KR101329180B1 - Multi-channel Active Sounddproof Wall and Active Noise Control Method Thereof - Google Patents

Abstract

The present invention relates to a multi-channel active sound barrier and an active noise control method thereof, the multi-channel active sound barrier according to the present invention is a sound barrier body; At least one speaker, at least one noise measuring means, and at least one error measuring means, which are formed at an upper end of the sound barrier wall main body and have an inner soundproof part and an outer soundproof part and have a Y-shaped soundproof wall upper part, forming a multi-channel. The soundproof wall upper end is installed; And an active noise controller for controlling the multi-channel such that control sound sources for attenuating the noises are radiated through the at least one speaker based on the noises input from the at least one noise measuring means and the at least one error measuring means. It includes.
Therefore, the sound insulation performance can be improved by applying the multi-channel active noise control technique to the Y-shaped structure having high sound insulation effect.

Description

Multi-channel Active Sounddproof Wall and Active Noise Control Method Thereof}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multi-channel active noise barrier and an active noise control method thereof, and more particularly, to a sound barrier and a method of controlling the active noise thereof using a multi-channel active noise control technique.

The research on the conventional sound barrier has focused on the material and structure of the sound barrier for passively removing noise. For example, studies have been made mainly to make a soundproof wall made of a material having high sound absorption, a structure in which sound is well reflected on the soundproof wall, and the shape of grooves of the soundproof wall.

On the other hand, among the methods for removing noise is a method using the Active Noise Control (ACN). This technology removes noise due to the overlapping effect between two wavelengths by artificially radiating a control sound source having a phase opposite to that of the noise source, and is mainly applied to products such as automobiles, airplanes, or earphones.

Recently, there is an example in which such an active noise control system is applied to a soundproof wall. However, in most cases, the active noise control technology of a single channel is merely implemented, but multi-channel active noise control is required because there are various noise sources and noise paths in a wide space or outdoor environment. In addition, there is a secondary path in the active noise control system that shows the transmission characteristics from the speaker to the microphone, and the identification of the secondary path has a great influence on the stability of the active noise control system. In general, secondary paths have time-varying characteristics that change with time. Therefore, when the identification of the secondary path is incomplete, there is a problem that the performance of the conventional active noise control system is lowered.

Therefore, an object of the present invention is to increase the sound insulation performance by applying a multi-channel active noise control technology to the Y-shaped top structure having a high sound insulation effect. In addition, it is to provide a highly stable active sound barrier and its active noise control method by estimating the secondary path in real time. In addition, it is to provide an active sound barrier and its active noise control method with improved performance by reducing the residual error signal through power scheduling.

The object is a multi-channel active sound barrier, the sound barrier body; At least one speaker, at least one noise measuring means, and at least one error measuring means, which are formed at an upper end of the sound barrier wall main body and have an inner soundproof part and an outer soundproof part and have a Y-shaped soundproof wall upper part, forming a multi-channel. The soundproof wall upper end is installed; And an active noise controller for controlling the multi-channel such that control sound sources for attenuating the noises are radiated through the at least one speaker based on the noises input from the at least one noise measuring means and the at least one error measuring means. It can be achieved by a multi-channel active sound barrier comprising a.

Here, the at least one speaker is arranged on the inner surface of the inner sound insulation; The at least one noise measuring means is installed on an outer surface of the inner sound insulation part; The at least one error measuring means may be installed on the inner surface of the outer sound insulation.

In addition, the at least one noise measuring means may be installed to adjust the installation angle of the microphone including a support member removably configured in the receiving portion provided in the inner soundproofing portion and a microphone installed on the support member.

The soundproof wall upper end may be inserted into the soundproof wall main body.

In addition, the multi-channel active soundproof wall may further include a sealing plate provided on both sides of the upper end of the soundproof wall.

The active noise controller may include a secondary path estimator that estimates a secondary path from the at least one speaker to the at least one error measuring means in real time.

The secondary path estimator may estimate the secondary path in real time using a variable step-size normalized least mean square (VSS-NLMS) technique.

