JPH07225585A - Noise controller - Google Patents

Abstract

PURPOSE:To provide a noise controller preventing an abnormal sound when the operation of the noise controller is stopped without giving a load to a process of a DSP(digital signal processor) and without enlarging a device scale. CONSTITUTION:When the operation of the noise controller is stopped, a renewal of a filter coefficient When) (i is a filter coefficient number) deciding a characteristic of an ADF(adaptive digital filter) 106 in the DSP 9 is performed by setting a value of a coefficient (gamma) in the following relation Wi(n+1)=gammaWi(n)+2mue(n)X(n-i) within the range of 0<gamma<1. Where, (n) is a sample number in a system, x(n) is a noise signal, e(n) is a synthesis sound wave signal obtained by detecting a synthesis sound between a noise N and a muffle interference sound S outputted from a speaker 102 for canceling the noise N, (mu) is a convergence coefficient and (gamma) is a control integer.

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,
More specifically, the present invention relates to a noise control device that eliminates or suppresses noise such as in-vehicle noise that occurs during operation of an automobile or the like, by means of a silencing interference sound from another separately provided sound source.

[0002]

2. Description of the Related Art Generally, various complex noises such as engine noises, gear noises, wind noises or running noises (road noises) are present inside a vehicle such as an automobile. Therefore,
In order to obtain a quiet and comfortable interior space, mechanical and physical noise countermeasures such as structural modification of the vehicle body and the use of soundproof materials are adopted.

Apart from such noise countermeasures, attention has been paid to the generated noise itself, and a technique for directly canceling the generated noise itself has been studied, and an active noise control device is known as such a technique.

FIG. 6 shows a conventional active noise control (hereinafter referred to as AN
It is called C (Active Noise Control). ) The principle of the device is shown. As shown in FIG. 6A, the ANC device includes a microphone A100 that collects noise N output from a primary sound source (for example, an engine of an automobile) and an amplifier that amplifies the noise N collected by the microphone A100. 101 and microphone A10
And a speaker 102 as a secondary sound source which is arranged at a distance of 0 half of the wavelength of the noise N and outputs the noise N amplified by the amplifier 101 as the silencing interference sound S. .

Next, the operation will be described with reference to FIGS. 6 (a) and 6 (b). In the principle shown in FIG. 6, it is assumed that the noise N has a single frequency component.

Now, when the noise N is collected by the microphone A100, the collected noise N is amplified by the amplifier 101 by the amount attenuated by the collected sound in the microphone A100, and becomes a noise-interference sound S for canceling the noise. It is output from the speaker 102.

Since the microphone A 100 and the speaker 102 are arranged apart from each other by a half wavelength of the wavelength of the noise N as described above, the silencing interference sound S output from the speaker 102 is shown in FIG. As shown in b) and b, the noise N has a waveform whose phase is opposite to that of the noise N.

Therefore, as shown in FIG. 6 (b), the noise N
Is at the position of the speaker 102 and is canceled by destructive interference due to the silencing interference sound S output from the speaker 102. The above is the principle of the ANC device, but in practical use, the noise N is a time-varying signal that includes various time-varying frequency components and also includes uncertain factors such as wavelength changes due to ambient temperature changes. Therefore, it is necessary to control the silencing interference sound S output from the speaker 102 according to the change of the characteristic of the noise N and the change of the surrounding environment.

This control is performed by an adaptive (adaptive) filter (hereinafter referred to as ADF) using a digital signal processor (hereinafter referred to as DSP (Digital Signal-Processor)).
(Adaptive Digital Filter). ) And an adaptive control algorithm which is a coefficient setter for optimizing the filter coefficient of the ADF.

Here, the ADF is a filter for filtering a digital input signal to obtain an output signal, and the filter coefficient that determines the filter characteristic is a predetermined optimum filter based on the control by the adaptive control algorithm. It is a filter that can obtain a desired optimum filter characteristic by being sequentially updated to have a coefficient.

