JP2989379B2 - Active control device using adaptive IIR digital filter - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an active control device using an adaptive IIR digital filter, and more particularly, to a signal control circuit in an active noise control device, an active vibration control device, an echo canceller, an adaptive equalizer, and the like. The present invention relates to an active control device using an adaptive IIR digital filter used for a signal control circuit in general control.

[0002]

2. Description of the Related Art Generally, the transfer function H (z) of an adaptive IIR digital filter is represented by the following equation (1).

[0003]

(Equation 01)

FIG. 9 shows a direct type configuration of the adaptive IIR digital filter. Adaptive II
The direct configuration of the R digital filter is a unit delay element 2
8, a multiplier 29, and an adder 30. The input signal u is added by the adder 30 via the unit delay element 28 and the multiplier 29 and output as an output signal y. a (i) and b (j) indicate filter coefficients. Here, the transfer function H _N (z) is formed by the filter coefficient a (i).

[0005]

(Equation 02)

[0006] The FIR digital filter part for realizing the above is replaced with the non-cyclic part 18 of the adaptive IIR digital filter 16.
, And is composed of a filter coefficient b (j), and has a transfer function H _R (z)

[0007]

[Equation 03]

[0008] A part for realizing the above will be referred to as a circulating unit 17.

Assuming that the input signal of the adaptive IIR digital filter 16 is u (n) and the output signal is y (n), the input / output relationship is expressed by Equation 4.

[0010]

[Equation 04]

Here, N is the number of taps of the non-repeating section 18, M
Indicates the number of taps of the traveling unit 17. There are various algorithms for updating the filter coefficients. One of them is a method for updating the filter coefficients according to the following equations (5) and (6).

[0012]

A (i, n + 1) = a (i, n) + αu (n-i) e (n)

[0013]

B (j, n + 1) = b (j, n) + βy (nj) e (n) where e (n) is the output of the desired response d (n) and the filter output y (n) The error d (n) -y (n) is shown, and α and β are step size parameters, and take small positive values. Further, in order to improve the stability and convergence of updating the filter coefficient, there is a method using f (n) shown in the following equation 7 instead of e (n) in the equations 5 and 6.

[0014]

[07]

Here, c (l) indicates the weight of the moving average and is a predetermined constant. L indicates the number of data to be averaged.

The adaptive IIR digital filter 36 shown in FIG. 10 is obtained by adding a filter coefficient update control unit to that shown in FIG. In the acyclic unit 18, the input signal u (n) and the output error signal e (n) are
It is connected so that a control signal is inputted to each multiplier 29 from the coefficient control unit 20 when inputted to 0. On the other hand, the output signal y (n) and the output error signal e (n) are also input to the coefficient control
Are connected so that a control signal is input to each of the multipliers 29. In the case where Equation 7 is used and in Equation 5 u (n−
i), when the following equations 8 and 9 are used instead of y (n-j) in equation 6, the configuration basically becomes as shown in FIG. Can be processed.

[0017]

[Equation 08]

[0018]

[09]

The above-mentioned adaptive type is applied to the signal processing unit of the active noise reduction apparatus for suppressing noise by emitting sound waves of the same amplitude, which are 180 ° out of phase with the noise, from the sound-generating means for noise reduction and causing sound wave interference. IIR digital filter 1
FIG. 11 shows a configuration when 6, 36 are applied. FIG.
In the figure, reference numeral 12 denotes a container having an opening at one end, and noise from a noise source 11 arranged in the container 12 is detected by a noise detection microphone 13. A detection signal detected by the noise microphone 13 is digitally converted by an A / D converter 25 via a preamplifier 22 and input to the adaptive IIR digital filter 16. Noise detection microphone 13
Is calculated by the adaptive IIR digital filter 16 based on the equation (4), and the result y (n) of the input signal u (n) is used as a mute signal by the D / A converter 26 and the power amplifier 23. The sound is output from the muffling speaker 15 through the speaker.
The filter coefficient a (i) of the non-recursive part 18 of the adaptive IIR digital filter 16 and the filter coefficient b of the recursive part 17
(j) indicates the respective coefficient control units 2 via the preamplifier 24 and the A / D converter 27 in accordance with the silencing error signal -e (n) detected by the silencing error detection microphone 14 in the container 12.
0, 21 based on the signals input to the coefficient control unit 2
It is sequentially updated at 0 and 21. Noise microphone 13
Is input to the coefficient control unit 20 via the digital filter 32 without passing through the adaptive IIR digital filter 16, and the mute signal is output without passing through the mute error detection microphone 14. Direct Adaptive II
R digital filter 16 to digital filter 31
Is input to the coefficient control unit 21 via the.

