JP2020071473A - Noise control apparatus, noise control method and program - Google Patents

Abstract

To accurately obtain a noise reduction effect of a target at a control point without being affected by noise of the outside of the target.SOLUTION: A noise control apparatus includes a noise detector, a control filter for processing a detected noise signal using a control factor, a speaker for reproducing an output signal of the control filter, an error microphone for detecting a residual noise at a control point where interference between the noise and a reproduced control sound occurs, a transmission characteristic correction filter for processing the noise signal using a transmission characteristic of the sound from the speaker to the error microphone, a control factor updater for updating the control factor to minimize an error signal using the error signal indicating the residual noise and an output signal of the transmission characteristic correction filter, a correction filter for processing the output signal of the control filter using the transmission characteristic, a subtractor for subtracting the output signal of the correction filter from the error signal, and an effect measurement unit for measuring a noise reduction effect at the control point based on a difference between a control-off signal that is an output signal of the subtractor and a control-on signal that is the error signal.SELECTED DRAWING: Figure 1

Description

  The present disclosure relates to a noise control device, a noise control method, and a program that reduce noise.

  BACKGROUND ART Conventionally, there is known a technique of canceling noise by reproducing a control sound having a phase opposite to that of noise with a speaker. Further, in Patent Document 1, a control sound to be reproduced by a speaker is controlled based on an engine sound so that noise collected by an error microphone installed in the vehicle interior is minimized, so that the engine propagates the sound into the vehicle interior. Techniques for reducing noise caused by noise have been proposed.

  When these conventional techniques are applied to a space where a large number of passengers such as an aircraft exist, it is necessary to perform multipoint control for reducing noise at the position where each passenger exists. For example, in Patent Document 2, in order to reduce the running noise (road noise) of the vehicle in the vehicle, a plurality of sensors are provided in the suspension portion near the tire, and the sound is collected by each of a plurality of error microphones installed in the vehicle. A technique has been proposed in which a control sound to be reproduced by a plurality of speakers is controlled based on a sound detected by the plurality of sensors so that the sound generated by the plurality of speakers is minimized.

JP 2004-20714 A JP-A-6-59688

  However, for example, it is assumed that noise that is not a target to be reduced (hereinafter, noise that is not a target) is generated because the driver largely changes the reproduced sound of the car audio in response to the request of another passenger. In this case, in the above-described related art, the error microphone collects not only the noise to be reduced but also the noise including the noise outside the target. Therefore, the control sound is controlled so as to minimize the noise including the noise outside the target, and only the target noise cannot be accurately reduced.

  The present disclosure is to solve the above-described conventional problems, and a noise control device capable of accurately obtaining a target noise reduction effect at a control point without being affected by noise outside the target. The purpose is to provide.

  In order to solve the conventional problems, a noise control device according to an aspect of the present disclosure has a noise detector that detects noise generated in a noise source and a noise signal that indicates the noise detected by the noise detector. A control filter that performs signal processing using the control coefficient of, a speaker that reproduces an output signal of the control filter as a control sound, and noise that propagates from the noise source and control sound that is reproduced by the speaker cause interference. An error microphone installed at a control point for detecting residual noise remaining at the control point due to the interference, and a transfer characteristic correction filter for processing the noise signal by using a transfer characteristic of sound from the speaker to the error microphone. And an error signal indicating the residual noise detected by the error microphone and an output signal of the transfer characteristic correction filter so as to minimize the error signal. A coefficient updater for updating the control coefficient, a correction filter for signal processing the output signal of the control filter using the transfer characteristic of the sound from the speaker to the error microphone, and an output signal of the correction filter from the error signal Subtracting the subtracter and the output signal of the subtractor as a control-off signal that represents the noise before control due to the interference, and the error signal as a control-on signal that represents the noise after control due to the interference. An effect measurement unit that measures a noise reduction effect at the control point based on a difference between a signal and the control-on signal.

  According to the above aspect, it is possible to accurately obtain the effect of reducing the target noise at the control point without being affected by the noise outside the target.