The adaptive control algorithm is an algorithm (software) for automatically optimizing the filter coefficient of the ADF. The adaptive control algorithm adapts to a change in a signal such as a control signal input from the outside to filter the ADF. The coefficient is updated sequentially and optimized.

FIG. 7 shows an example of a conventional ANC device using an adaptive transversal filter (also called a FIR (Finite Impulse Response) filter) as the ADF and an LMS (Least Mean Square) algorithm as an adaptive control algorithm. Show.

In the conventional ANC device shown in FIG. 7, the signal based on the noise N collected by the microphone is converted into A
A synthesized sound wave detection signal based on a synthesized sound obtained by synthesizing the noise N and the silencing interference sound S for canceling the noise N is used as a digital input signal of the DF as a control signal for the LMS algorithm from the outside. Then, the filter characteristics of the ADF are optimized by the LMS algorithm so that a silencing interference sound signal for obtaining the silencing interference sound S that minimizes the sound pressure level of the synthetic sound wave detection signal can be obtained as an output signal.

In the following description, n is a discrete time and represents a sample number in the ANC device. In the ANC device shown in FIG. 7, the microphone A100 collects the noise N and outputs a noise sound wave detection signal.

The microphone B 103 is provided with the noise and the speaker 1.
A synthetic sound with the silencing interference sound S outputted from No. 02 is collected to output a synthetic sound wave detection signal. DSP105 is A /
Based on the noise sound wave detection signal x (n) A / D converted by the D converter 104 and the synthesized sound sound wave detection signal e (n) A / D converted by the A / D converter 108, the ADF 1
By the operation of 06 and the LMS algorithm 107, the silencing interference sound signal d (n) that minimizes the synthesized sound wave detection signal e (n) is generated.

According to the LMS algorithm 107, AD
The i-th filter coefficient at the sample number n in F106 is W _i (n) (i is the filter coefficient number),
If the i-th filter coefficient at the updated sample number n + 1 is W _i (n + 1), then W _i (n) and W _i
The relationship with (n + 1) is given by the following equation.

W _i (n + 1) = W _i (n) +2 μe
(N) x (n-i) where μ is a convergence coefficient. W _i obtained by the above equation
(N) is the LMS algorithm 107 to the ADF 106
Output to characterize the ADF 106.

The D / A converter 109 and the amplifier 110 are
The muffling interference sound signal d (n) is respectively D / A converted and amplified to generate an output signal to be output from the speaker 102 as the muffling interference sound S.

Considering the operation of an automobile, for example, engine noise generated during sudden deceleration or sudden acceleration has a sudden change in its frequency component and amplitude. Then, such a change in noise may be so rapid that it exceeds the control processing capability of the DSP 105.

That is, the time-invariant principle (the principle that the characteristics of the system constituting the system is invariable within the sampling time of the system) which is a prerequisite for the normal operation of the DSP 105 is not satisfied. The engine speed may change suddenly.

In such a case, if the conventional ANC device described above is operated as it is, the ADF1
The update of the filter coefficient in 06 is the noise sound wave detection signal x
(N) In the case of a car (for example, an ignition signal of an engine in the case of an automobile), it becomes impossible to follow the change, which makes control impossible and, conversely, causes an abnormal noise from the speaker 102.

In order to prevent this, the sampling time may be shortened so that the time-invariant principle holds even if the engine speed changes rapidly. In that case, the processing capacity of the DSP 105 It often exceeds the limit of.

Therefore, as a next measure, there is a method of stopping the operation of the ANC device while the rotation of the engine is rapidly changing. Therefore, conventionally, the operation of the ANC device is generally stopped by the method shown in FIG.

That is, in the first stopping method, the value of the convergence coefficient μ in the LMS algorithm 107 is set to 0, and AD
This is a method of stopping the output of digital data from the F106 to the D / A converter 109.

In the second stopping method, a switch S1 is provided between the D / A converter 109 and the amplifier 110 as shown in FIG.
This is a method of instantaneously shutting off the output of (n).

