JPH0540485A - Noise controller - Google Patents

Abstract

PURPOSE:To execute a follow-up even if a frequency of a noise is varied steeply, with regard to the noise controller for erasing a noise by outputting a signal being out-of-phase and having equal acoustic pressure to a noise detected by a microphone, from a loudspeaker. CONSTITUTION:In the noise controller having an adaptive type filtering means 4 for forming a compensating signal for erasing a noise to an electric signal/ acoustic wave converter 2 by controlling a filter coefficient, based on a signal of an acoustic wave/electric signal converter 3, a period detection control means 5 for detecting a noise period of a noise generation source 1m predicting a variation of the noise period, and changing a filter characteristic of the means 4 in accordance with a prediction variation of the noise period is provided. As for the means 5, it is also allowable to update and set of a mulfiplication coefficient of a multiplier contained in the means 4, and also, it is allowable to shift a tap of a delaying device, and it is also allowable that the multiplication coefficients of plural multipliers are subjected to vector display and the period variation detecting prediction is executed, and the multiplication coefficients of plural multipliers are updated and set.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a noise control device for canceling noise detected by a microphone by outputting a signal having a sound pressure of opposite phase and an equal phase negative pressure from a speaker. Even so, the purpose is to be able to follow.

[0002]

2. Description of the Related Art Conventionally, a passive muffling device such as a muffler has been used to reduce noise generated from an internal combustion engine, etc., but improvements have been made from the viewpoint of size and muffling characteristics. On the other hand, conventionally, there has been proposed an active noise control device that cancels noise by outputting from a speaker a compensating sound having an opposite phase and equal sound pressure to the noise generated from a sound source.

However, the frequency characteristics, stability, etc. of the active noise control device itself are not sufficient, and its practical application has been delayed. As a result of the recent development of signal processing technology using digital circuits and expansion of the frequency range that can be handled, many practical noise control devices have been proposed (for example, Japanese Patent Laid-Open No. 63-311396).

This is the result of noise being detected by a noise source microphone installed upstream of the duct, and a signal processing circuit outputting a signal of opposite phase and equal sound pressure to the noise from a speaker installed downstream of the duct, and the result being silenced. This is a so-called two-microphone / one-speaker type active noise control device in which a feedforward system and a feedback system for detecting and feeding back the noise by a microphone for a sound deadening point are combined.

[0005]

However, in the conventional active noise control device, when the noise period of the noise source changes abruptly, the drawback of the feedback system is that at least the sound transmission characteristic from the speaker to the microphone is exceeded. Since there is a delay, there was a problem that the noise reduction effect was reduced.

Therefore, the present invention has been made in view of the above problems.
An object of the present invention is to provide a noise cycle control device that can follow a sudden change in the noise cycle.

[0007]

FIG. 1 is a diagram showing the principle configuration of the present invention. In order to solve the above problems, the present invention provides an electric signal for eliminating noise from the noise source 1.
The sound wave converter 2, the sound wave / electric signal converter 3 for converting the residual sound of the noise erased by the sound wave from the electric signal / sound wave converter 2 into an electric signal, and the sound wave / electric signal converter 3 A noise control device having an adaptive filtering means 4 for forming a compensation signal for canceling noise to the electric signal / sound wave converter 2 based on the signal, in the adaptive filtering means 4 according to the predicted change of the noise period. A cycle detection control means 5 for changing the filter characteristic is provided.

The cycle detection control means 5 detects the noise cycle of the noise generating source 1, predicts a change in the noise cycle, and is included in the adaptive filtering means 4 according to the predicted change in the noise cycle. The multiplication coefficient set in the plurality of multipliers is updated and set. Further, the cycle detection control means 5
Detects a noise cycle in the noise source 1, predicts a change in the noise cycle, and moves taps of a plurality of delay devices included in the adaptive filtering means 4 according to the predicted change in the noise cycle. You may make it so.

Further, the period detection control means 5 forms a multidimensional vector of multiplication coefficients of a plurality of multipliers included in the adaptive filtering 4, detects and predicts a change in the vector, The multiplication coefficients of the plurality of multipliers may be updated and set according to the predicted change of the vector.

