JPH07319478A - Active type noise control device - Google Patents

Abstract

PURPOSE:To prevent a device from becoming a state in which the device does not work effectively because a control system diverges. CONSTITUTION:A control sound generating means 10 generates a control sound reducing noise by interfering with noise based on a driving signal, a residual noise measuring means 12 outputs a residual noise signal by measuring the residual noise after the interference with the control sound. A simulated control sound signal calculating means 14 calculates a simulated control sound equivalent to the control sound generated by the control sound generating means 10 based on the driving signal and also calculates the simulated control sound signal at the time of assuming that the simulated control sound is measured by the residual noise measuring means 12 and outputs the simulated control sound signal. A driving signal calculating means 16 calculates the driving signal based on the simulated control sound signal, the residual noise signal and a filter coefficient and a filter coefficient adjusting means 28 adjusts the filter coefficient so that the gain of the one turn transfer function of a closed loop composed of the control sound generating means, the residual noise measuring means, the simulated control sound signal calculating means and the driving signal calculating means becomes less than one.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an active noise control device, and more particularly to an active noise control device that actively reduces noise in a predetermined space such as a vehicle compartment or a factory.

[0002]

2. Description of the Related Art FIG. 2 shows a conventional active noise control device (Japanese Patent Laid-Open No. 5-503622) based on the feedback control technique which is the basis of the present invention.

The conventional active noise control device is applied when a noise reference signal cannot be measured, and includes a control sound generating means 10 for generating a control sound that interferes with the noise and reduces the noise based on the drive signal. , Residual noise measuring means 12 for measuring residual noise after interference and outputting a residual noise signal, and control sound generating means 1 based on the drive signal.
A simulated control sound corresponding to a control sound generated by 0 is calculated, and a simulated control sound signal is calculated on the assumption that the simulated control sound is measured by the residual noise measuring means 12,
Simulated control sound signal calculation means 14 for outputting this simulated control sound signal, and drive signal calculation means 1 for calculating a drive signal based on the simulated control sound signal, the residual noise signal and the filter coefficient.
6 and filter coefficient adaptation means 18 for adjusting the filter coefficient so as to reduce the residual noise signal. In addition, 11 is a combination point of control sound and noise,
Reference numeral 15 is an addition point of the residual noise signal and the simulated control sound signal.

Here, z ^-1 is used as a delay operator, the drive signal calculation means 16 is composed of a variable coefficient FIR filter having a characteristic of C (z ^-1 ), and the simulated control sound signal calculation means 14 has a characteristic of H ( z ^-1 ) fixed coefficient FIR filter, the transfer characteristic from the drive signal to the residual noise signal is G (z ^-1 ), the residual noise signal is y (k), the drive signal is u (k), and the noise is w
(k), simulated control sound signal z (k), simulated control sound signal z (k)
The simulated noise signal synthesized from the residual noise signal y (k) is e (k).

Further, the characteristic C (z ^-1 ) of the FIR filter,
The i-th coefficient of H (z ^-1 ) is C _i , H _i , the number of taps of each FIR filter is p, m, the delay due to the control period, and the closed loop of the transfer characteristic G (z ^-1 ) dead time, etc. Assuming that the dead time possessed by z ^−L , the characteristics H (z ⁻¹ ) and C (z ⁻¹ ) are expressed by the following equations (1) and (2). However, it is assumed that this dead time is included in G (z ^-1 ) and H (z ^-1 ).

H (z ^-1 ) = (H ₀ z ^-0 + H ₁ z ^-1 + ... + H _m-1 z ^{-m + 1} ) z ^-L (1) C (z ^-1 ) = C ₀ z ^-0 + C ₁ z ^-1 + ... + C _P-1 z ^{-P + 1} (2) As a method of minimizing the residual noise signal y (k), one of the steepest descent methods, the least mean square (LMS) The algorithm and the Filtered-X algorithm represented by the following expressions (3) and (4) are used.