[0026]

The waveform of the silencing interference sound S output from the speaker 102 when the operation of the ANC device is stopped by the above two stopping means is shown in FIG.
As shown in (b), the waveform becomes discontinuous when the operation is stopped (time t ₁ ). Due to the discontinuity of the waveform, an abnormal sound (pop sound) is generated from the speaker 102, which causes a discomfort to a driver or other personnel, and goes against the original purpose of noise elimination.

As a method of preventing an abnormal sound (pop sound) due to such a waveform discontinuity, the timing when the absolute value of the amplitude of the silencing interference sound S becomes substantially zero (for example, FIG. 8).
There is a method of utilizing so-called zero cross mute, which controls to stop the operation of the ANC device at time t ₂ ) in (b).

However, this zero cross mute is applied to the DSP.
In order to realize by software processing in 105, processing as shown in FIG. 9 must be executed. That is, it is confirmed whether the operation is stopped (step S
100), if the operation is not stopped (step S100N)
o) With switch S2 turned on (step S1)
01), the following processing is executed. If the operation is stopped (Yes in step S100), the absolute value of the amplitude of the silencing interference sound signal d (n) is calculated (step S102), and it is confirmed whether the calculated absolute value of the amplitude is less than or equal to a predetermined threshold value. (Step S103). If it is equal to or more than the predetermined threshold value (No in step S103), the following process is executed with the switch S2 kept on. Below a predetermined threshold (step S
If Yes, the switch S2 is turned off (step S104) to cut off the output of the muffling interference sound signal d (n) from the ADF unit 106 and stop the operation of the ANC device.

When the above processing is attempted to be executed by the DSP 105, problems such as time delay occur. In addition, if it is attempted to be realized by hardware other than the DSP 105, components such as a comparator, a mute element, and an external latch are required. Further, since a port for outputting a control signal from the DSP 105 is also required, there arises a problem that the whole device becomes complicated and large-scaled, and a production cost becomes high.

An object of the present invention is to provide a device which prevents abnormal noise when the operation of the ANC device is stopped without imposing a load on the processing of the DSP 105 and without enlarging the scale of the device.

[0031]

In order to solve the above-mentioned problems, the invention as set forth in claim 1 detects noise sound wave N by detecting noise sound wave N directly or indirectly as shown in FIG. A noise sound wave detection unit 1 that outputs a signal x (n), and a noise canceling sound wave signal d (n) for generating a noise canceling sound wave S that attenuates the noise sound wave N based on the noise sound wave detection signal x (n). ), Output means 2 for receiving the silencing interference sound wave signal d (n) and radiating the silencing interference sound wave S into the sound field space, and a synthesized sound of the noise sound wave N and the silencing interference sound wave S. Synthetic sound wave detection means 3 for detecting sound waves and outputting a synthetic sound wave detection signal e (n), the synthetic sound wave detection signal e (n) and the noise sound wave detection signal x.
Based on (n), the filter means 4 calculates the optimum filter coefficient so as to have a filter characteristic that minimizes the sound pressure level of the synthetic sound wave, and the filter coefficient signal W
coefficient setting means 5 for outputting _i (n) to the filter means 4, and the coefficient setting means 5 is provided when the frequency component of the noise sound wave detection signal x (n) changes abruptly.
It is characterized in that it has a continuity maintaining means 6 for maintaining the continuity of the waveform of the silencing interference sound wave S.

According to a second aspect of the present invention, in the noise control device according to the first aspect, the coefficient setting means 5 is an LMS.
Constructed by the algorithm, the i-th filter coefficient when W _i to (n + 1) of the sample number n + 1 in the filter unit, the γ as a control factor, when W _i (n) of the sample number n in the filter means The i-th filter coefficient, μ is the convergence coefficient, and e (n) is the synthetic sound wave detection signal at sample number n,
Let x (n−i) be the noise detection signal at sample number n−i, and W _i (n + 1) is the following formula W _i (n + 1) = γW _i (n) +2 μe (n) x (n−
In the case represented by i), the continuity maintaining means 6 changes the control coefficient γ to a positive number less than 1 when the frequency component of the noise sound wave detection signal x (n) changes abruptly. Configured as a feature.