[0010]

According to the noise control device in FIG. 1, the adaptive filtering means 4 for inputting the noise signal from the noise signal from the noise source 1 and the sound signal of the opposite phase equal sound pressure from the speaker 2 is input. The compensation signal is adjusted in amplitude and phase to muffle noise. Further, when the noise cycle changes suddenly, the cycle detecting means detects the noise cycle change, and predicts the change of the previous noise cycle in consideration of the transfer characteristic up to the muffling point via the electric signal / sound wave converter 2 and the like. The multiplication coefficients of the plurality of multipliers constituting the adaptive filtering means 4 are shift-controlled, and the cycle of the compensating sound wave from the electric signal / sound wave converter 2 coincides with the cycle of noise at the muffling point.
Therefore, it becomes possible to follow even if the noise cycle changes sharply.

Even if the tap of the delay device of the adaptive filtering means 4 is moved by the cycle detecting means 5, the same function can be obtained. Further, the cycle detecting means 5 vectorizes the multiplication coefficient of the multiplier of the adaptive filtering 4,
Since the vector change is closely related to the noise cycle, the noise cycle can be easily predicted by predicting the vector change, and the transfer characteristic is taken into consideration. Can be matched with.

[0012]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 2 is a diagram showing a noise control device according to an embodiment of the present invention. The configuration of this figure will be described. The noise control device shown in the figure has a speaker 2 for eliminating noise from a noise source 1 such as an automobile engine near a muffling point P (in the figure).
An amplifier 201 for amplifying the output to the speaker 2, and a D / A (Digital To Ana) for converting a digital signal into an analog signal in order to supply an analog signal to the amplifier 201.
log Converter) converter 202, a microphone 3 for converting the residual sound of the noise generated by canceling the noise from the noise source 1 with the sound wave from the speaker 2 into an electric signal, and amplifying the electric signal of the microphone 3. Amplifier for 30
1 and an A / D (Analog To Digital Converter) converter for converting an analog signal of the amplifier 301 into a digital signal
302 and an adaptive filtering means 4 for controlling a filter coefficient based on a signal from the A / D converter 302 to form a compensation signal for canceling noise to the speaker 2.
And input the timing signal from the noise source 1,
A noise signal from a microphone 6 or the like, which will be described later, or a noise reproduction signal from the difference signal calculating means 8 or the like is input, a noise cycle is detected, a cycle change is predicted, and the adaptive filtering means 4 is selected according to the cycle predicted change. Period detection control means 5 for controlling the abrupt change to control abrupt changes, a microphone 6 installed near the noise source 1, an amplifier 601 for amplifying the output of the microphone 6, and an amplifier 601.
A / D that converts the analog output signal of the
A transfer characteristic simulation for simulating a transfer characteristic Hd which is connected to the output of the converter 602 and the adaptive filtering means 4 and reaches the input of the difference signal calculating means 8 described later via the speaker 2 and the microphone 3 from the output point. Means 7, a difference signal calculating means 8 for calculating a difference signal between the output of the transfer characteristic simulating means 7 and the output of the A / D converter 302, and the input signal of the adaptive filtering means 4 alternatively. Switch means 9 for selecting. Here, the adaptive filtering means 4, the cycle detection control means 5, etc. are provided with a DSP (Digital Signal).
Processor)).

FIG. 3 is a diagram showing the structure of the cycle detection control means shown in FIG. The cycle detection control means 5 shown in the figure comprises a cycle detection section 501, a cycle prediction section 502, and a control section 503 for the coefficient of the adaptive filtering means 4 and the like. FIG. 4 is a diagram for explaining an example of the cycle detecting method of the cycle detecting unit 501 in FIG.
FIG. 1A shows a method of detecting ignition timing or rotation timing (rotation speed) of an automobile engine, a motor or the like as the noise generation source 1. A rectangular wave signal is input to the input of the cycle detection unit 501, the cycle T is obtained, and output to the cycle prediction unit 502. This is because abrupt noise changes in automobile noise occur due to changes in the engine speed of the automobile.