Here, if equations 5 and 6 are used, the update equation is expressed by the following equations 10 and 11.

[0021]

A (i, n + 1) = a (i, n) + αu ₁ (ni) e (n)

[0022]

B (j, n + 1) = b (j, n) + βy ₁ (nj) e (n) u ₁ (n) is a digital filter of the input signal u (n) of the adaptive IIR digital filter 16. 32, y ₁ (n) is the output signal y (n) of the adaptive IIR digital filter 16 corrected by the digital filter 31. The characteristics of the digital filters 31 and 32 are obtained from the output of the adaptive IIR digital filter 16 via the silence error detecting microphone 14 and the respective coefficient control units 20 and 2.
This is the transfer characteristic up to 1.

As still another conventional example, there is the one shown in FIG. When the desired response d is directly obtained, the input signal u is calculated in the non-cyclic section 18 and the cyclic section 17 of the adaptive IIR digital filter 16, and the calculation result y is output as a mute signal. Is compared with Then, a signal corresponding to the noise reduction error signal e is input to the respective coefficient control units 20 and 21, and the output signal y is transmitted to the coefficient control unit 2 via the circulating unit 17.
0, the input signal u is input to the coefficient control unit 21 via the cyclic unit 17, and the filter coefficient of the non-cyclic unit is updated. Thus, instead of u (ni) and y (nj) in Equations 5 and 6, u ₀ (i, n) and y ₀ (j, n) in Equations 8 and 9 are used.

[0024]

As shown in FIG. 10, in updating the filter coefficients of the adaptive IIR digital filter 36 having the circulating unit 17, the problems of the stability and the convergence of the filter always accompany. Even in the various filter coefficient updating methods described above, stability and convergence are not guaranteed. In the recursive unit 17 of the adaptive IIR digital filter 36, there is a problem that the output diverges in the process of updating the filter coefficient depending on the case. On the other hand, when the level of the output of the cyclic section or the output of the adaptive IIR digital filter 36 is minimized, the cyclic section 1
7, the divergence of the signal in the circulating unit 17 can be suppressed, but the input signal of the adaptive IIR digital filter 36 is periodic. When a typical signal is required as an output of the adaptive IIR digital filter 36, there is a problem that the required periodic signal is canceled by the circulating unit.

Further, when the equations (10) and (11) are used, the amount of calculation increases, and high-speed operation is required regardless of the amount of calculation. In application to active silence control,
As shown in FIG. 11, there is a problem that the computations of the digital filters 31 and 32 are also required as preprocessing of the coefficient control units 20 and 21, and an increase in the amount of computation is inevitable.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and improves the stability and convergence in updating the filter coefficients of an adaptive IIR digital filter.
The amount of calculation in the filter coefficient control unit is reduced, and IIR
It is an object of the present invention to provide an active control device using an adaptive IIR digital filter, which can ensure the stability of a recursive unit when updating a filter coefficient of a digital filter and can cope with a periodic input. And

[0027]

In order to achieve the above object, an active control apparatus using an adaptive IIR digital filter according to the present invention comprises an adaptive IIR (Infinite Infinity Rate) signal processing means.
Impulse Response) in the active control apparatus using a digital filter, the adaptive IIR de error signal to minimize between desired response and the response of the subject to be controlled by the output signal of the adaptive IIR digital filter
A first coefficient control unit for updating a filter coefficient of the cyclic portion and the non-recursive portion of the digital filter, the output signal of the adaptive IIR digital filter, or the output signal of the cyclic portion is inputted to minimize the signal level So that the tour
A second coefficient control unit that updates a filter coefficient of the repetition unit, and the filter coefficient of the repetition unit is configured to be updated by the first coefficient control unit and the second coefficient control unit. (1) In an active control device using an adaptive IIR digital filter as a signal processing means, a cyclic portion of the adaptive IIR digital filter is connected to a stage preceding a non-cyclic portion, and An adaptive control unit for a filter coefficient for the repetition unit to which an error signal between a response of a target controlled by the output signal of the adaptive IIR digital filter and a desired response is input; And (2) an adaptive control section for adaptively controlling a filter coefficient for the acyclic section to which an output signal and the error signal are input. In an active control device using a digital filter, a first coefficient for updating a filter coefficient so as to minimize an error signal between a target response controlled by an output signal of the adaptive IIR digital filter and a desired response. A control unit; a second coefficient control unit that receives an output signal of the adaptive IIR digital filter or an output signal of the recursive unit and updates a filter coefficient so as to minimize the signal level; An output signal of the IIR digital filter or a signal of the recursive unit is input and a level detecting unit for detecting a signal level thereof, and updating of the filter coefficient of the recursive unit is performed by the first coefficient control unit and the second coefficient control unit. The second coefficient control unit controls the filter coefficient update operation according to the signal level detected by the level detection unit. It is characterized in that it is configured to be allowed (3).