[0033]

According to the first aspect of the present invention, the noise sound wave detecting means 1 directly or indirectly detects the noise sound wave N and outputs the noise sound wave detection signal x (n).

The filter means 4 has a noise sound wave detection signal x.
Based on (n) and the filter coefficient signal W _i (n), a silenced interference sound wave signal d (n) for generating a silenced interference sound wave S that attenuates the noise sound wave N is generated. Output means 2
Receives the muffling interference sound wave signal d (n), and
To the sound field space.

The synthetic sound wave detecting means 3 detects the synthetic sound wave of the noise sound wave N and the silencing interference sound wave S and outputs a synthetic sound wave detection signal e (n). The coefficient setting means 5 is optimal so that the filter means 4 has a filter characteristic that minimizes the sound pressure level of the synthetic sound wave based on the noise sound wave detection signal x (n) and the synthetic sound wave detection signal e (n). The filter coefficient is calculated, and the filter coefficient signal W _i (n) is output to the filter means 4.

The continuity maintaining means 6 included in the coefficient setting means 5 calculates the optimum filter coefficient of the coefficient setting means 5 when the frequency component of the noise sound wave detection signal x (n) changes abruptly. The filter coefficient is set so that the waveform continuity of is obtained.

With the above operation, the noise sound wave detection signal x
It is possible to prevent the occurrence of abnormal sound due to the discontinuity of the waveform of the silencing interference sound wave S when the frequency component of (n) changes abruptly.

According to the invention of claim 2, claim 1
In the noise control device described in the paragraph 1, the coefficient setting means 5 is an LM.
It is composed of the S algorithm and has W _i (n + 1)
I for the sample number n + 1 in the filter means
The second filter coefficient, γ is the control coefficient, and W
Let _i (n) be the i-th filter coefficient for sample number n in the filter means, μ be the convergence coefficient, and e
Let (n) be the synthetic sound wave detection signal in sample number n, and x (n-i) be the noise sound wave detection signal in sample number n-i, and W _i (n + 1) is the following expression W _i (n + 1) = γW _i (N) +2 μe (n) x (n-
In the case represented by i), the continuity maintaining unit 6 changes the control coefficient γ to a positive number less than 1 when the frequency component of the noise sound wave detection signal x (n) changes abruptly.

With the above operation, the noise sound wave detection signal x
When the frequency component of (n) changes rapidly, W
_{Since i} (n) does not change abruptly and is updated so as to gradually converge, the continuity of the waveform of the silencing interference sound wave S can be maintained.

Therefore, since the continuity of the waveform of the silencing interference sound wave S is maintained only by changing the control coefficient γ, the noise sound wave detection signal x can be obtained without expanding the scale of the existing device.
It is possible to prevent abnormal sound caused by the discontinuity of the waveform of the silencing interference sound wave S when the frequency component of (n) changes abruptly.

[0041]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, preferred embodiments of the present invention will be described with reference to FIGS. The embodiment shown in FIG. 2 is an example in which the present invention is applied to control the engine noise N in an automobile. In this embodiment, the engine noise N collected by the microphone as the noise sound wave detection signal x (n).
The noise sound wave N is indirectly detected by utilizing the ignition signal of the engine instead of the signal based on.

Although the ignition signal of the engine does not include information indicating the amplitude of the engine noise N, the amplitude information of the engine noise N is not used in this embodiment.

In the description of the embodiment shown in FIG. 2, the same parts as those in FIG. 7 are designated by the same reference numerals, and detailed description thereof will be omitted. As shown in FIG. 2, the present embodiment removes unnecessary noise from the ignition signal from the engine 7, which is the source of the noise N, and shapes the waveform by a waveform shaping section 8.
, An A / D converter 104 for A / D converting the waveform-shaped ignition signal, and a speaker 10 as an output means 2 for outputting a muffling interference sound S for canceling the engine noise N.
2, a synthetic sound of engine noise N and silencing interference sound S (N +
S), a microphone B103 as a synthesized sound wave detection means 3 for outputting a synthesized sound wave detection signal, an A / D converter 108 for A / D converting the synthesized sound wave detection signal, and an A / D conversion. And the DSP 9 that generates and outputs the muffling interference sound signal d (n) based on the ignition signal x (n) and the synthetic sound wave detection signal e (n).
A / A converter 109 and amplifier 1 for A / A converting and amplifying and generating an output signal to be output from the speaker 102.
10 and.

