JP3382630B2 - Active noise and vibration 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 active noise and vibration control system for actively reducing noise in a passenger compartment of an automobile or a passenger compartment of an aircraft.
Regarding the device .

[0002]

2. Description of the Related Art Conventionally, as an active noise control device of this type, there is, for example, one shown in FIG. 5 of British Patent Publication No. 2149614.

This conventional device is applied to a cabin of an aircraft or a closed space similar to this, and is provided with loudspeakers 103a, 103b and 103c and microphones 105a, 105b, 105c and 105d in a closed space 101, and Control sounds that interfere with noise are generated by the speakers 103a, 103b, 103c, and residual signals (residual noise) are measured by the microphones 105a, 105b, 105c, 105d. The loudspeakers 103a, 103b, 103c and the microphones 105a, 105b, 105c, 105d are connected to a signal processor 107, and the signal processor 107 measures the fundamental frequency of the noise source measured by the fundamental frequency measuring means and the microphones 105a, 105b. , 105c, 105
and the loudspeakers 103a, 1a to receive the input signal from d and minimize the sound pressure level in the closed space 101.
The drive signal is output to 03b and 103c.

Here, in the closed space 101, three loudspeakers 103a, 103b, 103c and four microphones 105a, 105b, 105c, 105d are provided. 10
It is assumed that each of 3a and 105a is provided.
Now, the transfer function from the noise source to the microphone 105a is H, and the loudspeaker 103a to the microphone 10 are used.
Assuming that the transfer function up to 5a is C and the sound source information signal generated by the noise source is X _p , the noise signal E as residual noise observed by the microphone 105a is E = X _p · H + X _p · G · C Become. Here, G is a transfer function necessary for silencing. When the noise is completely canceled at the noise reduction target point (position of the microphone 105a), E = 0.
At this time, G is G = -H / C. Then, G that minimizes the microphone detection signal E is obtained, and the filter coefficient in the signal processor 107 is adaptively updated based on this G. As a method of obtaining the filter coefficient so as to minimize the microphone detection signal E,
LMS algorithm (Leas, which is a kind of steepest descent method)
t Mean Square) and the like.

When a plurality of microphones are installed as shown in FIG. 5, for example, each microphone 105
It is controlled so that the total sum of the signals detected by a, 105b, 105c, and 105d becomes the minimum.

Now, the LMS algorithm will be described more specifically. L-side microphone 105a
The noise signal detected by (105b ...) Is _el (n), and the l-th microphone 105a (105) when there is no control sound from the loudspeakers 103a, 103b, 103c.
b, ...) The noise signal detected by e _pl (n), the transfer function (FIR (finite impulse response)) between the m-th loudspeaker 103a (103b, ...) And the l-th evaluation point, that is, the working position. J) of the function (j = 0, 1, 2, ...,
The filter coefficient when the term I _c -1) is represented by a digital filter is C _{lm j} , the reference signal, that is, the sound source information signal x _p (n) and the reference signal x _p (n) are input, and the m-th loudspeaker 103a is input. If the i-th (i = 0, 1, 2, ..., I _K −1) coefficient of the adaptive filter that drives (103b, ...) Is W _mi ,

[0007]

[Equation 1]

[0008] is satisfied.

Next, the evaluation function (variable to be minimized) J
e

[0010]

[Equation 2]

[0011]

Then, in order to obtain the filter coefficient W _m that minimizes the evaluation function Je, the LMS algorithm is adopted. That is, the filter coefficient W _mi is updated with a value obtained by performing partial differentiation of the evaluation function Je with respect to each filter coefficient W _mi .

Therefore, from equation (2),

[0014]

[Equation 3]

From equation (1),

[0016]

[Equation 4]

Therefore, the right side of the equation (4) is r
_{If lm} (n−i) is written, the rewriting formula of the filter coefficient is obtained by the LMS algorithm of the following formula (5).

[0018]

[Equation 5]

In this LMS algorithm, the deviation between the coefficient C used in the calculation when the transfer function between the speaker and the microphone is represented by a digital filter and the coefficient C'in the actual acoustic space in the vehicle compartment is corrected. This system updates the filter coefficient of the adaptive digital filter.

