JP2019086570A - Active noise reduction device and active noise reduction method - Google Patents

Abstract

To provide an active noise reduction device and an active noise reduction method, which realize ANC operation that is hardly affected by a sound signal with a small processing amount.SOLUTION: An active noise reduction device comprises: a voice signal analyzer for determining a filter coefficient of a first adaptive filter applied to a sine wave and a cosine wave, which are synchronized with a frequency of a second voice signal generated from a sound source on the basis of an inputted first voice signal; a signal generator for generating a removal signal for removing a frequency component synchronized with the frequency of the second voice signal in the first voice signal on the basis of the filter coefficient of the first adaptive filter; and an antinoise generator for determining the filter coefficient of a second adaptive filter on the basis of a synthesis signal obtained by synthesizing a third voice signal detected by a voice detection device with the removal signal, and of a correction signal to which a transmission characteristic determined on the basis of the transmission characteristic of a route from a voice output device to the voice detection device is applied with respect to the sine wave and the cosine wave, and reducing the second voice signal on the basis of the filter coefficient of the second adaptive filter.SELECTED DRAWING: Figure 1

Description

  The present disclosure relates to an active noise reduction device and an active noise reduction method that actively reduce an engine booming noise generated in a vehicle interior along with engine vibration.

  Conventionally, as a technology for reducing noise signals at the microphone position, Active Noise Control (ANC) technology for reducing noise signals at the position of the microphone by emitting a signal in the opposite phase to the noise signal from the speaker It has been known.

  In addition, ANC technology is known which reduces engine muffled noise by using a SAN (Single Frequency Adaptive Notch) algorithm. The term "engine booming noise" refers to noise composed of the rotational frequency of the engine generated by combustion in the cylinder and its harmonic components.

In the SAN algorithm, sine waves sin (t) and cosine waves cos (t) are generated as reference signals in synchronization with the operating frequency (the frequency of the engine boom noise to be reduced) determined according to the engine speed. Then, a signal obtained by multiplying the sine wave sin (t) and the cosine wave cos (t) by the coefficients A and B of the adaptive filter and adding them is output from the speaker as an ANC output signal x (t). The ANC output signal x (t) is expressed by the following equation (1).
x (t) = A (t) × sin (t) + B (t) × cos (t) (1)

Next, the microphone detects a result of interference between the ANC output signal x (t) output from the speaker and the engine muffled sound n (t). The error signal e (t) detected by the microphone is the sum of the engine muffled sound n (t) and the signal obtained by multiplying the ANC output signal x (t) by the transfer characteristic C between the speaker and the microphone, It is represented by the following equation (2).
e (t) = n (t) + x (t) × C (2)

Then, using the error signal e (t) and the C ^ (C hat) signal obtained by modeling the transfer characteristic C, the coefficients A and B of the adaptive filter are updated based on the LMS (Least Mean Square) arithmetic expression Do. Specifically, a signal obtained by multiplying a sine wave sin (t) and a cosine wave cos (t) by C ^ obtained by modeling the transfer characteristic C and an error signal e (t) which is an input signal of a microphone are used. The coefficients A (t + 1) and B (t + 1) of the adaptive filter are obtained by Equations (3) and (4) which are arithmetic expressions of LMS.
A (t + 1) = A (t) -C ^ × sin (t) × e (t) (3)
B (t + 1) = B (t) -C ^ × cos (t) × e (t) (4)

  In this way, using the error signal e (t) obtained from the microphone, optimization processing is performed to cause the coefficients A and B of the adaptive filter to converge to the optimum value, and the coefficients of the adaptive filter obtained by performing the optimization processing A and B modulate the amplitude and phase of the sine wave signal generated based on the engine speed to generate an antiphase signal of the engine booming noise.

  When such an ANC operation and audio reproduction are performed simultaneously, the ANC operation also reduces the sound signal reproduced from the audio in addition to the engine muffled sound. In order to cope with such a situation, in Patent Document 1, an estimated voice path filter is used to remove a voice signal which is undesirable for ANC operation from an error signal. According to the method described in Patent Document 1, the optimization process can be performed using an error signal that does not include an audio signal, so that an ANC operation that is not affected by the audio signal can be performed.

Patent No. 5026495 gazette

  However, in the case described in Patent Document 1, it is necessary to perform convolution of an estimated speech path filter on the speech signal in order to remove all speech signals. Therefore, there is a problem that the amount of processing increases.

