JPH04274500A - Active type noise controller - Google Patents

Abstract

PURPOSE:To enable control wherein the influence of higher harmonic components which have constant phase shifts from a reference signal even when the higher harmonic component is included in a residual noise in a closed space. CONSTITUTION:The active noise controller is equipped with a control sound source 3 which reduces a noise at an evaluation point by generating a control sound interferring with the noise, a means which detects the residual noise at a specific position after the interference, a means which detects the reference signal regarding the noise generation state of a noise source, and a control means 2 which outputs a signal for driving the control sound source 3 by using the steepest fall algorithm according to the output signal of a residual noise detecting means 5 and the output reference signal of a noise generation state detecting means. Further, the controller A is equipped with a means which extracts a signal corresponding to the reference signal of the noise generation state detecting means from the output signal of the residual noise detecting means 5 and outputs it to a control means 7. Furthermore, the controller is equipped with a means which corrects the convergence term of the steepest fall algorithm corresponding to the phase shift from the reference signal.

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 for actively reducing noise in automobile cabins, aircraft cabins, etc.

0002

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

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

Here, three loudspeakers 103a, 103b, 103c and four microphones 105a, 105b, 105c, 105d are provided in the closed space 101, but for the sake of simplicity, each 10
It is assumed that one each of 3a and 105a is provided. Now, let us assume that the transfer function from the noise source to the microphone 105a is H, and from the loudspeaker 103a to the microphone 10
When the transfer function up to 5a is C and the sound source information signal generated by the noise source is Xp, the noise signal E as the residual noise observed by the microphone 105a is E=Xp・
It becomes H+Xp・G・C. Here, G is a transfer function necessary for silencing. When the noise is completely canceled at the point to be muted (the position of the microphone 105a), E=0. At this time, G becomes G=-H/C. Typically, this calculation is performed in the frequency domain using fast Fourier transform, and the result is inversely Fourier transformed to obtain an impulse response, and the signal processor 107
set as a filter coefficient. This filter coefficient is determined by determining G that minimizes the microphone detection signal E, and based on this G, the filter coefficient in the signal processor 107 is adaptively updated. As a method for finding filter coefficients to minimize the microphone detection signal E, the LMS algorithm (Least Mean
Square).

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

[0006] The LMS algorithm will now be explained in more detail. Microphone 105a on the lth side (
el (n) is the noise signal detected by the l-th microphone 105a (105b...) when there is no control sound from the loudspeakers 103a, 103b, 103c.
b,...) is the noise signal detected by epl(n), and the transfer function (FIR(
(finite impulse response) function) jth (j = 0, 1, 2
..., Ic -1) as a digital filter, the reference signal, that is, the sound source information signal ...) of the i-th adaptive filter that drives the i-th (i=0, 1, 2, 1...)
, IK -1), and let Wmi be the coefficient of

[0007]

[Math 1]

##EQU1## holds true.

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

0010

[Math 2]

[0011]

[0012] In order to find the filter coefficient Wm that minimizes the evaluation function Je, the LMS algorithm is adopted. In other words, the evaluation function Je is
The filter coefficient Wmi is updated with a value obtained by partial differentiation with respect to .

[0013] Therefore, from equation (2),

[0014]

[Math 3]

From equation (1),

[0016]

[Math 4]

Therefore, the right side of this equation (4) is rlm
(ni), the filter coefficient rewriting formula can be obtained by the LMS algorithm of formula (5) below.

[0018]

[Math 5]

As is clear from this format, the stability and convergence of this algorithm are

[0020]

[Math 6]

It is governed by the eigenvalue spread and the convergence coefficient α of .

[0022]

[Problems to be Solved by the Invention] By the way, in the above control, the microphone 105a (105b...)
The noise signal e1 (n) detected by the signal processor 10
The reference signal xp representing the noise source signal is directly input to the digital filter based on the LMS algorithm of No. 7, and the reference signal xp representing the noise source signal is directly input to the digital filter and the other digital filter, and the two signals e1
By rewriting the coefficients of the adaptive digital filter from (n) and xp, for example, a speaker output with an opposite phase corresponding to each frequency is generated at the microphone position.