[0020]

However, the deviation between the coefficient C used in the above calculation and the coefficient C'in the actual acoustic space becomes large, or the value of the predetermined convergence coefficient α of the LMS algorithm of the equation (5) is increased. Is too loud, the control sound generated from the loudspeaker is an uncomfortable intermittent sound (hereinafter referred to as “buzz”).
There is a risk that the sound will diverge in the worst case.
This phenomenon will be described with reference to FIG.

In this example, the coefficients of the applied digital filter are two, W _0, and W _{1 ,} for easy understanding. Figure 6
The quadratic curve (e) drawn by the solid line in (a) shows the square value of the sound pressure taken when the coefficients of the applied digital filter are W ₀ and W ₁ . The dotted line (2) is the equal squared sound pressure line. FIG. 6 (b) is a view obtained by cutting out and projecting FIG. 6 (a) by the equal squared sound pressure line (2).

Here, when the difference between the coefficient C used for calculation and the coefficient C'in the actual acoustic space is small, the values of W ₀ and W ₁ are calculated by the adaptive digital filter coefficient updating equation (5). Like the vector (a) in the middle, it draws a spiral on the quadric surface (e) and converges to the minimum point (c) of the square of the sound pressure. However, if the transfer characteristics in the vehicle interior change due to changes over time, the deviation becomes large, or the convergence coefficient α becomes relatively large, the minimum point (c) of the square of the sound pressure becomes like the vector (b). Instead of trying to converge, the orbit of a slender ellipse is drawn, but it rotates. When the rotary motion is performed in this manner, the filter coefficients W ₀ and W ₁ have the same phase,
Or, it changes in the opposite phase. The simulation result is shown in FIG. FIG. 7A shows a change in sound pressure input to the microphone when a "beat" occurs.
Further, FIG. 7B shows the filter coefficient W (mi) of the adaptive digital filter at that time. As mentioned above, m is the speaker number and i is the filter coefficient number. Therefore,
Looking at FIG. 7 (b), the filter coefficients W (11),
It can be seen that W (10) vibrates in the same phase.

There is a problem that an unpleasant "beat" is generated from the speaker due to such a change in the filter coefficient W (mi).

[0024] Therefore, the present invention is the ability that can be done "beat" effectively regulate, the accurately noise and vibration control
An object is to provide a dynamic noise and vibration control device .

[0025]

In order to solve the above-mentioned problems, the invention of claim 1 provides a control sound for interfering with noise and vibration.
・ A control source that generates control vibration to reduce noise / vibration at the evaluation point, a means to detect residual noise / vibration at a predetermined position after the interference, and a signal related to noise / vibration source noise / vibration generation state. means for creating a detected reference signal, the output of the residual noise and vibration detecting means on the basis of the output reference signal of the output signal and the reference signal producing means of the residual noise and vibration detecting means
Multiple filters in the digital filter to reduce the signal.
Output a signal to drive the control source by changing the filter coefficient
However, the phase shift of the plurality of filter coefficients due to the shift of control
Control noise and vibration of the control source
An active noise and vibration control device having a control means capable of becoming a control sound, which detects a beat state of the control sound or control vibration.
Buzzing state detection means and the control sound or control vibration
Detection of becomes state is characterized in that a beat state regulating means for regulating the beat state.

According to a second aspect of the present invention, there is provided a control source for reducing noise / vibration at an evaluation point by generating control sound / control vibration interfering with noise / vibration , and residual noise / vibration at a predetermined position after the interference.
Motion detecting means, means for detecting a signal related to the noise / vibration generation state of the noise / vibration source to create a reference signal, and residual noise detected by the residual noise / vibration detecting means.
.A / D converter for converting the vibration signal into a digital signal,
A digital filter for filtering the reference signal created by the reference signal creating means, a D / A converter for converting the output of the digital filter into an analog signal and outputting it to the control source , and a digital signal conversion by the A / D converter. active and control means for the to reduce residual noise and vibration changes the plurality of filter coefficients of the digital filter using a steepest descent algorithm
A noise and vibration control apparatus , comprising means for removing a DC component from the plurality of filter coefficients, means for calculating a product of the plurality of filter coefficients from which the DC component has been removed , and the calculation means.
And means for determining the DC component value, the absolute value of the DC component of the obtained product the control sound by exceeds a predetermined threshold value
Or a beat state detection hand that detects the beat state of controlled vibration
And a beat state of the control sound or control vibration.
It is characterized by comprising a beat state regulation means that is activated .

