KR100493146B1 - Multirate Estimator and Method - Google Patents

Abstract

According to the present invention, in order to prevent an error from occurring by estimating on the basis of a simplified model in a conventional multirate estimator, a plant modeled by a linear transformation function and a signal output from the plant are sampled and inputted and the analog inputted. A first converter for converting a signal into a digital signal, an interrupt generating means for generating an interrupt as an input of the output of the first converter, a predetermined time is counted at the time the interrupt is generated, and the interrupt is generated by the counted output Counter means for driving the means, estimation means for estimating the position information of the plant at the interruption time generated by the counter means, and a control signal for controlling the position of the plant based on the estimated position information as an analog signal. A second to convert and apply to the plant The present invention relates to a multi-rate estimating apparatus and a method for compensating in consideration of an error between a measurement value of a previous sample and an estimated value modeled during intermediate sampling.

Description

Multirate estimator and method

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multirate estimating apparatus in a digital control system. In particular, the present invention relates to an estimated position and speed of an expected track and an error of a position and speed of an actual track during intermediate sampling in a disk drive. A multirate estimating apparatus capable of accurately tracking a track, and a method thereof.

In general, a disk drive device is composed of a disk on which information is stored, a head for reading and writing information on the disk, and a driver for moving the head to a desired track. The head reads the position information from the track on the disk rotating at an isometric speed. The moving speed of the head is estimated directly from the state estimator because it is not directly measured.

This position and velocity information is used so that the head is accurately on track by a controller that controls the driver including the head. In the sector servo method that is widely used today, the sampling period is determined by the rotational speed of the disk and the number of sectors. The positional information according to the movement of the head is valid only for sector signals used in this sector servo method.

In addition, since the track density increases as the capacity of the hard disk increases, a high servo bandwidth is required for more precise control. That is, more servo sectors are needed for more accurate state estimation. Moreover, the discrete controller requires a faster sampling period, so more sectors are needed as the required servo bandwidth increases. However, it is undesirable to waste a lot of disk storage space for the servo sector. As a result, a multirate estimator is required, thereby reducing the servo sector, thereby saving disk storage space.

However, if the model used in the state estimator is not accurate, this results in poor state prediction performance, resulting in performance degradation of the multirate estimator.

Prior arts, "Multirate State-Space Digital Controller for Sector Servo System, IBM, 1990, by WW Chiang" and "Multirate Estimation and Control under Time-Varying Data Sampling with Application to Information Storages, UC Berkeley, 1995, by AM Phillip , M. Tomizuka ", and the like, have established a simple plant model and estimate the positional displacement according to the plant input.

Here, the plant model is a device modeled by a linear transfer function. In the present invention, a rotating disk, a plurality of magnetic heads for reading or writing information on the disk, and positions of the plurality of magnetic heads are described. Apply a hard disk drive composed of actuators to move.

However, in the above-described prior art, a non-modeled part of a disc, that is, a disc except for a disc, a magnetic head, an actuator, and a disc vibrates so that the head does not accurately track a track due to vibration of the disc. Since the factor was not taken into account, a difference was generated between the actual plant and the theoretical model, which caused an estimation error in which the position of the head moving the track consisting of sectors on the disk was not accurately estimated.

Therefore, the technical problem to be achieved by the present invention, in order to prevent the error is generated by estimating based on the simplified model in the prior art, the error between the measurement of the previous sample and the modeled estimate in the intersample (intersample) It is to provide a multi-rate estimating apparatus and method for compensating in consideration.

Multi-rate estimator for solving the above problems is a plant modeled by a linear transform function; A first converter sampling and inputting a signal output from the plant and converting the input analog signal into a digital signal; Interrupt generating means for generating an interrupt with an output of the first converter as input; Counter means for counting a predetermined time at the time when the interrupt is generated and driving the interrupt generating means by the counted output; Estimation means for estimating the positional information of the plant at the interrupt time generated by the counter means; And a second converter converting a control signal for controlling the position of the plant into an analog signal based on the estimated position information and applying the same to the plant.