On the other hand, when the noise source emits a strong periodic signal such as an automobile engine, the time history data of vehicle interior noise is as shown in FIG. 6(a), for example. When this is frequency-analyzed, it can be decomposed into a reference signal of the fundamental frequency and a sine wave signal with constant phase shifts τ1 and τ2, as shown in FIG. 6(b). These constant phase relationships τ1
, τ2 is included in the noise signal e1 (n) detected by the microphone 105a (105b...) in the vehicle interior as a harmonic of the fundamental frequency. That is, this harmonic component is included in el (n) of the above equation (5). In this case, if the phase shift of the harmonics is not constant with respect to the reference signal, α・Σel in equation (5)
(n) ・rlm(n-1) is gradually added or subtracted in the LMS algorithm to be averaged and canceled, but the phase shift τ1, τ2
is constant, they are added one after another, and the convergence coefficient α
gradually becomes incompatible with each other, and the convergence characteristics of the LMS algorithm become extremely slow.If the conditions worsen further, the sound pressure at the microphone position may increase, leading to a so-called divergence state.

Therefore, the present invention provides an active noise control device that can perform accurate noise control even when residual noise contains harmonic components with a constant phase shift with respect to a reference signal. purpose.

[0025]

[Means for Solving the Problems] In order to solve the above problems, the invention of claim 1 provides a control sound source that generates a control sound that interferes with noise to reduce noise at an evaluation point, and a control sound source that generates a control sound that interferes with noise to reduce noise at an evaluation point, and means for detecting residual noise at a position; means for detecting a reference signal regarding the noise generation state of the noise source; and a steepest descent algorithm based on the output signal of the residual noise detection means and the output reference signal of the noise generation state detection means. and a control means for outputting a signal for driving the control sound source using the control means, the active noise control device comprising: a control means for outputting a signal for driving the control sound source, the active noise control device comprising: a signal according to a reference signal of the noise generation state detection means from an output signal of the residual noise detection means; The present invention is characterized by comprising means for extracting and outputting the extracted information to the control means.

[0026] The invention according to claim 2 also includes: a control sound source that generates a control sound that interferes with noise to reduce noise at an evaluation point; and means for detecting residual noise at a predetermined position after the interference;
means for detecting a reference signal regarding the noise generation state of the noise source; and a signal for driving the control sound source using a steepest descent algorithm based on the output signal of the residual noise detection means and the output reference signal of the noise generation state detection means. and a control means for outputting the steepest drop according to a phase shift of a signal other than a signal corresponding to the reference signal among the output signals of the residual noise detection means with respect to the reference signal. It is characterized by having means for correcting the convergence term of the algorithm.

[0027]

In the first aspect of the invention, the control means outputs a signal for driving the control sound source using a steepest descent algorithm based on the output signal of the residual noise detection means and the output reference signal of the noise generation state detection means. Thereby, the control sound source can generate a control sound that interferes with the noise, thereby reducing the noise at the evaluation point. In this case, the signal extraction means extracts a signal corresponding to the reference signal of the noise generation state detection means from the output signal of the residual noise detection means and outputs it to the control means, so that noise control can be performed appropriately.

Further, in the invention of claim 2, in the noise control, the correction means performs the steepest descent algorithm according to the phase shift of the signal other than the reference signal among the output signals of the residual noise detection means with respect to the reference signal. Since the convergence term is corrected, the convergence characteristic can be maintained and appropriate noise control can be performed.

[0029]

[Embodiments] Examples of the present invention will be described below. Note that the explanation will be given using a vehicle interior space as an example.

FIG. 1 shows a block diagram of an active noise control device according to an embodiment of the present invention, in which a loudspeaker 3 as a control sound source and a residual noise detecting means are installed in a vehicle interior 1, which is a closed space. A microphone 5 is provided, and each of the microphones is connected to a controller 7 as a control means. The controller 7 has a digital filter 9 and an adaptive digital filter 11, the microphone 5 is connected to the digital filter 9 via an analog low-pass filter 13, and the loudspeaker 3 is directly connected to the adaptive digital filter 11. This is what is being done.