  An object of the present disclosure is to provide an active noise reduction device and an active noise reduction method that enable ANC operation insensitive to audio signals with a small amount of processing.

One form of the present disclosure is
A speech signal analyzer for determining the filter coefficients of a first adaptive filter applied to sine waves and cosine waves synchronized with the frequency of a second speech signal generated from a sound source based on a first speech signal inputted When,
A signal generator for generating a removal signal for removing a frequency component synchronized with the frequency of the second audio signal in the first audio signal based on the filter coefficient of the first adaptive filter;
Based on the transfer characteristic of the path from the voice output device to the voice detection device with respect to the sine wave and the cosine wave with respect to the sine wave and the cosine wave, and a composite signal obtained by combining the third voice signal detected by the voice detection device and the removal signal. The filter coefficient of the second adaptive filter is determined based on the correction signal to which the determined transfer characteristic is applied, and the second audio signal is reduced based on the filter coefficient of the second adaptive filter. And an anti-noise generator that generates a signal.

In addition, one mode of the present disclosure is
Determining a filter coefficient of a first adaptive filter applied to a sine wave and a cosine wave synchronized with a frequency of a second audio signal generated from a sound source based on the input first audio signal;
Generating a removal signal for removing a frequency component synchronized with the frequency of the second audio signal in the first audio signal based on the filter coefficient of the first adaptive filter;
Based on the transfer characteristic of the path from the voice output device to the voice detection device with respect to the sine wave and the cosine wave with respect to the sine wave and the cosine wave, and a composite signal obtained by combining the third voice signal detected by the voice detection device and the removal signal. The filter coefficient of the second adaptive filter is determined based on the correction signal to which the determined transfer characteristic is applied, and the second audio signal is reduced based on the filter coefficient of the second adaptive filter. Generating a signal. An active noise reduction method.

  According to the present disclosure, since the frequency components that affect the ANC operation in the audio signal are intensively removed, it is possible to enable the ANC operation that is not easily influenced by the audio signal with a small amount of processing.

Block diagram of active noise reduction device according to first embodiment of the present disclosure A conceptual diagram showing an audio signal included in the adaptive filter updating signal Block diagram of active noise reduction device according to second embodiment of the present disclosure

  Hereinafter, an active noise reduction device according to an embodiment of the present disclosure will be described in detail with reference to the drawings. The embodiment described below is an example, and the present disclosure is not limited to this embodiment.

First Embodiment
FIG. 1 is a block diagram showing a configuration of an active noise reduction device 1 according to a first embodiment of the present disclosure. In FIG. 1, the engine 2 is a noise source (an example of a “sound source”) that generates noise to be reduced (an example of the “second sound signal”; hereinafter may be referred to as a “noise signal”). The active noise reduction device 1 operates to reduce periodic noise (engine muffled noise) emitted from the engine 2.

  The active noise reduction device 1 comprises an audio signal analyzer 10, a signal generator 20 and an anti-noise generator 30.

  The audio signal analyzer 10 includes an adaptive filter 11 (an example of “a first adaptive filter for audio signal analysis”), an adaptive filter 12 (an example of a “second adaptive filter for audio signal analysis”), and an adder 13 (“an adaptive filter”). Example of first audio signal analysis adder), adder 14 (example of second audio signal analysis adder), adaptive control algorithm calculator 15 (adaptive control algorithm calculator for audio signal analysis) Example) and the estimated path filter 16. The adaptive filter 11 and the adaptive filter 12 may be collectively referred to as a "first adaptive filter".

  The signal generator 20 includes a transfer element 21, a transfer element 22, a coefficient multiplier 23 (an example of “first coefficient generation coefficient multiplier”), a coefficient multiplier 24 (“second signal generation coefficient multiplier”) And an adder 25 (an example of a “signal generation adder”).

  The anti-noise generator 30 includes an adaptive filter 31 (an example of a “first anti-noise generation adaptive filter”), an adaptive filter 32 (an example of a “second anti-noise generation adaptive filter”), and an adder 33 (“an And an adaptive control algorithm computing unit 34 (an example of an "adaptive control algorithm computing unit for anti noise"). The adaptive filter 31 and the adaptive filter 32 may be collectively referred to as a "second adaptive filter".

