JPH03204354A - Active type noise control device - Google Patents

Abstract

PURPOSE:To exhibit a satisfactory noise reducing effect by forming the noise correlation detecting signal of an internal combustion engine or its multiplied signal as a sampling clock, and conducting operations such as A/D conversion, D/A conversion and filter processing by use of the sampling clock. CONSTITUTION:When an engine is started, a crank angle detection signal (x) as a noise correlation signal corresponding to double of the engine rotating speed is outputted from a sensor 8 and inputted to a processor unit 10, and a residual noise detection signal (e) detected by a microphone 7 is also inputted to the unit 10. Therefore, in the unit 10, the signals (x) and (e) are digitalized in converters 14, 15 by a clock signal SC in which the signal (x) outputted from a sampling clock generating circuit 13 is increased by four times. Then, the conversion signals X, E corresponding to the signals (x), (e) converted into a frequency area by the converting parts 27, 28 of a body 16 are subjected to filter processing by an adaptive filter part 19, and the result is converted into a time area signal by an inverse Fourier transforming part 21 and outputted to a speaker 5.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an active noise control device in a vehicle cabin, and in particular to an active noise control device for controlling noise transmitted from an internal combustion engine using a frequency domain control algorithm. The control sound generated from the control sound source is caused to interfere with the drive signal generated by the control sound source, thereby reducing noise.

[Conventional technology]

Conventionally, this type of active noise control device has been described in, for example, British Publication Patent No. 2149614 (hereinafter referred to as the first conventional example) and US Patent No. 4490841 (hereinafter referred to as the second conventional example). There are things that exist.

The first conventional example operates under conditions such as an aircraft cabin or a similar closed space where sound including harmonics f, -f, etc. is generated. Specifically, a plurality of loudspeakers (secondary sound sources) and microphones installed in a closed space, and the frequency f of the noise source. A signal processor that supplies a signal having an opposite phase to the detected frequency f. to a plurality of loudspeakers based on the output signals of the plurality of microphones and the detection signal of the frequency detection means. The secondary stage generated from the loudspeaker and the primary sound transmitted from the noise source interfere with each other to minimize the sound pressure level in the closed space. The processor can perform processing in the frequency domain by performing a process of Fourier transforming the output of the microphone and determining the output of the loudspeaker so as to minimize the sum of squares of the resulting Fourier coefficients.
When the frequency of the noise source changes and high-speed processing is required, such as in the internal combustion engine of an automobile, a multi-channel LMS algorithm that performs real-time calculations is applied.

In addition, the second conventional example detects the residual vibration generated as a result of interference between the original vibration from the repetitive vibration source and the secondary vibration caused by the driven actuator, and combines the detected residual vibrations. into a set of independent elements, each of which represents the residual vibrations at different locations in the region, and each of which is synchronized with the repeating period of the original repeating vibration from the source. Then, each element of each set of elements is individually converted, the converted element is re-transformed as an actuator drive signal, and the conversion of the elements is controlled so as to reduce the power or amplitude of the residual vibration. This allows control to be performed in the frequency domain.

[Problem to be solved by the invention]

However, in the first conventional example described above, the multi-channel LMS algorithm is suitable when high-speed processing is required due to changes in the frequency of the noise source, such as in an automobile engine. However, the algorithm in this case evaluates the sound pressure at the noise reduction position, determines the filter coefficients of the digital filter for outputting the secondary stage, and generates the output signal all using time-series data, so it is difficult to solve the problem in the time domain. This is called the LMS algorithm, and if the current sampling frequency of a noise source is significantly different from the frequency of the previous sampling, the processing mode will change greatly depending on the frequency fluctuation, and the processing method will change significantly due to the frequency fluctuation. There was an unresolved problem with the vehicle engine, which was that sufficient effects could not be obtained.

In addition, in the second conventional example, control in the frequency domain is possible, but since fast Fourier transform is used, a large number of sampled values (several tens to hundreds, or more) are required for one calculation. Therefore, when the engine speed changes from moment to moment, the frequency of the secondary unit that should output must also change from moment to moment, which is an unresolved problem in that it is not possible to obtain sufficient follow-up speed. was there.

Therefore, this invention was made by focusing on the unresolved problems of the conventional example, and is based on the frequency domain LMS algorithm, and synchronizes the sampling clock with a correlation detection signal that is correlated with the internal combustion engine. It is an object of the present invention to provide an active noise control device that can solve the problems of the above-mentioned conventional examples.

