4, F5) in aneighth computing block68, and the disturbance suppression drive value of theactuator1 is acquired, as shown in Expression (3).
• The estimated values b[k], z1[k] and z2[k] of Expression (1) from theaddition block66, on the other hand, are multiplied by an estimated gain Ad1 (gain of b[k] of Expression (5) and gain of 2×2 matrix A) in aninth computing block64, the result is input to thedelay block62, and estimated values b[k+1], z1[k+1] and z2[k+1] of the next sample are acquired.
• And in anaddition block70, the disturbance suppression drive value is subtracted from the drive value u[k], and an output drive value uout[k] of Expression (3) is acquired.
• In other words, the estimated gain L is separated between the controller model and disturbance model, and the feedback gain F is separated between the controller model and disturbance model, so as to design the controller model and disturbance model separately. The design of the disturbance observer will be described in detail later.
• Now as an input to an adaptive control system30-1, the estimated position error e[k] of the observer is supplied. The estimated position error e[k] of the observer is a difference value between an actual position error (r−y[k]) of acomputing block40 and the estimated position x[k] of the observer.
• The disturbance suppression adaptive control system has a ω estimation section30-1 for estimating a disturbance frequency according to an adaptive rule, and a table30-2 (32) for storing estimated gains L and A and a target gain according to an estimated frequency (angular frequency ω in this case). Aω estimation section24 calculates an estimated angular frequency ω1[k] from an estimated error e[k] using the following adaptive Expression (6).
• $\begin{array}{cc}\omega 1\left[k\right]=\omega 1\left[k-1\right]+\mathrm{Ka}·\frac{L5·z1\left[k\right]-L4·z2\left[k\right]}{z{1\left[k\right]}^2+z{2\left[k\right]}^2}e\left[k\right].& \left(6\right)\end{array}$
• This adaptive expression has an integration form for adaptively correcting an estimated angular frequency ω1[k−1] one sample before using estimated disturbance gains L4 and L5, estimated disturbance values z1[k] and z2[k] and estimation position error e[k]. Ka is a predetermined gain.
• Based on the estimated value of theaddition block66, the estimated disturbance values z1[k] and z2[k] are acquired, and output to the ω estimation section30-1. The ω estimation section30-1 has a computing section for computing the second term (Ka - - - e[k] of the ω adaptive expression in Expression (6), a delay section for delaying the estimated ω[k] by one sample, and an addition section for adding the delayed ω(ω[k−1]) and the computing result of the second term in the computing section. In other words, the adaptive expression of Expression (6) is computed.
• A table30-2, on the other hand, stores L1, L2, L3, L4 and L5 according to the value of the estimated disturbance frequency Fdist (estimated angular frequency ω), a11, a12, a21 and a22, and values of the target gain, as shown inFIG. 14. Based on L1, L2, L3, L4 and L5 of this table30-2, L1, L2, L3, L4 and L5 of the state estimation blocks42 and60 are changed according to the estimated angular frequency. Also based on a11, a12, a21 and a22 of this table30-2, a11, a12, a21 and a22 of the ninth computing block64 (see Expression (5)) are changed according to the estimated angular frequency.
• In other words, according to the disturbance (angular) frequency ω, the disturbance model and the estimated gain are changed, without changing the state feedback gain F. Here all the estimated gains of the observer are influenced, not only the disturbance model for shaping the frequency characteristic in a notch filter form. In other words, if the disturbance frequency ω or the disturbance model changes, not only the disturbance estimated gains L4 and L5 of Expression (1), but all of the gains L1, L2 and L3 of position, velocity and bias are influenced. Particularly this influence is major if the ζ2 value is high in the pole assignment when the disturbance model is designed in the form of a shaping filter, that is if the width of the suppression area in the notch filter form is wide in the frequency characteristic. Therefore all the estimated gains from L1 to L5 must be changed according to the disturbance frequency. The values of the estimated gains are calculated by the pole assignment method and stored in the table30-2 in advance.
• On the other hand, just likeFIG. 12, thegain calibration block34 applies a sine wave disturbance SD having a predetermined frequency according to the gain calibration instruction, detects signals in the loop before and after applying the sine wave disturbance, and calculates a gain of thegain multiplication block26. The sine wave disturbance SD for measurement is supplied from theaddition block28 to the feedback loop for a position which is an input of thecontroller20 constructed with the disturbance observer, and positions before and after applying the sine wave disturbance are observed.
• Here the target gain table is integrated with the parameter table30-2, as shown inFIG. 14. In other words, the parameter table30-2 stores a target gain G corresponding to a disturbance frequency Fdist, supplies the target gain G corresponding to the disturbance frequency Fdist from the ω estimation section30-1 to thegain calibration block34, and uses it as a reference of gain calibration in thegain calibration block34.
