JP3499885B2 - Adaptive active noise control device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an adaptive active noise control device.
It is about. [0002] 2. Description of the Related Art The noise of air conditioners and the noise of automobile engines are known.
Microphone for detecting noise such as sound and microphone for detecting residual noise
With the noise at the control point
And radiate an additional sound with the same amplitude.
For example, a technology for silencing by mutual interference between noise and noise
Reference 1: JP-A-3-40598 (H04R3 / 00)
It has been disclosed. [0003] Also used in an adaptive active noise control device.
Adaptive Digital Filter
r: ADF), the VS (Variable Step) algorithm
This algorithm is used, for example, in Reference 2:
IEEE TRANSACTIONS ON ACOUSTICS, SPEECH, AND SIGNAL P
ROCESSING, VOL, ASSP-34.NO.2, APRIL 1986 PP309-316 "A
Variable Step (VS) Adaptive Filter Algorithm "
Well explained. [0004] In the conventional VS algorithm, noise
Detection microphone and coherence for residual sound detection
It may not be 1.0 due to the effect of wind, or
Affected, for example, Document 3: IEICE
Literature A Vol.J70-A No.12 pp1858-1860 December 1987
Speed-up of Variable Step Adaptive Digital Filter "
It is explained as follows. That is, a basic adaptive function as shown in FIG.
Noise control device (here, ANC: Adaptive Nois Canc)
eller) at time k, main input d (k) of the ANC, reference
The input x (k), the vector X is set with the order of the ADF being L._k Is
It is defined as Equation 1. [0006] (Equation 1) The (p + 1) tap at time k
Weight coefficient of the filter_p(k), the weighting factor
Kuturu W_k Becomes like Equation 2. [0008] (Equation 2) This input vector X_k And filter weights
Number vector W_k Is used, the output y (k) of the ADF becomes
Is given as [0010] (Equation 3) In the VS algorithm, according to the following equation 4,
Modify the weighting factor of the filter. [0012] (Equation 4) The error correction coefficient μ at the (p + 1) -th tap_p(k)
Is μ_max And μ_min To change between. Also μ_max Following
Is set so that the expression 5 is satisfied. [0014] (Equation 5) Further, μ_min The required adaptive error after convergence is
Take this into consideration. That is, even if the adaptive error after convergence is large,
If good, μ_min Large, and the adaptive error after convergence is small.
If you want to_min Set smaller. Thus, μ_pIn setting (k), the error signal
Power E [e^Two(k)] weighting factor w of the filter_pfor (k)
∇_{p, k} (= -2e (k) x (kp))
You. ∇ _{p, k} Is m₀ If it changes consecutively,
Μ to reduce the step width_pmultiply (k) by 1 / α (α> 1)
You. Also_{p, k} Is m₁ If the same for consecutive times,
Μ to stop convergence speed_p(k) is multiplied by α. As described above, m₀ , M₁ Increase the value of
By setting, the system diverges due to the influence of wind and noise
I tried not to make noise control faster
There was a problem that efficiency could not be achieved. [0018] SUMMARY OF THE INVENTION The present invention relates to the above-mentioned prior art.
It was made in view of the problem of operation, and the noise detection
Between the microphone and the residual noise detection microphone
Poor coherence due to disturbance and wind turbulence
Efficient VS algorithm without diverging even if there is a frequency
To provide an adaptive active noise control device that can apply
aimed to. [0019] SUMMARY OF THE INVENTION The present invention is directed to a method for detecting noise.
Microphone and microphone for residual noise detection
And a noise signal based on the noise detection microphone.
Adaptive sound to create additional sound to cancel noise
Filter and noise signal from the noise detection microphone.
Signal and the signal detected by the microphone for residual noise detection.
Control unit for controlling the coefficient of the adaptive filter
Based on the additional sound signal from the adaptive filter.
Speaker for generating an additional sound, and a microphone for detecting the noise
Cohere between microphone and microphone for residual noise detection
A first band-pass filter in the form of a
Around which is the target of the detection signal by the sound detection microphone
A second bandpass filter for band limitation for limiting a wavenumber band
Filter and the first and second bandpass filters.
Of the detection signal of the residual noise detection microphone
An arithmetic control unit for calculating the reliability based on the difference;
A noise signal from a detection microphone and the second
The residual noise detection microphone passed through the bandpass filter
Changes in the sign of the product of the detection signals of the
The control unit based on the reliability obtained by the control unit
Algorithm that computes the step width of coefficient correction in VS
A control unit, wherein the control unit controls the VS algorithm.
Based on the calculation result of the rhythm control unit, the adaptive filter
It is characterized by modifying the coefficient of Lutha Microphone for noise detection and residual noise detection
Pre-measurement of coherence with
In accordance with the coherence waveform in the frequency band
Create one bandpass filter. [0021] A microphone for detecting a residual noise is input.
Pass the signal through a first bandpass filter. And target
Passed through a second bandpass filter representing the frequency band
The difference is calculated by the arithmetic and control unit, and for example,
A is defined by such a function, and this A is set between 0.0 and 1.0.
In addition, the signal picked up by the residual noise detection microphone
The reliability at that time. [0022] (Equation 6) With such a configuration, m₀ ,
m₁ VS algorithm is no longer necessary
M in the₀ , M₁ Control is stable and
Can be performed at high speed. [0024] BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an adaptive active noise control apparatus according to the present invention.
Example applied to an apparatus for controlling noise of a machine
This will be described in detail with reference to the drawings. FIG. 1 shows an adaptive active noise control device for an air conditioner.
FIG. 1 is a diagram showing a configuration, wherein reference numeral 1 denotes a noise for detecting noise of a duct or the like.