[Means to solve the problem]

In order to achieve the above object, the active noise control device according to claim (1) causes the control sound generated from the control sound source to interfere with the noise transmitted from the internal combustion engine in the vehicle interior, thereby reducing the residual noise energy of both. is detected by a residual noise detection means, and
An active device comprising: a control means for detecting a signal correlated with the noise of the internal combustion engine with a noise correlation signal detection means, and forming a drive signal for the control sound source based on the detected residual noise detection signal and the noise correlation detection signal. In the type noise control device, the control means includes a sampling clock forming means for forming a noise correlation detection signal or a sampling clock obtained by multiplying the noise correlation detection signal, and uses the sampling clock of the sampling clock forming means to perform arithmetic processing on the drive signal. It is characterized by the fact that

Further, in the active noise control device according to claim (2), the control means forms the drive signal using a frequency domain control algorithm including a fast Fourier transform means and an inverse fast Fourier transform means. It is characterized by being configured.

[Effect]

In the active noise control device according to claim (1) 6ζ,
Since the signal processing in the control means that forms the drive signal for the control sound source so as to reduce the residual noise detection signal of the residual noise detection means is performed using the noise correlation detection signal of the internal combustion engine or the sampling clock obtained by multiplying the noise correlation detection signal of the internal combustion engine, Even if there is a frequency fluctuation in the engine, the control means can perform processing in synchronization with this frequency fluctuation to obtain a drive signal.
The number of calculations required by the control means can be significantly reduced,
A good noise reduction effect can be exhibited.

Further, in the active noise control device according to claim (2), the control means forms the drive signal using a frequency domain control algorithm including a fast Fourier transform means and an inverse fast Fourier transform means. Therefore, it is possible to exhibit a good noise reduction effect even with respect to frequency fluctuations of the noise source that cannot be followed by a time-domain control algorithm, and it is also possible to control the order components of the noise.

〔Example] Embodiments of the present invention will be described below based on the drawings.

FIG. 1 is a schematic diagram showing an embodiment of the present invention applied to a front engine front wheel drive vehicle.

In the figure, 1 is a vehicle body, a four-cylinder engine 3 is disposed in the front side of the vehicle interior 2, a front seat 4F and a rear seat 4R are located in the vehicle interior 2, and the front seat 4F is located in front of the front seat 4F. For example, a loudspeaker 5 as a control sound source is arranged below the dash panel (not shown), and a microphone 7 as a residual noise detection means is arranged in the headrest 6 of the front seat 4F, which is the driver's seat. There is.

A crank angle sensor 8 is attached to the engine 3 as a noise correlation signal detection means that outputs a detection signal correlated with the engine noise. A noise correlation signal X consisting of pulses is output.

Then, the residual noise detection signal e outputted from the microphone 7 and the noise correlation signal X outputted from the crank angle sensor 8 are inputted to the processor unit 10, and the residual noise detection signal e is set to zero in this processor unit 10. A drive signal y is generated, which drive signal y is supplied to the loudspeaker 5 to generate a control sound for the noise.

As shown in FIG. 2, the processor unit 10 includes a multiplier circuit 11 to which the noise correlation detection signal X from the crank angle sensor 8 is input, and a waveform shaping circuit 12 that shapes the waveform of the multiplied output of the multiplier circuit 11. an A/D converter 14 that converts the noise correlation detection signal X from the crank angle sensor 8 into a digital signal, and an A/D converter that converts the residual noise detection signal e from the microphone 7 into a digital signal. 15, a processor main body 16 into which the digital conversion signals from both A/D converters I4 and 15 are input, and a D/A converter 17 that converts the drive signal output from the processor main body 16 into an analog signal. The drive signal y output from the D/A converter 17 is supplied to the loudspeaker 5. The sampling clock SC output from the sampling clock forming circuit 13 is input to the A/D converters 14 and 15, the processor main body 16, and the D/A converter 17.

Here, the sampling clock forming circuit 13 uses its multiplier circuit 11 to extract only the harmonic component, for example, the fourth-order component (corresponding to the eighth-order component of the engine speed) of the noise correlation detection signal X input from the crank angle sensor 8. The signal is extracted, amplified, and output to the waveform shaping circuit 12.