Another multirate estimating method for solving the above problem is a method for estimating a position in a digital control system using a system modeled by a linear transformation function, the method comprising: (a) obtaining the intermediate sample time from a signal sensed by the position; Calculating; (b) intersampling to estimate the location by the calculated time; (c) calculating a difference between an update value of the multirate estimator and a value predicted by the modeling according to the intermediate sample in step (b); (d) calculating a difference between the predicted value and the update value of the multirate estimator obtained by the position measurement value at the k-1 th reference sample for the position estimation at the k th intermediate sample; (e) comparing the difference value calculated in step (d) with a predetermined value for tuning the modeled system and setting the gain to " 0 " if the predetermined value is smaller than the calculated difference value; If less than the difference value calculated in step (d), obtaining a predetermined gain; And (f) obtaining a position estimate value during the intermediate sample according to the predetermined gain obtained in step (e), and controlling the system according to the estimate value.

1 is a block diagram showing a multi-rate estimating apparatus according to the present invention.

The multirate estimating apparatus according to the present invention includes a plant 10, a first converter 11, an interrupt generator 18, a multirate estimator 14, a timer 16, and a second converter 12. It is composed.

Plant 10 is a system that is modeled as a linear transformation function, preferably a disk drive.

The first converter 11 is an analog / digital converter in which a signal output from the plant 10 is sampled and input, and converts the input analog signal into a digital signal.

The interrupt generator 18 generates an interrupt by the output of the timer 16 and the first converter 11, and accordingly drives the multirate estimator 14.

The multirate estimator 14 takes an interrupt output from the interrupt generator 18 as an input, estimates the time between the input interrupts, estimates the position of the head in the plant 10, and estimates the plant ( And a controller for controlling the position of the head according to the estimation.

The timer 16 counts a predetermined time on the basis of the signal generated by the interrupt in the multirate estimator 14, and drives the interrupt generator 18 to generate an interrupt based on the counted output.

The second converter 12 is a digital / analog converter for converting the digital signal output from the multirate estimator 14 into an analog signal and applying it to the plant 10.

2 is a flowchart illustrating a multirate estimating method according to the present invention.

The multirate estimation method according to the present invention configured as described above will be described with reference to FIG.

The present invention uses the timer 16 to make an intermediate sample, and starts counting the timer 16 using a signal resulting from the start of a sector of a disk (not shown) provided in the plant 10 as a reference signal.

At this time, the timer 16 is set to be 1 / n times the reference sampling time. Accordingly, the interrupt generator 18 generates an interrupt at the 1 / n time after the reference sampling.

In order to accurately estimate the head position of the linearly transformed plant 10 in a digital control system using a system modeled by the linear transformation function, the beginning of the sector is first converted into the first converter 11, the interrupt generator 18, and the multirate. When detected through the estimator 14, the counting of the timer 16 is started and the position signal of the head is sampled. The start of the sector is a gray code and a position error signal (PES). By detecting this, the beginning of the sector is detected.

As described above, the position change of the head may be detected at the beginning of the sector, but once the beginning of the sector is detected, the detection of the information of the sector accordingly may be detected. Possible at the beginning of the sector following the detected sector (step 20).

Accordingly, in order to detect the position of the head in the middle of the sector, the time due to the operation of the timer 16 is calculated based on the beginning of the sector (step 21).

By intersample by the above calculation of time, the multirate estimator estimates the position of the head.

In order to estimate the position of the head in the plant 10 during the intermediate sample, the following algorithm is used in the multirate estimator 14.

First, considering the discrete time state equation is as follows.

[Equation 1]

[Equation 2]

Where Φ is the system matrix, T is the input matrix, C is the output matrix, x (k) is the k-th state value, x (k + 1) is the k + 1th state value, and u (k) is the plant input. And y (k) is the plant output.

In addition, in order to estimate the state value in the multirate estimator 14, the following equation is generally used (step 22).

[Equation 3]

[Equation 4]

here, Is the status updated value, Is the state prediction value and L _c is the multirate estimator gain.