Detection microphone, 2 add noise canceling
Adaptive filter to create sound, 3 is the frequency of the additional sound
Bandpass filter for first band limiting for limiting several bands
Filter, 4 is a speaker that outputs the created additional sound, 5 is
The microphone for detecting the residual noise at the point
The noise signal obtained by the noise detection microphone 1
Second band limiting bar that limits the frequency band to be signaled.
Filter and FIR filter, 8 is Filtered-
X algorithm control unit, 9 is a microphone for detecting the residual noise.
The target frequency of the residual noise signal obtained by the lophone 5
Third band limiting bandpass filter for limiting wavenumber band
10 is a VS algorithm control unit, and 11 is
The arithmetic control unit 12 for obtaining the reliability A using the
Outgoing microphone 1 and residual noise detection microphone
Coherence band according to the form of coherence with 5
It is a pass filter. In the configuration shown in FIG.
Of microphone 1 and microphone 5 for residual noise detection
Measurement. The waveform diagram of FIG. 2 shows the waveform diagram of this result.
Is shown. And in the frequency band of interest in the device
Coherence band
A dope filter 12 is created. The detection signal of the noise detection microphone 1 is
After passing through the band limiting filter 6 and the FIR filter 7,
ltered-X algorithm control unit 8 and VS algorithm
Each is supplied to the control unit 10. On the other hand, detection of the residual noise detection microphone 5
The outgoing signal passes through the bandpass filter 9
ltered-X algorithm control unit 8, VS algorithm control
What is supplied to the control unit 10 and the arithmetic control unit 11
The arithmetic control unit 1 passes through the coherence bandpass filter 12
It is divided into those supplied to 1. In the arithmetic control unit 11, the band limiting band
The residual noise signal passed through the pass filter 9 and the coherence bar
Between the residual noise signal passed through the
An operation based on the notation 6 is performed, and the obtained value of A is
It is supplied to the algorithm control unit 10. In the VS algorithm control unit 10, (the number
ERR at 6₁ ) × X (X: the band limiting file)
(A noise signal that has passed through the filter 6 and the FIR filter 7)
Looking at the number of consecutive changes in the code,
By counting the number of consecutive times as A times instead of counting once,
Use the result. In other words, when reliability is low,
Do not count as a continuation. As a result,
M as explained in the technical section₀ , M₁ Large in advance
There is no need to take it. FIG. 4 shows such a VS algorithm control unit 10.
5 is a flowchart showing the flow of the calculation in the
In step S1, the calculation result A obtained by the calculation control unit 11 is acquired.
Mi, ∇ _{p, k} The change in the sign of is determined in step S2,
If it is continuous, the obtained A is added to the input x (k) and a new
In step S3, an operation for setting X (k) is performed.
If it changes, add the obtained A to the output y (k).
In step S4, an operation for newly setting y (k) is performed. This
In step S5, x is m₀ Greater or greater
Io y is m₁ Only change the number of steps if
New and execute the next operation of A with X and y set to 0 respectively.
This is repeated hereafter. In the Filtered-X algorithm control unit 8,
Based on the result of the VS algorithm control unit 10,
A signal for correcting the weight coefficient of the adaptive filter 2 is generated.
And supplies it to the adaptive filter 2. So
Adaptive filter 2 is based on the filter coefficients.
To create an additional sound signal with the same amplitude and opposite phase as the noise signal
This is passed through the band-pass filter 3 for band limitation.
Supply to peaker 4. The speaker 4 receiving this will make noise
Outputs additional sound to cancel. [0033] According to the present invention, as described above, the conventional VS
M needed for algorithm₀, M₁ No to setting
Because noise is eliminated, noise can be reduced while maintaining control stability.
Adaptive Active Noise Control System with High Speed Cancellation
Can be obtained.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a configuration of an adaptive active noise control device according to the present invention. FIG. 2 is a waveform diagram showing coherence between the noise detection microphone and the residual noise detection microphone in FIG. FIG. 3 is an input / output diagram for explaining the principle of a general adaptive active noise control device. FIG. 4 is a flowchart illustrating an operation in the VS algorithm. [Description of Signs] 1 Microphone for noise detection 2 Adaptive filters 3, 6, 9 Bandpass filter 4 for band limitation Speaker 5 Speaker for residual noise detection 7 FIR filter 8 Filtered-X algorithm control unit 10 VS algorithm control unit 11 Operation control Part 12 Coherence band pass filter

Claims (1)

(57) [Claims 1] A microphone for detecting noise, a microphone for detecting residual noise, and a microphone for detecting noise
Additional sound to cancel noise based on noise signal from phone
And an adaptive filter for creating, said adaptive filter based on a detection signal from the residual noise detecting microphone and the noise signal by the noise detecting microphone
A control unit that controls the coefficient of the sound, a speaker that generates an additional sound based on the additional sound signal from the adaptive filter, and a coherence between the noise detection microphone and the residual noise detection microphone. 1 band-pass filter and the residual sound detection microphone.
A second band-pass filter for limiting a frequency band to be detected by the first and second band-pass filters ;
Calculates the reliability based on the difference between the detection signals of the residual noise detection microphone that has passed through the band-pass filter of FIG.
Calculation control unit, and the noise detection microphone
Passing through the noise signal and the second bandpass filter
Of the product of the detection signals of the microphone for detecting residual noise.
The sign change and the reliability determined by the arithmetic and control unit.
A coefficient correction step in the control unit based on the degree
A VS algorithm control unit for calculating a width, wherein the control unit calculates the calculation result of the VS algorithm control unit.
The adaptive active noise control device, wherein the coefficient of the adaptive filter is corrected based on the following .