JP2006266758A - Correction system for encoder output signal waveform - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To eliminate the wave form distortions in the 2 phase sine wave signal with a comparatively simple operational process. <P>SOLUTION: The wave form correction part 30 comprises the adaptive spectral line enhancers 31A and 31B, the higher harmonic components estimation part 32, and the 1st subtractors 33A and 33B. The adaptive spectral line enhancers 31A and 31B are input with the digital signals A1 and B1 of A phase and B phase output from A/D converter 20A and 20B, and eliminate the white noise component contained in the digital signals A1 and B1. The harmonics estimation part 32 estimates the harmonic components contained in the output signals A3 and B3 from the adaptive spectral line enhancer 31A and 31B and form and output the estimated harmonic signals A3 and B3. The 1st subtractors 33A and 33B output the waveform corrected output signals A2 and B2 by subtracting the estimated harmonic signals A3" and B3" from the output signals A3 and B3 of the adaptive spectral line enhancers 31A and 31B. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an encoder output signal waveform correction apparatus that corrects a waveform of a two-phase sinusoidal signal of an encoder that detects position, angle, velocity, angular velocity, and the like.

  Since there is a processing limit on the interval of the grating formed on the encoder scale, in order to measure an interval finer than the scale grating, the spatial period of the phase change of the sinusoidal signal output from the encoder is further subdivided and interpolated. There is a need. For this reason, various interpolation circuits are conventionally used. An interpolation circuit by digital processing is, for example, an A / D converter that samples A and B phase sinusoidal signals output from an encoder with different 90 ° phases and converts them into digital data, and this A / D converter. And a memory storing a look-up table for obtaining the phase angle data PH at each sampling point based on the digital data DA and DB obtained by the above. The lookup table is created based on PH = ATAN (DB / DA) using an arc tangent function (ATAN).

The A and B phase sinusoidal signals output from the encoder do not completely match in amplitude and offset, and the phase difference is not completely 90 °. For example, when the amplitudes of the A and B phase sinusoidal signal voltages are different, the Lissajous waveform becomes an ellipse, and the Lissajous waveform becomes a circle or ellipse waveform deviated from the origin due to the offset value of each signal voltage. When the phase difference is not completely 90 ° but includes an error, the Lissajous waveform is an ellipse whose major axis or minor axis is 45 ° with respect to the coordinate axis. Since the interpolation circuit is made on the assumption that the A and B phase sine wave signals are sine waves, the deviation from the ideal sine wave adversely affects the interpolation accuracy. For this reason, for example, Patent Document 1 proposes an apparatus for correcting the amplitude error, phase error, and offset in the A and B phase sinusoidal signals. This apparatus uses an estimation model of A-phase and B-phase sine wave signals having n coefficients, and samples and samples at least n sets of A-phase and B-phase sine wave signals including errors at the same time. Based on the measurement value, n coefficients of the estimation model are corrected in real time, and the output signal of the estimation model or the information derived from the output signal of the estimation model is output as position information or angle information. .
JP 2004-219333 A, paragraphs 0024-0026, FIG.

  However, the A and B phase sinusoidal signals output from the encoder are usually not perfect sine waves, and are distorted because waveform distortion components such as white noise and odd harmonic components are superimposed. Such waveform distortion causes a decrease in estimation accuracy in the technique of Patent Document 1. In particular, in the above technique, only 4 or 5 points are sampled in consideration of real-time characteristics, and therefore it is difficult to avoid estimation accuracy deterioration due to waveform distortion.

  The present invention has been made in view of such points, and effectively removes waveform distortion components such as white noise components and harmonic components contained in a sinusoidal signal, thereby improving the interpolation accuracy in the interpolation circuit. An object of the present invention is to provide an encoder output signal waveform correction apparatus capable of improving the measurement accuracy and, consequently, the measurement accuracy.

  To achieve the above object, an encoder output signal waveform correction apparatus according to the present invention inputs an encoder output signal in which a waveform distortion component including a white noise component and a harmonic component is superimposed on a sinusoidal fundamental wave signal. An adaptive line spectrum emphasizing unit that suppresses a white noise component contained in the output signal of the encoder, and a harmonic component contained in an output signal from the adaptive line spectrum emphasizing unit is estimated to generate and output an estimated harmonic signal. Harmonic component estimating means and first subtracting means for subtracting the estimated harmonic signal from the output signal of the encoder whose white noise component has been suppressed by the adaptive line spectrum enhancing means to output a waveform-corrected output signal It is characterized by comprising.