Here, the secondary path estimator may include an additional noise power control unit for calculating the power of the additional noise for the secondary path estimation.

On the other hand, the object of the multi-channel active soundproof wall in accordance with the present invention, the soundproof wall body is provided with at least one error measuring means; An auxiliary soundproof wall protruding from one side of the soundproof wall main body, and having at least one speaker and at least one noise measuring means installed together with the at least one error measuring means to form a multichannel; And an active noise controller for controlling the multi-channel such that control sound sources for attenuating the noises are radiated through the at least one speaker based on the noises input from the at least one noise measuring means and the at least one error measuring means. It can also be achieved by a multi-channel active sound barrier comprising a.

The multi-channel active sound barrier may further include a sealing plate installed at both side ends of the auxiliary sound barrier and the upper end of the sound barrier body.

On the other hand, the object of the active noise control method of the active soundproof wall provided with at least one noise measuring means, at least one speaker, and at least one error measuring means for forming a multi-channel according to the present invention, the at least one Receiving a noise signal from the noise measuring means and converting the noise signal into a control sound source through an adaptive filter; Generating an additional random noise signal to estimate in real time a secondary path from said at least one speaker to said at least one error measuring means; Calculating the filtered noise signal by passing the noise signal from the at least one noise measuring means through the estimated secondary path; And updating the control sound source based on the calculated filtered noise signal and the error signal from the error measuring means.

In addition, the secondary path may be updated by a variable step-size normalized least mean square (VSS-NLMS) technique.

The method of controlling active noise of the active sound barrier may further include adjusting power of the additional random noise signal according to whether the secondary path is converged.

As described above, according to the present invention, the sound insulation performance can be improved by applying the multi-channel active noise control technique to the Y-shaped structure having a high sound insulation effect. In addition, an active noise barrier having high stability and an active noise control method thereof are provided by estimating a secondary path in real time. In addition, an active noise barrier having improved performance by reducing residual error signals through power scheduling and an active noise control method thereof are provided.

1 is a schematic diagram of a multi-channel active sound barrier according to a first embodiment of the present invention,
2 is a control block diagram of a multi-channel active sound barrier according to the present embodiment,
3 is a view from above of FIG.
FIG. 4 illustrates an example of online secondary path modeling of a 1 × 2 × 2 multichannel of a multichannel active sound barrier according to an embodiment of the present invention.
5 is a schematic diagram of a multi-channel active sound barrier according to a second embodiment of the present invention,
6 and 7 show modeling errors of S ₁₁ (z) and S ₂₁ (z), respectively, and are graphs comparing the present method with the conventional method.
8 is a graph showing the mean square error of the conventional method and the active noise control system according to the present invention.

Hereinafter, specific embodiments of the present invention will be described with reference to the drawings.

1 is a schematic diagram of a multi-channel active sound barrier according to a first embodiment of the present invention, and FIG. 2 is a control block diagram of a multi-channel active sound barrier according to the present embodiment.

Referring to FIG. 1, the multi-channel active sound barrier according to the present embodiment includes a sound barrier wall body 100 and a sound barrier wall upper end 200. The soundproof wall body 100 and the soundproof wall upper end 200 may be fitted to each other.

The active sound barrier according to the present invention is a multi-channel active noise system is applied, the upper end of the sound barrier 200, the inner soundproofing portion 210 and the outer soundproofing portion 220 is formed to have a Y-shaped overall, forming a multi-channel A plurality of noise measuring means 10, a plurality of error measuring means 20, a plurality of speakers 30, and an active noise control unit 40 is provided.

There are I noise measuring means 10, K error measuring means 20, J is a speaker 30. Therefore, since IxJxK only needs to form a multichannel, it is not necessary that each of them be a plurality. Since the active sound barrier according to the present invention forms a multi-channel, the noise source and the noise path are effective in various outdoor environments.