• The present embodiment also has a switch SW for stopping the input of a position error of thecontroller20 to the ω estimation section30-1 during gain calibration. By this, the ω estimation section30-1 maintains the estimated disturbance frequency before starting gain calibration during gain calibration. Therefore during gain calibration, the disturbance suppression control is executed, but adaptive control is interrupted so that unnecessary adaptive control is not performed by a sine wave disturbance for measuring the gain.
• In this way, a desired disturbance suppression function can be easily added to the controller by constructing the controller with the disturbance observer. Also the disturbance suppression adaptive control can be easily implemented by changing the estimated gain.
• FIG. 15 shows another parameter table, andFIG. 16 is a diagrams depicting the operation thereof. Compared with the table30-2 inFIG. 14, a column of measurement frequency Fcal is created for each disturbance frequency Fdist in this table30-2, in addition to the estimated gain and target gain. In other words, the measurement frequency can also be changed according to the disturbance frequency Fdist.
• Here, asFIG. 16 shows, a frequency deviated from the disturbance frequency Fdist by α is set for the measurement frequency. Therefore the measurement frequency is shifted from the disturbance frequency without overlapping, and open loop gain can be accurately measured.
• FIG. 17 shows another parameter table, andFIG. 18 is a diagram depicting the operation thereof. Compared with the table30-2 inFIG. 14, a column of measurement frequency Fcal is created for each disturbance frequency Fdist in this table30-2, in addition to the estimated gain and target gain.
• InFIG. 17, just likeFIG. 11, if the disturbance frequency Fdist is the measurement frequency fsd, the measurement frequency is changed to fsd+α (α≠0) or fsd−α. Also, asFIG. 18 shows, the measurement frequency is set to be higher than the disturbance frequency Fdist (e.g. fsd+α) if the disturbance frequency is lower than this measurement frequency fsd, and if the disturbance frequency is higher than this measurement frequency fsd, the measurement frequency is set to be lower than the disturbance frequency Fdist (e.g. fsd−α).
• In other words, the measurement frequency is shifted from the disturbance frequency, but if the shifting range is inappropriate, the gain calibration becomes difficult. For example, if the measurement frequency is set low, the measurement frequency extends beyond the lower limit of the servo band, and if the measurement frequency is set high, the measurement frequency extends beyond the upper limit of the servo band, so measurement itself becomes difficult.
• Therefore in an area where the disturbance frequency is low, the measurement frequency is set to higher than the disturbance frequency Fdist (e.g. fsd+α), and in an area where the disturbance frequency is high, the measurement frequency is set to lower than the disturbance frequency Fdist (e.g. fsd−α).
• And in the column of the disturbance frequency Fr*K(=fsd), two measurement frequencies, (fsd+α) and (fsd−α), are set, and the corresponding target gains GK1 and GK2 are set. By this, the interpolation of the gain can be switched at the boundary.
Designing Disturbance Observer
• Now the design procedure of the disturbance observer will be described. The observer control system when theactuator1 is a double integral model is given by the following analog expression, Expression (7).
• $\begin{array}{cc}\begin{array}{c}s\left(\begin{array}{c}x\\ v\end{array}\right)=\left(\begin{array}{cc}0& 1\\ 0& 0\end{array}\right)\left(\begin{array}{c}x\\ v\end{array}\right)+\frac{\mathrm{Bl}}{m}\left(\begin{array}{c}0\\ 1\end{array}\right)u+\left(\begin{array}{c}L1\\ L2\end{array}\right)\left(y-x\right)\\ y=\left(\begin{array}{cc}1& 0\end{array}\right)\left(\begin{array}{c}x\\ v\end{array}\right)\\ u=-\left(\begin{array}{cc}\mathrm{Fx}& \mathrm{Fv}\end{array}\right)\left(\begin{array}{c}x\\ v\end{array}\right)\end{array}\}& \left(7\right)\end{array}$
• In Expression (7), ‘s’ is a Laplace operator, ‘x’ is an estimated position, ‘v’ is an estimated velocity, ‘y’ is a current position, ‘r’ is a target position, L1 and L2 are estimated gains of position and velocity respectively, ‘u’ is a drive current, and B1/m is a force constant of theactuator1.
• This control system has asensitivity function 1/(1+CP), and the disturbance suppression for this sensitivity function is defined by the second degree filter in the following Expression (8).
• $\begin{array}{cc}\frac{{\mathrm{<figure-callout\; label="s"\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">s</figure-callout>}}^2+2{\mathrm{<figure-callout\; label="\varsigma "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\varsigma </figure-callout>}}_1{\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1s+{\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1^2}{{\mathrm{<figure-callout\; label="s"\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">s</figure-callout>}}^2+2{\mathrm{<figure-callout\; label="\varsigma "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\varsigma </figure-callout>}}_2{\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_2s+{\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_2^2}& \left(8\right)\end{array}$