When the timing signal as shown in FIG. 3A cannot be obtained, the noise waveform is detected from the microphone or the vibrometer 6 in the vicinity of the automobile engine or the like and the time waveform peak is obtained from the waveform of FIG. It shows that the period T of the noise signal is obtained. In this signal processing, if a rectangular wave is generated when a certain noise signal level exceeds a constant noise signal level, the cycle T can be obtained as in the case of FIG.

FIG. 1C is a BP for obtaining the noise period T after digitizing the noise signal input to the microphone.
An F (Band Pass Filter) peak detection method is shown. In this method, a plurality of band pass filters 1, 2, ..., N, an absolute value conversion unit (ABS) connected to each band pass filter 1, 2, ..., N, and an absolute value conversion unit are connected. Averaging unit (LP
F) and a maximum band detection unit that detects the maximum value of each averaging unit, detects the maximum frequency band of the noise level, and sets the period of the maximum frequency band as the period of the noise signal.

FIG. 3D shows a cycle detecting method using an adaptive filter, which is a delay device for inputting the difference signal S _R of the difference signal calculating means 8 and an output of the delay device. An adaptive filter (ADF), an adder that takes the difference signal between the output of the adaptive filter and the through input signal, and a least squares method for the difference signal of the adder to determine the coefficient of the adaptive filter. It consists of a square method processing unit (LMS) and calculates the period of the noise signal from the coefficient of the adaptive filter.

FIG. 5 is a diagram for explaining a method of predicting the periodic change amount based on the detection period. If the cycle predicting unit 502 changes so that the cycle becomes smaller at time (t ₀ ) where the initial cycle is constant as shown in this figure, the cycle detecting unit 501 detects the cycle change amount as shown in the figure. .. On the other hand, in the conventional technique, the transfer characteristic Hd is delayed at the position of the microphone 3 as shown in the figure. Here, for the sake of simplicity of description, the transfer characteristics and the like of the signal processing unit such as the adaptive filtering unit 4 are ignored. The cycle predicting unit 502 considers the transfer characteristic Hd and calculates data for changing the cycle earlier as shown by the curve in the figure with respect to the curve in the figure. Although the change of the period is shown as a straight line with respect to time in this figure, it may be a curve, and in that case, the curve in the figure may be determined by fitting a function and fitting. In the curve of this figure obtained in this way, the period T ₂ predicted with respect to the period T _{1 at} the present time (t ₁ ) is obtained. The control unit 503 such as the ADF coefficient shown in FIG. 3 will be described later.

Next, the adaptive filtering means 4
Briefly explained. The switch means 9 operates the difference signal calculation means 8.
When selected, the noise S of the noise source 1_NAnd then
Its transfer characteristics up to 3_NOISEAnd the adaptive fi
The compensation signal of the filtering means 4 is S_CAnd adaptive fill
Microphone from the dialing means 4 through the speaker 2
The transfer characteristic of the system up to 3 is H_SPAnd microphone 3
The transfer characteristic of the system from the differential signal calculating means 8 to H_micage,
Regarding the transfer characteristic Hd1 of the transfer characteristic simulating means 7,
= H_SP・ H_mic= Hd, from microphone 3
Output residual sound signal S_MIs S_M= S_N・ H_NOISE
+ S_C・ H_spBecomes Therefore, the calculation in the difference calculation unit 8
The resulting difference signal S_RIs S_R= S_M・ H_mic-S_C・
Hd1 = S_N・ H_NOISE・ H_mic+ S_C・ H_sp・ H_mic
-S_C・ H_SP・ H_mic= S_N・ H _NOISE・ H_micTona
, The signal when only the noise is detected by the microphone 3
It will be calculated. Also, adaptive filtering means
And a control signal for changing the coefficient of the adaptive filter of 4.
Output S of the A / D converter 302_EIs given.
In the adaptive filtering means 4, this control signal becomes zero.
The coefficient is changed as_E= S_M・ H_mic
Because S_EWhen = 0, S_M= 0. Therefore, the difference
Difference signal S from the signal calculation means 8_RIs suitable as a controlled signal
It is input to the adaptive filtering means 4 and A is used as a control signal.
Output S of the / D converter 302_EBy typing
The adaptive filtering means is S_ECompensate to be = 0
Signal S_CIs calculated. Switch means 9 is a microphone
If 6 is selected, the microphone 6 is suitable as an input signal.
The adaptive filtering means 4 uses the compensation signal S_CIs calculated.