  An engine pulse, which is an electrical signal synchronized with the rotation of the engine 2, is input to the waveform shaper 3. The engine pulse input to the waveform shaper 3 is subjected to waveform shaping after the noise and the like superimposed thereon are removed by the waveform shaper 3.

  The output signal of the waveform shaper 3 is input to the sine wave generator 4 and the cosine wave generator 5. The sine wave generator 4 and the cosine wave generator 5 generate sine waves and cosine waves as reference signals which are determined according to the number of revolutions of the engine 2 and which are synchronized with the frequency of the noise to be reduced. In the following description, the output signal of the sine wave generator 4 is referred to as “reference sine wave signal sin (t)”, and the output signal of the cosine wave generator 5 is referred to as “reference cosine wave signal cos (t)”. The engine pulse may be directly input to the sine wave generator 4 and the cosine wave generator 5 without being input to the waveform rectifier 3. Further, instead of the engine pulse, the rotation speed information of the engine 2 obtained through CAN (Controller Area Network) may be used.

  The reference sinusoidal signal sin (t) is multiplied in the speech signal analyzer 10 with the coefficient Am of the adaptive filter 11. Similarly, the reference cosine wave signal cos (t) is also multiplied by the coefficient Bm of the adaptive filter 12.

The output signal of the adaptive filter 11 and the output signal of the adaptive filter 12 are added by the adder 13. The output signal xm (t) of the adder 13 is expressed by the following equation (5).
xm (t) = Am (t) × sin (t) + Bm (t) × cos (t) (5)

Furthermore, the output signal xm (t) of the adder 13 is added by the adder 14 to the audio signal M (an example of the “first audio signal”). The adaptive filter updating signal em (t) which is the output signal of the adder 14 is expressed by the following equation (6).
em (t) = xm (t) + M (6)

  The reference sine wave signal sin (t) and the reference cosine wave signal cos (t) are input to the estimated path filter 16 and are respectively multiplied by the coefficient Cm of the estimated path filter 16. The coefficient Cm may be a value reflecting the characteristics of a predetermined path in the audio signal analyzer 10, or may be an arbitrary value determined in advance.

  The adaptive control algorithm computing unit 15 receives the output signal of the estimation path filter 16 and the adaptive filter updating signal em (t).

The adaptive control algorithm computing unit 15 updates the coefficient Am of the adaptive filter 11 according to the following equation (7).
Am (t + 1) = Am (t) -Cm × sin (t) × em (t) (7)

Further, the adaptive control algorithm computing unit 15 updates the coefficient Bm of the adaptive filter 12 according to the following equation (8).
Bm (t + 1) = Bm (t) -Cm × cos (t) × em (t) (8)

  Thus, the coefficient Am of the adaptive filter 11 and the coefficient Bm of the adaptive filter 12 converge to the optimal value.

  As described above, using the adaptive filter updating signal em (t) including the audio signal M, optimization processing is performed to converge the coefficient Am of the adaptive filter 11 and the coefficient Bm of the adaptive filter 12 to the optimum value.

  Then, the coefficient Am of the adaptive filter 11 and the coefficient Bm of the adaptive filter 12 obtained by such optimization processing are output to the signal generator 20.

  In the signal generator 20, the transfer element 21 receives the reference sine wave signal sin (t), and is multiplied by the coefficient Cgen of the transfer element 21. The output signal of the transfer element 21 is multiplied by the coefficient Am of the adaptive filter 11 in the coefficient multiplier 23.

  Further, the reference cosine wave signal cos (t) is input to the transfer element 22 and is multiplied by the coefficient Cgen of the transfer element 22. The output signal of transfer element 22 is multiplied by coefficient Bm of adaptive filter 12 in coefficient multiplier 24.

The output signal of the coefficient multiplier 23 and the output signal of the coefficient multiplier 24 are added by the adder 25. The output signal x'm (t) of the adder 25 is expressed by the following equation (9).
x'm (t) = Cgen x Am (t) x sin (t) + Cgen x Bm (t) x cos (t) (9)

  The output signal x'm (t) of the adder 25 is a signal for removing a frequency component synchronized with the frequency of the noise signal in the audio signal M (hereinafter, may be referred to as a "removal signal"). It is. That is, the adder 25 generates a removal signal.

  In the anti-noise generator 30, the reference sinusoidal signal sin (t) is multiplied by the filter coefficient A of the adaptive filter 31. Further, the reference cosine wave signal cos (t) is similarly multiplied by the filter coefficient B of the adaptive filter 32.