It is a block diagram of the noise control device according to the first embodiment. It is a figure which shows an example of a structure of an effect measurement part. It is a figure which shows an example of the noise reduction effect measured in the effect measurement part. It is a figure which shows another example of the noise reduction effect measured in the effect measurement part. It is a figure which shows another example of the noise reduction effect measured in the effect measurement part. It is a figure which shows an example of another structure of an effect measurement part. It is a block diagram of the noise control apparatus which concerns on Embodiment 2. It is a flowchart which shows the flow of an adaptive operation. It is a block diagram of an adaptable state determination part. It is a figure which shows an example of the determination conditions which an adaptable state determination part uses. It is a figure which shows the distance from a sensor in a noise control device to an error microphone, and the distance from a speaker to an error microphone. It is a figure which shows another example of the noise reduction effect measured in the effect measurement part. It is an operation | movement flowchart which shows the flow of the design operation | movement of a control coefficient based on the result of having determined the noise reduction effect in the effect measurement part. It is an operation | movement flowchart which shows the flow of the design operation | movement of the control coefficient in the noise control device whole. It is an operation | movement flowchart which shows the flow of the design operation | movement of the control coefficient in the noise control device whole. It is a block diagram of the noise control apparatus for reducing the engine sound of the vehicle which concerns on a prior art example. It is a top view which shows the structure in the motor vehicle in which the noise control apparatus for reducing the road noise which concerns on a prior art example is arrange | positioned. It is a side view which shows the structure in the motor vehicle in which the noise control apparatus for reducing the road noise which concerns on a prior art example is arrange | positioned. It is a block diagram of the noise control apparatus for reducing the road noise which concerns on a prior art example. It is a figure which shows the effect of the noise control of the road noise by the noise control apparatus which concerns on a prior art example. FIG. 6 is a configuration diagram showing a modified example of the noise control device according to the first embodiment.

(Findings that form the basis of this disclosure)
BACKGROUND ART Conventionally, there is known a technique of canceling noise by reproducing a control sound having a phase opposite to that of noise from a speaker. The technology has already been put to practical use in headphones and inner headphones (hereinafter, earphones). The headphones and earphones are generally known as noise canceling headphones. Headphones and earphones are worn directly on the ear. Therefore, when the above-mentioned conventional technique is applied to a headphone or an earphone, noise transmitted to an extremely small space inside the ear closed by the headphone or the earphone may be controlled.

  On the other hand, it is assumed that the above conventional technique is applied to a space where there are several passengers such as automobiles and aircrafts. In this case, since it is necessary to perform multi-point control so as to reduce noise at the position where each passenger is present, the control becomes complicated and practical application is difficult. In particular, it has been difficult to apply the above-mentioned conventional technique to a large space where there are many passengers such as an aircraft.

  However, in recent years, simple noise control specialized for the engine sound of automobiles has been put into practical use. FIG. 14 is a configuration diagram of a noise control device 1000a for reducing the engine sound of the automobile 100 according to the conventional example. For example, as shown in FIG. 14, in the noise control device 1000a, the tacho pulse generator 110 outputs a pulse signal synchronized with the engine speed when the engine 101 of the automobile 100 is running. This pulse signal is converted by a low-pass filter (hereinafter, LPF) 111 into a cosine wave having a frequency equal to a predetermined frequency which causes a problem as noise inside the vehicle. The cosine wave output from the LPF 111 is input to the first phase shifter 112 and the second phase shifter 113.

  The first phase shifter 112 is set so that the phase characteristic advances by π / 2 (rad) with respect to the second phase shifter 113. Therefore, the output signal of the first phase shifter 112 becomes a cosine wave signal having a frequency equal to the noise frequency (hereinafter referred to as a reference cosine wave signal). On the other hand, the output signal of the second phase shifter 113 is a sine wave signal having a frequency equal to the noise frequency (hereinafter referred to as a reference sine wave signal). The reference cosine wave signal and the reference sine wave signal are converted into digital signals and then input to the microcomputer 200.

  The reference cosine wave signal input to the microcomputer 200 is multiplied by the filter coefficient W0 in the coefficient multiplier 211 of the adaptive notch filter 210. The reference sine wave signal input to the microcomputer 200 is multiplied by the filter coefficient W1 in the coefficient multiplier 212 of the adaptive notch filter 210. Then, in the adder 213, the output signal of the coefficient multiplier 211 and the output signal of the coefficient multiplier 212 are added, and then reproduced as a control sound by the speaker 160.

  The control sound reproduced from the speaker 160 interferes with the noise propagated from the engine at the control point where the error microphone 150 is installed. This reduces noise at the control points. At this time, the noise remaining at the control point that has not been reduced (hereinafter, residual noise) is detected as an error signal by the error microphone 150. The error signal detected by the error microphone 150 is input to the two LMS calculators 207 and 208.

  In the transfer element 201, the reference cosine wave signal output from the first phase shifter 112 is convoluted with a coefficient simulating the transfer characteristic C0 of the sound from the speaker 160 to the error microphone 150. In the transfer element 202, the reference sine wave signal output from the second phase shifter 113 is convolved with a coefficient simulating the transfer characteristic C1 of the sound from the speaker 160 to the error microphone 150. In the transfer element 203, the reference sine wave signal output from the second phase shifter 113 is convoluted with a coefficient simulating the transfer characteristic C0 of the sound from the speaker 160 to the error microphone 150. In the transfer element 204, the reference cosine wave signal output from the first phase shifter 112 is convoluted with a coefficient simulating the transfer characteristic −C1 opposite to the transfer characteristic C1 of the sound from the speaker 160 to the error microphone 150.

  Then, the output signal of the transfer element 201 and the output signal of the transfer element 202 are added by the adder 205 and then input to the LMS calculator 207. Further, the output signal of the transfer element 203 and the output signal of the transfer element 204 are added by the adder 206 and then input to the LMS calculator 208.

  The LMS calculator 207 uses a known coefficient update algorithm such as an LMS (Least Mean Square) algorithm (least squares method) so that the error signal input from the error microphone 150 is minimized. Calculate W0. Similarly, the LMS calculator 208 calculates the filter coefficient W1 used by the coefficient multiplier 212 so that the error signal input from the error microphone 150 is minimized.

  As described above, the filter coefficients W0 and W1 used by the coefficient multipliers 211 and 212 of the adaptive notch filter 210 are recursively updated so as to minimize the error signal input from the error microphone 150, and converge to the optimum value. I will do it. That is, the filter coefficients W0 and W1 are recursively updated at the installation location of the error microphone 150 so that the noise propagated from the engine is minimized, and converge to the optimum values.

  Therefore, the conventional noise control device 1000a shown in FIG. 14 reduces the noise propagated from the engine at the control point where the error microphone 150 is installed by using the inexpensive microcomputer 200 without using the expensive DSP. can do.

  However, since the noise control device 1000a uses the cosine wave signal and the sine wave signal based on the noise generated in the engine as the signal referred to by the adaptive notch filter 210, the noise propagated from a noise source other than the engine is not detected. It cannot be reduced.

  Therefore, a plurality of sensors are used to reduce the running noise (hereinafter, road noise) of the vehicle including the engine sound.

  FIG. 15A is a plan view showing a configuration inside a vehicle 100 in which a noise control device 1000b for reducing road noise according to a conventional example is arranged. FIG. 15B is a side view showing a configuration inside a vehicle 100 in which a noise control device 1000b for reducing road noise according to a conventional example is arranged. FIG. 16 is a configuration diagram of a noise control device 1000b for reducing road noise according to a conventional example.

  As shown in FIGS. 15A and 15B, in the suspension portion near the tire of the automobile 100, four sensors (noise detectors) 1a for detecting road noise (noise) generated in the suspension portion (noise source), 1b, 1c and 1d are installed. Specifically, each of the sensors 1a, 1b, 1c, 1d detects the vibration of the suspension section when the automobile 100 is running as road noise.

  As shown in FIG. 16, the vibration signals detected by the sensors 1a, 1b, 1c, 1d are input to the control filters 20aa, 20ab, 20ba, 20bb, respectively. For convenience of description, FIG. 16 shows two sensors 1a and 1b, two speakers 3a and 3b, and two error microphones 2a and 2b provided in the front half of the automobile 100. Only are shown.

  However, in reality, the noise control device 1000b has two sensors 1c and 1d, two speakers 3c and 3d, and two error microphones 2b and 2c in the rear half of the automobile 100. Further prepare. The noise control device 1000b similarly performs control for reducing road noise in the front half of the automobile 100 and the rear half of the automobile 100. Therefore, in the following, only the control in which the noise control device 1000b shown in FIG. 16 reduces the road noise in the front half of the automobile 100 will be described in detail.

  As shown in FIG. 16, the noise control device 1000b performs four control filters 20aa, 20ab, 20ba, 20bb and two sensors 1a when performing control for reducing road noise in the front half of the automobile 100. 1b, two adders 30a and 30b, two speakers 3a and 3b, two (one or more) error microphones 2a and 2b, and eight LMS calculators (coefficient updaters) 61aaa and 61aaab. , 61aba, 61abb, 61baa, 61bab, 61bba, 61bbb, and eight transfer characteristic correction filters 62aaa, 62aab, 62aba, 62abb, 62baa, 62bab, 62bba, 62bbb.

  The noise control device 1000b includes a microcomputer (not shown) including a CPU, a memory such as a RAM and a ROM, and the like. Control filter 20aa, 20ab, 20ba, 20bb, adders 30a, 30b, LMS arithmetic unit 61aaa, 61aab, 61aba, 61abb, 61baa, 61bab, 61bba, 61bbb and transfer characteristics compensation filter 62aaa, 62aab, 62aba, 62abb, 62baa, 62bab , 62bba, 62bbb are configured by the CPU executing a program stored in advance in the ROM.

  The two control filters 20aa and 20ab perform convolution processing (signal processing, first signal processing) on the vibration signal (noise signal) indicating the vibration detected by the sensor 1a using a predetermined control coefficient. The two control filters 20ba and 20bb perform convolution processing using a predetermined control coefficient on the vibration signal indicating the vibration detected by the sensor 1b.

  The adder 30a adds the output signal of the control filter 20aa and the output signal of the control filter 20ba, and outputs the signal after the addition to the speaker 3a. The adder 30b adds the output signal of the control filter 20ab and the output signal of the control filter 20bb, and outputs the signal after the addition to the speaker 3b.

  The speaker 3a reproduces a signal obtained by adding the output signal of the control filter 20aa and the output signal of the control filter 20ba by the adder 30a as a control sound. The speaker 3b reproduces a signal obtained by adding the output signal of the control filter 20ab and the output signal of the control filter 20bb by the adder 30b as a control sound.

  The two error microphones 2a and 2b are installed in a region where the road noise propagated from the suspension section into the vehicle and the control sound reproduced by the speakers 3a and 3b interfere with each other. The two error microphones 2a and 2b detect residual noise remaining at the control point, which is their installation location, due to the interference.

  The error microphone 2a outputs an error signal indicating the detected residual noise to the four LMS calculators 61aaa, 61aba, 61baa, 61bba. The error microphone 2b outputs an error signal indicating the detected residual noise to the four LMS calculators 61aab, 61abb, 61bab, 61bbb. On the other hand, the sensor 1a outputs a vibration signal indicating the detected vibration to the four transfer characteristic correction filters 62aaa, 62aab, 62aba, 62abb.

  Here, the transfer characteristic correction filter 62aaa uses the coefficient that approximates the transfer characteristic C11 of the sound from the speaker 3a to the error microphone 2a to convolve the vibration signal output from the sensor 1a (signal processing, second signal processing). And outputs the signal after the convolution processing to the LMS calculator 61aaa. The transfer characteristic correction filter 62aab performs convolution processing on the vibration signal output from the sensor 1a using a coefficient that approximates the transfer characteristic C12 of the sound from the speaker 3a to the error microphone 2b, and performs LMS calculation on the signal after the convolution processing. Output to the device 61aab. Similarly, the transfer characteristic correction filters 62aba and 62abb perform convolution processing on the vibration signal output from the sensor 1a using coefficients that approximate the transfer characteristics C21 and C22 of the sound from the speaker 3a to the error microphones 2a and 2b, respectively. And outputs the signal after the convolution processing to the LMS calculators 61aba and 61abb.

  The LMS calculator 61aaa executes the LMS algorithm by using the signal input from the transfer characteristic correction filter 62aaa and the error signal input from the error microphone 2a, thereby detecting the error signal input from the error microphone 2a. The control coefficient of the control filter 20aa is updated so as to be minimized. The LMS calculator 61aab executes the LMS algorithm by using the signal input from the transfer characteristic correction filter 62aab and the error signal input from the error microphone 2b, thereby detecting the error signal input from the error microphone 2b. The control coefficient of the control filter 20aa is updated so as to be minimized.

  Similarly, the LMS calculators 61aba and 61abb execute the LMS algorithm using the signals input from the transfer characteristic correction filters 62aba and 62abb and the error signals input from the error microphones 2a and 2b. As a result, the LMS calculators 61aba and 61abb update the control coefficient of the control filter 20ab so as to minimize the error signal input from the error microphones 2a and 2b.

  Similarly, the sensor 1b outputs a vibration signal indicating the detected vibration to the four transfer characteristic correction filters 62baa, 62bab, 62bba, 62bbb. The transfer characteristic correction filters 62baa and 62bab perform a convolution process on the vibration signal output from the sensor 1b using a coefficient that approximates the transfer characteristics C11 and C12 of the sound from the speaker 3a to the error microphones 2a and 2b, and perform the convolution process. The latter signal is output to the LMS calculators 61baa and 61bab. The transfer characteristic correction filters 62bba and 62bbb perform convolution processing on the vibration signal output from the sensor 1b using a coefficient that approximates the transfer characteristics C21 and C22 of the sound from the speaker 3b to the error microphones 2a and 2b, and perform the convolution processing. The latter signal is output to the LMS calculators 61bba and 61bbb.

  The LMS calculators 61baa and 61bab execute the LMS algorithm using the signals input from the transfer characteristic correction filters 62baa and 62bab and the error signals input from the error microphones 2a and 2b. As a result, the LMS calculators 61baa and 61bab update the control coefficient of the control filter 20ba so as to minimize the error signal input from the error microphones 2a and 2b. Similarly, the LMS calculators 61bba and 61bbb execute the LMS algorithm using the signals input from the transfer characteristic correction filters 62bba and 62bbb and the error signals input from the error microphones 2a and 2b. As a result, the LMS calculators 61bba and 61bbb update the control coefficient of the control filter 20bb so as to minimize the error signal input from the error microphones 2a and 2b.

  As described above, finally, the road noise caused by the vibration signal indicating the vibration detected by the sensors 1a, 1b, 1c, and 1d is the speaker at the control point where the error microphones 2a, 2b, 2c, and 2d are installed. It is reduced by interfering with the reproduced control sound by 3a, 3b, 3c and 3d.

  By the way, normally, a driver adjusts the traveling speed of the vehicle and the rotational speed of the engine according to the situation by changing the opening degree of the accelerator according to the traveling state of the vehicle when the vehicle is driven. Therefore, the frequency and level of the engine sound change frequently while the vehicle is running. Therefore, in the control for reducing the engine sound, it is necessary to constantly adapt the control sound reproduced by the speaker to the traveling state. That is, even if the frequency of the engine sound (engine speed) once converges, it is necessary to continue the operation of updating the control coefficient (hereinafter, adaptive operation). As described above, the control for reducing the engine sound can be realized simply and at low cost because the control for continuing the adaptive operation only for the engine sound may be performed.

  On the other hand, road noise is noise with strong randomness generated from a plurality of noise sources and has a wide frequency band. Therefore, in the control for reducing the road noise, the tap length of the control coefficient may be set longer and a plurality of sensors for detecting the noise generated from the plurality of noise sources may be provided. Further, in order to appropriately reduce road noise at a plurality of locations inside the vehicle, a plurality of speakers and a plurality of error microphones may be provided, and the adaptive operation may be continued. In this case, by continuing to update each control coefficient so as to minimize the residual noise collected by each error microphone, road noise can be reduced at the control point where each error microphone is installed. ..

  As described above, road noise generally has a wide frequency band with strong randomness. Therefore, for example, the control coefficients of the control filters 20aa and 20ab shown in FIG. 16 converge according to the transmission characteristics of the sound when the road noise generated in the suspension section near the sensor 1a is transmitted to the error microphones 2a and 2b. To do. That is, when reducing the road noise, once it converges to the control coefficient according to the transfer characteristic, it is possible to maintain a constant noise reduction effect without continuing the adaptive operation.

  Specifically, it is assumed that the control coefficient converges to a control coefficient that reduces road noise in the band of 100 to 500 Hz by 10 dB in a certain traveling state (for example, when traveling at 60 km / h). In this case, if the control coefficient is used, the road noise in the band of 100 to 500 Hz can be reduced by 10 dB even in another traveling state (for example, when traveling at 100 km / h).

  In this way, unlike the control that reduces the engine noise, the control that reduces the road noise does not depend on the change in the traveling speed (or the engine speed) of the vehicle, and even if the control coefficient is fixed, a constant noise A reduction effect can be obtained. Therefore, when the noise control device 1000b is applied to an automobile to reduce road noise, it is conceivable to set an initial value of the control coefficient and fix the control coefficient to the initial value. Hereinafter, a specific example of such a method of determining the initial value of the control coefficient will be described.