FIG. 6 is a diagram showing the adaptive filtering means. The adaptive filtering means of this figure is configured by a non-recursive filter, specifically, a series of delay devices 401 that delays one sampling period, and a plurality of multipliers 402 connected to each delay device 401, A plurality of adders 403 for adding the outputs of the multipliers 402, and a coefficient updating means 404 for controlling the multiplication coefficient of each of the multipliers 402 so that the output of the microphone 3 is minimized by the least square method. Including.

The series of delay units 401 may be composed of a random access memory (RAM). In this case, the input sampling data is sequentially shifted to the next address for each sampling, or the address to which the sampling data is input. It suffices that the value of is sequentially shifted for each sampling. The multiplication coefficients g ₁ , g ₂ , ..., G _n of the multiplier 402 of the adaptive filtering means 4 shown in this figure.
About ADF coefficient control unit of the cycle detection control means 5
Reconfiguration by 503 will be explained.

FIG. 7 is a diagram for explaining the shift of the multiplication coefficient of a plurality of multipliers constituting the adaptive filtering. This figure (a) shows typically the signal which passes the delay device 401. Normally, each multiplier is based on the signal from the microphone 3.
The multiplication coefficients (g ₁ , g ₂ , ... G _n ) of 402 are set, but when the cycle prediction unit 502 predicts a change from a short cycle to a long cycle, the control unit 503 such as the coefficient of the ADF etc. ,
The multiplication coefficients (g ₁ , g ₂ , ... G _n ) of each multiplier 402 are (g ′ ₀ , g ₁ , g ₂ , ..., G _n−1 ), ..., (g ′ _-8 , g ′).
_{-7, ..., g '0,} g 1, g 2, ..., such that the g _n-9), that is, shifted toward the n-th multiplier (delay unit). As a result, the delay amount becomes longer and the cycle can be lengthened.

Contrary to the above, FIG.
Is predicted to change from a long cycle to a short cycle in AD
The coefficient of F, etc. is controlled by the control unit 503 to determine the multiplication coefficient of each multiplier 402.
(G ₁, G₂,,, g_n) Is (g₂, G₃,,, g_n，
g '_{n + 1}), ..., (g_Ten, G₁₁,,, g_n, G ′_{n + 1}, G ′
_{n + 2},…, G ′_{n + 9}), ..., that is, the 0th
It is shifted towards the multiplier (delayer). By this
The delay amount can be shortened and the cycle can be shortened.
It However, g ′ is an arbitrary optimum value (for example, 0).

FIG. 8 is a diagram for explaining the tap movement of the delay device constituting the adaptive filtering means, which is a modification of that shown in FIG. In this figure (a), the taps (T ₁ , T ₂ , ..., T _n ) of the delay device 401 are normally set, but the cycle prediction unit 502 predicts a change from a short cycle to a long cycle. And the taps (T ₁ , T ₂ , ..., T _n ) are (T ′ ₀ , T ₁ , T ₂ ,
…, T _n-1 ),…, (T ′ _{−1 0} ,…, T ′ ₋₁ , T ′ ₀ ,
_{T 1, T 2, ...,} T n-9), is shifted towards the other words in the n-th delay units so that ... a. As a result, the delay amount becomes longer and the cycle can be lengthened.

Contrary to the above, FIG.
Is predicted to change from a long cycle to a short cycle in AD
The taps (T ₁ , T ₂ , ..., T _n ) of each delay device 401 are (T ₂ , T ₃ , ..., T _n ,
T ′ _{n + 1} ), ... (T ₁₀ , T ₁₁ , ..., T _n , T ′ _{n + 1} , T ′
_{n + 2} , ..., _{T'n + 9} ), ..., That is, toward the 0th multiplier. As a result, the delay amount becomes short and the cycle can be shortened. However,
T'is an arbitrary optimum value (for example, 0).