[0028]

In the silencer shown in FIG. 11, when the conventional filter coefficient updating method is used, the signal may diverge in the circulating unit 17. According to the configuration described in (1) of the present invention, since the filter coefficient of the recurring unit is updated in parallel with the filter coefficient of the recurring unit so as to minimize the output level, The divergence of the signal in the traveling section is suppressed. That is, from the patrol section
So that the output level of the second coefficient is minimized.
When the filter coefficient of the recurring unit is updated by the
The stationary periodic component of the input signal to the cyclic portion
Is canceled and the periodic component of the output from the circulating unit is
It will approach zero. However, the tour
The random component of the input signal to the loop is canceled.
And remains in the output from the circulating unit. Therefore,
Controlled by a first coefficient control unit and a second coefficient control unit
What is the update of the filter coefficient of the circulating unit
The control for the random component is performed. Emits a noise source
The noise generated by the noise source is usually a random sound, and if the noise generated by the noise source is a random sound, even if the filter coefficient of the traveling unit is updated so that the output level of the traveling unit is minimized, the noise signal may be reduced. Is not canceled by the traveling unit. At this time, in the traveling unit, the sound radiated from the sound deadening speaker cancels a part of the signal component fed back to the noise detection microphone.

According to the configuration described in (2) above,
For updating the filter coefficient of the non-recursive part of the adaptive IIR digital filter, the F
This can result in updating coefficients of an IR (Finite Impulse Response) digital filter. The stability and convergence of the update of the filter coefficient of the FIR digital filter are guaranteed. Therefore, the stability and convergence of the update of the filter coefficient of the non-recurring portion are guaranteed. Further, in the case of updating the filter coefficients using Expressions 8 and 9, in the present invention, the calculation corresponding to Expression 8 has already been performed in the cyclic unit, and there is no need to perform the calculation again. Further, even if the order of the cyclic portion and the non-cyclic portion is switched as in the present invention, it is equivalent as a filter. Therefore, the filter coefficient of the recurring unit can be updated using the output of the non-recurring unit and the output error signal corresponding thereto.

Further, according to the configuration described in the above (3),
When the output level of the recurring unit is equal to or higher than a predetermined reference for the filter coefficient of the recurring unit, filter coefficient updating that minimizes the output level of the recurring unit can be performed in parallel. Since the level detector is provided, divergence of the signal of the traveling unit is suppressed. Even when the noise generated by the noise source is a periodic sound, if the output level of the traveling unit does not exceed a predetermined reference, the filter coefficient for minimizing the output level of the traveling unit is not updated and is required. The periodic signal is not completely canceled in the cyclic portion, and it is possible to more effectively cope with periodic sounds.

[0031]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an active control device using an adaptive IIR digital filter according to the present invention will be described below. Note that components having the same functions as those of the conventional example are denoted by the same reference numerals.

FIG. 1 is a block diagram showing an embodiment of an active control device using an adaptive IIR digital filter according to the present invention. A noise source 11 is disposed in a container 12 having an opening at one end, and a noise detection microphone 13 and a silencing error detection microphone 14 are disposed in the container 12 so as to face the noise source 11. Is provided with a muffling speaker 15. The noise detected by the noise detection microphone 13 is converted into a digital signal by an A / D converter 25 via a preamplifier 22, and the converted noise signal u (n) is processed by an adaptive IIR digital filter 16. The result of the calculation is output to the D / A converter 26 as a muffling signal y (n). The muffling signal converted to an analog signal by the D / A converter 26 is output from the speaker 15 via the power amplifier 23. The mute result is the mute error detection microphone 1
4 through a preamplifier 24 and an A / D converter 27 to obtain a noise reduction error signal -e (n) as coefficient control units 20 and 21.
It is taken in. The input u (n) of the acyclic unit 18 is also input to the digital filter 32, and the output u ₁ (n) of the digital filter 32 is input to the coefficient control unit 20. The output y (n) of the circulating unit 17 is also input to the digital filter 31, and the output y ₁ (n) of the digital filter 31 is also input to the coefficient control unit 21. Then, the filter coefficients a (i) and b (j) of the adaptive IIR digital filter 16 are updated so that the noise reduction error signal is minimized. The updating formula at this time is expressed by the following equations (10) and (11).