• The disturbance model, of which denominator is a numerator of the shaping filter, is given by the following Expression (9).
• $\begin{array}{cc}\frac{1}{s^2+2{\mathrm{<figure-callout\; label="\varsigma "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\varsigma </figure-callout>}}_1{\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1s+{\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1^2}& \left(9\right)\end{array}$
• There are three possible methods to set up this disturbance model in the observer of the original controller (Expression (7)).
• The first method is setting up the disturbance model in Expression (9) as is. In other words, this is a second degree filter, so if the estimated state quantities of a disturbance are z1 and z2, and the estimated gains of disturbance are L3 and L4, the observer control system is expression by Expression (10).
• $\begin{array}{cc}\begin{array}{c}s\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)=\left(\begin{array}{cccc}0& 1& 0& 0\\ 0& 0& 1& 0\\ 0& 0& 0& 1\\ 0& 0& -{\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1^2& -2{\mathrm{<figure-callout\; label="\varsigma "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\varsigma </figure-callout>}}_1{\omega }_1\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)+\frac{\mathrm{Bl}}{m}\left(\begin{array}{c}0\\ 1\\ 0\\ 0\end{array}\right)u+\left(\begin{array}{c}L1\\ L2\\ L3\\ L4\end{array}\right)\left(y-x\right)\\ y=\left(\begin{array}{cccc}1& 0& 0& 0\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)\\ u=-\left(\begin{array}{cccc}\mathrm{Fx}& \mathrm{Fv}& K& 0\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)\\ K=\frac{m}{\mathrm{Bl}}\end{array}\}& \left(10\right)\end{array}$
• The second method is dispersing the term of the square of ω1, and Expression (11) is acquired by transforming Expression (10).
• $\begin{array}{cc}\begin{array}{c}s\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)=\left(\begin{array}{cccc}0& 1& 0& 0\\ 0& 0& 1& 0\\ 0& 0& 0& {\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1\\ 0& 0& -{\omega }_1& -2{\mathrm{<figure-callout\; label="\varsigma "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\varsigma </figure-callout>}}_1{\omega }_1\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)+\frac{\mathrm{Bl}}{m}\left(\begin{array}{c}0\\ 1\\ 0\\ 0\end{array}\right)u+\left(\begin{array}{c}L1\\ L2\\ L3\\ L4\end{array}\right)\left(y-x\right)\\ y=\left(\begin{array}{cccc}1& 0& 0& 0\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)\\ u=-\left(\begin{array}{cccc}\mathrm{Fx}& \mathrm{Fv}& K& 0\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)\\ K=\frac{m}{\mathrm{Bl}}\end{array}\}& \left(11\right)\end{array}$
• The third method is inverting the sign of ω1 in Expression (11), and is given by Expression (12).
• $\begin{array}{cc}\begin{array}{c}s\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)=\left(\begin{array}{cccc}0& 1& 0& 0\\ 0& 0& 1& 0\\ 0& 0& 0& -{\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1\\ 0& 0& {\omega }_1& -2{\mathrm{<figure-callout\; label="\varsigma "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\varsigma </figure-callout>}}_1{\omega }_1\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)+\frac{\mathrm{Bl}}{m}\left(\begin{array}{c}0\\ 1\\ 0\\ 0\end{array}\right)u+\left(\begin{array}{c}L1\\ L2\\ L3\\ L4\end{array}\right)\left(y-x\right)\\ y=\left(\begin{array}{cccc}1& 0& 0& 0\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)\\ u=-\left(\begin{array}{cccc}\mathrm{Fx}& \mathrm{Fv}& K& 0\end{array}\right)\left(\begin{array}{c}x\\ v\\ z1\\ z2\end{array}\right)\\ K=\frac{m}{\mathrm{Bl}}\end{array}\}& \left(12\right)\end{array}$
• Design is possible by any of these methods. The second and third methods are effective particularly when the model is transformed into a digital control system. In other words, the two state variables z1 and z2 are balanced, and the values of the estimated gains L3 and L4 of the observer for the two state variables are not very far apart.
• At this time, the values of the estimated gains L1, L2, L3 and L4 are designed by specifying the poles combining the pole of the shaping filter of Expression (8) (derived from denominator=0 in Expression (8)) and the poles used for designing the original observer control system.
• The observer control system combining the second degree filter shaping and the conventional steady state bias estimation is given by the following Expression (13).
• $\begin{array}{cc}\begin{array}{c}s\left(\begin{array}{c}x\\ v\\ b\\ z1\\ z2\end{array}\right)=\left(\begin{array}{ccccc}0& 1& 0& 0& 0\\ 0& 0& 1& 1& 0\\ 0& 0& 0& 0& 0\\ 0& 0& 0& 0& {\mathrm{<figure-callout\; label="\omega "\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1\\ 0& 0& 0& -{\omega }_1& -2{\mathrm{\varsigma \omega }}_1\end{array}\right)\left(\begin{array}{c}x\\ v\\ b\\ z1\\ z2\end{array}\right)+\frac{\mathrm{Bl}}{m}\left(\begin{array}{c}0\\ 1\\ 0\\ 0\\ 0\end{array}\right)u+\left(\begin{array}{c}L1\\ L2\\ L3\\ L4\\ L5\end{array}\right)\left(y-x\right)\\ y=\left(\begin{array}{ccccc}1& 0& 0& 0& 0\end{array}\right)\left(\begin{array}{c}x\\ v\\ b\\ z1\\ z2\end{array}\right)\\ u=-\left(\begin{array}{ccccc}\mathrm{Fx}& \mathrm{Fv}& K& K& 0\end{array}\right)\left(\begin{array}{c}x\\ v\\ b\\ z1\\ z2\end{array}\right)\\ K=\frac{m}{\mathrm{Bl}}\end{array}\}& \left(13\right)\end{array}$