FIG. 9 is a diagram showing another modification of the cycle detection control means of FIG. Cycle detection section 501 of cycle detection control means 5
Inputs the multiplication coefficient of the multiplier 402 of the adaptive filtering means 4 and forms the next n-dimensional vector.

[0026]

[Equation 1]

The adaptive filtering means 4 sequentially calculates the multiplication coefficients (g ₁ , g ₂ ) as shown in the figures (a), (b) and (c).
, G _n ) is updated, the period prediction unit 502
, The vector after t time is predicted, and when this prediction is made, the multiplication coefficient (g ₁ , g ₂ , ..., G _n ) is calculated from this vector. Is calculated, and these multiplication coefficients are set in the multiplier 402 by the control unit 503 such as the ADF coefficient. In this way, the filter characteristic of the adaptive filtering unit 4 can be changed by changing the multiplication coefficient of the multiplier 402 included in the adaptive filtering unit 4 or by moving the output tap of the delay unit 401. it can.

[0028]

As described above, according to the present invention, the noise cycle of the noise generating source is detected, and the cycle is controlled by predicting the future from the characteristics of the noise cycle. It became possible to follow.

[Brief description of drawings]

FIG. 1 is a diagram showing a principle configuration of the present invention.

FIG. 2 is a diagram showing a noise control device according to an embodiment of the present invention.

FIG. 3 is a diagram showing a configuration of a cycle detection control means of FIG.

FIG. 4 is a diagram illustrating a cycle detection method of a cycle detection unit in FIG.

FIG. 5 is a diagram illustrating a method of predicting a periodic change amount.

FIG. 6 shows the adaptive filtering means of FIG.

FIG. 7 is a diagram for explaining shifting of multiplication coefficients of a plurality of multipliers that constitute adaptive filtering means.

FIG. 8 is a diagram for explaining tap movement of a plurality of delay units that constitute adaptive filtering means.

FIG. 9 is a diagram showing another modification of the cycle detection control means in FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Noise generating source 2 ... Electric signal / sound wave converter 3 ... Sound wave / electric signal converter 4 ... Adaptive filtering means 5 ... Cycle detection control means

Claims (4)

[Claims] 1. An electric signal / sound wave converter (2) for eliminating noise from a noise generating source (1), and residual sound of noise eliminated by sound waves from the electric signal / sound wave transducer (2). Wave / electric signal converter (3) for converting sound into electric signals
And an adaptive filtering means (4) for controlling a filter coefficient based on the signal of the sound wave / electrical signal converter (3) to form a compensation signal for canceling noise to the electric signal / sound wave converter (2). ) And a noise control device including: a noise control unit that detects a noise cycle of the noise generation source (1), predicts a change in the noise cycle, and, in accordance with the predicted change in the noise cycle, the adaptive filtering unit (4). A noise control device comprising: a cycle detection control means (5) for changing the filter characteristic of the above. 2. The cycle detection control means (5) detects a noise cycle of the noise generating source (1), predicts a change in the noise cycle, and adaptively detects the noise cycle according to the predicted change in the noise cycle. 2. The noise control device according to claim 1, wherein the multiplication coefficient of the multiplier included in the filtering means is updated and set. 3. The cycle detection control means (5) detects the noise cycle of the noise source (1), predicts a change in the noise cycle, and adaptively responds to the predicted change in the noise cycle. 2. The noise control device according to claim 1, wherein an output tap of a delay unit included in the mold filtering unit 4 is moved. 4. The cycle detection control means (5) forms a multidimensional vector of multiplication coefficients of a plurality of multipliers included in the adaptive filtering (4), detects a change in the vector, and The noise control device according to claim 1, wherein the noise control device predicts and updates the multiplication coefficients of the plurality of multipliers according to the predicted change of the vector.