The invention of claim 3 is characterized in that the regulation means is means for stopping the control by the control means.

The invention of claim 4 is characterized in that the regulating means is means for correcting the steepest descent algorithm.

[0029]

According to the first aspect of the present invention, the control means has a residual noise.
A signal for driving the control source is output based on the output signal of the vibration detecting means and the reference signal of the reference signal generating means. As a result, the control source generates control sound / control vibration that interferes with noise / vibration to reduce the noise / vibration at the evaluation point. In such control, the detection means is controlled by a control sound or a beat of control vibration.
When the state is detected, the regulating means operates so as to regulate the growling state .

According to the second aspect of the invention, the DC component obtaining means removes the DC component from the plurality of filter coefficients of the digital filter, and obtains the DC component of the product of the filter coefficients from which the DC component has been removed, When the absolute value of the product of the obtained DC components exceeds a predetermined threshold value, the detection means detects that the control sound / control vibration is in a beat state.

According to the invention of claim 3, a control sound or
When the control vibration is in the beat state, the control by the control means is stopped by the stop means.

Further, according to the invention of claim 4, a control sound or
When the controlled vibration is in a beat state, the correction means corrects the steepest descent algorithm to regulate the beat state.

[0033]

Embodiments of the present invention will be described below.

The description will be given taking the vehicle interior space as an example.

FIG. 1 is a block diagram of an active noise control system according to an embodiment of the present invention. As a loudspeaker 3 which also serves as a control sound source and a residual noise detecting means in a vehicle interior 1 which is a closed space. The microphones 5 are provided and are respectively connected to the controller 7 as the control means. The controller 7 includes an adaptive controller 9 and an adaptive digital filter 1
1 and the microphone 5 is connected to the adaptive controller 9 via the A / D converter 13, and the loudspeaker 3 is connected to the adaptive digital filter 11 via the D / A converter 14. Has been done.

Noise in the passenger compartment 1 is generated by, for example, the power plant 1
5 is a noise source, and this power plant 1
In FIG. 5, an engine, a transmission as a power transmission device, and a differential gear are integrally housed. A crank angle signal sensor 17, for example, is used as a means for detecting a signal related to the noise generation state of the noise source, and the detected crank angle signal is used as a reference signal generation device 19
It is designed to be input to. Reference signal generator 19
The output of is output to the adaptive controller 9 and the adaptive digital filter 11.

The input signal to the reference signal generating device 19 may be a signal relating to the noise generation state of the noise source, for example, an output signal of a vibration sensor provided on the outer surface of the engine,
It is also possible to use an engine ignition pulse signal, a rotational speed signal obtained by detecting the rotational speed of the crankshaft with a sensor, or the like.

Further, in the embodiment of the present invention, a DC component obtaining device 21, a detector 23 and a switch 25 are provided.

The DC component detector 21 and the detector 23 constitute means for detecting the intermittent "beat" of the control sound. Then, the DC component obtaining device 21 constitutes means for removing the DC component from the filter coefficient of the adaptive digital filter 11 which is changed by using the steepest descent algorithm, and for obtaining the DC component of the product of the filter coefficients from which this DC component has been removed. ing. That is, the DC component determiner 21 includes a high pass filter and a low pass filter. The high pass filter and the low pass filter can be easily realized by an IIR filter.

Further, the detector 23 constitutes means for detecting the "beat" of the control sound when the absolute value of the DC component of the detected product exceeds a predetermined threshold value.

Further, the switch 25 constitutes means for stopping the control by the controller 7 as the control means, and constitutes means for regulating the intermittent state by detecting the "beat" of the control sound.

Next, the operation will be described.

The device shown in FIG. 1 basically operates in the same manner as the device shown in FIG.
The noise signal e (n) detected by is converted into a digital signal by the A / D converter and input to the adaptive controller 9. or,
The ignition pulse signal of the engine is input from the crank angle signal sensor 15 to the reference signal creating device 19 as a signal relating to the main noise component, and the reference signal creating device 19 creates the reference signal x _p by the ignition pulse signal and adaptively controls it. It is input to the device 9 and the adaptive digital filter 11. In the adaptive controller 9, the reference signal x _p and the noise signal e (n)
Based on the LMS algorithm using and

[0044]

[Equation 6]

The calculation of is performed. At this time, the filter coefficient C _lm when the transfer function between the loudspeaker 3 and the microphone 5 is represented by a digital filter is not displaced from the coefficient of the actual transfer function, and the convergence coefficient α is appropriate. If so, the LMS algorithm surely converges, and the filter coefficient W _mi of the adaptive digital filter 11 is rewritten by the equation (5).

Therefore, the adaptive digital filter 11
Is a reference signal x input from the crank angle signal sensor 17
A drive signal for the loudspeaker 3 is generated using _p, and the loudspeaker 3 outputs a speaker having an opposite phase to the engine noise drive signal transmitted by the transmission unit. Therefore, it is possible to perform accurate noise control.

At the same time (see FIG. 2), a plurality of filter coefficients W _mi of the adaptive digital filter 11 calculated in the update formula (5) of the LMS algorithm are calculated by the DC component determiner 31.
Input to the high pass filter of and removes each DC component.

Next, the product U of the filter coefficient W _mi ′ from which the DC component has been removed by the high pass filter is calculated. From the product U, an absolute value | U '| of only the DC component is taken out by a low pass filter.

Next, the absolute value | U '| of only the DC component is input to the detector 23 and compared with a predetermined threshold value _U'max . In this case, if the filter coefficient C _lm used in the calculation deviates from the coefficient of the actual transfer function due to a change with time or the like, | U ′ | becomes larger than the threshold value U ′ _max , and a signal is output to the switch 25 to output the controller. The control by 7 is stopped. Moreover, | U '| is threshold U' filter coefficients of the adaptive digital filter 11 in the LMS algorithm does not exceed the _max is updated.

Therefore, by such control, the "beat" of the control sound from the loudspeaker 3 can be regulated, and the noise control can be appropriately performed without causing the passenger an uncomfortable feeling.

FIG. 3 illustrates the method of detecting the "beat", and the filter coefficient W ₀ ,
I have two W ₁ . As shown in (a), the filter coefficient is input to the DC component obtaining unit 21, and the DC component is removed, resulting in (b). Multiply this (c)
And the absolute value of only the DC component is extracted.
The absolute value | U '| predetermined threshold as in the detector 23 (d) | U' | is being compared with _{ma x.}

FIG. 4 shows another example of the "beat" regulation control. When the absolute value | U '| of only the DC component is larger than _U'max , as shown in FIG. A signal is output to the filter 11, and the convergence coefficient α of the LMS algorithm is changed by α = Sα (S: a constant of 1 or less). Even if the filter coefficient C _lm used for calculation deviates from the coefficient of the actual transfer function by such control, the convergence coefficient α becomes correspondingly small, so that the LMS algorithm converges appropriately and the same as above. It is possible to achieve various operational effects. Moreover, in this example, since the control by the controller 7 is not stopped, continuous control can be performed.

For the sake of simplification of explanation, the description has been given assuming that one microphone and one loudspeaker are provided, but it is also possible to control by providing a plurality of each. Further, even if the evaluation point for noise reduction and the microphone position are spatially distant from each other, the residual noise at the evaluation point can be estimated based on a predetermined ratio and the control can be performed. It can also be applied to vibration control.

It is also conceivable that not only the absolute value of the product U and the absolute value of the DC component of the product U but also the moving average value of the product U is obtained and compared with a predetermined threshold value U _max for control.