• In this way, the filter form for shaping is considered first, then the disturbance model is added to the observer in designing. Therefore forms can be freely shaped without being limited by the physical response characteristic of the original disturbance model.
• Thus far description dealt with analog design. But in order to design a digital control system, on the other hand, the disturbance model is created in an analog space and an enlarged model is constructed, then after the enlarged model is transformed (digitized) in the digital space, the pole assignment is specified in the digital space.
• When the disturbance model has the characteristic of the second degree filter, if the enlarged model is transformed into a discrete system, both of the two variables z1 and z2 of the disturbance model in matrix A for designing the estimated gain of the observer influence theactuator1.
• Therefore correction is made so that only one of the variables of the disturbance model influences theactuator1, more specifically, so that only a variable, the same as that of analog design, influences theactuator1. In other words, after digitizing, the enlarged model is corrected.
• Specifically, when the analog model in the form of Expression (11), using the second degree filter, is digitized (that is z-transformation is performed and the result is converted into SI units), the following Expression (14) is established.
• $\begin{array}{cc}\begin{array}{c}z\left(\begin{array}{c}x\left[k\right]\\ v\left[k\right]\\ z1\left[k\right]\\ z2\left[k\right]\end{array}\right)=\left(\begin{array}{cccc}1& T& A13& A14\\ 0& 1& A23& A24\\ 0& 0& A33& A34\\ 0& 0& A43& A44\end{array}\right)\left(\begin{array}{c}x\left[k\right]\\ v\left[k\right]\\ z1\left[k\right]\\ z2\left[k\right]\end{array}\right)+\frac{\mathrm{Bl}}{m}\left(\begin{array}{c}{\mathrm{<figure-callout\; label="T"\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">T</figure-callout>}}^2/2\\ \mathrm{<figure-callout\; label="T"\; filenames="US20080065240A1-20080313-D00005.png"\; state="\{\{state\}\}">T</figure-callout>}\\ 0\\ 0\end{array}\right)u\left[k\right]\\ y=\left(\begin{array}{cccc}1& 0& 0& 0\end{array}\right)\left(\begin{array}{c}x\left[k\right]\\ v\left[k\right]\\ z1\left[k\right]\\ z2\left[k\right]\end{array}\right)\end{array}\}& \left(14\right)\end{array}$
• In Expression (14), z is a Z transformer and T is a sampling cycle. Here matrix A, that is A13, A14, A23 and A24, is focused on. Neither A14 nor A24 become “0” merely by digitizing. In other words, both of the two variables z1 and z2 of the disturbance model in matrix A for designing the estimated gain of the observer influence theactuator1.
• Therefore after digitizing the analog model, coefficients with which the state variables z1 and z2 of the disturbance model in matrix A influence theactuator1, are replaced.
• In the case of the examples of Expression (14), matrix A is corrected as the following Expression (15).
• $\begin{array}{cc}\begin{array}{c}\begin{array}{c}A14=A24=0\\ A13={\mathrm{<figure-callout\; label="T"\; filenames="US20080065240A1-20080313-D00000.png,US20080065240A1-20080313-D00001.png"\; state="\{\{state\}\}">T</figure-callout>}}^2/2\end{array}\\ A23=T\end{array}\}& \left(15\right)\end{array}$
• In the digital control system, the unit of distance is a track, the current value is normalized with the maximum current as “1”, and the velocity and acceleration are not in second units, but must be normalized by a sampling frequency.
• In the same way, if the observer in analog format in Expression (13) is transformed into the format of the current observer, Expression (16) is established.
• $\begin{array}{cc}\begin{array}{c}\left(\begin{array}{c}x\left(k\right)\\ v\left(k\right)\\ b\left(k\right)\\ z1\left(k\right)\\ z2\left(k\right)\end{array}\right)=\left(\begin{array}{c}x\left(k\right)\\ v\left(k\right)\\ b\left(k\right)\\ z1\left(k\right)\\ z2\left(k\right)\end{array}\right)+\left(\begin{array}{c}L1\\ L2\\ L3\\ L4\\ L5\end{array}\right)\left(y\left(k\right)-x\left(k\right)\right)\\ u\left(k\right)=-\left(\begin{array}{ccccc}F1& F2& F3& F4& F5\end{array}\right)\left(\begin{array}{c}x\left(k\right)\\ v\left(k\right)\\ b\left(k\right)\\ z1\left(k\right)\\ z2\left(k\right)\end{array}\right)\\ \left(\begin{array}{c}x\left(k+1\right)\\ v\left(k+1\right)\\ b\left(k+1\right)\\ z1\left(k+1\right)\\ z2\left(k+1\right)\end{array}\right)=\left(\begin{array}{ccccc}1& 1& 1/2& 1/2& 0\\ 0& 1& 1& 1& 0\\ 0& 0& 1& 0& 0\\ 0& 0& 0& a11& a12\\ 0& 0& 0& a21& a22\end{array}\right)\left(\begin{array}{c}x\left(k\right)\\ v\left(k\right)\\ b\left(k\right)\\ z1\left(k\right)\\ z2\left(k\right)\end{array}\right)+\frac{\mathrm{Bl}}{m}\frac{1}{\mathrm{Lp}}T^2\left(\begin{array}{c}1/2\\ 1\\ 0\\ 0\\ 0\end{array}\right)u\left(k\right)\end{array}\hspace{1em}\}& \left(16\right)\end{array}$