[0033]

A (i, n + 1) = a (i, n) + αu ₁ (ni) e (n)

[0034]

Equation 11] b (j, n + 1) = b (j, n) + βy 1 (nj) e (n) where alpha, and β denotes a step size parameter and takes a small positive value. In updating the filter coefficient b (j), f (n) in Equation 7 can be used instead of e (n) in Equations 10 and 11. Also, y in Equation 11
Instead of ₁ (nj), y ₂ (j, n) of the following Expression 12 can be used.

[0035]

(Equation 12)

It should be noted that the same result can be obtained even if the calculation unit of the formula 12 is provided before the digital filter 31 and the order is changed. Further, the filter coefficient b
(j) is also updated from the coefficient control unit 19. The coefficient control unit 19 receives the output y (n) of the cyclic unit 17 and outputs the output y (n)
The filter coefficient b (j) is updated according to the following equation 13 so that the level of?

[0037]

[Mathematical formula-see original document] b (j, n + 1) = b (j, n)-[gamma] y (nj) y (n) Here, [gamma] indicates a step size parameter, and takes a small positive value. Updating of the filter coefficients by the equations (11) and (13) is performed in parallel or time division.

In this embodiment, the filter coefficients of the recurring unit 17 are updated in parallel with the filter coefficients of the recurring unit 17 so as to minimize the output level. Divergence can be suppressed.
Further, when the noise is a random sound having no periodicity, the noise signal is not canceled by the traveling unit 17 even if the filter coefficient of the traveling unit 17 is updated so that the output level of the traveling unit 17 is minimized. At this time, the sound emitted from the muffling speaker 15 cancels a part of the signal component fed back to the noise detection microphone 13, and a further muffling effect can be expected.

FIG. 2 is a block diagram showing a second embodiment of an active control device using an adaptive IIR digital filter according to the present invention. The hardware configuration of this embodiment is different from that of the adaptive IIR digital filter shown in FIG. Circulating section 17 in IIR digital filter 16
And the acyclic unit 18 are merely interchanged, and a detailed description of the hardware configuration is omitted here.

In this embodiment, the filter coefficient a (i) is updated by the coefficient controller 20 according to the following equation (14).

[0041]

A (i, n + 1) = a (i, n) + αv ₁ (ni) e (n) The filter coefficient b (j) is updated by the coefficient controller 21 according to the following equation 11.

[0042]

Equation 11] b (j, n + 1) = b (j, n) + βy 1 (nj) e (n) Further, the cyclic portion 1 filter coefficients b (j) is a coefficient control unit 19
7 so as to minimize the level of the output v (n)
5 is updated.

[0043]

[Mathematical formula-see original document] b (j, n + 1) = b (j, n)-[gamma] v (nj) v (n) Updating of the filter coefficients by the equations (11) and (15) is performed in parallel as in the first embodiment. , Or time-divisionally.

In this embodiment, the filter coefficients of the recurring unit 17 are updated in parallel with the filter coefficients of the recurring unit 17 so as to minimize the output level. Divergence can be suppressed.
Further, by changing the arrangement of the non-cyclic unit 18 and the cyclic unit 17, it is possible to improve the stability of updating the filter coefficient of the non-cyclic unit 18. Further, as in the case of the above-described embodiment, when the noise is a random sound having no periodicity, even if the filter coefficient of the traveling unit 17 is updated so that the output level of the traveling unit 17 is minimized, the noise signal Is the traveling part 1
7, the sound emitted from the muffling speaker 15 at this time cancels a part of the signal component fed back to the noise detection microphone 13, and a further muffling effect can be expected.