C _i (k + 1) = C _i (k) + a (y (k) x (k−i)) ... (3) x (k + 1) = H ₀ e (k−L) + H ₁ e (k− 1-L) + ... + _Hm- 1e (k-m + 1-L) (4) However, i = 0, ..., p-1 and a is a weight for filter coefficient adjustment.

In feedback control in which a residual noise signal is fed back to form a closed loop, stability of the closed loop becomes a problem. Therefore, in the above-mentioned conventional active noise control device, the simulated control sound signal is subtracted from the residual noise signal and the above algorithm is applied to substantially avoid the divergence of the closed loop, that is, to guarantee the stability.

However, since these algorithms change the filter coefficient successively only for the purpose of minimizing the residual noise signal, the stability of the closed loop and the convergence of the parameters are not always guaranteed. The problem is that it may diverge. In particular, when the characteristics of the sound field between the control sound generating means and the residual noise measuring means fluctuate, the stability of the closed loop is not guaranteed.

Therefore, an object of the present invention is to provide an active noise control device which prevents the control system from diverging and the device from not working effectively.

[0011]

In order to achieve the above object, the invention of claim 1 is a control sound generating means for generating a control sound which interferes with noise and reduces the noise based on a drive signal, and a control sound. Residual noise measuring means for measuring residual noise after interference with and outputting a residual noise signal, and calculating and simulating a simulated control sound corresponding to the control sound generated by the control sound generating means based on the drive signal. Simulated control sound signal calculation means for calculating a simulated control sound signal in the case of assuming that the control sound is measured by the residual noise measurement means, and outputting the simulated control sound signal, the simulated control sound signal and the residual noise signal. Drive signal calculation means for calculating the drive signal based on a filter coefficient, the control sound generation means, the residual noise measurement means, the simulated control sound signal calculation means, and the drive signal calculation Gain of one-round transfer function of the closed loop composed of steps is configured to include a filter coefficient adjusting means for adjusting the filter coefficients to be 1 or less, the.

According to a second aspect of the present invention, the one-cycle transfer function of the first aspect of the invention is the transfer function from the driving signal to the simulated noise signal synthesized from the simulated control sound signal and the residual noise signal, and the driving signal. It is characterized in that it is represented by the product of the transfer function of the calculation means.

According to a third aspect of the invention, in the first or second aspect of the invention, the filter coefficient adjusting means may diverge the filter coefficient based on the residual noise signal and the simulated control sound signal. If there is, a dead zone for stopping the adjustment of the filter coefficient is further provided.

[0014]

According to the present invention, the control sound generating means generates a control sound that interferes with the noise and reduces the noise based on the drive signal, and the residual noise measuring means causes the residual noise after the interference with the control sound. Is measured and the residual noise signal is output. The simulated control sound signal calculation means calculates a simulated control sound corresponding to the control sound generated by the control sound generation means based on the drive signal, and the simulated control when it is assumed that the simulated control sound is measured by the residual noise measuring means. The sound signal is calculated and the simulated control sound signal is output. The drive signal calculation means calculates the drive signal based on the simulated control sound signal, the residual noise signal, and the filter coefficient, and the filter coefficient adjustment means is the control sound generation means,
The filter coefficient is adjusted so that the gain of the closed loop one-cycle transfer function composed of the residual noise measuring means, the simulated control sound signal calculating means, and the driving signal calculating means becomes 1 or less.

In the control theory, if the gain of the closed loop open loop transfer function is 1 or less, the closed loop is guaranteed by the small gain theorem. According to the invention of claim 1, the closed loop open loop transfer function is obtained. Gain of 1
Since it is set as follows, it is possible to prevent the divergence of the closed loop, prevent the divergence of the filter coefficient, and reduce the residual noise.

According to the second aspect of the present invention, the closed loop one-loop transfer function is transferred from the drive signal to the simulated noise signal synthesized from the simulated control sound signal and the residual noise signal, and the driving signal computing means. Since it is expressed by the product of the transfer function,
By obtaining beforehand the change in conditions that may occur in the sound field, the change in the transfer function from the drive signal to the simulated noise signal becomes known, so that the filter coefficient can be easily adjusted.