  The car manufacturer should know in advance the running condition of the car, such as where the user will drive the car, how many people will be driving other than the driver, or whether the car will be played while playing music etc. Is impossible. For example, even if the traveling position of the automobile can be determined based on the information stored in the navigation system, the state of the road surface on which the automobile is traveling cannot be accurately grasped or predicted. For example, it is difficult to accurately grasp or predict that the road surface on which an automobile is running is not a certain asphalt road surface such as a road surface with a lot of unevenness or a manhole surface. Is.

  In addition, it is difficult to accurately grasp or predict that the road surface on which the automobile is running is the road surface before and after the completion of road construction, and the road surface suddenly changed from a bumpy road to a flat road surface of new asphalt. is there. Further, it is difficult to accurately grasp or predict that the road surface on which the automobile is running is wet or wet with rain or snow, or is not dry. Furthermore, when the road surface condition of the main lane and the road surface condition of the overtaking lane are not the same, whether the car is traveling on the main line or the overtaking lane or whether the car is changing lanes It is difficult to grasp or predict.

  Therefore, a car maker usually drives a car on a test course in which the road surface condition is constantly controlled. Then, the car maker obtains the control coefficient by keeping the traveling state of the automobile constant, for example, traveling at a speed of 60 km / h without reproducing anything by the car audio. Then, the car manufacturer fixes the control coefficient to the obtained control coefficient, and sets the road noise when the vehicle is traveling in a predetermined effect measurement section (for example, a straight section of the test course) for a certain period (for example, 10 seconds). Average and measure each time.

  FIG. 17 is a diagram showing an effect of noise control of road noise by the noise control device 1000b according to the conventional example. Then, the car maker derives a control-off characteristic indicating the relationship between the frequency of the measured road noise and the level (sound pressure) of the measured road noise, as shown by the solid line portion in FIG. 17, for example. Specifically, if the noise control device 1000b is not caused to perform the adaptive operation and the road noise when the vehicle is traveling in the effect measurement section is averaged and measured every fixed period (for example, 10 seconds). Good. Next, the car maker averages and measures the road noise when the automobile is traveling in the effect measurement section at regular intervals (for example, 10 seconds) while causing the noise control device 1000b to perform the adaptive operation. Then, the car maker derives a control-on characteristic indicating a relationship between the frequency of the measured road noise and the level of the measured road noise, as shown by a broken line portion in FIG. 17, for example.

  Next, the car manufacturer calculates the difference between the levels of the respective frequencies in the derived control-off characteristic and control-on characteristic, and confirms whether or not the road noise reduction effect indicated by the difference achieves a predetermined target value. .. Thereby, the car maker confirms whether or not the control coefficients have converged. The car manufacturer determines that the control coefficient has not converged when the road noise reduction effect indicated by the difference does not reach the predetermined target value. In this case, the car manufacturer causes the noise control device 1000b to perform the adaptive operation again, causes the automobile to travel in the effect measurement section, and derives the control-on characteristic again as described above. Then, the car maker repeats this until the reduction effect reaches a predetermined target value.

  After that, the car manufacturer determines that the control coefficient has converged when the reduction effect reaches a predetermined target value, and thereafter determines that the control coefficient can be fixed. Then, the car manufacturer stores the control coefficient at the time of convergence in the ROM in advance as an initial value of the control coefficient.

  However, with this method, the car manufacturer has to design the control coefficients one by one, so that it takes a great deal of time to sell a large number of cars. It is assumed that the initial value of the control coefficient determined using one vehicle is set as the representative value and the representative value is set as the initial value of the control coefficient in another vehicle. However, in this case, the transmission characteristics of the road noise are not necessarily the same in all vehicles, and therefore, there is no guarantee that the desired reduction effect will be obtained.

  In particular, the speaker is generally managed by allowing the output characteristic to have a variation of about 10 to 20%. Further, it is considered that the output characteristics of the speaker installed in the automobile will further vary. In addition to this, it is considered that the characteristics of circuits such as a microphone, a microphone amplifier, and a power amplifier also vary. Therefore, even if the initial value of the control coefficient determined by using one vehicle is set as the representative value and the initial value of the control coefficient in other vehicles is set, the effect of reducing road noise is desired in all vehicles. Achieving the target value cannot be guaranteed. In some cases, the noise control device 1000b may oscillate.