FIG. 3 is a block diagram showing a third embodiment of an active control device using an adaptive IIR digital filter according to the present invention. The hardware configuration of this embodiment is the same as that of the adaptive IIR shown in FIG. The input of the coefficient control unit 9 of the digital filter is changed from the output v (n) of the cyclic unit to the output y (n) of the non-cyclic unit, and a detailed description of the hardware configuration is omitted here. I do.

The coefficient control section 19 outputs the output y (n) of the non-cyclic section 18
The filter coefficient b (n) is updated by the above equation 13 so as to minimize the level of.

In this embodiment, as in the case of the first and second embodiments described above, the filter coefficient is set so that the output level of the cyclic unit 17 is minimized with respect to the filter coefficient of the cyclic unit 17. Since the updating is performed in parallel, the divergence of the signal of the traveling unit 17 can be suppressed. Further, when the noise is a random sound having no periodicity, the noise signal is not canceled by the traveling unit 17 even if the filter coefficient of the traveling unit 17 is updated so that the output level of the traveling unit 17 is minimized. At this time, the sound emitted from the muffling speaker 15 cancels a part of the signal component fed back to the noise detection microphone 13, and a further muffling effect can be expected.

FIG. 4 is a block diagram showing a fourth embodiment of an active control apparatus using an adaptive IIR digital filter according to the present invention. A signal u (n) is input to an IIR digital filter 16 and y (n) is output. IIR
The internal configuration of the digital filter 16 is as follows: a circulating section 17 having a filter coefficient b (j) at the preceding stage;
(i) is located. The coefficient control unit 20 updates the filter coefficient a (i),
The coefficient controller 20 receives the output v (n) and the output error e (n) of the cyclic unit 17. The coefficient control unit 21 has II
The output y (n) of the R digital filter 16 and the output error e
(n) is entered. The output error e (n) is the desired response d
(n) and the output y (n) of the IIR digital filter 16. The coefficient control unit 20 updates the filter coefficient a (i) according to the following equation 16 so as to minimize the mean square value of the output error e (n).

[0049]

A (i, n + 1) = a (i, n) + αv (ni) e (n) In the coefficient control unit 21, the filter coefficient b (j) is calculated by the following equation (1).
Update according to 1.

[0050]

[Mathematical formula-see original document] b (j, n + 1) = b (j, n) + [beta] y (nj) e (n) Here, [alpha] and [beta] indicate step size parameters, and take small positive values. In updating the filter coefficient b (j), y0 (j, n) in the following _equation 9 can be used instead of y (nj) in the equation 11.

[0051]

[09]

In the present embodiment, the updating of the filter coefficient of the non-cyclic section 18 can be reduced to the updating of the coefficient of the FIR digital filter using the output of the cyclic section 17 as an input signal. , Its stability and convergence are guaranteed. Further, when the filter coefficients are updated using equations (8) and (9), the operation corresponding to equation (8) has already been performed in the circulating unit 17, so that there is no need to perform the calculation again, and the amount of calculation or the circuit scale is reduced. Can be achieved.

FIG. 5 is a block diagram showing a fifth embodiment of the active control apparatus using the adaptive IIR digital filter according to the present invention. In the embodiment shown in FIG. This is carried out by the unit 21 and is provided outside the coefficient control unit 21 in the configuration shown in FIG. The recursive digital filter 19 in FIG. 5 corresponds to this. In updating the filter coefficient b (j),
The following f which is the moving average of e (n) instead of e (n) of 1
(n) can also be used.

[0054]

[07]

Here, c (l) indicates the weight of the moving average, and L indicates the number of data to be averaged.

In this embodiment, as in the case of the above-described fourth embodiment, the update of the filter coefficient of the non-repeating section 18 is performed by using the FIR using the output of the recurring section 17 as an input signal.
Since the result of updating the coefficient of the digital filter can be reduced and the processing of Equation 9 is performed by the digital filter 19, the stability and convergence of the updating of the filter coefficient of the acyclic unit 18 are further ensured. Further, Equation 8,
When the filter coefficient is updated using equation (9), the operation corresponding to equation (8) has already been performed in the circulating unit 17, and there is no need to perform the calculation again, so that the amount of calculation or the circuit scale can be reduced. it can.

FIG. 6 is a block diagram showing a sixth embodiment of the active control device using the adaptive IIR digital filter according to the present invention. A noise source 11 is disposed in a container 12 having an opening at one end.
A microphone 13 for noise detection and a microphone 14 for noise reduction error detection are arranged opposite to each other.
5 are installed. Detected by the noise detection microphone 13,
The noise signal u (n) converted into a digital signal by the A / D converter 25 via the preamplifier 22 is processed by the adaptive IIR digital filter 16, and the calculation result is converted to a D / A signal as a muffling signal y (n). Output to the converter 26. D /
The mute signal converted into an analog signal by the A converter 26 is output from the speaker 15 via the power amplifier 23.
The silenced result is detected by the silence error detection microphone 14,
The coefficient is received by the coefficient control unit 20 and the coefficient control unit 21 via the preamplifier 24 and the A / D converter 27 as a silencing error signal -e (n). The output v (n) of the circulating unit 17 is also input to the digital filter 32, and the digital filter 32
Output v ₁ (n) is input to the coefficient control unit 20. The output y (n) of the acyclic unit 18 is also input to the digital filter 31, and the output y ₁ (n) of the digital filter 31 is also input to the coefficient control unit 21. Then, the filter coefficients a (i) and b (j) of the adaptive IIR digital filter 16 are updated so that the noise reduction error signal is minimized. The updating formula at this time is shown in the following Expressions 14 and 11.

[0058]

A (i, n + 1) = a (i, n) + αv ₁ (ni) e (n)

[0059]

Equation 11] b (j, n + 1) = b (j, n) + βy 1 (nj) e (n) where alpha, and β denotes a step size parameter and takes a small positive value. In updating the filter coefficient b (j), instead of e (n) in Equation 11, f (n) in Equation 7 is used.
Can also be used. Also, instead of y ₁ (nj) in Equation 11, y ₂ (j, n) in Equation 12 below can be used.

[0060]

(Equation 12)

In this embodiment, as in the case of the above-described fifth embodiment, the update of the filter coefficient of the non-cyclic unit 18 is performed by using the FIR with the output of the cyclic unit 17 as an input signal.
Since the result can be reduced to the update of the coefficient of the digital filter, the stability and convergence of the update of the filter coefficient of the acyclic unit 18 are guaranteed. Furthermore, the same result can be obtained even if the calculation unit of Expression 25 is placed before the digital filter 21 and the order is changed. When the filter coefficients are updated using the equations (8) and (9), the operation corresponding to the equation (8) has already been performed in the circulating unit 17, so that there is no need to perform the calculation again, and the amount of calculation or the circuit scale is reduced. Can be achieved.

FIG. 7 is a block diagram showing a seventh embodiment of an active control device using an adaptive IIR digital filter according to the present invention. A noise source 11 is disposed in a container 12 having an opening at one end, and a noise detection microphone 13 and a silencing error detection microphone 14 are disposed in the container 12 so as to face the noise source 11. Speaker 15 for noise reduction
Is installed. The noise signal u (n) detected by the noise detection microphone 13 and converted to a digital signal by the A / D converter 25 via the preamplifier 22 is processed by the adaptive IIR digital filter 16, and the calculation result is converted to a mute signal. It is output to the D / A converter 26 as y (n). D / A
The silencing signal converted to an analog signal by the converter is output from the silencing speaker 15 via the power amplifier 23. The silenced result is detected by the silence error detection microphone 14 and is taken into the coefficient control unit 20 and the coefficient control unit 21 via the preamplifier 24 and the A / D converter 27 as a silence error signal -e (n). The input u (n) of the non-repeating unit 18 is input to the digital filter 32, and the output u ₁ (n) of the digital filter 32 is input to the coefficient control unit 20. Circuit 17
The output y (n) is also input to the digital filter 31,
The output y ₁ (n) of the digital filter 31 is
Is input to Then, the filter coefficients a (i) and b (j) of the adaptive IIR digital filter 16 are updated so that the noise reduction error signal is minimized. The updating formula at this time is shown in the following equations 10 and 11.

[0063]

A (i, n + 1) = a (i, n) + αu ₁ (ni) e (n)

[0064]

Equation 11] b (j, n + 1) = b (j, n) + βy 1 (nj) e (n) where alpha, and β denotes a step size parameter and takes a small positive value. In updating the filter coefficient b (j), f (n) in Equation 7 can be used instead of e (n) in Equations 10 and 11. Also, y in Equation 11
Instead of ₁ (nj), y ₂ (j, n) of the following equation 12 can be used.