The output signal of the adaptive filter 31 and the output signal of the adaptive filter 32 are added by the adder 33. The output signal x (t) output from the adder 33 is expressed by the following equation (10).
x (t) = A (t) × sin (t) + B (t) × cos (t) (10)

The output signal x (t) output from the adder 33 is added to the audio signal M by the adder 6 and output from the speaker 7 (an example of “audio output device”) as an ANC output signal x ′ (t) . The ANC output signal x ′ (t) output from the speaker 7 is expressed by the following equation (11).
x '(t) = x (t) + M (11)

  The residual signal that can not be silenced due to the interference between the ANC output signal x ′ (t) output from the speaker 7 and the engine muffled sound n (t) is an error signal e (t) (“third audio signal”). As an example, it is detected by the microphone 8 (an example of a "voice detection device").

The error signal e (t) is obtained by adding the engine muffled sound n (t) and a signal obtained by multiplying the transfer characteristic C between the speaker 7 and the microphone 8 by the ANC output signal x ′ (t). The error signal e (t) is expressed by the following equation (12).
e (t) = n (t) + x '(t) × C (12)

In the adder 9, the output signal x′m (t) of the adder 25 is added to the error signal e (t), and the adaptive filter updating signal e anc (t) (an example of “combined signal”) is obtained. It is generated. The adaptive filter updating signal e anc (t) output from the adder 9 is expressed by the following equation (13).
e anc (t) = e (t) + x'm (t) (13)

  FIG. 2 is a conceptual diagram showing an audio signal included in the adaptive filter updating signal. According to the first embodiment, as shown in FIG. 2, frequency components synchronized with the frequency of the noise signal are intensively removed from the audio signal.

  The adaptive control algorithm computing unit 34 updates the coefficient A of the adaptive filter 31 and the coefficient B of the adaptive filter 32 using the adaptive filter updating signal e anc (t) thus obtained.

  The reference sine wave signal sin (t) and the reference cosine wave signal cos (t) are input to the estimated path filter 40, and are respectively multiplied by the coefficient C ^ of the estimated path filter 40. Here, ^ is a coefficient that models the above-described transfer characteristic C.

  The adaptive control algorithm computing unit 34 receives the output signal of the estimated path filter 40 (an example of the “correction signal”) and the adaptive filter updating signal e anc (t) output from the adder 9.

In the adaptive control algorithm computing unit 34, the coefficient A of the adaptive filter 31 is updated according to the following equation (14).
A (t + 1) = A (t) -C ^ × sin (t) × eanc (t) (14)

Further, in the adaptive control algorithm computing unit 34, the coefficient B of the adaptive filter 32 is updated by the following equation (15).
B (t + 1) = B (t) -C ^ × cos (t) × eanc (t) (15)

  Thus, the coefficient A of the adaptive filter 31 and the coefficient B of the adaptive filter 32 converge to the optimum value.

  As described above, the adaptive filter 31 uses the adaptive filter updating signal e anc (t) from which the frequency component synchronized with the frequency of the noise signal in the music signal M is intensively removed from the error signal e (t). An optimization process is performed to converge the coefficient A of and the coefficient B of the adaptive filter 32 to an optimal value.

  Then, using the coefficient A of the adaptive filter 31 and the coefficient B of the adaptive filter 32 obtained by such optimization processing, the amplitude and phase of the sine wave signal generated based on the engine rotational speed are modulated to obtain an engine It generates an antiphase signal of muddy sound.

  As described above, according to the present embodiment, the active noise reduction device is configured to receive the sine wave and the cosine wave synchronized with the frequency of the second audio signal generated from the sound source based on the input first audio signal. Synchronized to the frequency of the second speech signal in the first speech signal based on the speech signal analyzer for determining the filter factor of the first adaptive filter applied to the first speech filter and the filter factor of the first adaptation filter Signal generator for generating a removal signal for removing the frequency component, a synthetic signal obtained by combining the third audio signal detected by the voice detection device and the removal signal, the sine wave and the cosine wave On the other hand, the filter coefficient of the second adaptive filter is determined on the basis of the correction signal to which the transfer characteristic determined based on the transfer characteristic of the path from the audio output device to the voice detection device is applied. And, based on the filter coefficients of said second adaptive filter, and a anti-noise generator for generating a signal for reducing the second audio signal.