According to the third aspect of the invention, when it is determined that the filter coefficient may diverge based on the residual noise signal and the simulated control sound signal, the adjustment of the filter coefficient is stopped, The divergence of the filter coefficient can be prevented, and the convergence of the feedback system can be more surely guaranteed.

[0018]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 6 is a block diagram of the active noise control device according to the present embodiment. The active noise control device includes a control sound generating means 10 that generates a control sound that interferes with noise and reduces the noise based on a drive signal, and a residual noise measuring means that measures residual noise after interference and outputs a residual noise signal. 12 and a signal processing device 30.

The control sound generating means 10 is composed of a speaker 10A and an amplifier 10B, and the residual noise measuring means 12 is composed of a microphone. In addition, the signal processing device 3
Reference numeral 0 is composed of a digital signal processing device DSP 32, an A / D converter 34, and a D / A converter 36, which are configured by a computer or the like and execute signal processing described below in accordance with a prestored program. Speaker 10A
Is connected to the D / A converter 36 of the signal processing device 30 via the amplifier 10B, and the microphone is connected to the A / D converter 34 of the signal processing device 30.

When the signal processing of the DSP 32 is shown by a functional block, as shown in FIG. 1, a simulated control sound corresponding to the control sound generated by the control sound generating means 10 is calculated based on the drive signal, and the simulated control sound is generated. Residual noise measuring means 12
The simulated control sound signal calculation means 14 for calculating the simulated control sound signal and outputting the simulated control sound signal, and the drive signal based on the simulated control sound signal, the residual noise signal, and the filter coefficient. Is represented by the drive signal computing means 16 for computing the above, and the filter coefficient adjusting means 28 for adjusting the filter coefficient of the drive signal computing means 16. In addition,
In FIG. 1, the A / D converter 34 and the D / A converter 36 are omitted.

The simulated control sound signal calculating means 14 has a fixed filter coefficient and F of H (z ^-1 ) whose characteristic is represented by the above equation (1).
The drive signal calculating means 16 is composed of an IR filter, and receives the drive noise signal u (k) as an input to the drive noise signal e (k) synthesized from the drive noise signal z (k) and the residual noise signal y (k). )
It is composed of a C (z ⁻¹ ) FIR filter whose characteristic is expressed by the above-mentioned equation (2) and whose filter coefficient for generating The coefficient of the fixed FIR filter having the characteristic of H (z ^-1 ) was previously identified before the control.

The coefficient of the variable FIR filter having the characteristic C (z ⁻¹ ) is the adaptive circuit 2 of the filter coefficient adjusting means 28.
It is sequentially adjusted by 6.

C (z ^-1 ) = C ₀ (k) z ^-0 + C ₁ (k) z ^-1 + ... +
C _P-1 (k) z ^{p + 1 When} the signal processing of the filter coefficient adjusting means 28 is shown by a functional block, as shown in FIG. 3, the control sound generating means 10,
Residual noise measuring means 12, simulated control sound signal calculating means 14
1 of the closed loop composed of the drive signal calculating means 16
A closed-loop divergence prevention circuit 20 that makes the gain of the cyclic transfer function 1 or less, a filter-coefficient divergence prevention circuit 22 that stops adjusting filter coefficients based on the residual noise signal and the simulated control sound signal, a closed-loop divergence prevention circuit 20, and a filter. It is represented by a divergence prevention law synthesis circuit 24 that synthesizes the output of the coefficient divergence prevention circuit 22 and an adaptive circuit 26 that adjusts the filter coefficient based on the output of the divergence prevention law synthesis circuit 24.

Next, the operation of this embodiment will be described. The adaptive circuit 26 adjusts the filter coefficient of the variable FIR filter C (z ⁻¹ ) so as to minimize the residual noise signal y (k) in the region where the closed loop is stable and the filter coefficient does not diverge. As a method for minimizing the residual noise signal, an algorithm based on the LMS algorithm, the Filtered-X algorithm, or the like described in the related art can be used, but in the present embodiment, the algorithms of the above formulas (3) and (4) are used. The case of use will be described.