Here, n is a sample number in the system constituting this embodiment. The DSP 9 includes a rotation speed change rate detection unit 10 that detects a change rate of the engine rotation speed from the ignition signal x (n), and an ignition signal x.
Based on (n) and the synthetic sound wave detection signal e (n), the filter coefficient W of the ADF 106 that outputs the silencing interference sound signal d (n) that minimizes the synthetic sound wave detection signal e (n). The leaky LMS algorithm 11 as the coefficient setting means 5 for determining _i (n), and the ignition signal x under the control of the leaky LMS algorithm 11.
It is configured by the ADF 106 as the filter means 4 that generates and outputs the silencing interference sound signal d (n) based on (n) and the filter coefficient signal W _i (n).

The rotation speed change rate detecting section 10 is a frequency-voltage converter which obtains a DC signal having a level corresponding to the rotation speed of the engine 7 by converting the waveform-shaped signal into a half DC.
LPF (Low-Pass Filter) having the characteristics shown in FIG.
r) 12 and the DC signal obtained by the LPF 12 are compared at regular intervals, and the comparison and detection unit 13 detects the rate of change of the rotation speed from the difference.

The Leaky LMS algorithm 11 is an algorithm for determining the filter coefficient W _i (n) which characterizes the ADF 106 when generating the silencing interference sound signal d (n), and this filter coefficient W _i (n) Are updated over time. This algorithm is expressed by the following equation (1).

W _i (n + 1) = γW _i (n) +2 μe (n) x (n−i) (1) where μ is a convergence coefficient and γ is a control coefficient. Further, i is a filter coefficient number, and W _i (n) represents the i-th filter coefficient in the n-th sample.

In the present embodiment, when the operation of the ANC device is stopped, the control coefficient γ is set in the range of 0 <γ <1 by the continuity maintaining means 6, and the filter coefficient W is set.
The update of _i (n) is executed by the above equation (1). In this case, the specific value of the control coefficient γ is usually 0.
98 to 0.99 is desirable, which means
Confirmed experimentally.

When the operation of the ANC device is continued, the value of the coefficient γ in the above equation (1) is the Leaky LMS.
It is set to 1 by the continuity maintaining means 6 of the algorithm 11, and the filter coefficient W _i (n) is updated by the conventional LMS algorithm.

The continuity maintaining means 6 is represented by the algorithm shown in FIG. That is, whether or not the operation of the ANC device is stopped is determined by the setting signal from the rotation speed change rate detection unit 10 (step S1). If the operation is not stopped (step S1 No), the value of the control coefficient γ is set to 1. (Step S2), the following processing is performed. When the operation is stopped (step S1 Yes), the value of the control coefficient γ is set to 0 <γ
After setting <1 (step S3), the following processing is performed.

Next, the operation will be described with reference to FIGS. As shown in FIG. 2, the ignition signal output from the engine 7 is waveform-shaped by the waveform shaping unit 8, digitized by the A / D converter 104, and converted into the ignition signal x (n), and then rotated in the DSP 9. Number change rate detector 10, Leaky LMS algorithm 11 and A
It is input to the DF 106.

In the rotational speed change rate detecting section 10, the rotational speed change rate of the engine 7 is detected based on the input ignition signal x (n). When the rate of change exceeds a predetermined threshold value, the control coefficient γ in the above equation (1) is set to 0 in order to stop the ANC operation if there is a rapid change in the rotational speed, that is, abrupt acceleration or deceleration.
The setting signal for setting the range of <γ <1 is Leaky L
It is output to the continuity maintaining means 6 in the MS algorithm 11.

In the continuity maintaining means 6, the value of the control coefficient γ is set within the range of 0 <γ <1 based on this setting signal. The setting signal is not output when the rate of change of the rotational speed does not exceed the predetermined threshold value.