The noise measuring means 10 is for measuring the noise generated from the noise source, and is preferably installed close to the sound source. In the present embodiment, the noise measuring means 10 includes a support member 10a configured to be insertable and detachable from an accommodating portion provided on an outer surface side of the inner sound insulation part 210, and a microphone 10b installed on the support member 10a. It is composed. At this time, since the support member 10a is rotatably installed, the installation angle of the microphone 10b can be adjusted. This is to achieve the best performance according to the noise environment by adjusting the angle of the microphone 10b because the height of each soundproof wall, distance and angle with the noise source is different.

The error measuring means 20 measures residual noise at a target point to remove noise, and measures the residual noise after the emission of the control sound source and feeds it back to the active noise controller 40 to be described later. In FIG. 1, the error measuring means 20 is provided on an inner surface of the outer sound insulation part 220 of the soundproof wall upper end 200, and is installed at a position to measure noise and residual noise after offsetting of the control sound source radiated from the speaker 30. .

The active noise control unit 40 includes a plurality of speaker 30 having control sound sources for attenuating noises based on the noises input from the plurality of noise measuring means 10 and the plurality of error measuring means 20 forming a multi-channel. Control to radiate through.

The active noise control unit 40 is a noise processing module for processing signals input from the I noise measuring means 10 and the K error measuring means 20, and outputs the signals of the noise source to the control sound sources by reversing the phase. The ANC controller may include a control sound source processing module for processing the control sound sources and outputting the plurality of speakers 30.

Specifically, the noise processing modules may include an amplifier for amplifying input noise, a low pass filter for filtering the amplified signal, and an A / D converter for converting the filtered signal into a digital signal, and the control sound source processing module may be digitally controlled. A D / A converter for converting a sound source into an analog, a low pass filter for filtering the converted signal, and an amplifier for amplifying the filtered signal and outputting it to the speaker 30. The ANC controller inverts the phase of the signal input from the noise measuring means 10 and outputs it to the control sound source, and receives residual noise inputted from the error measuring means 20 and checks and corrects an error such as a time delay. The level of control is controlled to be below the reference value. Here, the ANC controller includes an adaptive filter that converts the signal received from the noise measuring means 10 into a control sound source and outputs it through the speaker 30.

These voice processing modules and the controller is preferably composed of a single board is built in the soundproof wall upper end 200. However, a connection terminal may be provided to allow the active noise control unit 40 to be accessed wirelessly or by wire from the outside, and in some cases, may be configured in an external form.

On the other hand, the multi-channel active soundproof wall according to the present embodiment may further include a sealing plate 300 which is installed at both ends of the soundproof wall upper end 200.

3 is a view of FIG. 1 seen from above. 3 shows that the sealing plate 300 is installed at one end of the soundproof wall upper end 200, but the other sealing plate 300 is omitted to clearly show another configuration, and both sealing plates ( 300) is installed.

In the multi-channel active sound barrier according to the present invention, the multi-channel is applied, but if there are too many channels, the control system is complicated, so that a certain number of channels are set, and a sealing plate 300 is installed between each of the sound barrier walls, thereby providing the Prevent interference

On the other hand, the active noise control unit 40 according to the present embodiment includes a secondary path estimator for estimating the secondary path from the plurality of speakers 30 to the plurality of error measuring means 20 in real time. In the active noise control system, there is a secondary path representing the characteristics from the speaker 30 to the microphone, and generally, a FxLMS (Filtered-x Least Mean Square) algorithm is widely used to compensate for the influence of the secondary path. However, in case of FxLMS, the performance decreases when the secondary path is changed. In the present invention, the secondary path is estimated online to minimize the influence of the time-varying characteristics of the secondary path.

4 illustrates an example of online secondary path modeling of 1 × 2 × 2 multichannels of a multichannel active sound barrier according to an exemplary embodiment of the present invention. In FIG. 4, the secondary path estimation unit includes a secondary path estimation filter, and the VSS NLMSs 41a, 41b, 41c, and 41d are applied to the estimation filter.