  Here, the adaptive line spectrum enhancement means includes, for example, an adaptive filter that processes the output signal of the encoder, and a second subtraction means that subtracts the output signal of the encoder from the output of the adaptive filter and outputs an error signal. And first adaptive algorithm executing means for adjusting the adaptive filter so as to minimize the error signal from the second subtracting means based on the adaptive algorithm.

  Further, the harmonic component estimation means, for example, based on the detected fundamental wave frequency, the fundamental wave frequency detection means for detecting the fundamental wave frequency that is the frequency of the fundamental wave signal included in the output signal of the encoder Fundamental wave signal estimating means for generating and outputting an estimated fundamental wave signal; third subtracting means for subtracting the estimated fundamental wave signal from the output signal of the adaptive line spectrum enhancing means; and third harmonics from the fundamental frequency. The error between the third harmonic frequency calculation means for obtaining the third harmonic frequency which is the frequency of the component and the output signal of the third subtraction means at the same frequency as the calculated third harmonic frequency is minimized. Thus, it is possible to provide an adaptive sine wave detecting means for generating an estimated third-order harmonic signal whose amplitude and phase are adjusted and outputting the estimated third-order harmonic signal as the estimated harmonic signal.

  Further, the adaptive sine wave detection means, for example, a first sine wave signal generating means for generating a first sine wave signal having the same frequency as the third harmonic frequency, and for the first sine wave signal, Second sine wave signal generating means for outputting a second sine wave signal having a phase difference of 90 °, and first and second amplitudes for adjusting the amplitudes of the outputs from the first and second sine wave signal generation means, respectively. Two amplitude adjusting means, an adding means for adding the outputs from the first and second amplitude adjusting means and outputting them as the estimated harmonic signal, and an output signal from the third subtracting means for the adding means. A fourth subtracting means for subtracting the output to output an error signal; and adjusting the first and second amplitude adjusting means to minimize the error signal from the fourth subtracting means based on an adaptive algorithm. Second adaptive al It can be configured with a rhythm execution unit.

  According to the present invention, the white noise is suppressed from the output signal of the encoder in which waveform distortion components such as white noise and harmonic components are superimposed on the sinusoidal fundamental wave signal by the adaptive line spectrum enhancer. The harmonic component is estimated from the suppressed signal by the harmonic component estimating means, and the estimated harmonic component estimated by the harmonic component estimating means is calculated from the output signal of the encoder whose white noise component is suppressed by the adaptive line spectrum enhancer. Since it is configured to subtract, it is possible to obtain an accurate sinusoidal fundamental wave signal in which white noise and harmonic components are suppressed from the output signal of the original encoder, improving the interpolation accuracy in the interpolation circuit, As a result, the measurement accuracy can be improved.

Next, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a block diagram showing a basic configuration of an encoder output signal correction apparatus according to an embodiment of the present invention. The encoder output signal correction apparatus includes A / D converters 20A and 20B that perform A / D conversion on the A phase output signal A0 and the B phase output signal B0 output from the encoder 10 at predetermined sampling frequencies, respectively. Waveform correction unit 30 for removing waveform distortion components including white noise and harmonic components from outputs A1 and B1 of / D converters 20A and 20B, and A-phase and B-phase sinusoidal waveforms output from waveform correction unit 30 The signals A2 and B2 are input, and the offset error, amplitude error and phase error included in the sine wave signals A2 and B2 are corrected, and final corrected A phase and B phase sine wave signals A4 and B4 are output. And an offset / amplitude / phase correction unit 40.

  The detection principle of the encoder 10 does not matter, but is, for example, a photoelectric type, a magnetic type, or a capacitance type. The two-phase sinusoidal signals A and B having a phase difference of 90 ° output from the encoder 10 usually include waveform distortion, amplitude error, phase error, offset, and the like.

  The two-phase sinusoidal signals A0 and B0 are sampled at predetermined frequencies by the A / D converters 20A and 20B, converted into digital signals A1 and B1, and input to the waveform correction unit 30. The digital signals A1 and B1 input to the waveform correction unit 30 include waveform distortion components such as white noise and third harmonic components, but the digital signals A1 and B1 are corrected by the waveform correction unit 30. The offset / amplitude / phase correction unit 40 corrects the offset, amplitude, and phase.

Here, the waveform correction unit 30 is a part that characterizes the present invention, and is configured, for example, as shown in FIG.
That is, the waveform correction unit 30 includes adaptive line enhancers (ALE) 31A and 31B, a harmonic component estimation unit 32, and first subtractors 33A and 33B. The adaptive line spectrum enhancers 31A and 31B receive the A-phase and B-phase digital signals A1 and B1 output from the A / D converters 20A and 20B, respectively, and the white noise components included in the digital signals A1 and B1. Repress. The harmonic component estimation unit 32 estimates the harmonic components included in the output signals A3 and B3 from the adaptive line spectrum enhancers 31A and 31B, and generates and outputs estimated harmonic signals A3 ″ and B3 ″. Then, the first subtracters 33A and 33B subtract the estimated harmonic signals A3 ″ and B3 ″ from the output signals A3 and B3 of the adaptive line spectrum enhancers 31A and 31B, and use the output signals A2 and B2 whose waveforms are corrected. Output.

Next, the operation of the waveform correcting unit 30 configured as described above and the detailed configuration of each unit will be described.
In the digital signals A1 and B1 output from the A / D converters 20A and 20B, a waveform distortion component is superimposed on a sinusoidal fundamental wave signal. The adaptive line spectrum enhancers 31A and 31B remove white noise components from the waveform distortion components from the digital signals A1 and B1.
For example, as shown in FIG. 3, the adaptive line spectrum enhancers 31A and 31B include a delay unit 311 that delays the digital signals A1 and B1, and an adaptive filter 312 that includes an FIR filter that performs a filtering process on the output of the delay unit 311. The second subtracter 313 that subtracts the input digital signals A1 and B1 from the output of the adaptive filter 312 and outputs an error signal, and the second subtractor 313 outputs the error signal based on a predetermined adaptive algorithm. And a first adaptive algorithm execution unit 314 that adjusts the coefficient of the adaptive filter 312 so as to minimize the error signal. The delay unit 311 is provided in order to eliminate the correlation of white noise, thereby enabling noise removal even if the frequency of the sine wave included in the input signal is unknown. As an adaptive algorithm, for example, a least mean square algorithm (Least Mean Square: LMS) by the steepest descent method can be used.

  Actually, the adaptive line spectrum enhancers 31A and 31B are configured by using the delay unit 311 as a 5-sample delay, the adaptive filter 312 as a 100th-order FIR adaptive filter, and the LMS as an adaptive algorithm, and inputting a sinusoidal signal including white noise. The experimental results are shown in FIG. As shown in FIG. 4, it can be seen that the white noise component is significantly suppressed by the adaptive line spectrum enhancement processing described above.

Next, the harmonic component estimation unit 32 will be described. The harmonic component estimation unit 32 is configured as shown in FIG.
The output signals A3 and B3 of the adaptive line spectrum enhancers 31A and 31B are signals in which a harmonic component is superimposed on a fundamental wave component. Generally, the relationship (fundamental wave signal >> harmonic component) is established. , The ω ₁ detector 321 detects the fundamental frequency ω ₁ included in the output signals A3 and B3. Since the fundamental frequency ω ₁ changes depending on the moving speed of the scale of the encoder 10, it is necessary to detect it in real time with high accuracy. The ω ₁ detector 321 that satisfies such a requirement can be configured as shown in FIG. 5, for example. That is, output signals A3 and B3 including fundamental wave signals that can be regarded as substantially the same as sin (ω ₁ k) and cos (ω ₁ k) by adjusting the amplitude of the AB phase in advance are input to the arctangent conversion unit (ATAN) 3211. And converted into an angle ω ₁ k. The difference between the angle ω ₁ k and the angle ω ₁ (k−1) delayed by one sample in the unit delay circuit 3212 is obtained by the subtractor 3213, so that the angular frequency (fundamental wave frequency) ω ₁ of the fundamental wave signal is obtained. Is required.

Further, in the above example, it is assumed that the two-phase signals A3 and B3 have a phase difference of 90 °. However, there may be a case where the phase difference between the two signals is not exactly 90 °. In addition, the phase of one output signal A3 is shifted by 90 ° by the Hilbert transformer 3214, and the processing time in the Hilbert transformer 3214 is compensated by the delay circuit 3215, thereby obtaining a signal having an accurate 90 ° phase difference. Therefore, the fundamental wave frequency ω ₁ may be obtained by inputting it to the Atan unit 3211.

The fundamental wave frequency ω ₁ obtained in this way is input to the A-phase waveform estimation unit 322A and the B-phase waveform estimation unit 322B, where it is used for estimation of the fundamental wave signal. The A-phase waveform estimation unit 322A and the B-phase waveform estimation unit 322B can be configured as shown in FIG. First, first and second sine wave signals having the same frequency as the fundamental frequency ω ₁ and 90 ° phases are generated from the first sine wave signal generator 3221 and the second sine wave signal generator 3222. These sinusoidal signals are amplitude-adjusted by the first and second amplitude adjusting units 3223 and 3224 and added by the adder 3225. The output signals A3 and B3 from the adaptive line spectrum enhancers 31A and 31B are subtracted from the output of the adder 3225 by the fourth subtractor 3226, and the difference between them is error signal ε _A2 and ε _B2 , and the adaptive algorithm execution unit 3227. Is given to. The adaptive algorithm execution unit 3227 adjusts the amplitude adjustment units 3223 and 3224 so as to minimize the output of the subtractor 3226 according to a predetermined algorithm such as LMS. As a result, the output of the adder 3225 is output as the estimated fundamental wave signals A2 'and B2'.

In FIG. 2, the output signals A3 and B3 from the adaptive line spectrum enhancers 31A and 31B contain the fundamental wave signal and the harmonic component, but the second subtracters 323A and 323B use the estimated fundamental wave signal. By subtracting A2 ′ and B2 ′, harmonic components A3 ′ and B3 ′ are extracted. On the other hand, the fundamental wave frequency ω ₁ detected by the ω ₁ detector 321 is also given to the third harmonic frequency calculator 324, where a third harmonic frequency ω ₃ obtained by multiplying ω ₁ by _three is obtained. This ω ₃ is supplied to the adaptive sine wave detectors 325A and 325B together with the harmonic components A3 ′ and B3 ′.

Specifically, the adaptive sine wave detectors 325A and 325B are configured in substantially the same manner as the A-phase and B-phase waveform estimation units 322A and 322B, as shown in FIG. That is, the adaptive sine wave detectors 325A and 325B generate and output first and second sine wave signals having the same frequency as the _third harmonic frequency ω ₃ and a phase difference of 90 ° from each other. The sine wave signal generators 3251 and 3252, the first and second amplitude adjustment units 3253 and 3254 that adjust the amplitudes of the outputs from the sine wave signal generators 3251 and 3252, and the amplitude adjustment units 3253 and 3254, respectively. An adder 3225 that adds the outputs and outputs them as estimated third harmonic components A3 "and B3", and an estimated third harmonic A3 "from the harmonic components A3 'and B3' from the third subtractors 323A and 323B. , the error signal epsilon _A3 subtracts B3 ", epsilon and fourth subtractor 3256 which outputs a _B3 error signal epsilon _A3 from the fourth subtracter 3256 based on the adaptive algorithm, epsilon _{B 3} and a second adaptive algorithm execution unit 3257 that adjusts the first and second amplitude adjustment units 3253 and 3254 so that ₃ is minimized.

  The estimated third harmonic components A3 ″ and B3 ″ obtained in this way are subtracted from the outputs of the adaptive line spectrum enhancers 31A and 31B including the harmonic components in the first subtracters 33A and 33B of FIG. By doing so, it is possible to obtain the output signals A2 and B2 that almost reproduce the sinusoidal fundamental wave signal from which the white noise and the third harmonic are removed.

Specifically, each of these units can be configured by hardware such as a DSP (Digital Signal Processor) and CPU, and software executed by the DSP and CPU. FIG. The flowchart in the case of implement | achieving is shown.
That is, although waveform distortion components are superimposed on the fundamental wave signal in the output signals A0 and B0 of the encoder 10, first, white noise is removed from the AB phase encoder output signal by the adaptive line spectrum enhancer (S1). . Next, the frequency ω ₁ of the AB phase fundamental wave signal is detected, and the fundamental wave signal is estimated (S2). Subsequently, the estimated fundamental wave signal is removed from the signal from which the white noise has been removed, and a harmonic component including the third harmonic component is extracted (S3). Next, the third harmonic component is estimated based on the harmonic component extracted in step S3 and the third harmonic frequency that is three times the frequency of the fundamental wave signal detected in step S2 (S4). Then, the AB phase signal is updated by subtracting the estimated third harmonic component from the most recently updated AB phase signal (S5). Thereby, output signals A2 and B2 having no waveform distortion can be obtained.
Note that ω ₁ in step S2 is estimated using the output signal of step S1 as an input until the output signal of step S5 is obtained. When the output signal of step S5 is obtained, this correction signal is used. if Te as an estimate of the omega _1, enables more highly accurate omega ₁ estimation.

FIG. 10 is a graph showing the results of simulating the power spectrum before (a) and after (b) removal of the third harmonic component. As shown in the figure, it was confirmed that the above-described configuration has a high third-order harmonic removal effect.
The adaptive signal processing in the above embodiment can be realized with the number of product-sum operations of about 500 to 700, and high-speed processing is possible. In the above embodiment, the least squares average algorithm based on the steepest descent method is used as the adaptive algorithm. However, in order to realize high convergence and efficient calculation, a recursive least square method (Recursive Least Square, More advanced processing can be realized by using RLS) or Least Squares Lattice (LSL) algorithm.

In addition, in order to handle the AB phase signal independently, the Hilbert transform circuit applicable from the DC component is necessary due to the property of the encoder that the fundamental frequency ω ₁ depends on the relative speed of the moving body with respect to the fixed body. This can be realized by frequency modulation. That is, the Hilbert transformer can be easily realized by converting the band signal from DC to fx into the band of fdc to ffx by frequency modulation. When it is not necessary to handle a signal from a DC component, the realization is further facilitated.
As the adaptive filter, it is possible to extract only the fundamental wave signal component using a narrow band pass filter.

It is a block diagram which shows the basic composition of the encoder output signal correction apparatus which concerns on one Embodiment of this invention. It is a circuit diagram which shows the specific structure of the waveform correction | amendment part in FIG. It is a circuit diagram which shows the specific structure of the adaptive line spectrum enhancer in FIG. It is a graph which shows the effect of an adaptive line spectrum enhancer. FIG. 3 is a circuit diagram illustrating a specific configuration of an ω ₁ detection unit in FIG. 2. FIG. 4 is a circuit diagram illustrating another specific configuration of the ω ₁ detection unit in FIG. 2. It is a circuit diagram which shows the specific structure of the AB phase waveform estimation part of FIG. It is a circuit diagram which shows the specific structure of the adaptive sine wave detector of FIG. It is a flowchart which shows the process of the same waveform correction part. It is a graph which shows the effect of the waveform correction part.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Encoder, 20A, 20B ... A / D converter, 30 ... Waveform correction part, 31A, 31B ... Adaptive line spectrum enhancer, 32 ... Harmonic component estimation part, 33A, 33B ... 1st subtractor, 40 ... Offset / amplitude / phase correction unit.

Claims (4)

Adaptive line spectrum enhancement means for inputting an output signal of an encoder in which a waveform distortion component including a white noise component and a harmonic component is superimposed on a sinusoidal fundamental wave signal and suppressing the white noise component contained in the output signal of the encoder When,
Harmonic component estimation means for estimating the harmonic component contained in the output signal from the adaptive line spectrum enhancement means and generating and outputting an estimated harmonic signal;
First subtracting means for subtracting the estimated harmonic signal from the output signal of the encoder whose white noise component is suppressed by the adaptive line spectrum enhancing means and outputting an output signal whose waveform is corrected. Encoder output signal waveform correction device. The adaptive line spectrum enhancement means includes:
An adaptive filter for processing the output signal of the encoder;
Second subtracting means for subtracting the output signal of the encoder from the output of the adaptive filter and outputting an error signal;
The encoder according to claim 1, further comprising first adaptive algorithm executing means for adjusting the adaptive filter so as to minimize an error signal from the second subtracting means based on an adaptive algorithm. Output signal waveform correction device. The harmonic component estimation means includes
Fundamental wave frequency detection means for detecting a fundamental wave frequency that is a frequency of a fundamental wave signal included in the output signal of the encoder;
Fundamental wave signal estimating means for generating and outputting an estimated fundamental wave signal based on the detected fundamental wave frequency;
Third subtracting means for subtracting the estimated fundamental wave signal from the output signal of the adaptive line spectrum enhancing means;
A third harmonic frequency calculating means for obtaining a third harmonic frequency which is a frequency of a third harmonic component from the fundamental frequency;
An estimated third harmonic signal having the same frequency as the calculated third harmonic frequency and an amplitude and phase adjusted so as to minimize an error from the output signal of the third subtracting unit is generated, and is generated. The encoder output signal waveform correction apparatus according to claim 1, further comprising: an adaptive sine wave detection unit that outputs an estimated harmonic signal. The adaptive sine wave detecting means includes
First sine wave signal generating means for generating a first sine wave signal having the same frequency as the third harmonic frequency;
Second sine wave signal generating means for outputting a second sine wave signal having a phase difference of 90 ° with respect to the first sine wave signal;
First and second amplitude adjusting means for adjusting the amplitudes of the outputs from the first and second sinusoidal signal generating means, respectively;
Adding means for adding the outputs from the first and second amplitude adjusting means to output as the estimated harmonic signal;
A fourth subtracting means for subtracting the output of the adding means from the output signal of the third subtracting means to output an error signal;
And second adaptive algorithm executing means for adjusting the first and second amplitude adjusting means so as to minimize the error signal from the fourth subtracting means based on the adaptive algorithm. 4. The encoder output signal waveform correction apparatus according to claim 3, wherein

Images