• In this way, if the disturbance model is designed to be a separate configuration, Expression (16) can be set up with the disturbance model separately, as shown inFIG. 12.
• In other words, in the comparison of Expression (16) with Expression (1) to Expression (5), Expressions (2) and (4) are Expression (16), wherein the model of the controller is independent, and Expression (3) and (5) are Expression (16), wherein thedisturbance model50 is separated.
Fourth Embodiment
• FIG. 19 is a block diagram depicting the fourth embodiment of the position control system for suppressing disturbance which theMCU14 inFIG. 1 executes. This position control system is a control system for detecting a disturbance frequency and suppressing the sine wave disturbance having a predetermined frequency. InFIG. 19, composing elements the same as those inFIG. 4 are denoted with the same reference symbols.
• A position error e between a target position r supplied from an interface circuit11-1 (in HDC11) and an observation position y is computed in acomputing unit24. This position error e is input to a controller20 (Cn) which performs feedback control. Thecontroller20 outputs a control current value Un by a known PID control, PI control+LeadLag, and observer control.
• A frequency estimation unit (ω estimation)30 for estimating the frequency of disturbance, and a compensator (Cd)20-1 for suppressing the disturbance having a specific frequency by adaptive control, are added to thiscontroller20.
• A sum U of an output Un of the controller20 (Cn) and an output Ud of the compensator20-1 (Cd) is determined in an addition block20-2, and is supplied to a plant22 (1,3) via again multiplication block26. By this, the position of thehead3 driven by theactuator1, which is the control target22, is controlled so as to follow up the disturbance. In other words, the device is vibrated by the distance, so the position of thehead3, with respect to themagnetic disk4, is also controlled to follow up the disturbance, so the position relationship of thehead3 and themagnetic disk4 does not change.
• Thisfrequency estimation unit30 estimates an angular frequency ω (=2πf) of the disturbance based on a position error e, as shown inFIG. 12, and supplies it to the transfer function of the disturbance frequency suppression of the compensator20-1. The compensator20-1 calculates the recurrence formula (adaptive control expression) of the sine wave based on the position error e and this estimated angular frequency ω, and a compensating current output Ud is calculated.
• In this way, in order to handle disturbance with an unknown frequency in a certain range, the frequency of disturbance is detected and the unknown frequency is suppressed. As a method for estimating an unknown frequency and suppressing disturbance of the unknown frequency, assuming the recurrent formula of a sine wave, or the above mentioned method of supplying adaptive rule based on the error signal e and correcting the drive quantity of the control target, can be used. Also a method of estimating an unknown frequency based on an error signal e, generating a disturbance suppression signal on the position level, correcting the error signal, and inputting it into the controller, can also be used.
• Here the interface circuit11-1 receives a disturbance suppression frequency from the outside, and sets it in thefrequency estimation unit30 as an initial value of the frequency estimation unit30 (initial value of the angular frequency of disturbance). Therefore the compensator20-1 performs adaptive control from this initial value.
• In other words, the initial value of thefrequency estimation unit30, which is normally based on the assumption that the disturbance frequency is unknown, is set at the center of the follow up range, with a position error e, and gradually reaches the disturbance frequency, but in the present embodiment, a known disturbance frequency is set as the initial value, so position control starts directly with the known disturbance frequency, and even if the frequency changes thereafter, the estimated frequency follows up from there.
• In this position control system where disturbance frequency is set from the outside, again calibration block34 and a target gain table32 are installed. Thegain calibration block34 applies a sine wave disturbance SD having a predetermined frequency according to a gain calibration instruction, detects signals in a loop before and after applying in the sine wave disturbance, and calibrates the gain of thegain multiplication block26. InFIG. 19, the sine wave disturbance SD for measurement is supplied from theaddition block28 to the feedback loop according to a position which is an input of the controller, and positions before and after applying the sine wave disturbance are observed.
• The target gain table32 stores a target gain G corresponding to a disturbance frequency Fdist, supplies the target gain G corresponding to the disturbance frequency Fdist from theω estimation section30 to thegain calibration block34, and uses it as a reference of gain calibration in thegain calibration block34.