  As a result, it is possible to intensively remove from the error signal the frequency component that affects the ANC operation in the audio signal.

  Therefore, it is possible to enable an ANC operation that is not easily influenced by the audio signal with a small amount of processing.

Second Embodiment
Although the above-described first embodiment has been described by way of an example having a single microphone, the present invention is not limited to this. As a second embodiment, one having a plurality of microphones will be described. FIG. 3 is a block diagram showing a configuration of an active noise reduction device 101 according to a second embodiment of the present disclosure. In addition, the same code | symbol is attached | subjected to the component same as the active noise reduction apparatus 1 shown in 1st embodiment.

  The active noise reduction device 101 comprises an audio signal analyzer 110, a signal generator 120 and an anti-noise generator 130.

  The audio signal analyzer 110 includes an adaptive filter 111 (an example of a “first adaptive filter for audio signal analysis”), an adaptive filter 112 (an example of a “adaptive filter for second audio signal analysis”), and an adder 113 (“an adaptive filter”). Example of a first audio signal analysis adder), adder 114 (an example of a second audio signal analysis adder), adaptive control algorithm calculator 115 (audio signal analysis adaptive control algorithm calculator) Example) and an estimated path filter 116. The configuration of the audio signal analyzer 110 is the same as that of the above-described audio signal analyzer 10, so the detailed description will be omitted.

  The signal generator 120 includes transfer elements 121, 121a and 121b, and transfer elements 122, 122a and 122b. Also, the signal generator 120 includes coefficient multipliers 123, 123a and 123b (an example of "first coefficient generating coefficient multiplier"), and coefficient multipliers 124, 124a and 124b ("second signal generating coefficient". Example of “multiplier”. Furthermore, the signal generator 120 includes adders 125, 125a and 125b (an example of a "signal generation adder").

  The anti-noise generator 130 includes an adaptive filter 131 (an example of a “first anti-noise generation adaptive filter”), an adaptive filter 132 (an example of a “second anti-noise generation adaptive filter”), and an adder 133 (An example of an "anti-noise generation adder") and adaptive control algorithm calculators 134, 134a and 134b (an example of an "anti-noise generation adaptive control algorithm calculator").

  The optimization method of the coefficient Am of the adaptive filter 111 and the coefficient Bm of the adaptive filter 112 performed in the speech signal analyzer 110 is the same as that of the first embodiment, and thus the detailed description will be omitted.

In the signal generator 120, the transfer element 121,121a and 121b are input reference sine wave signal sin (t), respectively, it is multiplied by the transfer element 121,121a and 121b of the coefficients _Cgen 11, Cgen ₁₂ and Cgen ₁₃ Ru. The output signals of transfer elements 121, 121a and 121b are multiplied by coefficient Am of adaptive filter 111 in coefficient multipliers 123, 123a and 123b.

In addition, the reference cosine wave signal cos (t) is input to the second signal generation transfer elements 122, 122a and 122b, and the coefficients Cgen ₁₁ , Cgen ₁₂ and Cgen _{13 of} the transfer elements 121, 121a and 121b, and It is multiplied. The output signals of transfer elements 122, 122a and 122b are multiplied by coefficients Bm of adaptive filter 112 in coefficient multipliers 124, 124a and 124b.

The output signal of the coefficient multiplier 123 and the output signal of the coefficient multiplier 124 are added by the adder 125. The output signal x′m ₁₁ (t) of the adder 125 is expressed by the following equation (16).
x'm ₁₁ (t) = Cgen ₁₁ x Am (t) x sin (t) + Cgen ₁₁ x Bm (t) x cos (t) (16)

The output signal of the coefficient multiplier 123a and the output signal of the coefficient multiplier 124a are added by the adder 125a. The output signal x'm ₁₂ (t) of the adder 125a is expressed by the following equation (17).
x'm ₁₂ (t) = Cgen ₁₂ x Am (t) x sin (t) + Cgen ₁₂ x Bm (t) x cos (t) (17)

The output signal of the coefficient multiplier 123b and the output signal of the coefficient multiplier 124b are added by the adder 125b. The output signal x′m ₁₃ (t) of the adder 125 b is expressed by the following equation (18).
x'm ₁₃ (t) = Cgen ₁₃ x Am (t) x sin (t) + Cgen ₁₃ x Bm (t) x cos (t) (18)

  In the anti noise generator 130, the reference sine wave signal sin (t) is multiplied by the filter coefficient A of the adaptive filter 131. Similarly, the reference cosine wave signal cos (t) is also multiplied by the filter coefficient B of the adaptive filter 132.