[0065]

(Equation 12)

It should be noted that the same result can be obtained even if the calculation unit of Expression 12 is placed before the digital filter 31 and the order is changed. Further, the filter coefficient b
(j) is also updated from the coefficient control unit 19. The coefficient control unit 19 receives the output y (n) of the cyclic unit 17 and outputs the output y (n)
The filter coefficient b (j) is calculated by the following equation (1) so that the level of
Update with 3.

[0067]

[Mathematical formula-see original document] b (j, n + 1) = b (j, n)-[gamma] y (nj) y (n) where [gamma] is a step size parameter and takes a small positive value. Here, the operation of updating the filter coefficient by the equation 13 controlled by the coefficient control unit 19 is based on the output signal y of the circulating unit 17.
(n). That is, the traveling unit 1
7 is input to the level detector 33, and the power P of the output signal y (n) is calculated. Coefficient control unit 19
Performs the filter coefficient update operation only when the input power P is larger than a predetermined reference power. At this time, the updating of the filter coefficients by the equations (11) and (13) is executed in parallel or in a time-division manner.

In the above-described embodiment, the operation of the coefficient controller 19 is permitted only when the signal level detected by the level detector 33 is equal to or higher than a predetermined reference value. Divergence is suppressed, and the adaptive IIR digital filter 16 can realize stable updating of filter coefficients. Furthermore, the coefficient control unit 19 operates only when the signal level detected by the level detection unit 33 is equal to or higher than a predetermined reference value.
Thus, an appropriate silencing process can be performed without being canceled.

FIG. 8 is a block diagram showing an eighth embodiment of an active control device using an adaptive IIR digital filter according to the present invention. The active control device of the adaptive IIR digital filter 16 shown in FIG. 17 shows a configuration in which the non-repeating unit 18 and the non-circulating unit 18 are interchanged. At this time, the filter coefficient a (i)
Is updated by the coefficient control unit 20 according to the following equation.

[0070]

A (i, n + 1) = a (i, n) + αv ₁ (ni) e (n) The coefficient control unit 21 updates the filter coefficient b (j) according to equation (11).

[0071]

Equation 11] b (j, n + 1) = b (j, n) + βy 1 (nj) e (n) Further, filter coefficient b (j) is the output v of the cyclic part 17 by the coefficient control unit 19 (n ) To minimize the level of
Will be updated by

[0072]

[Mathematical formula-see original document] b (j, n + 1) = b (j, n)-[gamma] v (nj) v (n) The output signal v (n) of the circulating unit 17 is input to the level detection unit 33, and the output signal v The power P of (n) is calculated. The coefficient control unit 19 executing Equation 15 performs the filter coefficient update operation only when the input power P is larger than the predetermined reference power. At this time, the update of the filter coefficients by the equations (11) and (15) is executed in parallel or in a time-division manner.

In the above-described embodiment, the operation of the coefficient control unit 19 is performed only when the signal level detected by the level detection unit 33 exceeds a predetermined reference value, as in the case of the above-described embodiment. By permitting, the divergence of the signal in the circulating unit 17 can be suppressed, and the adaptive IIR digital filter 16 can realize stable updating of the filter coefficient. Further, the coefficient control unit 19 operates only when the signal level detected by the level detection unit 33 is equal to or higher than a predetermined reference value, so that the cyclic unit 17 can perform not only a random sound but also a periodic signal. Appropriate silencing processing can be performed without being canceled.

[0074]

As described above in detail, in the active control apparatus using the adaptive IIR digital filter according to the present invention, according to the first aspect of the present invention, the coefficient control section controls the coefficient of the filter. By performing the updating in parallel, it is possible to prevent the divergence of the signal in the circulating unit, and to realize the stable updating of the filter coefficient of the adaptive IIR digital filter.

According to the second aspect of the present invention, with regard to updating the coefficient of the filter of the non-cyclic part of the adaptive IIR digital filter, the coefficient updating of the FIR digital filter using the output of the cyclic part as an input signal. Therefore, the stability and the convergence of the update of the filter coefficient of the non-repeated portion are guaranteed. In addition, the number 8, when updating the filter coefficients using equation 9, the calculations corresponding to the number 8 in this configuration has been made previously by the cyclic portion of the front stage, there is no need to again calculate, calculated Volume reduction,
Alternatively, the circuit scale can be reduced.