Assuming the output measurement is valid every T _s sample time, k = T _s n (n = 0, 1, 2, ...), but for intermediate samples between sectors where no signal is detected, i.e. , k = T _s (n + l) {where 0 <l <1} the output measured value is invalid.

Therefore, since the output measurement is not valid because there is no reference signal detected at the start of the sector by the head at the intermediate sampling in the multirate estimator 14, the updated value is Is not valid and is the expected value Only valid.

That is, in order to increase the predictability by y (k) which is the plant output value, And previous predicted values, Calculate using to improve the predictability.

First, an error value estimated by the intermediate sample output from the multirate estimator 14 ( To calculate (e (k)), the difference between the updated value of the multirate estimator and the value predicted by the modeling is calculated (Equation 5) or the difference between the actual measured value and the predicted value is calculated (Equation 6). Use the following formula (step 23).

[Equation 5]

( ; error, ; Update value by multirate estimator)

here, Is an error, Is the value updated by the state estimator, Is a value predicted by modeling.

[Equation 6]

(e (k); error)

In addition, the following equation is used to estimate the position of the head during intermediate sampling.

[Equation 7]

( ; Plant output, C; Output matrix, ; Value predicted by modeling)

Substituting Equation 7 into Equation 3 yields the following equation.

[Equation 8]

[Equation 9]

Equations 8 and 9 are expressions described by reference as a process for obtaining the following equation.

[Equation 10]

(L _c ; multirate estimator gain)

In Equation 10 Is not valid (step 24).

In addition, the equation used in intermediate sampling is defined as follows.

[Equation 11]

The following equation is obtained from the above equation (11).

[Equation 12]

[Equation 13]

Equations 11 and 12 are expressions described by reference as a process for obtaining Equation 13.

In Equation 13 Is not valid. That is, equations (10) and (13) are equations for calculating the gain of the multirate estimator using the output measurement signal of the k th sample (step 24).

However, Equation 10 is not valid because it does not know y (k), which is a value measured during intermediate sampling in Equation 13. Therefore, the following estimation method according to the present invention is used.

There are two cases of the multirate estimating method used in the present invention. The first method is as follows.

That is, according to equations (5) and (7), the sampling time is fast enough. end Since it is a value that is very close to, it is assumed to be approximately the same.

[Equation 14]

Shown in Equation 7 Is a predetermined small value (δ) that is the tuning parameter of the plant system, as in the equation ) And the size are determined (step 25).

[Equation 15]

end Especially very small in equation (14) obtained under the assumption In the value the sign The probability of different signs is high. If the signs are different Approximation May cause a larger error than the actual value. Therefore, Equation 14 is The smaller the value is, the more unsuitable. In this case, β (k) is set to "0", and if it is not small, the predetermined gain of β (k) as defined by Equation (14) is obtained (step 26).

The second method of the multirate estimating method used in the present invention is as follows.

The magnitude of e (k-1) shown in Equation (11) is determined by the following equation with a predetermined small value δ _{e which} is a tuning parameter of the plant system (step 25).

[Equation 16]

That is, suppose that e (k-1) is approximately equal to e (k) at a sufficiently fast sampling time according to equations (6) and (11). It becomes

Since e (k-1) is very close to e (k), the probability that the sign is different from e (k) is particularly high under the assumption that it is approximately equal. If the sign is different, the value e (k-1) approximating e (k) may cause a larger error than the actual value. Therefore, smaller e (k-1) is not suitable, in this case Is set to "0" and has a predetermined gain of "1" if not small (step 26).

By estimating the position of the head by the intermediate sample method as described above, the intermediate sample time is estimated by calculating the time counted by the timer 16 in order to precisely control the disk drive. By controlling the gain 10, the plant 10 is more precisely controlled (step 27).

3 is a view showing a position estimation result according to the present invention. 4 is a view showing a speed estimation result according to the present invention. 5A to 5C are diagrams illustrating measurement results of a speed, current, and position error signal PES obtained by moving a head 1900 track according to the present invention.