Here, the filter coefficient adjustment weight a is fixed in a general algorithm, but as will be described below, in the present embodiment, the results of the closed loop divergence prevention circuit 20 and the filter coefficient divergence prevention circuit 22 are determined. Based on this, the divergence prevention law combining circuit 24 adjusts.

That is, in the control theory, if the gain of the closed loop open loop transfer function is 1 or less, the small gain theorem guarantees that the closed loop is stable. Therefore, the closed-loop divergence prevention circuit 20 adjusts the parameters as follows so that the active noise control device satisfies the small gain theorem.

Using the symbols explained in FIG. 1, in the active noise control system of this embodiment, the closed loop has transfer functions H (z ^-1 ), G (z ^-1 ), and C as shown in FIG. (z ^-1) is constituted by a transmission element of the difference between the transfer function H (z ^-1) and the transfer function G (z ^-1) and R (z ^-1), 1 round transfer function of the closed loop Is a transfer function R from the drive signal to the simulated noise signal synthesized from the simulated control sound signal and the residual noise signal.
(z ^-1) and the product R (z ^-1) of the transfer function C of the drive signal calculating means 16 (z ^-1) can be expressed by C (z ^-1).

Here, as shown in the following equations (5) and (6), D is a control frequency of the transfer function R (z ^-1 ) (for example, 50).
The maximum value of the gain within the range of 1 Hz to 1 kHz, and F is defined as the sum of the absolute values of the coefficients of the taps of the signal processing device 30.

D = sup ω ‖ R (e- ^j ω ‖ (ω represents a frequency) (5) F (k) = ‖C ₀ (k) ‖ + ‖C ₁ (k) ‖ + ... + ‖C _P-1 (k) ‖… (6) From this definition, D ≧ ‖R (z ^-1 ) ‖, and F (k) = ‖C ₀ (k) ‖ + ‖C ₁ (k) ‖ ＋… ＋ ‖C _P-1 (k) ‖ ＝ ‖C ₀ (k) ‖‖ z ^-0 ‖ ＋ ‖C ₁ (k) ‖‖ z ^-1 ‖ ＋… ＋ ‖C _P-1 (k) ‖ z ^(P-1) ‖ ≧ ‖C ₀ (k) z ^-0 + C ₁ (k) z ^-1 +… + C _P-1 (k) z- ^(P-1) ‖ = ‖C ( z ^-1 ) ‖ ... (7).

From the above relational expression, DF (k) ≧ ‖C (z ⁻¹ ) ‖‖R (z ⁻¹ ) ‖ ≧ ‖C (z ⁻¹ ) R (z ⁻¹ ) ‖… ( 8)

Therefore, if the characteristic C (z ^-1 ) of the FIR filter is adjusted so that DF (k) is 1 or less, the gain of the closed loop open-loop transfer function C (z ^-1 ) R (z ^-1 ) is obtained. Becomes 1 or less, and the closed loop is always stable.

Since the adaptive circuit 26 uses the algorithms of equations (3) and (4), the characteristic that DF (k) becomes 1 or less by adjusting the filter coefficient adjustment weight a is C (z
^-1 ) FIR filter can be obtained.

Specifically, the closed loop divergence prevention circuit 20.
Has a maximum value D of the error between the transfer function H (z ^-1 ) of the simulated control signal calculation means 14 and the characteristic G (z ^-1 ) from the drive signal to the residual noise signal, and DF (k) < To be 1,
The filter coefficient adjustment weight a is determined according to the filter coefficient weight adjustment rule. In the present embodiment, the following adjustment rule is applied as the filter coefficient weight adjustment rule.

[0034]

[Equation 1]

However, the initial value of the weight a is defined as a _def .

Further, the maximum error value D is, for example, in the case of noise reduction in the passenger compartment, after the fixed filter H (z ^-1 ) is determined, the actual operating condition such as when there is an occupant or when there is no occupant. For all, drive signal calculation means 16 is used as a drive signal.
Instead of the output signal, a transfer signal from the drive signal to the pseudo noise signal is used, using a signal such as a random signal of 50 to 500 Hz including a frequency component of the control band or a sine wave signal whose frequency is temporally changed. The maximum value of the gain of F
The maximum value among the maximum values of the respective gains measured by the FT analyzer or the like is set as D.

A method of configuring the signal processing device so that the gain of the closed loop open loop transfer function is 1 or less is not limited to the method shown here, but the filter coefficient of C (z ^-1 ) is fixed. In this case, there are various methods such as the H∞ control system design method and the LQG control system design method. Further, if D is set in consideration of the change of the characteristic G (z ⁻¹ ), the closed loop can be always kept stable in the change of the considered characteristic.

In this case, since the change in the transfer function R (z ^-1 ) from the drive signal to the simulated noise signal is known by previously obtaining the change in conditions that may occur in the sound field, the filter coefficient is It can be easily adjusted.

[0039] Here, the transfer function G (z ^-1) or H (z ^-1), the function B (z ^-1) separating the dead time possessed by the closed loop,
Each A (z ^-1 ) is defined by the following equation.

B (z ⁻¹ ) = H (z ⁻¹ ) z ^L (9) A (z ⁻¹ ) = G (z ⁻¹ ) z ^L (10) Function B (z ^{− 1} ) and A (z ⁻¹ ), the block diagram of FIG. 4 can be rewritten as the block diagram of FIG.

Furthermore, the following assumption is introduced. However, these assumptions often hold.

A (z ⁻¹ ) z ^L C (z ⁻¹ ) ≈C (z ⁻¹ ) A (z ⁻¹ ) z ^L (11) (A (z ⁻¹ ) −B (z ⁻¹ )) u (k−L) = (A (z ⁻¹ ) −B (z ⁻¹ ) z ^L C (z ⁻¹ ) e (k) ≈C (z ⁻¹ ) A (z ⁻¹ ) −B (z ⁻¹ )) e (k−L) (12) Using the above assumption, the residual noise signal y (k) can be rewritten as follows.

Y (k) = G (z ^-1 ) u (k) + w (k) = A (z ^-1 ) z ^-L C (z ^-1 ) e (k) + w (k) = C (z ^-1 ) A (z ^-1 ) e (k-L) + w (k) = C (z ^-1 ) B (z ^-1 ) e (k-L) + C (z ^-1 ) (A (z ^-1 ) -B (z- ¹ )) e (k-L) + w (k) = C (z- ¹ ) + B (z- ¹ ) ^-1 ) B (z- ¹ ) e (k-L) -e ( k−L) + C (z ⁻¹ ) (A (z ⁻¹ ) −B (z ⁻¹ )) e (k−L) + e (k) − (A (z ⁻¹ ) −B (z ⁻¹ )) ) u (k-L) = (C (z- ¹ ) + B (z- ¹ ) ^-1 ) B (z- ¹ ) e (k-L) + (e (k) -e (k-L)) (13) Further, from the equations (4) and (9), x (k) = B (z ^-1 ) e (k-L) (14) and the characteristic is C (z ^-1 ). The number of taps p of the FIR filter is approximately F (F) with the characteristic of B (z ^-1 ) ^-1
Assuming that the characteristic can be described by an IR filter and is large to some extent, this characteristic B (z ⁻¹ ) ⁻¹ can be described by the following equation.

B (z ⁻¹ ) ⁻¹ = −B ₀ z ⁻⁰ −B ₁ z ⁻¹ −... −B _p-1 z ^{−p + 1} (15) Further, the following variables are defined here. (However, T represents a transposed matrix).

V (k) = e (k) -e (k-L) (16) Z (k) = (z- ^0x (k), z ^- 1x (k), ..., Z- ^{p +1} x (k)) ^T (17) P ^* = (B ₀ , B ₁ , ..., B _p-1 ) ^T (18) P (k) = (C ₀ (k), C ₁ (k ), ..., C _p-1 (k)) ^T (19) ΔP (k) = P (k) −P ^* (20) Using the variables defined in this way, equation (13) and
Expression (3) can be described by the following expressions.

Y (k) = ΔP (k) ^T Z (k) + v (k) (21) P (k) = P (k−1) −aZ (k) y (k) (22) Based on the above preparation, the convergence condition of the filter coefficient will be derived below by using Lyapunov's theory of stability.