[0031] The noise inside the vehicle compartment 1 is caused by, for example, the power plant 1.
5 is a noise source, and this power plant 15 integrally houses an engine, a transmission as a power transmission device, and a differential gear. As the noise generating state detection means, for example, a crank angle signal sensor 17 is used, and the detected crank angle signal is output to the digital filter 9 and the adaptive digital filter 11. Note that the noise generation state detection means only needs to be able to detect a reference signal related to the noise generation state of the noise source, and the signal may be, for example, an output signal of a vibration sensor provided on the outer surface of the engine,
It is also possible to use an ignition pulse signal of the engine, a rotation speed signal obtained by detecting the rotation speed of the crankshaft using a rotation speed sensor, or the like.

On the other hand, the analog low-pass filter 13 constitutes signal extraction means in this embodiment,
A signal corresponding to the reference signal of the noise generation state detection means is extracted from the output signal of the residual noise detection means. Specifically, harmonic components are cut from the output signal of the microphone 5, and a signal corresponding to the reference signal of the crank angle signal sensor 17 is extracted and output to the digital filter 9.

[0033] Here, the harmonic component is, for example, in the case of engine noise, assuming that the maximum rotational speed is 6000 rpm, the second-order component is 200 Hz, so the system is configured to cut off any harmonic components higher than this.

Next, the operation will be explained.

The device shown in FIG. 1 basically operates in the same manner as the device shown in FIG. 9 is input. A reference signal xp representative of the noise source signal is input from the crank angle signal sensor 15 to the digital filter 9 and the adaptive digital filter 11. The digital filter 9 uses the reference signal xp and the noise signal e(n) to calculate α・e(n)・r(n−i) based on the LMS algorithm.
...Perform the calculation (7). At this time, the noise signal e(n) is analog low-pass filter 1
3, harmonic components of, for example, 200Hz or higher are cut, and the reference signal x is sent to the digital filter 9.
Only the primary signal corresponding to p is input. For this reason,
The LMS algorithm reliably converges, and the filter coefficient Wi of the adaptive digital filter 11 becomes Wi(n+1
)=Wi(n)+
α・e(n)・γ(n−i)……(8) is rewritten.

Therefore, the adaptive digital filter 11 receives the reference signal xp input from the crank angle signal sensor 17.
is used to generate a driving signal for the loudspeaker 3, and the loudspeaker 3 outputs a speaker output in the opposite phase to the noise signal. Therefore, the noise signal e detected by the microphone 5
(n), there is a phase shift τ1 with respect to the reference signal xp,
Even if a constant harmonic of τ2 is included, its influence can be removed and accurate noise control can be performed. Furthermore, since this embodiment uses the analog low-pass filter 13, phase delay can be substantially ignored, resulting in an extremely simple structure.

FIG. 2 shows another embodiment.

In this embodiment, a digital filter 19 is used as the signal extraction means. Since the digital filter 19 normally has a phase lag, the reference signal xp to the adaptive digital filter 11 is
The configuration is such that the input is also performed via the digital filter 21. Therefore, in this embodiment as well, harmonic components can be removed from the noise signal e(n), and substantially the same effects as in the above embodiment can be achieved.

Although the digital filters 19 and 21 have the same amplitude and phase characteristics, the digital filter 21 may have the same phase characteristics as long as it has a delay so as not to be affected by the phase delay of the digital filter 19. It is fine as long as it does.

Furthermore, in the example of FIG. 2, the digital filter 19 may be a bandpass filter or a tracking filter. In this case, the other filter 21
In this case, a filter having at least the same phase characteristics as one filter 19, which is a band pass filter or a tracking filter, is used.