[0055]

As is apparent from the above, according to the configuration of the present invention, the control sound or control vibration output from the control source is
When a growling state or the state becomes a growling state, this is regulated, and proper noise / vibration control can be performed without causing discomfort.

[Brief description of drawings]

FIG. 1 is a block diagram according to an embodiment.

FIG. 2 is a control flowchart for beat regulation.

FIG. 3 is an explanatory diagram showing a beat detection method.

FIG. 4 is a control flowchart of beat regulation according to another embodiment.

FIG. 5 is a block diagram according to a conventional example.

FIG. 6 is an explanatory diagram of a mechanism of beat generation.

FIG. 7 is an explanatory diagram of a simulation result.

[Explanation of symbols]

3 loudspeaker (control sound source) 5 Microphone (residual noise detection means) 7 Controller (control means) 9 Adaptive controller (control means) 11 Adaptive digital filter (digital filter) 13 A / D converter 21 DC component detector (DC component obtaining means) 23 Detector (Detection means) 25 switch (regulating means, stopping means) 11 Adaptive digital filter (correction means)

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kenji Sato, 2520 Takaba, Katsuta, Ibaraki Prefecture, Takaba, Ltd. Automotive equipment division, Hitachi, Ltd. (56) Reference JP-A-58-197927 (JP, A) JP 59-160335 (JP, A) JP-A-2-285799 (JP, A) JP-A-3-35295 (JP, A) JP-A-4-34599 (JP, A) JP-A-4-282695 (JP , A) Japanese Unexamined Patent Publication No. 5-27780 (JP, A) Actual Development No. 3-70490 (JP, U) (58) Fields investigated (Int.Cl. ⁷ , DB name) G01K 11/178 B60R 11/02 G05D 19/02 H03H 17/02 601 H03H 21/00

Claims (4)

(57) [Claims] 1. A control sound / control vibration for interfering with noise / vibration.
Motion source to reduce noise / vibration at the evaluation point, means for detecting residual noise / vibration at a predetermined position after the interference, and signal for noise / vibration source noise / vibration generation state are detected. means for creating a reference signal, based on said output reference signal of the output signal and the reference signal generating means of the residual noise and vibration detecting means and the remaining noise and vibration detection
Of the digital filter to reduce the output signal of the means
Drive the control source by changing a plurality of filter coefficients
Outputting a signal, the plurality of filter elements due to control deviation
The control sound / control vibration of the control source is affected by the phase change of the number.
Noise and vibration with control means that can be
A control device , which detects a beat state of the control sound or control vibration.
The state detection means , the detection of the beat state of the control sound or control vibration
Active noise vibration control apparatus being characterized in that a beat state regulating means for regulating the beat state. 2. Noise·vibrationControl sound to interfere with・ Control vibration
Generating rating noise·vibrationTry to reduceControl sourceWhen, Residual noise at a predetermined position after the interference·vibrationMeans to detect
When, noise·vibrationSource noise·vibrationDetects signals related to the occurrence state
And a means for creating a reference signal, The residual noise·vibrationResidual noise detected by the detection means
sound·vibrationA / D converter for converting digital signals into digital signals
When, The reference signal created by the reference signal creating means is
A digital filter for filtering, The output of the digital filter is converted to analog andControl
SourceD / A converter that outputs to Residue converted to digital signal by the A / D converter
noise·vibrationUse steepest descent algorithm to reduce
Change the filter coefficients of the digital filter.
Control means for convertingActive noise and vibration control deviceso
There Remove DC component from the plurality of filter coefficientsmeans
When, The product of a plurality of filter coefficients from which the DC component has been removedPlayed
Means to calculate, Of the calculated value Means for determining the DC component, The absolute value of the calculated direct current component of the product exceeds the specified threshold.
By the control soundOr the beat state of control vibration is detected
Beat state detection means, The control soundOr activated by detecting the beat state of control vibration
Howling control meansIt is characterized by having andNoh
Dynamic noise and vibration control device. 3. The active noise / vibration control device according to claim 1, wherein the regulation means is means for stopping control by the control means. 4. The active noise vibration control apparatus according to claim 2, wherein the regulation means is means for correcting the steepest descent algorithm.