First, the positive definite value function V (k) is defined by the following equation. V (k) = ΔP (k) ^T ΔP (k) (23) The function ΔV (k) indicating the time change of this positive stationary function is given by the following equation.

[0048] △ V (k) = V ( k) -V (k-1) = △ P (k) T △ P (k) - △ P (k-1) T △ P (k-1) = [ ΔP (k) -ΔP (k-1)] ^T [ΔP (k) + ΔP (k-1)] = [-aZ (k) y (k)] ^T [2ΔP (k) + aZ (k) y (k ) ] ^{= -a 2 Z (k) T} Z (k) y (k) 2 -2a (Y (k) -v (k)) y (k) ... (24) where , Lyapunov's theory of stability shows that the positive stationary function V (k)
On the other hand, if -ΔV (k) is a quasi-definite function, it is guaranteed that the filter coefficient P (k) will converge. (24)
From the formula, a sufficient condition for −ΔV (k) to be a quasi-definite function can be calculated as a = 0 ... (25) when | y (k) | <| v (k) | . This condition is equivalent to setting a dead zone for the residual noise signal.

Therefore, when the filter coefficient divergence prevention circuit 22 of the present embodiment determines that the filter coefficient may diverge based on the residual noise signal and the simulated control sound signal, the adjustment of the filter coefficient is stopped. There is a dead zone to

The condition for preventing the filter coefficient divergence can be similarly obtained for the filter coefficient adjustment rules other than the filter adjustment rules represented by the equations (3) and (4).

Since the filter coefficient divergence prevention circuit 22 uses the Filtered-X algorithm in the adaptive circuit 26, the method described above is used as it is. That is, the dead zone was set for the error output as in the following equation.

If │y (k) │ <│v (k) │ a2 = 0 (26) else a2 = a _def (27) Thus, the filter coefficient is based on the residual noise signal and the simulated control sound signal. When it is determined that the filter coefficient may diverge, the adjustment of the filter coefficient is stopped, so that the filter coefficient can be prevented from diverging and the convergence of the feedback system is further ensured.

However, if the dead time of the closed loop is sufficiently smaller than the control band and the high frequency component is sufficiently small with respect to this dead time, or if the residual noise signal can be sufficiently removed by a low-pass filter or the like, v ( k) = e (k)-
Since e (k−L) ≈0, this condition, that is, the dead zone can be eliminated, and thus the filter coefficient divergence prevention circuit 22 can be omitted.

The closed loop divergence prevention circuit 20 and the filter coefficient divergence prevention circuit 22 output different filter coefficient adjustment weights. Therefore, the divergence prevention law combining circuit 24
The filter coefficient adjustment weights output from the two closed-loop divergence prevention circuits 20 and the filter coefficient divergence prevention circuit 22 are combined according to the law shown in the following equation, and the result is output to the adaptation circuit 26. Note that min (al, a2)
Means that the smaller one of al and a2 is adopted.

A = min (a output from the closed loop divergence prevention circuit, a output from the filter coefficient divergence prevention circuit) = min (al, a2) (28) However, the filter coefficient divergence prevention circuit 22 is not used. In this case, the divergence prevention law synthesis circuit 24 is equivalent to the output al of the closed loop divergence prevention circuit 20.

In this embodiment, the noise at the evaluation point is reduced without the noise reference signal by avoiding the divergence of the closed loop, which is a problem in the conventional active noise control device based on the feedback control.

FIG. 7 shows the sound pressure of the residual noise when the active noise control device is not used. FIG. 8 shows the results using the conventional method. In the conventional method, due to the divergence of the closed loop or the filter coefficient, the sound pressure of the residual noise at the evaluation point is higher than when the active noise control device is not used. FIG. 9 shows the result of using the active noise control device of the invention. From FIG. 9, the sound pressure is reduced by a few dozen dB at a frequency having a high frequency of sound pressure when uncontrolled, and this embodiment solves the problems of the closed loop of the conventional method and the divergence of the filter coefficient. It can be understood that the residual noise sound pressure can be reduced.