  The output signal of the adaptive filter 131 and the output signal of the adaptive filter 132 are added by the adder 133. The output signal x (t) output from the adder 133 is added to the audio signal M by the adder 106, and is output from the speaker 107 (an example of “audio output device”) as an ANC output signal x ′ (t). .

The residual signal that can not be silenced due to the interference between the ANC output signal x ′ (t) output from the speaker 107 and the engine muffled sound n (t) is subjected to the first error signal e ₁₁ (t) by the microphone And detected as a second error signal e ₁₂ (t) by the microphone 108 a and detected as a third error signal e ₁₃ (t) by the microphone 108 b. The microphones 108, 108a and 108b are examples of the "voice detection device".

The adder 109 adds the output signal x′m ₁₁ (t) of the adder 125 to the first error signal e ₁₁ (t) to generate an adaptive filter updating signal e anc ₁₁ (t). Ru.

The adder 109a adds the output signal x'm ₁₂ (t) of the adder 125a to the second error signal e ₁₂ (t) to generate the adaptive filter updating signal e anc ₁₂ (t). Ru.

The adder 109 b adds the output signal x′m ₁₃ (t) of the adder 125 b to the third error signal e ₁₃ (t) to generate the adaptive filter update signal e anc ₁₃ (t). Ru. Note that the adaptive filter updating signals e anc ₁₁ (t), e anc ₁₂ (t), and e anc ₁₃ (t) are examples of the “combined signal”.

The reference sine wave signal sin (t) and the reference cosine wave signal cos (t) are input to the estimated path filter 140, and are respectively multiplied by the coefficient ^ ₁₁ of the estimated path filter 140. Note that C ₁₁ is a coefficient that models the transfer characteristic C ₁₁ between the speaker 107 and the microphone 108.

The reference sine wave signal sin (t) and the reference cosine wave signal cos (t) are input to the estimated path filter 140a, and respectively multiplied by the coefficient C ^ ₁₂ of the estimated path filter 140a. Incidentally, _{C ^ 12} is a coefficient that models the transfer characteristics _{C 12} between speaker 107 and microphone 108a.

The reference sine wave signal sin (t) and the reference cosine wave signal cos (t) are input to the estimated path filter 140b, and are respectively multiplied by the coefficient C ^ ₁₃ of the estimated path filter 140b. C C ₁₃ is a coefficient that models the transfer characteristic C ₁₃ between the speaker 107 and the microphone 108b.

The output signal of the estimated path filter 140 and the adaptive filter updating signal e anc ₁₁ (t) output from the adder 109 are input to the adaptive control algorithm computing unit 134.

The adaptive control algorithm computing unit 134a receives the output signal of the estimated path filter 140a and the adaptive filter updating signal ean c ₁₂ (t) output from the adder 109a.

The output signal of the estimated path filter 140b and the adaptive filter updating signal eanc ₁₃ (t) output from the adder 109b are input to the adaptive control algorithm computing unit 134b. The output signals of the estimated path filters 140, 140a and 140b are an example of the “correction signal”.

  In the second embodiment, the update of the coefficient A of the adaptive filter 131 is performed based on the calculation results of the adaptive control algorithm calculators 134, 134a and 134b.

  Similarly, updating of the coefficient B of the adaptive filter 132 is performed based on the calculation results of the adaptive control algorithm calculators 134, 134a and 134b.

  Thus, the coefficient A of the adaptive filter 131 and the coefficient B of the adaptive filter 132 converge to the optimum value.

  As described above, in the second embodiment, a plurality of adaptive filter updating signals are used in which a signal corresponding to the frequency of the noise signal in the music signal M is removed from the plurality of error signals detected by the plurality of microphones. An optimization process is performed to converge the coefficient A of the adaptive filter 131 and the coefficient B of the adaptive filter 132 to optimal values.

  According to the active noise reduction device and the active noise reduction method according to the present disclosure, it is possible to enable an ANC operation that is not easily influenced by an audio signal with a small amount of processing, and is suitable for on-vehicle applications.