Specifically, x _i (n) is a signal input to the noise measuring means 10, the output y _j (n) is generated through the J speakers (30). The output y _j (n) through which the input x _i (n) has passed through the adaptive filter w _ji (n) can be expressed as follows. Here, the adaptive filter w _ji (n) represents a path between the i-th noise measuring means 10 and the j-th speaker 30. The adaptive filter w _ji (n) converts a signal received from the noise measuring means 10 into a control sound source and converts the speaker ( Means the adaptive filter to output through 30).

Where * denotes a convolution operation, and the adaptive filter residual error e _k (n) and the second-order path modeling error f _k (n) can be expressed as follows.

In Equation 2, e _k (n) represents the signal measured by the error measuring means 20 by a formula, d _k (n) is the noise transmitted from the noise measuring means 10 to the error measuring means 20 It means a signal. y ' _kj (n) refers to a signal that appears after the output y of the adaptive filter passes through the second path existing from the j-th adaptive filter output to the k-th error measuring means 20, and v' _kj (n) It refers to a signal that appears after an artificially injected random signal for estimating the secondary path passes through the secondary path between the j-th adaptive filter and the k-th error measuring means 20. In this case, when the sum of y ' _kj (n) coincides with a signal of d _k (n) and v' _kj (n) becomes 0, the error signal becomes 0, which is the ideal noise reduction.

을 빼서 얻은 2차경로 추정필터 오차 신호를 의미한다. 이값은 2차경로 추정필터

를 추정하는 VSS-NLMS식에서 사용된다.In equation (3), f _k (n) is a signal obtained by passing a random signal from the error signal measured by the k-th error measuring means 20 through the second-order path estimation filter.

It means the second-order estimation filter error signal obtained by subtracting. This value is the second-order path estimation filter

It is used in the VSS-NLMS equation to estimate.

The active controller updates the adaptive filter w _ji (n) using the measured residual error and the multiple error FxLMS (MeFxLMS) algorithm, and the update equation can be expressed as follows.

In Equation 4, w _ji (n + 1) denotes the MeFxLMS update equation of the adaptive filter w representing the path between the i-th noise measuring means 10 and the j-th speaker 30, and μ _w denotes the adaptive filter w. Constant value indicating the step-size of the LMS update for.

는 i번째 소음측정수단(10)으로부터 얻어진 신호를 추정한 2차경로 추정필터

을 통과시켜 얻은 filtered-x 신호를 의미하며, 적응필터 w의 MeFxLMS 업데이트 식에서 사용된다. 여기서,

는 2차경로의 임펄스 응답

의 추정값이다.In Equation (5)

Is a second-order path estimation filter for estimating a signal obtained from the i-th noise measuring means 10

It means the filtered-x signal obtained by passing through and used in the MeFxLMS update equation of the adaptive filter w. here,

Impulse response of the second path

Is an estimate of.

를 업데이트 하는 VSS-NLMS 식에서 사용되는 variable step-size 값을 구하는 식은 다음과 같다.As described above, the secondary path estimation unit of the active control unit according to the present invention uses the VSS-NLMS algorithm. That is, the quadratic path modeling filter coefficients in FIG. 4 are updated using the VSS-NLMS algorithm. Specifically, secondary path

The formula to obtain the variable step-size value used in the VSS-NLMS expression to update the equation is as follows.

In Equation 6, μ _max is the maximum value of step-size, p is a power factor for obtaining a variable step-size value, and C is a constant value of 1 / SNR. In Equation 7, α has a smoothing factor of 0.9 <α <1.

And, the second-path estimate can be estimated as follows:

In Equation 8, sigma is a regularization factor, which means a small constant value, and is intended to prevent the expression from being diverted by denominator being zero when norm of vj (n) becomes zero.

As described above, the multi-channel active sound barrier according to the present invention minimizes the influence of time-varying characteristics of the secondary path by applying the VSS-NLMS multi-channel online secondary path estimation.

On the other hand, the active noise control unit 40 according to the present embodiment, as shown in Figure 4, includes an additional noise power control unit for calculating the power of the additional noise for secondary path estimation. In Figure 4, the additional noise power control unit is represented by POWER SCHEDULING (43a, 43b).

Since the random signal v _k (n) is an artificially injected signal for estimating the secondary path, after convergence of the secondary path, the random signal v _k (n) is injected into the error measuring means 20 without affecting the secondary path estimation. Since the signal is generated, an additional noise power control unit is applied to reduce the power of the random signal v _k (n) as the secondary path converges.

In order to calculate the power of noise for the second path estimation, the power ratio ρ _k (n) of the adaptive filter error and the second path modeling error can be calculated by the following equation.

ρ _k (n) represents the power ratio of the k-th secondary path modeling error signal and the k-th adaptive filter error signal, which is necessary for power scheduling. At the beginning of the active noise control operation, ρ _k (n) has a value close to 1 since the numerator and the denominator have almost similar values, and when the active noise control converges, the f _k (n) signal is zero. Since ρ _k (n) converges to 0, it has a small value close to zero.

That is, ρ _k (n) has a value close to 1 near n = 0 and approaches 0 as the active noise control system operates. Using this characteristic, the power of noise applied to the second-order path estimation filter can be adjusted as follows.

는 주입하는 랜덤신호가 가질 수 있는 최소 분산 값으로, zero-mean을 가정하므로 랜덤신호의 최소 파워값과 동일하다.

는 주입하는 랜덤신호가 가질 수 있는 최대 분산 값으로, zero-mean을 가정하므로 랜덤신호의 최대 파워값과 동일하다. 이 두 개의 값은, 실험적으로 결정된 적절한 값으로 선택된다. 여기서

은 단위 분산값과 평균이 0인 백색 잡음이다. here,

Is the minimum variance value of the random signal to be injected. Since zero-mean is assumed, it is equal to the minimum power value of the random signal.

Is the maximum variance value of the random signal to be injected. Since zero-mean is assumed, it is equal to the maximum power value of the random signal. These two values are chosen to be appropriate values determined experimentally. here

Is white noise with unit variance and mean zero.

Since the value described in Equation 9 is in the form of a convex combination, the power of the random signal has a minimum value starting from the maximum value at the beginning of the active noise control operation and proceeding toward the convergence direction as the active noise control system operates. That is, at n = 0, σ ² _vk is close to σ ² _vmax , so fast estimation is possible.If the second path is estimated almost as the active noise control system progresses, the residual error can be minimized as it approaches σ ² _vmin . have.

5 is a schematic diagram of a multi-channel active sound barrier according to a second embodiment of the present invention. Referring to FIG. 4, the multi-channel active sound barrier according to the present embodiment includes a sound barrier wall body 100 and an auxiliary sound barrier wall 400 installed in the sound barrier wall body 100.

The auxiliary soundproof wall 400 protrudes on one side of the soundproof wall main body 100, and includes a plurality of speakers 30 and at least one noise measuring means 10 that form a multi-channel with a plurality of error measuring means 20. This is installed. A plurality of error measuring means 20 is installed in the soundproof wall body 100.

The two embodiments are common in that they have a Y-shape as a whole only with a difference in their upper structures. Since other configurations are the same, the detailed description thereof is replaced with the description of the above-described embodiment.

Experiment result

In this experiment, a multi-channel active noise control system having one noise measurement means 10, two speakers 30, and two error measurement means 20 is considered as shown in FIG. The paths P ₁ (z) and P ₂ (z) used in the experiment were 256, S ₁₁ (z), S ₁₂ (z), S ₂₁ (z), and S ₂₂ (z) each having a length of 128. Modeled with a finite impulse response (FIR) filter.

The input signal x (n) consists of the sum of 100Hz, 200Hz and 300Hz sinusoids and adds a 30dB SNR white noise signal. The adaptive filter and the second-order path modeling filter were selected as FIR filters having lengths of 192 and 128, respectively. For step-size and noise power, the following parameters were selected:

In the conventional VSS-LMS method, μ _w = 6 × 10 ^-7 , μ _smax = 7 × 10 ^-3 , μ _smin = 2 × 10 ^-3 , σ _v = 0.01,

And active noise control according to the present invention was tested with μ _w = 6 × 10 ^-7 , μ _smax = 0.01, σ _vmin = 1 × 10 ^-6 , σ _vmax = 0.01.

사이의 모델링 오차는 다음의 식으로 정의하였다.Secondary path estimated with existing second path S _kj (z)

The modeling error between was defined by the following equation.

사이의 오차를 계산하는 식으로서, 계산된 값은 dB 단위로 표현된다.In the above equation, △ S _kj (z) is the secondary path estimated with the existing secondary path S _kj (z).

As an equation for calculating the error between, the calculated value is expressed in dB.

6 and 7 show modeling errors of S ₁₁ (z) and S ₂₁ (z), respectively. As can be seen in Figures 6 and 7, it can be seen that the performance of the active noise control unit 40 according to the present invention shows faster convergence performance than the conventional method.

In addition, Figure 8 shows the mean square error of the active noise control system, the active noise control through the additional noise power control according to the present invention shows a more improved performance than the conventional method.

Although several embodiments of the present invention have been shown and described, those skilled in the art will appreciate that various modifications may be made without departing from the principles and spirit of the invention . It is intended that the scope of the invention be defined by the claims appended hereto and their equivalents.

10: noise measuring means 20: error measuring means
30: speaker 40: active noise control unit
41: second path estimation unit 43: additional noise power control unit
100: soundproof wall main body 200: soundproof wall upper end
210: inner sound insulation unit 220: outer sound insulation unit
300: sealing plate 400: auxiliary sound barrier

Claims (13)

In a multi-channel active sound barrier,
Soundproof wall body;
At least one noise measuring means, at least one speaker, and at least one error measuring means, which are formed at an upper end of the soundproof wall main body and have an inner soundproofing part and an outer soundproofing part and have a Y-shaped soundproofing wall upper part, forming a multi-channel. The soundproof wall upper end is installed; And
An active noise controller for controlling the multi-channel such that control sound sources for attenuating the noises are radiated through the at least one speaker based on the noises input from the at least one noise measuring means and the at least one error measuring means. Including;
The multi-channel active sound insulation wall further includes a sealing plate provided on both sides of the upper end of the sound insulation wall;
And the active noise controller comprises a secondary path estimator for estimating in real time the secondary path from the at least one speaker to the at least one error measuring means.
The method of claim 1,
The at least one speaker is arranged on an inner surface of the inner sound insulation unit;
The at least one noise measuring means is installed on an outer surface of the inner sound insulation part;
The at least one error measuring means is a multi-channel active soundproof wall, characterized in that installed on the inner surface of the outer soundproofing portion
3. The method of claim 2,
The at least one noise measuring means is installed so as to adjust the installation angle of the microphone, including a support member and a microphone installed on the support member that can be inserted into the receiving portion provided in the inner soundproofing portion Channel active sound barrier.
The method of claim 1,
Multi-channel active soundproof wall, characterized in that the upper end of the soundproof wall is inserted into the soundproof wall body.
delete delete The method of claim 1,
And the second path estimator estimates the second path in real time through a variable step-size normalized least mean square (VSS-NLMS) technique.
The method of claim 7, wherein
And the second path estimator includes an additional noise power controller for calculating power of additional noise for the second path estimation.
In a multi-channel active sound barrier,
A soundproof wall body having at least one error measuring means installed therein;
An auxiliary soundproof wall protruding from one side of the soundproof wall main body, and having at least one speaker and at least one noise measuring means installed together with the at least one error measuring means to form a multichannel;
An active noise controller for controlling the multi-channel such that control sound sources for attenuating the noises are radiated through the at least one speaker based on the noises input from the at least one noise measuring means and the at least one error measuring means; And
A sealing plate installed at both side ends of the auxiliary soundproof wall and the soundproof wall body;
And the active noise controller comprises a secondary path estimator for estimating in real time the secondary path from the at least one speaker to the at least one error measuring means. delete delete delete delete

Images