In this embodiment, since the filter coefficient is adjusted so that the gain of the closed loop is 1 or less, the convergence of the feedback system is guaranteed, and one of the closed loop loop transfer functions has a known change. Since the function is a function, the filter coefficient can be easily adjusted. Further, since the adjustment of the filter coefficient is stopped due to the dead zone, it is possible to prevent the divergence of the filter coefficient.

[0059]

As described above, according to the invention of claim 1, since the gain of the closed loop one-round transfer function is set to 1 or less, the divergence of the closed loop is prevented and the divergence of the filter coefficient is prevented. In addition, the effect that the residual noise can be reduced can be obtained.

According to the second aspect of the present invention, the closed loop open loop transfer function of the drive signal to the transfer function from the drive signal to the simulated noise signal synthesized from the simulated control sound signal and the residual noise signal, and the drive signal calculating means. Since it is expressed by the product of the transfer function,
By obtaining beforehand the change in conditions that may occur in the sound field, the change in the transfer function from the drive signal to the simulated noise signal becomes known, and the filter coefficient can be easily adjusted.

Further, according to the invention of claim 3, when it is judged that the filter coefficient may diverge based on the residual noise signal and the simulated control sound signal, the adjustment of the filter coefficient is stopped, The effect that the divergence of the filter coefficient can be prevented and the convergence of the feedback system is further ensured can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing functional blocks of an embodiment of the present invention.

FIG. 2 is a block diagram of a conventional feedback type active noise control device.

FIG. 3 is a block diagram specifically showing a filter coefficient adjusting unit according to the exemplary embodiment of the present invention.

FIG. 4 is a block diagram for explaining an operation of a closed loop divergence prevention circuit.

FIG. 5 is a block diagram for explaining the operation of a filter coefficient divergence prevention circuit.

FIG. 6 is a block diagram showing an embodiment of the present invention.

FIG. 7 is a diagram showing the sound pressure of residual noise when there is no control (when the active noise control device is not used).

FIG. 8 is a diagram showing the sound pressure of residual noise when no control is performed and when an active noise control device according to a conventional method is used.

FIG. 9 is a diagram showing the sound pressure of residual noise when no control is performed and when the active noise control device according to the present embodiment is used.

[Explanation of symbols]

 10 Control Sound Generating Means 12 Residual Noise Measuring Means 28 Filter Coefficient Adjusting Means

Front Page Continuation (72) Hideo Nakai Inventor Hideo Nakai 1 in 41, Yokochi, Nagakute-cho, Aichi-gun, Aichi Prefecture Toyota Central Research Institute Co., Ltd. (72) Inventor Taiji Kimura, Toyota-cho, Aichi Prefecture Toyota-cho Within the corporation

Claims (3)

[Claims] 1. A control sound generating means for generating a control sound that interferes with noise and reduces the noise based on a drive signal, and residual noise for measuring residual noise after interference with the control sound and outputting a residual noise signal. Measuring means, and a simulated control in the case of assuming that the simulated control sound corresponding to the control sound generated by the control sound generating means is calculated based on the drive signal and the simulated control sound is measured by the residual noise measuring means. Simulated control sound signal calculation means for calculating a sound signal and outputting a simulated control sound signal; drive signal calculation means for calculating the drive signal based on the simulated control sound signal, the residual noise signal and a filter coefficient, The closed loop one-loop transfer function having the control sound generating means, the residual noise measuring means, the simulated control sound signal calculating means and the driving signal calculating means has a gain of 1 or more. Active noise control system comprising, a filter coefficient adjusting means for adjusting the filter coefficients such that. 2. The one-cycle transfer function is represented by a product of a transfer function from a drive signal to a simulated noise signal synthesized from a simulated control sound signal and a residual noise signal and a transfer function of a drive signal computing means. The active noise control device according to claim 1. 3. A dead band for stopping the adjustment of the filter coefficient when the filter coefficient may diverge based on the residual noise signal and the simulated control sound signal. The active noise control device according to claim 1 or 2, characterized by having.