Next, the detection of the rate of change of the rotation speed will be described with reference to FIG.
Will be explained. The ignition signal output from the engine 7 has a waveform as shown in FIG. A signal obtained by waveform-shaping this by a monostable flip-flop circuit forming the waveform shaping unit 8 has a waveform shown in FIG.

The output signal (ignition signal) of the waveform shaping section 8 is A / D converted by the A / D converter 104 and then guided to the LPF 12 of the rotational speed change rate detecting section 10. Here, in the waveform shown in FIG. 3, the time of the off-duty portion changes in accordance with the change in the engine speed. Therefore, the change of the time of this portion is output to the LPF 12 as the change of the engine speed, that is, the change of the frequency of the ignition signal.

In the LPF 12, when a signal having a long off-duty time, that is, a signal indicating that the engine speed is low (for example, indicated by a in FIG. 3B), its output voltage is changed. The level becomes a low value (indicated by G _a in FIG. 3B). Further, when a signal in which the time of the off-duty portion is short, that is, a signal indicating that the engine speed is high (for example, indicated by b in FIG. 3B) is input, the level of the output voltage is high ( Figure 3 shows a medium G _{b (b).)} become. As a result, a DC signal having a signal level corresponding to the rotation speed of the engine 7 is obtained.

In the comparison / detection unit 13, the DC signal output from the LPF 12 is compared at regular intervals, the temporal change thereof is calculated, and the rate of change of the engine speed is detected.
When the rate of change becomes larger than the predetermined threshold value, it is determined that there is a sudden deceleration or a sudden acceleration, and the control coefficient γ is set within the range of 0 <γ <1 so as to stop the operation of the ANC device. The setting signal for doing so is output to the continuity maintaining means 6 in the Leaky LMS algorithm 11.

In the Leaky LMS algorithm 11, the synthesized sound wave detection signal e (n) and the ignition signal x which are A / D converted by the A / D converter 108.
The filter coefficient W of the ADF 106 based on (n)
_i (n) is updated by the above equation (1) and then ADF10
Output to 6.

In the ADF 106, under the control of the Leaky LMS algorithm 11, the ignition signal x
A silence interference sound signal d (n) is generated and output by filtering (n).

The silencing interference sound signal d (n) output from the ADF unit 106 is the D / A converter 109 and the amplifier 110.
Thus, the D / A conversion and amplification are performed, and the sound-muffling interference sound S is output from the speaker 102.

As described above, according to this embodiment, the ANC device is
The control coefficient in equation (1) when stopping the operation
Leaky LMS algorithm that sets the value of γ in the range of 0 <γ <1
The filter coefficient is updated by the algorithm. As a result,
The sound interference sound signal d (n) does not suddenly attenuate
As shown in FIG. 5, the silencing interference sound S is generated when the operation is stopped (t
₁) Gradually decreases without generating discontinuity of waveform
Decline. Therefore, the scale of the ANC device can be increased and the DSP
Stop the operation of the device without burdening the processing of 9.
In this case, the abnormal sound caused by the discontinuity of the output sound waveform is prevented.
You can stop.

[0062]

As described above, according to the invention described in claim 1, the continuity of the waveform of the silencing interference sound wave S is maintained when the frequency component of the noise sound wave detection signal x (n) changes abruptly. Since it is maintained, it is possible to prevent the generation of abnormal sound due to the discontinuity of the waveform of the silencing interference sound wave S.

According to the invention of claim 2, claim 1
In the noise control device according to [1], the noise sound wave detection signal x
When the frequency component of (n) suddenly changes, the filter
Coefficient W _i(N) is updated by the following formula W_i(N + 1) = γW_i(N) +2 μe (n) x (n-
i)_i(N) does not change rapidly and gradually
It is updated in the direction of convergence, and the waveform of the silencing interference sound wave S continues.
Discontinuity of the waveform of the muffling interference sound wave S, because the soundness is maintained.
It is possible to prevent the generation of abnormal sound due to

Therefore, since the continuity of the waveform of the silencing interference sound wave S can be maintained only by changing the control coefficient γ, the noise sound wave detection signal x (n) can be obtained without expanding the size of the existing device. It is possible to prevent an abnormal sound due to the discontinuity of the waveform of the silencing interference sound wave S when the frequency component changes rapidly.