• The present embodiment also has a switch SW for stopping the input of a position error to theω estimation section30 during gain calibration. By this, theω estimation section30 maintains the estimated disturbance frequency before starting gain calibration during gain calibration. Therefore during gain calibration, the distance suppression control is executed, but adaptive control is interrupted so that unnecessary adaptive control is not performed by a sine wave disturbance for measuring the gain.
• In this way, the present invention can also be applied to the position control system where disturbance frequency is set from the outside.
Fifth Embodiment
• FIG. 20 is a block diagram depicting the fifth embodiment of a position control system for suppressing disturbance which theMCU14 inFIG. 1 executes. This position control system is a control system for detecting a disturbance frequency and suppressing sine wave disturbance having a predetermined frequency. InFIG. 20, composing elements the same as those inFIG. 4,FIG. 12 andFIG. 19 are denoted with the same reference symbols.
• InFIG. 20, a position error e between a target position r supplied from an interface circuit11-1 (in HDC11) and an observation position y is computed in acomputing unit24. This position error e is input to a controller20 (Cn) which performs feedback control. Thecontroller20 outputs a control current value Un by a known PID control, PI control+LeadLag, and observer control.
• Afrequency converter30 for converting a frequency of disturbance to a corresponding angular frequency, and a compensator (Cd)20-1 for suppressing disturbance having a specific frequency by adaptive control are added to thiscontroller20.
• A sum U of an output Un of the controller20 (Cn) and an output Ud of the compensator20-1 (Cd) is supplied to a plant22 (1,3) via an addition block20-2 and again multiplication block26. By this, position of thehead3 driven by theactuator1, which is the control target22, is controlled so as to follow up the disturbance. In other words, the device is vibrated by the disturbance, so the position of thehead3 with respect to themagnetic disk4 is also controlled to follow up the disturbance, and the position relationship of thehead3 and themagnetic disk4 does not change.
• Thisfrequency converter30 supplies an angular frequency ω (=2πf) to a transfer function of the disturbance frequency suppression of the compensator20-1. The compensator20-1 calculates the recurrence formula (adaptive control expression) of the sine wave based on the position error e and this estimated angular frequency ω, and a compensating current output Ud is calculated.
• In this way, in order to handle disturbance with a frequency in a certain range, disturbance frequency which changes is suppressed according to the frequency of disturbance. As this method, assuming a recurrence formula of a sine wave, or the above mentioned method of introducing an adaptive rule based on the error signal e and correcting the drive quantity of the control target, can be used. Also a method of estimating an unknown frequency based on an error signal ‘e’, generating a disturbance suppression signal on the position level, correcting the error signal, and inputting it into the controller, can also be used.
• In the present embodiment, the interface circuit11-1 receives a disturbance suppression frequency from the outside, and sets it in thefrequency converter30. Therefore the compensator20-1 performs adaptive control from this initial value (angular frequency).
• Since a known disturbance frequency is set as an initial value, position control starts directly with the known disturbance frequency, and even if the frequency changes thereafter, the compensating current Ud of the compensator20-1 follows up from there.
• In this way, this position control system has a means of changing an internal constant (angular frequency in the case ofFIG. 5 andFIG. 6), or the configuration according to the set value of the disturbance frequency to be selectively suppressed, and the disturbance frequency can be referred to or set from the outside via the interface11-1.
Other Embodiment
• In the above embodiments, the position control device was described using an example of a head positioning device of a magnetic disk device, but the present invention can also be applied to other medium storage devices, such as an optical disk device or other devices to control the position of an object. The number of disturbance frequencies can be arbitrary according to necessity, and the number of disturbance models to be used can be arbitrary accordingly. The embodiments were described using a second degree filter, but a first degree filter or a combination of a first degree filter and a second degree filter may be used according to the frequency which need be suppressed.
• The present invention was described using embodiments, but the present invention can be modified in various ways within the scope of the essential character thereof, and these variant forms shall not be excluded from the scope of the present invention.
• Because of disturbance suppression control, open loop gain is calibrated using a target gain according to a disturbance frequency even if a loop characteristics of a feedback controller changes, so the open loop gain can be calibrated without interrupting the disturbance suppression control. Therefore the open loop gain can be accurately calibrated without being affected by disturbance, and accurate position control is possible.

