JPH086575A - Noise control device - Google Patents

Abstract

PURPOSE:To allow the improvement of an off-line training compensating a simulated transmission system to be performed. CONSTITUTION:A noise control part 4-1 having a transmission characteristic simulating part simulating a space transmission characteristic with which a sound transmits a space between a speaker 2 and a microphone 3 is provided in a noise control device having the speaker 2 outputting a canceling sound and the microphone 3 detecting an error signal to form a canceling signal. An off-line training part 4-2 having an initial value storage part consisting of a nonvolatile memory storing initial values of space transmission characteristics outputs a reference signal having a white noise from the speaker 2 to the space and calculates a space transmission characteristic by learnings from a caught signal using the initial value to update the space transmission charascteristic of the transmission characteristic simulating part. A stater delaying circuit 110 delays the timing of a starting switch performing a current supply from a battery to driving the starter starting an engine for a short time and the off-line training part 4-2 is made to be executed in this delay time.

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 for generating a canceling sound having a phase opposite to that of noise and having the same amplitude. The present invention relates to one having off-line training for correcting a simulated transmission system separately from a noise control device when performing.

[0002]

2. Description of the Related Art Conventionally, noise control devices have detected noise and vibration from an intake pipe and an exhaust pipe of an internal combustion engine (engine) including a vehicle, a ship, etc. with a microphone or a similar sensor, and use digital signal processing to detect the noise and vibration. There is a method in which a canceling sound or vibration having an opposite phase and the same amplitude as noise is generated from a speaker or an actuator similar thereto to effectively reduce the noise vibration. Since the speaker, the microphone, and the like that constitute the noise control device change over time, the simulated transfer characteristic is corrected by off-line training according to this change over time. In the off-line training of the noise control device, when the number of products is small and regular maintenance is guaranteed, the noise control device is stopped and off-line training is performed each time. This off-line training is realized by learning which improves self-adaptiveness, but it also depends on the sampling period used for digital signal processing, for example this is 3 kHz.
In the case of z, about 10,000 learnings are required, and this learning requires about 3 to 4 seconds.

[0003]

By the way, the number of products of the noise control device is large, regular maintenance is not guaranteed, and offline training is performed for each model at the time of product shipment, and the result is stored in a non-volatile memory. ROM (Read On
may be stored in memory and used. However, since the same model uses the same ROM, there is a problem that the noise reduction effect is reduced due to variations in products. Further, there is a problem that the noise reduction effect is deteriorated due to the secular change of the control noise device itself in the long-term use and the secular change of its surroundings.

Therefore, in view of the above problems, the present invention makes it possible to perform off-line training against product variations and aging even when the number of products is large and regular maintenance is not guaranteed. An object is to provide a control device.

[0005]

In order to solve the above problems, the present invention provides a noise control device having the following configuration. That is, a cancel signal output to the speaker is output to a speaker that outputs a cancel sound having a phase opposite to that of the noise and the same amplitude, and a noise control device that has a microphone for detecting and feeding back an error signal from the residual sound of the canceled noise. In order to form the above, a noise control section having a transfer characteristic simulating section for simulating a space transfer characteristic in which sound transmits the space between the speaker and the microphone is provided.

An offline training unit having a non-volatile memory for storing the initial value of the space transfer characteristic outputs a white noise reference signal to the space from the speaker and captures it, and uses the initial value from the captured signal. Then, the spatial transfer characteristic is obtained by learning, and the spatial transfer characteristic of the transfer characteristic simulation unit is updated. The starter delay circuit delays the timing of the start switch for supplying current from the battery to drive the starter for starting the engine for a short time, and causes the offline training unit to execute within this delay time.

In place of the initial value storage unit, the off-line training unit is provided with a previous value storage unit for storing the previously determined spatial transfer characteristic in a rewritable memory, and the previously calculated spatial transfer characteristic is set to the next initial value storage unit. It may be used as a value. Instead of the starter delay circuit, a periodical instruction unit may be provided that periodically executes the offline training unit when the engine is stopped based on time information, ignition information, and noise level.

The off-line training section stores the previously obtained spatial transfer characteristic in a rewritable memory instead of the initial value storage section, and uses the previously obtained spatial transfer characteristic as the next initial value. A value storage unit and, instead of the starter delay circuit, a periodic instruction unit that periodically executes the offline training unit based on time information, ignition information, and noise level when the engine is stopped. Good.

Instead of the starter delay circuit, based on time information, ignition information, and noise level, a regular instructing section for periodically executing the off-line training section when the engine is stopped, and the off-line training section. The amplification multiplication unit and the attenuation multiplication unit, which are provided in the preceding stage and the subsequent stage, respectively, may be provided. Instead of the starter delay circuit, based on time information, ignition information, and noise level, a regular instruction unit that periodically executes the offline training unit when the engine is stopped, and a front stage of the offline training unit and Instead of the amplification multiplication unit and the attenuation multiplication unit respectively provided in the subsequent stage and the initial value storage unit of the offline training unit, the space transfer characteristics previously obtained are stored in a rewritable memory, and the space previously obtained is stored. A previous value storage unit that uses the transfer characteristic as the next initial value may be provided.

Further, the speaker is outputted to a noise control device having a speaker for outputting a cancel sound having a phase opposite to that of the noise and having the same amplitude and a microphone for detecting and feeding back an error signal from the residual sound of the canceled noise. In order to form a cancellation signal for the noise control section, a noise control section having a transfer characteristic simulating section for simulating a space transfer characteristic in which sound is transmitted through the space between the speaker and the microphone, and a nonvolatile memory for storing an initial value of the space transfer characteristic. A white noise reference signal is output from the speaker to the space and captured, the spatial transfer characteristic is obtained by learning from the captured signal using the initial value, and the space of the transfer characteristic simulating unit is acquired. A noise reduction effect based on the level of the error signal during the execution of the online training unit that updates the transfer characteristic and the noise control unit. The online training unit if it is determined that the better may comprise a determination unit to execute at the same time.

[0011]

According to the noise control device of the present invention, the spatial transfer characteristic is obtained by learning using the initial value at the time of product shipment, so that the learning time can be shortened significantly as compared with the case without this initial value. become. Therefore, it becomes possible to delay the timing of the start switch for starting the engine for a short time and execute the offline training unit within this delay time. That is, even when the number of products is large and regular maintenance is not guaranteed, it is possible to perform offline training against product variations and aging.

Instead of the initial value storage unit composed of a ROM, a previous value storage unit that stores the previously obtained space transfer characteristic in a rewritable memory improves the accuracy of the initial value and further shortens the learning time. Therefore, the delay time of the timing of the start switch can be shortened. Instead of the starter delay circuit, it becomes possible to regularly execute the offline training unit based on time information, ignition information, and noise level, and the range of application of offline training becomes wider. .

Since the learning time is shortened by providing the previous value storage unit and the periodical instruction unit together, the time for outputting the noise as the reference signal from the speaker can be shortened and the influence on the environment can be reduced, and at the same time. Wide application range.
The periodic instruction section and the amplification multiplication section and the attenuation multiplication section, which are provided in the front and rear stages of the off-line training section, respectively, are combined to widen the applicable range and output noise as a reference signal from the speaker. The level to be used can be reduced and the impact on the environment can be reduced.

By providing the periodic instruction section, the amplification multiplication section and the attenuation multiplication section, and the previous value storage section together, the applicable range is widened, and the level of noise as the reference signal is output from the speaker. The time can be shortened and the impact on the environment can be reduced. Further, while the noise control unit is executing, if it is determined from the level of the error signal that the noise reduction effect is good, the online training unit is also executed at the same time, so that the heat of exhaust gas or the like changes the sonic velocity, and the transmission system It is possible to respond to changes in the above and prevent the reduction of the noise reduction effect.

[0015]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing an example of a noise control device according to an embodiment of the present invention, which is installed in an automobile. As shown in the figure, the noise from the engine 1 of the automobile is transmitted through the exhaust pipe, the noise is reduced by the muffler, and the noise is discharged to the atmosphere. The muffler space is provided with a speaker 2 that outputs a canceling sound having an opposite phase and the same amplitude as the noise to further reduce the noise emitted to the atmosphere.

A microphone 3 for detecting a residual sound whose noise has been canceled is provided near an exhaust port through which the noise is discharged to the atmosphere, and is fed back as an error signal.
The control unit 4 that forms the cancel signal output to the speaker 2 inputs the error signal from the microphone 3 as a feedback signal. FIG. 2 is a diagram illustrating the configuration of the noise control device of FIG. As shown in the figure, the noise control device includes the speaker 2 and the microphone 3,
Further, a power amplifier 5 that drives the speaker 2 is provided.

The low pass filter 6 in the front stage of the power amplifier 5 removes unnecessary high frequency components. Low pass filter 6
D / A converter 7 (Digital to Analog Converte
r) converts a digital signal into an analog signal. An amplifier 8 is provided in the microphone 3, and a low-pass filter 9 in the latter stage of the amplifier 8 removes signal aliasing. The A / D converter 10 (Analog t
o Digital Converter) converts analog signals to digital signals. Switches 11 and 12 are provided at the front stage of the D / A converter 7 and the rear stage of the A / D converter, respectively. A noise control unit 4-1 and an off-line training unit 4-2 are provided between the switches 11 and 12, and these are selectively switched.

FIG. 3 is a diagram showing the configuration of the noise control section 4-1 shown in FIG. As shown in this figure, the noise control unit 4-1 is
FIR (Finite Impulse Res) with variable filter coefficient
PONSE) filter as an adaptive filter 41.
The adaptive filter 41 forms a cancel signal Sc (n) (n: indicates the nth sampling signal) and outputs it to the switch 11.

The coefficient updating unit 42 for updating the variable filter coefficient of the adaptive filter 41 forms the filter coefficient so as to minimize the level of the error signal from the switch 12. Assuming that the acoustic transfer characteristic Hd of the space between the speaker 2 and the microphone 3 is Hd, the first transfer characteristic simulating unit 43 that simulates the acoustic transfer characteristic Hd corrects the cancellation signal of the adaptive filter 41.

The adding section 44 adds the inverted signal of the output signal of the first transfer characteristic simulating section 43 and the error signal from the switch 12 to form the noise reproduction signal q (n), and the noise reproduction signal q (n ) Is input to the adaptive filter 41 and used as a controlled signal. Further, the second noise control signal q (n) is input.
The transfer characteristic simulating unit 45 has the same configuration as the first transfer characteristic simulating unit 43, and its correction signal is input to the coefficient updating unit 42. First transmission simulation unit 43 and second transmission simulation unit 4
5 is composed of an FIR filter, and its filter coefficient is
It is updated by the offline training unit 4-2 described later.

The relationship between the signals of the noise controller will be described. When the noise is Sn and the transfer function from the engine 1 which is the noise source to the microphone 3 is G, the signal Sm detected by the microphone 3 is as follows. Sm (n) = Sn (n) .G + Sc (n) .Hd (1) Since the filter coefficient of the adaptive filter 42 is changed so that the control signal Sm (n) becomes zero, the adaptive filter 42 The compensation signal Sc (n), which is the output signal of 121, is determined from the above equation (1) with Sm (n) = 0 as follows.

Sc (n) = − Sn (n) · G / Hd (2) The input signal q (n) of the adaptive filter 41 is q (n) = − Sc (n) + Sm (n) = ( -Sn (n) ・ G / Hd) ・
It is obtained as Hd + (Sn (n) * G + Sc (n) * Hd) = Sn (n) * G.

FIG. 3 is a diagram for explaining the filter coefficient of the adaptive filter 41. As shown in the figure, the filter coefficient Cp (n) of the adaptive filter 41 is updated by the coefficient updating unit 42 as follows. Cp (n + 1) = Cp (n) + α.q (np) .Sm (n) (p = 0,1,2,
…, K) where α is the convergence coefficient.

FIG. 5 shows the offline training unit 4 of FIG.
It is a figure which shows the structure of -2. As shown in the figure, the offline training section 4-2 has a noise generating means 51 for generating a reference signal. The noise generating means 51 generates white noise as the reference signal w (n) and outputs it to the switch 11. This reference signal w (n) is output from the speaker 2, propagates in the space, is distorted, and is received by the microphone 3 as w (n).
-It is captured as Hd. In this case, the engine 1 which is the noise source is stopped and the generation of noise from the engine is stopped.

The offline training section 4-2 has an adaptive filter 52 for realizing self-adaptiveness. The adaptive filter 52 is the adaptive filter 41 of FIG.
It is assumed that the reference signal w (n) from the noise generating means 51 is input as a controlled signal, processed, and output as w (n) .Hd '. The adding means 53 adds the signal w (n) .Hd from the switch 12 and the inverted signal -w (n) .Hd 'of the adaptive filter to obtain a difference signal w between them.
(n) (-Hd '+ Hd) is formed.

The coefficient updating unit 54 is provided with an initial value storage unit 55, but has the same configuration as the coefficient updating unit 42 of FIG. 3, and receives the reference signal w (n) from the noise generating unit 51 and receives the difference signal. The coefficient of the adaptive filter 52 is updated so that w (n) (-Hd '+ Hd) is minimized. In this way, the adaptive filter 52 performs the spatial transfer characteristic H by the above learning.
We obtain d '= Hd.

Since the initial value storage means 55 stores the filter coefficients of the first and second transfer characteristic simulating portions 43 and 45 of FIG. 3 measured at the time of shipping of the product, it is constituted by a non-volatile memory ROM. . When the coefficient updating unit 54 updates the coefficient using this initial value, the coefficient updating unit 54 takes about 30
Learning is required 00 times, and the learning time is about 1 second. As described above, if such an initial value is not used, it takes 3 to 4 seconds, so that the learning time can be greatly shortened.

Returning to FIG. 1, a starter delay circuit 110 for delaying the timing of supplying current from the battery 100 of the automobile to the starter 101 for starting the engine to drive the engine.
Will be explained. The starter delay circuit 110 includes a start switch (ST) 102 and a timer (TLR) 10 of about 1 second.
3 is provided, and a time-delayed operation contact (TLR-a) 104 is provided between the battery 100 and the starter 101.

Then, a voltage signal A between the start switch 102 and the timer 103 and an inverted signal of the voltage signal B of the time-delayed operation contact 104 and the starter 101 are inputted, and a logic for switching the switch 11 and the switch 12 by this output signal. A product circuit 105 is provided. FIG. 6 is a time chart explaining the operation of the starter delay circuit 110 of FIG. By the ON operation of the start switch 102 of FIG. 9A, the voltage signal A of FIG. 7C becomes “H (high)”.

The level of the voltage signal B in FIG. 3B becomes "H" after 1 second, and becomes "L" together with the level "L (low)" of the voltage signal A. Thus, the start switch 10
The starter 101 is driven 1 second after the ON operation of 2.
The level of the output signal shown in FIG.
2 becomes "H" only 1 second after the ON operation, and the switches 11 and 12 are switched to the off-line training section 4-2 side only during the period of this "H" signal.

Therefore, according to the present embodiment, the first and second transfer characteristic simulating units are updated for a short period of 1 second before the start switch 102 is turned on and the engine 1 is driven. Therefore, even if the number of products is large and regular maintenance is not guaranteed, it is possible to perform off-line training against product variations and aging without giving a feeling of strangeness at the start.

FIG. 7 is a diagram showing another configuration of the offline training section 4-2. As shown in the figure, the previous value storage unit 56 is provided in place of the initial value storage unit 55 in FIG. 5, and has an EEPROM (Electrically Erasab).
le Programmable ROM), adaptive filter 52
The filter coefficients formed by the above are stored for use in the next off-training. By doing so, the convergence time of the coefficient updating unit 54 is short, so the set time (1 second) of the timer 103 in FIG. 2 can be further shortened.

FIG. 8 is a diagram showing a second configuration of the noise control device of FIG. As shown in the figure, the periodical instruction unit 20
Are switches 11 and 12 instead of the relay circuit of FIG.
The instruction for periodically switching the switching is performed as follows. That is, the regular support unit 20 is
Timer, IGSw. , Is connected to the output of the A / D converter 10, and enables off-training to be performed regularly and automatically according to time information, engine stop information, and external noise level. It is possible to obtain the same effect as that of the first configuration.

The off-line training section 4-2 of the second configuration may be configured as shown in FIG. As a result, the time for offline training is shortened, so it is possible to perform offline training without worrying about the surroundings, even in the time to suppress noise such as at night or in a place such as hospital where noise is suppressed. System data can be acquired, and variations and changes over time can be handled.

FIG. 9 is a diagram showing a third configuration based on the noise control device of FIG. As shown in the figure, an amplification multiplication unit 21 for amplifying an input signal is provided in the front stage of the offline training unit 4-2, and an attenuation multiplication unit 22 for attenuating the output signal is provided in the subsequent stage. In this way, further offline training is possible without worrying about the surroundings.

FIG. 10 shows a fourth embodiment based on the noise control device of FIG.
It is a figure which shows the structure of. As shown in this figure, the online training 4-3 has the same configuration as the offline training 4-2 of FIG. The switches 31 and 32 are
It is provided instead of the switches 11 and 12 of FIG. 8, and is used in parallel with the noise control section 4-1 and the online training 4-3 under a certain condition.

Judgment unit 30 connected to A / D converter 10
Detects the degree of noise reduction of the noise control unit 4-1 from the level of the error signal, and the online training unit 4-3
Determine the execution of. That is, if the effect of noise reduction is good, online training is performed simultaneously with noise control. In this way, a good training effect can be obtained even when the speed of sound changes due to the heat of exhaust gas, etc., and thus the effect of noise reduction can be prevented.

[0038]

As described above, according to the present invention, the space transfer characteristic is obtained by learning using the initial value at the time of product shipment, so that the learning time is significantly shortened as compared with the case without this initial value. Will be done. Therefore, it becomes possible to delay the timing of the start switch for starting the engine for a short time and execute the offline training unit within this delay time. That is, even when the number of products is large and regular maintenance is not guaranteed, it is possible to perform offline training against product variations and aging.

[Brief description of drawings]

FIG. 1 is a diagram showing an example of a noise control device according to an embodiment of the present invention, which is mounted on an automobile.

FIG. 2 is a diagram illustrating a first configuration of the noise control device of FIG.

3 is a diagram showing a configuration of a control control unit 4-1 of FIG.

4 is a diagram illustrating filter coefficients of an adaptive filter 41 of FIG.

5 is a diagram showing a configuration of an offline training unit 4-2 of FIG.

6 is a time chart explaining the operation of the starter delay circuit 110 of FIG.

FIG. 7 is a diagram showing another configuration of an offline training unit 4-2 for vinegar 5.

FIG. 8 is a diagram showing a second configuration of the noise control device of FIG.

9 is a diagram showing a third configuration based on the noise control device of FIG.

10 is a diagram showing a fourth configuration based on the noise control device of FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Engine 2 ... Speaker 3 ... Microphone 4-1 ... Noise control part 4-2 ... Offline training part 4-3 ... Online training part 20 ... Regular instruction part 43, 45 ... Transfer characteristic simulation part 55 ... Initial value storage part 56 ... Previous value storage unit 100 ... Battery 101 ... Starter 110 ... Starter delay circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G10K 15/12 G11C 27/00 C H03H 17/02 L 8842-5J 21/00 8842-5J

Claims (7)

[Claims] 1. A noise control having a speaker (2) for outputting a canceling sound having an opposite phase and the same amplitude as the noise, and a microphone (3) for detecting and feeding back an error signal from the residual sound of the canceled noise. In the device, in order to form a cancel signal to be output to the speaker (2), a transfer characteristic simulating unit (43, which simulates a space transfer characteristic in which sound is transmitted through a space between the speaker (2) and the microphone (3). Noise control unit (4-1) having 45)
And a reference value of white noise is output from the speaker (2) to the space and captured, and an initial value storage unit (55) including a nonvolatile memory that stores the initial value of the space transfer characteristic is captured. From the signal, the spatial transfer characteristic is obtained by learning using the initial value, and the transfer characteristic simulation unit (43, 4
5) An offline training unit (4-2) for updating the space transfer characteristic and a start switch (102) for supplying a current from a battery (100) to drive a starter (101) for starting the engine (1). Is delayed for a short time, and the offline training unit (4
And a starter delay circuit (110) for executing -2). 2. The offline training unit (4-
In 2), instead of the initial value storage section (55), a rewritable memory is provided with a previous value storage section (56) for storing the previously obtained space transfer characteristic, and the previously obtained space transfer characteristic is The noise control device according to claim 1, wherein the noise control device is used as an initial value. 3. Instead of the starter delay circuit (110), the offline training section (4-2) is regularly performed based on time information, ignition information, and noise level when the engine (1) is stopped. The noise control device according to claim 1, further comprising a periodical instruction unit (20) for executing the following. 4. The offline training unit (4-
In 2), instead of the initial value storage unit (55), a previously writable space transfer characteristic is stored in a rewritable memory, and a previously stored value is stored to use the previously found space transfer characteristic as the next initial value. Instead of the section (56) and the starter delay circuit (110), the offline training section (4-2) is periodically performed based on time information, ignition information, and noise level when the engine (1) is stopped. ) The regular instruction part (2
0) and the noise control device according to claim 1. 5. The off-line training unit (4-2) is regularly used when the engine (1) is stopped based on time information, ignition information, and noise level instead of the starter delay circuit (110). And an amplification multiplication unit (21) and an attenuation multiplication unit (22) provided at the front stage and the rear stage of the offline training unit (4-2), respectively. The noise control device according to claim 1, wherein the noise control device is characterized. 6. Instead of the starter delay circuit (110), the offline training section (4-2) is regularly performed based on time information, ignition information, and noise level when the engine (1) is stopped. A regular instruction unit (20) for executing the above, an amplification multiplication unit (21) and an attenuation multiplication unit (22) respectively provided in a front stage and a rear stage of the offline training unit (4-2), and the offline training A previous value that stores the space transfer characteristic obtained last time in a rewritable memory instead of the initial value storage part (55) of the part (4-2) and uses the space transfer characteristic obtained last time as the next initial value. The noise control device according to claim 1, further comprising a storage unit (56). 7. A noise control having a speaker (2) for outputting a cancellation sound having a phase opposite to that of the noise and having the same amplitude, and a microphone (3) for detecting and feeding back an error signal from the residual sound of the canceled noise. In the device, in order to form a cancel signal to be output to the speaker (2), a transfer characteristic simulating unit (43, which simulates a space transfer characteristic in which sound is transmitted through a space between the speaker (2) and the microphone (3). Noise control unit (4-1) having 45)
And a non-volatile memory (55) for storing an initial value of the space transfer characteristic, outputting a reference signal of white noise from the speaker (2) to the space and capturing the white noise reference signal. The online training unit (4-3) for obtaining the spatial transfer characteristic by learning and updating the spatial transfer characteristic of the transfer characteristic simulating unit (43, 45) and the noise control unit (4-1) are executed. A noise control device comprising: a determination unit (30) which also executes the online training unit (4-3) at the same time when it is determined that the noise reduction effect is good from the level of the error signal.