More specifically, the noise sound wave detection signal x
Even when the operation of the ANC device is stopped when the frequency component of (n) suddenly changes, the silencing interference sound wave S is attenuated while maintaining its waveform continuity. It is possible to prevent the generation of abnormal sound due to the discontinuity of the waveform.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a diagram showing an example of the present invention.

FIG. 3 is a diagram showing a method for detecting a rate of change in engine speed according to the present invention.

FIG. 4 is an algorithm for determining a coefficient γ.

FIG. 5 is a diagram showing an effect of the present invention.

FIG. 6 is a diagram showing the principle of a noise control device.

FIG. 7 is a diagram showing an ANC device under adaptive control.

FIG. 8 is a diagram showing an operation stop of an ANC device according to a conventional technique.

FIG. 9 is an algorithm showing stop of operation by zero cross mute.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Noise sound wave detecting means 2 ... Output means 3 ... Synthetic sound wave detecting means 4 ... Filter means 5 ... Coefficient setting means 6 ... Continuity maintaining means 7 ... Engine 8 ... Waveform shaping section 9 ... DSP section 10 ... Rotation speed change rate Detection unit 11 ... Leaky LMS algorithm 12 ... LPF 13 ... Comparison detection unit 100 ... Microphone A 101 ... Amplifier 102 ... Speaker 103 ... Microphone B 104, 108 ... A / D converter 105 ... DSP section 106 ... ADF section 107 ... LMS algorithm 109 ... D / A converter 110 ... Amplifier x (n) ... Noise sound wave detection signal (ignition signal) e (n) ... Synthetic sound sound wave detection signal d (n) ... Silent interference sound signal _Wi (n) ... Filter coefficient signal

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Shinjiro Kato 25-25 Nishimachi, Yamada, Kawagoe, Saitama Prefecture 1 Pioneer Co., Ltd. Kawagoe Factory (72) Nobuo Tarui 25, Nishimachi, Yamada, Kawagoe, Saitama 1 Pioneer Corporation Kawagoe Factory

Claims (2)

[Claims] 1. A noise sound wave detecting means for directly or indirectly detecting a noise sound wave and outputting a noise sound wave detection signal, and a noise-canceling interference sound wave for attenuating the noise sound wave based on the noise sound wave detection signal. Filter means for generating a silencing interference sound wave signal, an output means for receiving the silencing interference sound wave signal and radiating the silencing interference sound wave into a sound field space, and detecting a synthesized sound wave of the noise sound wave and the silencing interference sound wave. Synthetic sound wave detection means for outputting a synthesized sound sound wave detection signal, and a filter characteristic for minimizing the sound pressure level of the synthesized sound sound wave based on the synthetic sound wave detection signal and the noise sound wave detection signal. And a coefficient setting unit that outputs the filter coefficient signal to the filter unit, the coefficient setting unit including: During a sudden change of the frequency components of the wave detection signal, the noise control device characterized by having a continuous maintenance means for maintaining the continuity of the silencing interference wave waveform. 2. The noise control device according to claim 1, wherein the coefficient setting means is constituted by an LMS algorithm, and W _i (n + 1) is an i-th filter coefficient at a sample number n + 1 in the filter means, Let γ be the control coefficient, W _i (n) be the i-th filter coefficient at the sample number n in the filter means, μ be the convergence coefficient, and e (n) be the synthetic sound wave detection signal at the sample number n. And x (n-i) is the sample number n
As the noise sound wave detection signal at −i, W _i (n +
1) is the following expression W _i (n + 1) = γW _i (n) +2 μe (n) x (n−
In the case represented by i), the continuity maintaining means changes the control coefficient γ to a positive number less than 1 when the frequency component of the noise sound wave detection signal changes abruptly. apparatus.