DESCRIPTION OF SYMBOLS 1, 101 Active noise reduction device 2 engine 3 waveform shaper 4 sine wave generator 5 cosine wave generator 6, 106 adder 7, 107 speaker 8, 108, 108a, 108b microphone 9, 109, 109a, 109b adder 10, 110 Speech signal analyzer 11, 111 Adaptive filter 12, 112 Adaptive filter 13, 113 Adder 14, 114 Adder 15, 115 Adaptive control algorithm calculator 16, 116 Estimated path filter 20, 120 Signal generator 21, 121 , 121a, 121b transfer element 22, 122, 122a, 122b transfer element 23, 123, 123a, 123b coefficient multiplier 24, 124, 124a, 124b coefficient multiplier 25, 125, 125a, 125b adder 30, 130 anti noise generation 31, 131 adaptation Ruta 32, 132 adaptive filter 33, 133 adder 34,134,134a, 134b adaptive control algorithm processor 40,140,140a, 140b estimated path filter

Claims (5)

A speech signal analyzer for determining the filter coefficients of a first adaptive filter applied to sine waves and cosine waves synchronized with the frequency of a second speech signal generated from a sound source based on a first speech signal inputted When,
A signal generator for generating a removal signal for removing a frequency component synchronized with the frequency of the second audio signal in the first audio signal based on the filter coefficient of the first adaptive filter;
Based on the transfer characteristic of the path from the voice output device to the voice detection device with respect to the sine wave and the cosine wave with respect to the sine wave and the cosine wave, and a composite signal obtained by combining the third voice signal detected by the voice detection device and the removal signal. The filter coefficient of the second adaptive filter is determined based on the correction signal to which the determined transfer characteristic is applied, and the second audio signal is reduced based on the filter coefficient of the second adaptive filter. An active noise reduction device comprising: an anti-noise generator for generating a signal. The voice signal analyzer
A first audio signal analysis adaptive filter applied to the sine wave;
A second audio signal analysis adaptive filter applied to the cosine wave;
A first audio signal analysis adder for combining an output signal of the first audio signal analysis adaptive filter with an output signal of the second audio signal analysis adaptive filter;
A second audio signal analysis adder that synthesizes the output signal of the first audio signal analysis adder and the first audio signal;
The filter coefficient of the first audio signal analyzing adaptive filter using the signal to which a predetermined transfer characteristic is applied to the sine wave and the cosine wave, and the output signal of the second audio signal analyzing adder And an adaptive control algorithm computing unit for voice signal analysis that determines filter coefficients of the second voice signal analysis adaptive filter.
An active noise reduction device according to claim 1. The signal generator
A first signal generating coefficient multiplier that multiplies the signal obtained by multiplying the sine wave by a predetermined coefficient with the filter coefficient of the first audio signal analysis adaptive filter;
A second signal generation coefficient multiplier that multiplies the signal obtained by multiplying the cosine wave by a predetermined coefficient with the filter coefficient of the second audio signal analysis adaptive filter;
And a signal generation adder that combines the output signal of the first adaptive filter for signal generation and the output signal of the second adaptive filter for signal generation.
An active noise reduction device according to claim 2. The anti-noise generator
A first anti-noise adaptive filter applied to the sine wave;
A second antinoise generating adaptive filter applied to the cosine wave;
An anti-noise generation adder for combining an output signal of the first anti-noise adaptive filter and an output signal of the second anti-noise adaptive filter;
Adaptive control algorithm operator for anti noise generation which determines the filter coefficient of the first adaptive filter for anti noise generation and the filter coefficient of the second adaptive filter for anti noise generation using the synthesized signal and the correction signal And
The active noise reduction device according to any one of claims 1 to 3. Determining a filter coefficient of a first adaptive filter applied to a sine wave and a cosine wave synchronized with a frequency of a second audio signal generated from a sound source based on the input first audio signal;
Generating a removal signal for removing a frequency component synchronized with the frequency of the second audio signal in the first audio signal based on the filter coefficient of the first adaptive filter;
Based on the transfer characteristic of the path from the voice output device to the voice detection device with respect to the sine wave and the cosine wave with respect to the sine wave and the cosine wave, and a composite signal obtained by combining the third voice signal detected by the voice detection device and the removal signal. The filter coefficient of the second adaptive filter is determined based on the correction signal to which the determined transfer characteristic is applied, and the second audio signal is reduced based on the filter coefficient of the second adaptive filter. Generating a signal. An active noise reduction method.

Images