Claims (21)

1. A position control method for controlling a position of an object to a predetermined position by an actuator, comprising:

a step of computing a position error based on a target position of the object and a current position of the object;

a step of computing a control value in which a disturbance frequency component is suppressed, using a predetermined feedback loop based on the position error and computing a drive value of the actuator by multiplying the control value by a loop gain;

a step of fetching a target loop gain according to the disturbance frequency from a table;

a step of adding disturbance of a measurement frequency to the feedback loop and measuring a loop gain of the feedback loop; and

a step of calibrating the loop gain of the control value computing step based on the measured loop gain and the target loop gain.

2. The position control method according toclaim 1, wherein the step of fetching the target loop gain comprises a step of fetching a measurement frequency according to the disturbance frequency, and

wherein the measurement step comprises a step of adding disturbance of the fetched measurement frequency to the feedback loop, and measuring a loop gain of the feedback loop.

3. The position control method according toclaim 1, wherein the step of fetching the measurement frequency comprises a step of fetching a measurement frequency which does not overlap the disturbance frequency.

4. The position control method according toclaim 1, wherein the drive value computing step comprises:

a step of estimating the disturbance frequency according to the position error based on an adaptive control;

a step of computing a control value in which the disturbance frequency component is suppressed, according to the estimated disturbance frequency; and

a step of computing a drive value of the actuator by multiplying the control value by a loop gain.

5. The position control method according toclaim 4, further comprising a step of interrupting the estimation of the disturbance frequency according to the position error during calibration of the loop gain.

6. The position control method according toclaim 4, wherein the drive value computing step comprises:

a step of estimating the disturbance frequency according to the position error based on an adaptive control, and changing a constant of a controller according to the estimated disturbance frequency; and

a step of computing a control value in which the disturbance frequency component is suppressed, using the changed controller according to the position error, and multiplying the result by a loop gain, so as to compute a drive value of the actuator.