The operation in this case will be explained using FIG. 3. FIG. 3(a) shows a frequency analysis of a noise signal detected by a microphone, and FIG. 3(b) shows the band of a bandpass filter or tracking filter. In contrast to the solid line in (a), the broken line shows the case where the engine rotational speed changes. Then, the band-pass filter is used to extract, for example, a second-order component of the engine and a first-order component of the gear as the order components that cause the noise inside the passenger compartment 1. Also, due to changes in engine speed, the engine secondary component or the gear primary component changes as shown in (a)
Therefore, it is also possible to use a tracking filter to change the band of the filter from the solid line position to the broken line position in (b) to track the problematic order component. In this way, high-level frequencies that are problematic can be reliably reduced.

FIG. 4 corresponds to the invention recited in claim 2, and is a block diagram in which the analog low-pass filter 13 is removed from the block diagram of FIG.
This also serves as means for correcting the convergence term of the steepest descent algorithm in accordance with the phase shift of the harmonics with respect to the reference signal. That is, as shown in FIG. 4(a), the harmonic phase shift is actually measured in advance as a transfer function of a digital filter. This is converted into an impulse response h as shown in (b). Then, the filter coefficient C when the transfer function between the loudspeaker 3 and the microphone 5 shown in equations (4) and (5) above is expressed by a digital filter is convolved with the impulse response h of the actually measured phase difference. By creating Cnew = h*C, the above (7) is applied according to the phase shift of the harmonics.
Correct the convergence term in the equation, that is, use Cnew to
(ni) can be calculated and accurate noise control can be performed. Therefore, this embodiment has the advantage that there is no increase in actual components.

Furthermore, in the embodiment shown in FIG. 2, the digital filters 19 and 21 are unified and used, and the signal from the microphone 5 and the signal from the crank angle signal sensor 17 are transmitted through one filter having the same transfer characteristic. It is also possible to construct such a configuration that the signals are respectively input to the digital filter 9 and the adaptive digital filter 11. This embodiment enables more advanced digital control in conjunction with the embodiment using a bandpass filter, a tracking filter, and the embodiment in which calculation is performed to obtain Cnew. In this case, calculation can be easily performed by delaying the calculation time by one step using one calculation element.

To simplify the explanation, the explanation has been made assuming that one microphone and one loudspeaker are each provided, but it is also possible to control a plurality of each. Further, even if the evaluation point for noise reduction and the microphone position are spatially separated, the residual noise at the evaluation point can be estimated and controlled based on a predetermined ratio.

[0045]

[Effects of the Invention] As is clear from the above, according to the configuration of the present invention, even when residual noise contains a harmonic component with a constant phase shift with respect to the reference signal, the influence of this harmonic component is removed. This makes it possible to perform accurate noise control.

[Brief explanation of the drawing]

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

FIG. 2 is a block diagram according to another embodiment.

FIG. 3 is an explanatory diagram of an operation according to still another embodiment.

FIG. 4 is an explanatory diagram of an operation according to still another embodiment.

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

FIG. 6 is an explanatory diagram of residual noise.

[Explanation of symbols]

3 Loudspeaker (control sound source) 5 Microphone (residual noise detection means) 7 Controller (control means)

Claims (2)

[Claims] 1. A control sound source that generates a control sound that interferes with noise to reduce noise at an evaluation point, means for detecting residual noise at a predetermined position after the interference, and a reference signal regarding the noise generation state of the noise source. and a control means for outputting a signal for driving the control sound source using a steepest descent algorithm based on the output signal of the residual noise detection means and the output reference signal of the noise generation state detection means. An active noise control device characterized by comprising means for extracting a signal corresponding to a reference signal of the noise generation state detection means from the output signal of the residual noise detection means and outputting it to the control means. type noise control device. 2. A control sound source that generates a control sound that interferes with noise to reduce noise at an evaluation point, means for detecting residual noise at a predetermined position after the interference, and a reference signal regarding the noise generation state of the noise source. and a control means for outputting a signal for driving the control sound source using a steepest descent algorithm based on the output signal of the residual noise detection means and the output reference signal of the noise generation state detection means. type noise control device, comprising means for correcting a convergence term of the steepest descent algorithm according to a phase shift of a signal other than a signal corresponding to the reference signal from the reference signal among the output signals of the residual noise detection means. An active noise control device characterized by: