WO1992022054A1

WO1992022054A1 - Noise control apparatus

Info

Publication number: WO1992022054A1
Application number: PCT/JP1992/000680
Authority: WO
Inventors: Masaaki Nagami; Kazuya Sako
Original assignee: Fujitsu Ten Limited
Priority date: 1991-05-30
Filing date: 1992-05-26
Publication date: 1992-12-10
Also published as: EP0598120A4; CA2086926C; EP0598120B1; EP0598120A1; CA2086926A1; DE69227252D1; DE69227252T2

Abstract

A noise control apparatus for eliminating noise capable of responding to abruptly changing noise frequency. The apparatus includes adaptive filtering means (6) for receiving a noise signal and controlling a filter coefficient on the basis on an error signal of an acoustoelectrical convertor (2) to produce a compensation signal for eliminating the noise to the convertor (3); transmission characteristics simulation means (4) for simulating transmission characteristics of a closed-loop system including the adaptive filtering means (6) and the acoustoelectrical convertor (3); difference signal operation means (5) for generating a noise signal from the difference between the error signal of the convertor (2) and the compensation signal sent from the adaptive filtering means (6) through the transmission characteristics simulation means (4); cycle detection unit (7) for detecting a noise cycle of a noise generation source (1), cycle adjustment unit (8) for changing the cycle of the noise signal of the difference signal operation means (5) in accordance with the change of the noise cycle, and cycle detection control means (10) for changing the filter coefficient in accordance with an estimated quantity of the noise cycle.

Description

Document noise control device

[Field of technology]

The present invention eliminates noise by outputting a signal having the same phase and the same sound pressure as the noise detected by the microphone from the speaker.Especially, in the present invention, the noise frequency changes rapidly. The present invention relates to a noise control device capable of following even when the noise control is performed.

(Background technology)

Conventionally, passive silencers such as mufflers have been used to reduce noise generated from internal combustion engines, etc., but improvements have been desired from the viewpoints of size, silencing characteristics, and the like.

On the other hand, an active noise control device that cancels the noise by outputting a complementary sound having the same phase and the same sound pressure as the noise generated from the sound source from the speaker has been proposed.

However, the frequency characteristics and stability of the active noise control device itself were not sufficient, and practical application was delayed.

In recent years, the development of signal processing technology using digital circuits has expanded the frequency range that can be handled, and as a result, many practical noise control devices have been proposed (for example, see Japanese Patent Application Laid-Open No. 63-3 / 1988). 1 1396).

The noise is detected by a microphone for noise source installed upstream of the duct, and the signal processing circuit sends a signal of opposite phase and equal sound pressure to the noise. The signal is output from a speaker installed downstream of the duct, and the residual sound resulting from the silencing is detected by microphone and fed back. This is a so-called 2-microphone 1-speaker type active noise control device that combines a voice system.

On the other hand, in order to obtain a noise reduction effect in a space where the noise source is unclear such as in the interior of a car, for example, only a feedback system that does not require the installation of a microphone in the noise source is used. O One microphone mouth phone used · It must be a device with one speaker configuration o

However, in an active noise control system consisting of one microphone '1 speaker with only a feedback system, if the noise period of the noise source changes sharply, the One of the drawbacks of the system is that it delays by at least the sound transmission characteristic from the speaker to the microphone, so that the noise reduction effect is reduced.

Accordingly, it is an object of the present invention to provide a sound control device capable of following a steep change in a noise cycle in view of the above problems.

[Disclosure of the Invention]

FIG. 1 is a diagram showing a first principle configuration of the present invention. In order to solve the above problems, the noise control device according to the present invention includes a sound / electric signal converter 2 for detecting noise and converting it into an electric signal, and an electric signal for outputting supplementary sound waves for eliminating noise. Sound wave transducer 3, a transfer characteristic simulation means 4, a difference signal calculation means 5, an adaptive filtering means 6, a cycle detection section 7, and a cycle adjustment section 8. The difference signal between the output of the filter 2 and the output of the adaptive filtering means 6 is calculated.

The transfer characteristic simulating means 4 is inserted between the adaptive filtering means 6 and the difference signal calculating means 5, and receives the electric signal / sound wave converter 3 and the electric signal / sonic wave from the adaptive filtering means 6. Simulates the transfer characteristics of the system that returns through the sound wave and electrical signal converter 2.

The cycle detector 7 detects the noise cycle of the noise source 1. The cycle adjuster 8 changes the cycle of the output signal of the difference signal calculator 5 in accordance with the amount of change in the noise cycle. The adaptive filtering means 6 is for outputting a supplementary sound wave to the electric signal / sound wave converter 3 based on the output signal means of the period adjusting unit 8 and the output of the sound wave / electric signal converter 2. Calculate the complement signal. The adaptive filtering means 6 may directly input a signal obtained by adjusting the period of the noise signal from the noise source, and in this case, the transfer characteristic simulating means 4 and The difference signal calculation means 5 is omitted.

According to the noise period control device shown in FIG. 1, the output of the transfer characteristic simulating means 4 and the output of the sound wave-to-electric signal converter 2 are converted into a difference signal by the difference signal calculating means 5 to form a reproduced noise signal. The amplitude and phase are adjusted by the adaptive filtering means 6 for inputting the input signal, and the supplementary signal causes the electric signal / sound wave converter 3 to output a supplementary sound wave to mute the noise. Further, the cycle detection unit 7 detects the noise cycle and monitors the fluctuation of the noise cycle. The period adjuster 8 adjusts the period of the output signal of the difference signal calculating means 5, that is, the period of the input signal of the adaptive filtering means 6, according to the fluctuation of the noise period. (4) The period of the sound wave coincides with the period of the noise at the silencing point. Therefore, even if the noise cycle changes sharply, it can follow.

FIG. 2 is a diagram showing a second principle configuration of the present invention. In order to solve the above-mentioned problems, the present invention provides an electric signal for eliminating noise from a noise source 1, a sound wave converter 3, and a residual sound eliminated by sound waves from the electric signal / sound wave converter 3. A sound wave for converting sound into an electric signal as an error signal, an electric signal converter 2, and for eliminating noise to the electric signal, sound wave converter 3 based on the signal of the sound wave, electric signal converter 2. The filter characteristics of the adaptive filtering means 6 are changed according to the predicted change of the noise period to the noise control device having the adaptive filtering means 6 for forming the supplementary signal of the noise. Cycle detection control means 10 is provided.

The cycle detection control means 10 detects the noise cycle of the noise source 1, predicts a change in the noise cycle, and includes the adaptive filtering means 6 in accordance with the predicted change in the noise cycle. Update the multiplication coefficients set in the multiple multipliers. Further, the cycle detection control means 10 detects a noise cycle in the noise generation source 1, predicts a change in the noise cycle, and according to the predicted change in the noise cycle, the adaptive filter. The taps of a plurality of delay units included in the switching means 6 may be moved.

Further, the period detection control means 10 is adapted to be adapted to the adaptive filter. The multiplication coefficients of the plurality of multipliers included in the ring 6 are formed into a multidimensional vector to detect and predict a change in the vector and to perform a prediction according to the predicted change of the vector. The multiplication coefficients of the plurality of multipliers may be updated and set.

According to the noise control device shown in FIG. 2, an adaptive filter that inputs this noise signal based on the difference signal between the noise from the noise source 1 and the sound wave with the opposite phase equal sound pressure from the speaker 3 The compensation signal of the ring means 6 is adjusted for the amplitude and the phase, and the noise is eliminated. Further, when the noise cycle changes suddenly, the change in the noise cycle is detected by the cycle detection means, and the noise cycle is determined based on the transmission characteristics to the sound-absorbing point via the electric signal, the sound wave converter 3, etc. By predicting the change, the multiplication coefficients of a plurality of multipliers constituting the adaptive filtering means 6 are shift-controlled, and the period of the electric signal and the sound wave used from the sound wave converter 3 is at the sound deadening point. It will coincide with the period of the noise. Therefore, it is possible to follow even if the noise cycle changes sharply.The same function can be obtained even if the tap of the delay unit of the adaptive filtering means 6 is moved by the cycle detection control means 10. Can be

Further, the period detection control means 10 vectorizes the multiplication coefficient of the multiplier of the adaptive filtering 6, and the vector change is closely related to the noise period. By predicting the vector change, the noise period can be easily predicted, and by considering the transfer characteristics, even if the period change is steep, the period of the supplementary sound wave can be matched at the sound deadening point. And become possible. [Brief description of drawings]

FIG. 1 is a diagram showing a first principle configuration of the present invention.

FIG. 2 is a diagram showing a second principle configuration of the present invention.

FIG. 3 is a diagram showing a noise period control device according to the first embodiment of the present invention.

FIG. 4 is a diagram for explaining a cycle detection method of the cycle detector of FIG.

FIG. 5 is a diagram showing a configuration of the cycle adjusting unit in FIG.

FIG. 6 is a diagram showing a relationship between input and output signals of the cycle adjusting unit in FIG.

FIG. 7 is a diagram showing the relationship between the period change amount and its control amount calculation.

FIG. 8 is a diagram for explaining the function of the delay amount control unit.

FIG. 9 is a diagram showing a noise period control device according to a second embodiment of the present invention.

FIG. 10 is a diagram showing a noise cycle control device according to a third embodiment of the present invention.

FIG. 11 is a diagram showing a noise period control device according to a fourth embodiment of the present invention.

FIG. 12 is a diagram showing a noise control device according to a fifth embodiment of the present invention.

FIG. 13 is a diagram showing a configuration of the cycle detection control means of FIG.

FIG. 14 is a diagram for explaining a cycle detection method of the cycle detector of FIG. FIG. 15 is a diagram for explaining a method of predicting the amount of periodic change. FIG. 16 is a diagram showing the adaptive filtering means of FIG.

FIG. 17 is a diagram for explaining the shift of the multiplication coefficient of a plurality of multipliers constituting the adaptive filtering means.

FIG. 18 is a diagram for explaining the tap movement of a plurality of delay units constituting the adaptive filtering means.

FIG. 19 is a diagram showing another modification of the cycle detection control means of FIG.

[Best mode for realizing the invention]

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 3 is a diagram showing a noise period control device according to the first embodiment of the present invention. The configuration of this drawing will be described. This figure shows a noise source 1 such as a car engine or motor, and a sound wave from the noise source 1 propagates, cancels the noise near the sound deadening point and is captured as a residual sound, and the electric signal is captured. Microphone 2 for converting to noise, speaker 3 for outputting supplementary sound waves for eliminating noise near the sound deadening point, and speaker 3 and microphone 3 from adaptive filtering means 6. A transfer characteristic simulating means 4 for simulating a transfer characteristic of a system from the microphone 2 to the difference signal calculating means 5, and an output of the microphone 2 and an output of the transfer characteristic simulating means 4. A difference signal calculating means 5 for calculating a difference signal; an adaptive filtering means 6 for calculating a compensation signal for outputting a compensation sound wave from the speaker 3 based on the calculation result of the difference signal calculating means 5; Cycle for detecting the noise cycle of the noise source 1 A detection unit 7, a period adjustment unit 8 that changes the period of the input signal of the adaptive filtering unit 6 according to the amount of change in the noise period, and an amplifier 101 of the microfin 2 (Analog To Digital Converter) AZD converter 102 that digitizes the output of amplifier 101 and outputs it to difference signal calculating means 5, and the output of adaptive filtering means 6 (D / A to digital converter) 103 that converts the analog signal into an analog signal, and an amplifier 104 that amplifies the output of the DZA converter 103 and outputs the amplified signal to the speaker 3. including. The adaptive filtering means 6 may be composed of a bandpass filter and an amplifier.

Further, the transfer characteristic simulating means 4, the difference signal calculating means 5, the adaptive filtering means 6, the cycle detecting section 7 and the cycle adjusting section 8 may be constituted by DSP (DigitalSinalProcessor).

FIG. 4 is a diagram for explaining a cycle detection method of the cycle detector of FIG. This figure (a) shows a method for detecting the rotation timing of a vehicle engine or motor as the sound source 1. A rectangular wave signal is input to the input の of the cycle detector 7, the cycle T is determined, and the output T is output to the cycle adjuster 8. This is because in the sound of a car, the sudden change in noise is caused by changes in the speed of the car engine.

In this figure (b), if the timing signal as shown in this figure (a) cannot be obtained, a microphone is installed near the car engine and the noise waveform is detected. Noise from the time waveform peak This shows that the signal period T is obtained. In this signal processing, a rectangular wave is generated when the noise signal level exceeds a certain level, and this rectangular wave is input to the period detection unit 7, so that the period T Is obtained.

This figure (c) shows the BPF (B and Pass Fi 1 ter) peak detection method for obtaining the noise period T after digitizing the noise signal input to the microphone. A plurality of node path filters 1, 2,…, n, and an absolute value unit (ABS) connected to each node path filter 1, 2,…, n It consists of an averaging unit (LPF) connected to each absolute value unit, and a maximum band detection unit that detects the maximum value of each averaging unit, and detects the maximum frequency band of the noise level, The cycle of the maximum frequency band is used as the cycle of the noise signal.

This figure (d) shows a period detection method using an adaptive filter, in which a delay signal (D e 1 ay) for inputting the difference signal of the difference signal calculating means 5 and an output of the delay signal are input. An adaptive filter (ADF), an adder for obtaining a difference signal between an output of the adaptive filter and a through input signal, and a least square method processing on the difference signal of the adder, to obtain an adaptive filter It consists of a least-squares method processing unit (LMS) that determines filter coefficients, and determines the period of the noise signal from the fixed coefficients of the adaptive filter.

FIG. 5 is a diagram showing a configuration of the cycle adjusting unit of FIG. The cycle adjusting unit 8 in the figure receives the difference signal of the difference signal calculating means 5, has M delay taps, and outputs a delay from the delay point to the adaptive filtering means 6 Memory 8 1 and delay memory 8 1 The delay amount control unit 82 that controls the delay amount by moving the delay point of the same, the period change amount detection unit 83 that detects the period change amount from the period data from the period detection unit 7, and the period change amount FIG. 6 is a diagram showing the relationship between the input and output signals of the cycle adjusting unit in FIG. This figure (a) shows that the period of the delay memory 81 input signal is T⑧, and this figure (t>) shows that the period of the delay memory output signal is T④.

FIG. 7 is a diagram showing the relationship between the period change amount and its control amount calculation. As shown in this figure, if the period changes so as to become smaller at the time (to) at which the initial period is constant, the period change detector 83 detects the period change as shown by ② in the figure. On the other hand, in the conventional technology, the position of the microphone mouth phone 2 is delayed by the transmission characteristic Hd as shown in the figure (2). Here, for the sake of simplicity, the transfer characteristics of the signal processing unit such as the adaptive filtering means 6 are ignored. The control amount calculating unit 84 calculates data for changing the cycle earlier as shown by the curve in the figure with respect to the curve in the figure, taking the transfer characteristic H d into consideration. In this figure, the change in the period is shown as a straight line with respect to time, but this may be a curve. In this case, the curve in the figure ④ is provided with a function, which is determined by fitting. You may. In the curve of Fig. 2 obtained in this way, the predicted period T④ for the period T③ of the current time (tt) is obtained.

FIG. 8 is a diagram for explaining a delay amount control unit. In this figure, the delay memory 81 is the input signal data at a fixed sampling period. The input signal cycle T _in and the output signal cycle T. _u, are translated displayed in ye number of taps, the period T input signal or et al cycle of _in T. _The delay controller 82 moves the delay point at a certain speed V in order to obtain an output signal of _u ,. In the figure, A represents the tap speed V in the case of the absolute amount of change. The input signal cycle T _itl = 30 _taps and the output signal cycle T. _{To make ut} = _{29 taps,} it is shown that the taps are moved in the direction of the input side at the speed of V = 1 taps / 29 samples as shown in the figure. T. _ut = _{28 taps} to get V = 2 taps Z 28 samples, T. _{To make u} = 27 taps V = 3 taps 27 samples,…, T. _u , = 15 taps to make V = l5 taps 15 samples, T. _{To make ut} = 14 V-16 taps 14 samples…, generally, the input signal period T i „is the output signal period T. To make _ut = T _in 1 n, V = n (T _; „-N), n: It is sufficient to set the period shift amount.

In B in the figure, the movement of the delay amount control unit is described for the rate of change, and the input signal period T _in = 30 taps and the output signal period T. _ul = (9/10) X 30 To make taps, tap speed V = 1 Z 9 Tap samples and T. _u , = (8/10) 30 = taps V = 2/8 taps / sample,…, T. _ut = ( _5/10 ) X 30 To make a tap V = 5 no 5 tap "sample, T. _u , = (4/10) X 30 To make a tap V = 6/4 taps / sample,…, generally, input signal period T _in is output signal period T. To make _u , = (k / 10) x T _in , V = (1 0 — k ) ZK, k / 10: Period shift ratio, and Just do it.

Next, the adaptive filtering means 6 will be briefly described. Strictly speaking, the transfer characteristics of electric signals must be considered, but they are not directly related to the present invention, and are ignored for simplicity of description. . And noise S _N of the noise source 1, the transfer characteristic up microphone B off O emissions 2 and H _N0ISE, the complement No.儐信adaptive filter-ring means 6 and S _c, adaptive Fi The transfer characteristic of the system from the filtering means 6 to the difference signal calculating means 5 via the speaker 3 and the microphone 2 is defined as H d. The transfer characteristic H dl of the transfer characteristic simulating means 4 is If H dl = H d, the microphone

The signal S _M output from 2 is S _M = SN-HNOISE + S c

Hd. Accordingly, the difference signal S _E is a calculation result of the difference calculation unit _{5, S E = SM - S c} · H d 1 = SM - S c - H d = S N * H N. _{I SE,} and the next and this for calculating a signal when detect only noise at microphone b off O emissions 2, enter the difference signal S _E to the suitable応型off I filter-ring means 6, S _M = The supplementary signal Sc is calculated so that it becomes zero.

FIG. 9 is a diagram showing a noise cycle control device according to a second embodiment of the present invention. The configuration of this drawing differs from that of the first embodiment in FIG. 3 in that the period detection unit 7 does not receive the signal for detecting the period from the noise source 1 and is shared with the period adjustment unit 8. The difference signal from the feedback signal from the difference signal calculation means 5 is input. This is because the control amount calculation unit 84 of the period adjustment unit 8 has a function of predicting a change in the period, and the delay amount control unit 82 uses the output of the period adjustment unit 8 to output the speaker 3. Through my crophone The reason for this is that the compensated tone corresponding to the preceding cycle and can be reproduced by the delay corresponding to H d with the transfer characteristic up to the silencing point of No. 2.

FIG. 10 is a diagram showing a noise cycle control device according to a third embodiment of the present invention. If the configuration of this diagram is different from that of the first embodiment in Fig. 2, it is connected to Microphone 105, which collects the noise signal directly from Noise Source 1, and Microphone 105. AZD converter 107 connected to amplifier 106, which is connected to amplifier 106, and serves as input of period adjustment unit 8, and either output of AZD converter 107 or difference signal calculating means 5 A switch unit 108 which selects one of them as an alternative and is connected as an input of the period detection unit 7. That is, even if the period adjustment unit 8 directly receives the noise signal from the noise source 1 and the period detection unit 7 receives the AZD converter 107 or the difference signal calculation unit 5, the same applies as described above. The operation and effect of the invention can be obtained.

FIG. 11 is a diagram showing a noise period control device according to a fourth embodiment of the present invention. The configuration of this diagram differs from that of the third embodiment in FIG. 9 in that the period detector 7 receives the timing signal of the noise source 1 as an input. In this configuration, the same operation and effect as described above can be obtained.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 12 is a diagram showing a noise control device according to a fifth embodiment of the present invention. The configuration of this drawing will be described.

The noise control device shown in this figure is a microphone for converting the noise from the noise source 1 such as the engine of a car into the electric signal by converting the residual sound of the noise, which is eliminated by the sound wave from the speaker 3 described below, into an electric signal. Amplifier for amplifying the electric signal of the microphone 2 and the microphone mouth 2 101, an A / D converter 102 for converting the analog signal of the amplifier 101 into a digital signal, and canceling the noise from the noise source 1 near a silencing point P (in the figure). , An amplifier 104 for amplifying the output to the speaker 3, and a DZA for converting a digital signal to an analog signal in order to supply an analog signal to the amplifier 104. An adaptive filter for controlling a filter coefficient based on a signal from the converter 103 and the AZD converter 102 to form an auxiliary signal for eliminating noise to the speaker 3; Filtering means 6 and a timing signal from the noise source 1 are input, and a sound signal from a microphone 105 or the like described later, or a noise from a difference signal calculating means 5 or the like is described later. Input the reproduction signal, detect the noise period, predict the period change, and A period detection control means 10 for controlling the adaptive filtering means 6 so as to be able to follow a steep change; a microphone 105 provided near the noise generating source 1; An amplifier 106 for amplifying the output of the clone 105, an A / D converter 107 for converting an analog output signal of the amplifier 106 to a digital signal, and the adaptive type amplifier. Simulated by the output of the filtering means 6, the transfer characteristic Hd from the output point to the input of the difference signal calculation means 5 to be described later via the speaker 3 and the microphone 2 Transfer characteristic simulating means 4, a difference signal calculating means 5 for calculating a difference signal between the output of the transfer characteristic simulating means 4 and the output of the A / D converter 102, and the adaptive filter And switch means 11 for selectively selecting an input signal of the ring means 6. Here, adaptive filtering means 6, period detection control means 10 Etc. are composed of DSP.

FIG. 13 is a diagram showing the configuration of the cycle detection control means of FIG. The cycle detection control means 5 shown in the figure comprises a cycle detection section 1001, a cycle prediction section 1002, and a control section 1003 for controlling the coefficients of the adaptive filtering means 6. .

FIG. 14 is a view for explaining an example of the cycle detection method of the cycle detector 1001 in FIG. This figure (a) shows the method of detecting the ignition timing of a vehicle engine or a motor or the tilling timing (rotation speed) as the noise source 1. A rectangular wave signal is input to the input of the period detection unit 1001, and the period T is obtained and output to the period prediction unit 1002. This is because a steep change in vehicle noise is caused by changes in the number of revolutions of the vehicle engine.

This figure (b) shows the noise waveform from the microphone or vibrometer 9 near the car engine when the timing signal as shown in this figure (a) cannot be obtained. Is detected, and the period T of the noise signal is obtained from the time waveform peak. In this signal processing, when a certain noise signal level exceeds a certain level of the noise signal level, a period T can be obtained in the same manner as in FIG.

This figure (c) shows the BPF peak detection method for obtaining the noise period T after digitizing the noise signal input to the microphone. In this method, a plurality of bandpass filters 1, 2,..., N, an absolute value unit (ABS) connected to each bandpass filter 1, 2,. Averaging section connected to absolute value section (LPF) and a maximum band detector that detects the maximum value of each averaging unit.The maximum frequency band of the noise level is detected, and the period of the maximum frequency band is used as the period of the noise signal. is there. This figure (d) shows shall apply in the period detecting method using an adaptive full I filter, a delay unit for inputting the difference signal S _R of the differential signal calculation means 5 and (D e 1 ay), the output of the delay device An adaptive filter (ADF) to be input, an adder for obtaining a difference signal between the output of the adaptive filter and a through input signal, and a least square method processing of the difference signal of the adder for adaptive processing. It consists of a least-squares method processing unit (LMS) that determines the coefficients of the type filter, and determines the period of the noise signal from the coefficients of the adaptive filter.

FIG. 15 is a diagram for explaining a method of predicting a period change amount based on a detection period. If the period prediction unit 1002 changes so that the period becomes smaller at the time (to) when the initial period is constant as shown in the figure, the period detector 1001 The cyclic change is detected as in ①. On the other hand, in the conventional technique, the position of microphone 2 is delayed by the transfer characteristic Hd as shown in the figure. Here, for the sake of simplicity, the transfer characteristics and the like of the signal processing unit such as the adaptive filtering means 4 are ignored. The cycle prediction unit 1002 calculates the data for changing the cycle earlier as shown by the curve ③ in the figure with respect to the curve ① in the figure in consideration of the transfer characteristic H d. In this figure, the change in the period is shown by a straight line with respect to time, but this may be a curve. In this case, the curve ③ in the figure is provided with a function, which is determined by fitting. Is also good. In the curve of Fig. 3 obtained in this way, the current Time (t,) the period T of the cycle T ₂ expected for sought. The control unit 1003 for the ADF coefficient and the like in FIG. 13 will be described later.

Next, the adaptive filtering means 6 will be briefly described. When the switch means 11 selects the difference signal calculation means 5, the noise _SN of the noise source 1 is set as H _{N0 1 SE,} and the transmission characteristic of the noise source 1 to the microphone 2 is set as H _{N0 1 SE.} the complement No.儐信the full i filter Li in g unit 6 and S _c, the transfer characteristics of the system leading to adaptive full i filter-ring means 6 crows Ma and through the speaker 3 i click b off O emissions 2 and H _SP, the transfer characteristics of the system ranging from the microstrip click port off O emissions 2 to differential signal calculation means 5 and H _{ra ics,} the transfer characteristic H d 1 of the transfer characteristics simulating means 4, H d 1 = H When _{SP · H mi c = H d} , the signal S _M of microphones b off O emissions 2 or we output the residual sound, S _M = SN 'HNOISE +

S c · Η, _Ρ . Therefore, the difference signal S _{R as the} calculation result in the difference calculation unit 5 is given by S _R = S _M · H _MI- 1 SC · H d 1 =

SN * HN 0 ISE * H m ί c + SC * HSP * 1 mic - SC * HSP * H mi c = S N - HNOISE ' Ri Do and H _mi c, detect only noise at microphone port off O emissions 2 The signal at this time is calculated. Further, as a control signal for changing the coefficients of the adaptive full I filter of the adaptive full I filter-ring unit 6 is given the output S _E of AZD converter 1 0 2. The adaptive plastic filtering means 6 changes the coefficient so that this control signal becomes zero, and S _E = SM

Since H _mi c, S _M = 0 when S _E = 0. Therefore, it inputs the difference signal S _R from the differential signal calculation means 5 in the adaptive full I filter-ring means 6 and the controlled signal, AZD varying as the control signal I'm in and the child to enter the exchanger 1 0 2 of the output S _E, adaptive off I filter-ring means 6 is arithmetic complement僙信No. S _c in such a way that S _E = 0. Sweep rate pitch means 1 1 mic B off O Selecting emissions 9 microphone B off O adaptive to input signals from the down 9 off I filter Li in g unit 6 calculates a complement儍信No. S _c I do.

FIG. 16 is a diagram showing the adaptive filtering means of FIG. The adaptive filtering means 6 in the figure is constituted by a non-recursive filter, and more specifically, a series of delay units 60 1 for delaying one sampling period, and each of the delay units 6 0 1 , A plurality of adders 603 for adding the outputs of the multipliers 602, and an output of the microphone 2 to the least squares method. And coefficient updating means 604 for controlling the multiplication coefficient of each of the multipliers 602 so as to minimize the above.

Note that the series of delay units 601 may be constituted by random access memory (RAM). In this case, the input sampling data is shifted to the next address sequentially for each sampling. The address value for inputting the sampling data may be shifted sequentially for each sampling.

For the multiplication coefficient, g ₂ , '·, g „of the multiplier 60 2 of the adaptive filtering means 6 shown in the figure, the control unit 100 of the ADF coefficient etc. of the cycle detection control means 10 3 Explain the reconfiguration ο

FIG. 17 is a diagram for explaining a shift of a multiplication coefficient of a plurality of multipliers constituting the adaptive filtering. This figure (a) schematically shows a coefficient sequence of the multiplier 602. Normally micro Multiplication factor of off O emissions each multiplier 6 Ri by the signal 2 0 2 (g ,, g 2 , ... g n) but is set, from the short cycle in the cycle prediction 1 0 0 2 to longer periods When the change is predicted, the multiplication coefficients (gi, g 2,

… Gn) force, (g '., G 1, 2,…, π-1),…, ('-g '-7,…,' 0, 1, g 2,…, g _n -S) , That is, shifted toward the n-th multiplier (delay unit). As a result, the delay amount becomes longer, and the period can be made longer.

In the figure (b), conversely, when a change from a long period to a short period is predicted by the period prediction unit 1002, each multiplication is performed by the control unit 1003 such as the ADF coefficient. Coefficient of the unit 60 2 (g:, g 2

…, Gn) force, '(g 2, g S.…, Gn, g', ₊ l), ···,

(g 1 0, g 1 1, ···, g „, g '„ ₊ 1,' _{n + 2} , ···, g '» ₊ s). (Delay device). As a result, the amount of delay is shortened, and the period can be shortened. Here, g 'is an arbitrary optimal value (for example, 0).

FIG. 18 is a diagram for explaining the tap movement of the delay device constituting the adaptive filtering means. This figure (a) is usually a delay unit

Although the tap (T,, T 2,…, T _n ) of ₆₀₀₁ is set, when the period prediction unit 1002 predicts a change from a short period to a long period, the ADF The taps (T,, T2,…, _Tn ) are changed to (T '., T1, T2,… -T…),. T '.,…, T'-], T'0, T !, T 2,…, Τ „-9),…, that is, toward the η-th delay unit, thereby increasing the amount of delay and extending the period.

In the figure (b), when the change from the long cycle to the short cycle is predicted by the cycle prediction section 1002, the delay of the ADF coefficient etc. is controlled by the control section 1003. vessel 6 0 1 of data-up (Τ, T 2, -, T n!) force _{"(: Τ 2, Τ 3} , ···, τ η, Τ 'η + ι), · ■ · (Τ 10 ,

Τ, ·, Τ „, Τ 'η ₊ Ι, Τ' _{π + 2} , '·,' _π ),…, that is, shifted toward the 0th multiplier. As a result, the amount of delay can be shortened and the period can be shortened, where Τ 'is an arbitrary maximum value (for example, 0).

FIG. 19 is a diagram showing another modification of the cycle detection control means of FIG. The cycle detection unit 1001 of the cycle detection control means 10 receives the multiplication coefficient of the multiplier 62 of the adaptive filtering means 6 and forms the following η-dimensional vector.

V (t) = gi (t) · g ₂ (t) 'i ₂ + ... 10 g „(t)' i _{n The} adaptive filtering means 6 sequentially draws (a), (b),

Since the multiplication coefficients (gi, g2,..., _Gn ) are updated as shown in (C), the period prediction unit 502 generates a vector such as t = 0, 1, 2, 2,. The vector after time t is predicted by tracking the torque, and when this prediction is made, the multiplication coefficient (g,, g2,…, g ») is calculated from this vector, and the ADF coefficient etc. These multiplication coefficients are set in the multiplier 62 by the control unit 1003 of the control unit. this As described above, by changing the multiplication coefficient of the multiplier 602 included in the adaptive filtering means 6, or by moving the output tap of the delay unit 601. Thus, the filter characteristics of the adaptive filtering means 6 can be changed.

As described above, according to the present invention, the noise period of the noise source is detected, and the period is controlled by predicting ahead from the characteristics of the noise period, so that it can follow a steep frequency change. It is now possible.

(Industrial applications)

Its application can be found in digital signal processing devices for canceling noise from engines and motors.

Claims

The scope of the claims

1. In a noise control device that outputs a compensatory sound wave of opposite phase and equal sound pressure to noise from a noise source (1) that generates noise and cancels the noise,

A sound wave * electric signal converter (2) for detecting a residual sound of the noise and the captured sound wave and converting the residual sound to an electric signal of an error signal; and an electric signal / sound wave converter for outputting the supplementary sound wave. An adaptive filtering means (6) for updating a filter coefficient with the error signal to form an auxiliary signal to obtain the auxiliary sound wave;

It is provided on the output side of the corresponding type filtering means (6), and outputs from the output through the electric signal / sound wave converter (3) and the sound wave electric signal converter (2) as the error signal. A transfer characteristic simulating means (4) for simulating the transfer characteristic of the system until returning;

The difference signal between the adaptive filtering means (6) and the supplementary signal passed through the transfer characteristic simulation means (4) and the error signal from the sound wave / electric signal converter (2) is calculated. Difference signal calculating means (5) for forming a reproduced noise signal

A period detector (7) for directly detecting a noise period from the noise source (1);

A period adjustment unit (5) that changes the period of the output signal of the difference signal calculation unit (5) according to the amount of change in the sound period, and outputs it as a reference signal to the adaptive filter unit (6). 8) A noise control device characterized by comprising:

2. The noise control device according to claim 1, wherein the cycle detection section (7) detects a noise cycle from a reproduced noise signal of the difference signal calculation means (5).

3. In a noise control device that outputs complementary sound waves with the same phase and opposite sound pressure as the noise from the noise source (1) that generates noise, and cancels the noise,

A sound / electric signal converter (2) for detecting a residual sound of the noise and the supplementary sound wave and converting the residual sound to an electric signal of an error signal;

An electric signal / sound wave converter (3) for outputting the captured sound wave; and an adaptive filter for forming a supplementary signal by updating a filter coefficient with the error signal to obtain the supplementary sound wave. Ring means (6);

Detecting the noise period of the noise source (1), predicting a change in the noise period, and filtering characteristics of the adaptive filtering means (6) according to the predicted change in the noise period. A noise control device characterized by comprising a period detection control means (10) for changing the noise.

4. The cycle detection control means (10) detects the J8 sound cycle of the noise source (1), predicts a change in the noise cycle, and, in accordance with the predicted change in the noise cycle, the adaptive filter. 4. The noise control device according to claim 3, wherein the multiplication coefficient of the multiplier included in the filtering means (6) is updated and set.

5. The cycle detection control means (10) detects a noise cycle of the noise source (1) and predicts a change in the noise cycle, The noise control according to claim 3, wherein an output tap of a delay unit included in the adaptive filtering means (6) is moved in accordance with the predicted change of the noise cycle. apparatus.

6. The cycle detection control means (10) forms a multidimensional vector by the multiplication coefficients of the plurality of multipliers included in the adaptive filtering (6), and 5. The noise control device according to claim 4, wherein a change in torque is detected and predicted, and a multiplication coefficient of the plurality of multipliers is updated and set in accordance with the predicted change in the vector.