According to the third aspect of the present invention,
The filter coefficient update operation by the second coefficient control unit is:
By permitting when the signal level detected by the level detection unit becomes equal to or higher than a predetermined reference value, it is possible to suppress the divergence of the signal in the cyclic unit, and realize stable updating of the filter coefficient. it can. Further, the second coefficient control unit operates only when the signal level detected by the level detection unit is equal to or higher than a predetermined reference, so that even the periodic signal is canceled by the cyclic unit. And appropriate processing can be performed.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing a first embodiment of a silencer as an active control device using an adaptive IIR digital filter according to the present invention.

FIG. 2 is a configuration diagram showing a second embodiment of a silencer as an active control device using an adaptive IIR digital filter according to the present invention.

FIG. 3 is a configuration diagram showing a third embodiment of a silencer as an active control device using an adaptive IIR digital filter according to the present invention.

FIG. 4 is a block diagram showing a fourth embodiment of the active control device using the adaptive IIR digital filter according to the present invention.

FIG. 5 is a configuration diagram showing a fifth embodiment of the active control device using the adaptive IIR digital filter according to the present invention.

FIG. 6 is a configuration diagram showing a sixth embodiment of a noise reduction device as an active control device using an adaptive IIR digital filter according to the present invention.

FIG. 7 is a configuration diagram showing a seventh embodiment of a noise reduction device as an active control device using an adaptive IIR digital filter according to the present invention.

FIG. 8 is a configuration diagram showing an eighth embodiment of a noise reduction device as an active control device using an adaptive IIR digital filter according to the present invention.

FIG. 9 is a circuit diagram showing a direct type configuration of an IIR digital filter.

FIG. 10 is a circuit diagram showing a conventional active control device using an adaptive IIR digital filter.

FIG. 11 is a schematic block diagram showing a silencer as an active control device using a conventional adaptive IIR digital filter.

FIG. 12 is a circuit diagram showing a conventional active control device using an adaptive IIR digital filter.

[Explanation of symbols]

 Reference Signs List 16 adaptive IIR digital filter 17 cyclic section 18 non-cyclic section 19 coefficient control section 20 coefficient control section 21 coefficient control section 32 level detection section

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Kozo Hiyoshi 22-22 Nagaike-cho, Abeno-ku, Osaka-shi, Osaka Inside Sharp Corporation (58) Field surveyed (Int.Cl. ⁶ , DB name) G10K 11/178

Claims (3)

(57) [Claims] 1. An adaptive IIR (Infinite) signal processing means.
Impulse Response) in the active control apparatus using a digital filter, the adaptive IIR said adaptive IIR digital filter to the error signal to a minimum between the digital filter response of the subject and which is controlled by the output signal of the desired response Tour of
A first coefficient control unit for updating a filter coefficient of times unit and non-cycling portion, the output signal of the adaptive IIR digital filter, or the signal level output signal is input of the cyclic part so as to minimize A second coefficient control unit that updates a filter coefficient of the repetition unit, wherein updating of the filter coefficient of the repetition unit is controlled by the first coefficient control unit and the second coefficient control unit. An active control device using an adaptive IIR digital filter. 2. An active control device using an adaptive IIR digital filter as a signal processing means, wherein a cyclic portion of the adaptive IIR digital filter is connected to a stage preceding a non-cyclic portion, and an output signal of the non-cyclic portion and the adaptive signal. An adaptive control unit for a filter coefficient for the repetition unit which receives an error signal between a target response controlled by an output signal of a type IIR digital filter and a desired response, and an output signal of the repetition unit and the error signal And an adaptive control unit for adaptively controlling a filter coefficient for the non-recurring unit which receives the above as an input. An active control device using an adaptive IIR digital filter. 3. An active control device using an adaptive IIR digital filter as a signal processing means, wherein an error signal between a target response controlled by an output signal of the adaptive IIR digital filter and a desired response is minimized. A first coefficient control unit that updates a filter coefficient so as to perform an output signal of the adaptive IIR digital filter;
Alternatively, a second coefficient control unit that receives the output signal of the recursive unit and updates the filter coefficient so as to minimize the signal level, and that the output signal of the adaptive IIR digital filter or the signal of the recursive unit is A level detection unit for detecting a signal level of the input signal and updating the filter coefficient of the cyclic unit by the first coefficient control unit and the second coefficient control unit; An active control device using an adaptive IIR digital filter, wherein an operation of updating a filter coefficient by the section is permitted according to a signal level detected by the level detection section.