Further, as shown in the functional block diagram of FIG. 3, the processor main body 16 includes a fast Fourier transform section 27 that performs fast Fourier transform (FFT) on the reference signal A fast Fourier transform unit 28 similarly performs fast Fourier transform on the residual noise detection signal e input from the converter 15, and an adaptive unit receives the frequency domain transform output X output from the fast Fourier transform unit 27 and performs adaptive filter processing. a calculation unit 20 that receives the filter unit 19 and the frequency domain transform outputs X and E of both the Fourier transform units 27 and 28 and updates the filter coefficients of the adaptive filter unit 19;
an inverse fast Fourier transform unit 2 that performs inverse fast Fourier transform (IFFT) on the drive signal Y output from the adaptive filter unit 19;
1.

As shown in detail in FIG.
4 points) order components X, 1 to Xn1024 are input, and the order component E output from the fast Fourier transform section 28 is input, and based on these, the filter of the adaptive filter section 19 is input according to the LMS algorithm in the frequency domain. Update the coefficient W n j. Here, as a frequency domain LMS algorithm, IEEE, 19
December 1978 issue vol. 1.66, pages 1658-1659 M, Dentino, J, McCool, and B
The LMS algorithm described in the paper entitled "Frequency domain adaptive filtering" by Widrotn et al. can be applied.

Next, the operation of the above embodiment will be explained. When the engine 3 is started, the crank angle sensor 8 outputs a crank angle detection signal X as a noise correlation signal equivalent to twice the rotational speed of the engine 3, which is input to the processor unit 10 and detected by the microphone 7. The residual noise detection signal e is also input to the processor unit 10. Therefore, in the processor unit 10, the crank angle detection signal
are digitized by A/D converters 14.15, and then fast Fourier transform section 27.
The converted signals X and E corresponding to the crank angle detection signal
fij and update the updated filter coefficients W, ,
.

1, the adaptive filter unit 19 performs filter processing based on
The processing result is converted into a time domain signal by the inverse Fourier transform section 21, which is converted into an analog signal by the D/A converter 17 and output to the loudspeaker 5. As a result, control sound is output from the loudspeaker 5, which interferes with noise such as muffled noise caused by engine rotation, thereby reducing the noise.

At this time, the filter coefficients W to be updated in the calculation unit 20,
When the engine noise to be reduced is decomposed into engine rotation order components, nj in , W is the j-th component from the bottom of the order components to be controlled. , represents that it is equivalent. Therefore, if, for example, a 1024-point fast Fourier calculation is performed using the fourth-order component of the crank angle detection signal
Assuming that nj corresponds to the radial component, that is, the first-order component of the crank angle detection signal X, nj in this case can be determined as follows.

In this embodiment, since the clock signal SC is a fourth-order component of the crank angle detection signal X, according to the sampling theorem, the highest frequency that can be analyzed is equal to two crank angles. Furthermore, in n-point fast Fourier calculation, the resolution is the value obtained by dividing the highest analyzable frequency by n/2.

Here, if the engine speed N is, for example, 3000 rpm, the clock signal SC, which corresponds to 4 gin of the crank angle, becomes the 8th engine component in a 4-cylinder engine, so its highest frequency is 3000/60X8=400)1z, and the resolution is 4001512=0.7 with n=1024
8125] (z. At this time, the engine rotation second component, that is, the first component of the crank angle detection signal X is 3000/6
0X2=1001 (z, so the corresponding nj
becomes 10010.78125=128. The clock signal SC is n-10 as the 8th component of the engine rotation speed.
When performing a 24-point fast Fourier calculation, nj, which generally corresponds to the m-th component of the engine rotation speed, is (frequency corresponding to the m-th component)/(where the engine rotation speed is N).
Resolution) −(Nxm/60)/(Nx8/601512)=
512 x m/8...
......It can be obtained using (1). As can be seen from this equation (1), nj is the engine rotation speed N
It is not influenced by

Therefore, in cases where a muffled sound is generated in the passenger compartment 2 due to the influence of combustion overvibration force of a specific cylinder, in a 4-cylinder engine vehicle, there is a rumble noise that is generated in the passenger compartment 2 due to the 0.5 order component of engine rotation. However, since nj in this case has an order of m-0°5, n2=12X0.5/8-32 from the above equation (1).

In addition, if the order components that cause problems in the noise inside the passenger compartment 2 are the 0.5th, 2nd, 3rd, and 4th orders of engine rotation, nl-32, n2= 128,
n3=192 and n4=256.