As shown in Figs. 3 and 4, in the tracking of a reference signal by the plant according to the present invention, especially when the sampling time is short, the estimation error of the previous sample is similar to the estimation error of the next intermediate sample. do. Therefore, since the estimation error in the intermediate sample can be reduced, it can be seen that the performance is improved by about 45% in the position estimation and the speed estimation of the disc head, compared with the conventional multirate estimator.

5A to 5B show measurement results of speed, current, and position error signals obtained by controlling the hard disk drive so that the head of the hard disk moves 1900 tracks using the multirate estimating method according to the present invention. .

According to the present invention, an intermediate sample is obtained by time calculation of a timer, and accordingly, the multirate estimating apparatus estimates the position of the intermediate sample, thereby obtaining more accurate head position and velocity estimation results.

1 is a block diagram showing a multi-rate estimating apparatus according to the present invention.

2 is a flowchart illustrating a multirate estimating method according to the present invention.

3 is a view showing a position estimation result according to the present invention.

4 is a view showing a speed estimation result according to the present invention.

5A to 5C are diagrams illustrating measurement results of a speed, current, and position error signal PES obtained by moving a head 1900 track according to the present invention.

Claims (15)

A plant modeled as a linear transformation function; A first converter sampling and inputting a signal output from the plant and converting the input analog signal into a digital signal; Interrupt generating means for generating an interrupt with an output of the first converter as input; Counter means for counting a predetermined time at the time when the interrupt is generated and driving the interrupt generating means by the counted output; Estimation means for estimating the positional information of the plant at the interrupt time generated by the counter means; And And a second converter converting a control signal for controlling the position of the plant based on the estimated position information into an analog signal and applying the same to the plant. The method of claim 1 wherein the counter means Multirate estimating apparatus, characterized in that the timer. The method of claim 1, wherein the estimating means And a multirate estimator. The method of claim 1, wherein the interrupt generated by the interrupt generating means is And the output of the first converter and the counter means. A method for estimating a position in a digital control system using a system modeled with a linear transformation function, (a) calculating the intermediate sample time from the signal sensed by the position; (b) intersampling to estimate the location by the calculated time; (c) calculating a difference between an update value of the multirate estimator and a value predicted by the modeling according to the intermediate sample in step (b); (d) calculating a difference between the predicted value and the update value of the multirate estimator obtained by the position measurement value at the k-1 th reference sample for the position estimation at the k th intermediate sample; (e) comparing the difference value calculated in step (d) with a predetermined value for tuning the modeled system and setting the gain to " 0 " if the predetermined value is smaller than the calculated difference value; If less than the difference value calculated in step (d), obtaining a predetermined gain; And (f) obtaining a position estimate value at the time of the intermediate sample according to the predetermined gain obtained in the step (e), and controlling the system according to the estimated value. 6. The method of claim 5, wherein the intermediate sampling in step (b) is defined by the following equation. ( ; Plant output, C; Output matrix, ; Value predicted by modeling) The method of claim 5, wherein the calculation of the difference in step (c) is defined by the following equation. ( Is an error, Is the value updated by the state estimator, Is the value predicted by modeling) The method of claim 5, wherein the calculation of the difference in step (c) is defined by the following equation. 6. The method of claim 5, wherein the calculation of the difference in step (d) is defined by the following equation. 6. The method of claim 5, wherein the calculation of the difference in step (d) is defined by the following equation. The method of claim 5, wherein the difference value calculated in step (d) When the predetermined gain in the step (e) is defined by the following equation. The method of claim 5, wherein the difference value calculated in step (d) In the case of, the predetermined gain in the step (e) is " 1 &quot;. 6. The method of claim 5, wherein the calculation of time in step (a) is made by counting a timer. 6. The method of claim 5, wherein the comparison in step (e) is defined by the following equation. ( ; Tuning parameters of the plant system) 6. The method of claim 5, wherein the comparison in step (e) is defined by the following equation. ( ; Tuning parameters of the plant system)