7. The position control method according toclaim 4, wherein the drive value computing step comprises:

a step of estimating the disturbance frequency according to the position error based on an adaptive control;

a step of changing a constant of a controller constructed with an observer according to the estimated disturbance frequency; and

a computing step of computing a control value in which the disturbance frequency component is suppressed according to the position error, using the changed observer, and multiplying the result by a loop gain, so as to compute a drive value of the actuator.

8. A position control device for controlling a position of an object to a predetermined position by an actuator, comprising:

a control unit for computing a position error based on a target position of the object and a current position of the object, computing a control value in which a disturbance frequency component is suppressed, using a predetermined feedback loop based on the position error, and multiplying the result by a loop gain, so as to compute a drive value of the actuator; and

a table for storing a target loop gain according to the disturbance frequency,

wherein the control unit fetches the target loop gain according to the disturbance frequency from the table, adds a disturbance of a measurement frequency to the feedback loop, measures a loop gain of the feedback loop, and calibrates the loop gain in the control value computing step based on the measured loop gain and the target loop gain.

9. The position control device according toclaim 8, wherein the control unit fetches a measurement frequency according to the disturbance frequency from the table, adds a disturbance of the fetched measurement frequency to the feedback loop, and measures a loop gain of the feedback loop.

10. The position control device according toclaim 8, wherein the control unit fetches a measurement frequency which does not overlap with the disturbance frequency from the table.

11. The position control device according toclaim 8, wherein the control unit estimates the disturbance frequency according to the position error based on an adaptive control, computes a control value in which the disturbance frequency component is suppressed according to the estimated disturbance frequency, and multiplies the result by a loop gain, so as to compute a drive value of the actuator.

12. The position control device according toclaim 11, wherein the control unit interrupts the estimation of the disturbance frequency according to the position error during calibration of the loop gain.

13. The position control device according toclaim 11, wherein the control unit estimates the disturbance frequency according to the position error based on adaptive control, changes a constant of a controller according to the estimated disturbance frequency, computes a control value in which the disturbance frequency component is suppressed, using the changed controller, according to the position error, and multiplies the result by a loop gain, so as to compute a drive value of the actuator.

14. The position control device according toclaim 11, wherein the control unit estimates the disturbance frequency according to the position error based on adaptive control, changes a constant of a controller constructed with an observer according to the estimated disturbance frequency, computes a control value in which the disturbance frequency component is suppressed according to the position error, using the changed observer, and multiplies the result by a loop gain, so as to compute a drive value of the actuator.

15. A medium storage device, comprising:

a head for at least reading data on a storage medium;

an actuator for positioning the head to a predetermined position on the storage medium;

a control section for computing a position error based on a target position of the head and a current position acquired from the head, computing a control value in which a disturbance frequency component is suppressed, using a predetermined feedback loop based on the position error, and multiplying the result by a loop gain, so as to compute a drive value of the actuator; and

a table for storing a target loop gain according to the disturbance frequency,

wherein the control section fetches the target loop gain according to the disturbance frequency from the table, adds a disturbance of a measurement frequency to the feedback loop, measures loop gain of the feedback loop, and calibrates the loop gain in the control value computing step based on the measured loop gain and the target loop gain.

16. The medium storage device according toclaim 15, wherein the control section fetches a measurement frequency according to the disturbance frequency from the table, adds a disturbance of the fetched measurement frequency to the feedback loop, and measures a loop gain of the feedback loop.

17. The medium storage device according toclaim 15, wherein the control section fetches a measurement frequency which does not overlap with the disturbance frequency from the table.

18. The medium storage device according toclaim 15, wherein the control section estimates the disturbance frequency according to the position error based on adaptive control, computes a control value in which the disturbance frequency component is suppressed, according to the estimated disturbance frequency, and multiplies the control value by a loop gain, so as to compute a drive value of the actuator.

19. The medium storage device according toclaim 18, wherein the control section interrupts the estimation of the disturbance frequency according to the position error during calibration of the loop gain.

20. The medium storage device according toclaim 18, wherein the control section estimates the disturbance frequency according to the position error based on adaptive control, changes a constant of a controller according to the estimated disturbance frequency, computes a control value in which the disturbance frequency component is suppressed, using the changed controller, according to the position error, and multiplies the result by a loop gain, so as to compute a drive value of the actuator.

21. The medium storage device according toclaim 18, wherein the control section estimates the disturbance frequency according to the position error based on adaptive control, changes a constant of a controller constructed with an observer according to the estimated disturbance frequency, computes a control value in which the disturbance frequency component is suppressed, according to the position error, using the changed observer, and multiplies the result by a loop gain, so as to compute a drive value of the actuator.

Images