Therefore, since the order components that are a problem in the noise generated in the passenger compartment 2 are known in advance through experiments, in order to attenuate these order components, the fast Fourier transform section 27, the adaptive filter 19, the calculation section 20, and the inverse fast Fourier transform unit 21
So, X3□, X1□8. It is only necessary to perform fast Fourier calculation, filter coefficient update calculation, filter processing calculation, and inverse fast Fourier calculation for the frequency components of X, 2, and X256, and various calculations can be omitted for the remaining 1020 points. This means that muffled noise caused by engine rotation can be effectively reduced by calculating only a few limited frequencies (order components).

This can be explained by considering that, as disclosed in the above-mentioned paper, if the filter order is on the order of several tens, the number of calculations between the time-domain LMS algorithm and the frequency-domain LMS algorithm is almost the same. This means that if you use , you can achieve a significant reduction in the number of calculations compared to either of them.

Furthermore, by using the ratio of the magnitudes of the order components to be controlled in the residual noise detection signal e from the microphone 7 as a judgment criterion, a control such as not controlling a certain order any more can be added. It is possible to control the order components by changing the amount of reduction for the order components, and it is possible to improve the sound quality of the noise.

In the above embodiment, a case has been described in which the crank angle sensor 8 is used as a sensor for detecting the rotational speed of the engine 3, but the invention is not limited to this, and the ignition pulse signal of the engine 3 or the crankshaft A rotation speed signal detected by a rotation speed sensor or the like can be used.

Further, in the above embodiment, a case has been described in which the crank angle detection signal X is multiplied by 4 in the multiplier circuit 11, but the invention is not limited to this, and any multiplication number greater than or equal to 1 may be selected.

Further, in the above embodiment, a case has been described in which a set of loudspeakers 5 and microphones 7 are arranged in the vehicle interior 2, but the present invention is not limited to this, and a plurality of N loudspeakers and a plurality of M loudspeakers microphones can be arranged and adaptive filtering can be performed using a multi-channel frequency domain LMS algorithm in the processor unit IO.

Further, in the above embodiment, a vehicle equipped with a four-cylinder engine has been described, but the invention is not limited to this, and the present invention can also be applied to a vehicle equipped with an engine of any number of cylinders, such as six cylinders or eight cylinders. can be applied.

〔Effect of the invention〕

As explained above, according to the active noise control device according to claim (1), the noise correlation detection signal of the internal combustion engine or its multiplied signal is formed as a sampling clock, and this sampling clock is used to generate an A/D signal. Since calculations such as conversion, D/A conversion, and filter processing are performed, digital signal processing can be performed in the frequency domain, and in this case, various calculations are performed only on predetermined order components that cause discomfort to passengers. The number of calculations can be significantly reduced compared to when using conventional time-domain LMS algorithms and frequency-domain LMS, and the ability to follow frequency fluctuations of the internal combustion engine is improved. The effect of exhibiting a good noise reduction effect can be obtained.

Further, according to the active noise control device according to claim (2), the processing by the control means uses a frequency domain control algorithm including fast Fourier transform means and inverse fast Fourier transform means. , it is possible to perform accurate noise reduction processing that follows frequency fluctuations in the rotational speed of the internal combustion machine.

[Brief explanation of drawings]

Fig. 1 is a schematic configuration diagram showing an embodiment of the present invention, Fig. 2 is a block diagram of a processor unit, Part 3 is a schematic functional configuration diagram of the processor main body, and Fig. 4 is a functional configuration showing details of the processor main body. It is a diagram. In the figure, ■ is the vehicle body, 2 is the passenger compartment, 3 is the engine, 5 is the loudspeaker (control sound source), 7 is the microphone (residual noise detection means), 10 is the processor unit (control means),
27 and 28 are fast Fourier transform sections, 19 is an adaptive filter section, 20 is an arithmetic section, and 21 is an inverse fast Fourier transform section.

Claims (2)

[Claims] (1) A control sound generated from a control sound source is caused to interfere with the noise transmitted from the internal combustion engine in the vehicle interior, and the residual noise energy of both is detected by a residual noise detection means, and the noise that is correlated with the noise of the internal combustion engine is detected. An active noise control device comprising a control means for detecting a signal by a noise correlation signal detection means and forming a drive signal for the control sound source based on the detected residual noise detection signal and the noise correlation detection signal, the control means The active device comprises a sampling clock forming means for forming a noise correlation detection signal or a sampling clock obtained by multiplying the noise correlation detection signal, and the driving signal is subjected to arithmetic processing using the sampling clock of the sampling clock forming means. type noise control device. 2. The active drive signal of claim 1, wherein the control means is configured to form the drive signal using a frequency domain control algorithm including fast Fourier transform means and inverse fast Fourier transform means. type noise control device.