  Therefore, it is conceivable that the user who purchases the automobile sets the initial value of the control coefficient. However, it is difficult for the user to derive the difference between the control-off characteristic and the control-on characteristic under stable traveling conditions such as the above-described car maker's test course. Therefore, it can be said that it is difficult for the user to determine an appropriate initial value of the control coefficient.

  From the above, the present inventor considered that it is difficult to continue reducing the road noise while fixing the control coefficient. Therefore, the present inventor performs the adaptive operation when the desired reduction effect cannot be obtained while performing the fixed operation of fixing the control coefficient, and then obtains the desired reduction effect. Under the circumstances, it was considered to fix the control coefficient to the control coefficient at that time and to perform the fixed operation again.

  However, for example, it is assumed that noise that is not a target to be reduced (hereinafter, noise that is not a target) is generated because the driver largely changes the reproduced sound of the car audio in response to the request of another passenger. In this case, in the above-described related art, the error microphone collects not only the noise to be reduced but also the noise including the noise outside the target.

  Therefore, the control sound is controlled so as to minimize the noise including the noise outside the target, and only the target noise cannot be accurately reduced. That is, in the above-mentioned conventional technique, it was not possible to accurately grasp the noise reduction effect of the reduction target. Therefore, the present inventor has come up with the present disclosure as a result of earnest studies on accurately grasping the noise reduction effect of the reduction target.

  In the embodiment according to the present disclosure, a noise detector that detects noise generated in a noise source, a control filter that performs signal processing using a predetermined control coefficient for a noise signal indicating noise detected by the noise detector, The speaker that reproduces the output signal of the control filter as a control sound is installed at a control point where the noise propagated from the noise source interferes with the control sound reproduced by the speaker, and remains at the control point due to the interference. An error microphone that detects residual noise, a transfer characteristic correction filter that performs signal processing of the noise signal using the transfer characteristic of sound from the speaker to the error microphone, and an error signal that indicates residual noise detected by the error microphone. And a coefficient updater for updating the control coefficient so as to minimize the error signal using the transfer signal and the output signal of the transfer characteristic correction filter; Correction filter that performs signal processing on the output signal of the control filter using the transfer characteristic of the sound from the speaker to the error microphone, a subtractor that subtracts the output signal of the correction filter from the error signal, and the output of the subtractor A signal is a control-off signal that represents noise before control due to the interference, and the error signal is a control-on signal that represents noise after control due to the interference, based on the difference between the control-off signal and the control-on signal. A noise control device is provided, which comprises: an effect measurement unit that measures a noise reduction effect at the control point.

  Further, in the embodiment according to the present disclosure, the computer of the noise control device detects the noise generated by the noise source by using the sensor, and uses the predetermined control coefficient for the noise signal indicating the noise detected by the sensor. One signal processing is performed, the signal after the first signal processing is reproduced as a control sound by a speaker, and the noise propagated from the noise source and the control sound reproduced by the speaker are interfered with each other. An error microphone is used to detect residual noise remaining at the control point due to the interference, and a second signal processing is performed on the noise signal using a transfer characteristic of sound from the speaker to the error microphone to obtain the error microphone. Using the error signal indicating the residual noise detected and the signal after the second signal processing, the control coefficient is updated so as to minimize the error signal, and the speaker outputs the control signal. The third signal processing is performed on the signal after the first signal processing by using the transfer characteristic of the sound to the Lara microphone, the signal after the third signal processing is subtracted from the error signal, and the signal after the subtraction is subjected to the interference. According to the control off signal representing the noise before the control, the error signal as a control on signal representing the noise after the control due to the interference, based on the difference between the control off signal and the control on signal, at the control point A noise control method for measuring the noise reduction effect is provided.

  Moreover, in the embodiment according to the present disclosure, a program for causing a computer to execute the noise control method is provided.

  Further, in the embodiment according to the present disclosure, a noise detector that detects noise generated in a noise source, and a control filter that performs signal processing using a predetermined control coefficient for a noise signal indicating the noise detected by the noise detector. And a speaker that reproduces an output signal of the control filter as a control sound, and a control point at which interference between the noise propagated from the noise source and the control sound reproduced by the speaker occurs, and the control point is provided by the interference. An error microphone that detects residual noise remaining in the output signal, a correction filter that performs signal processing on the output signal of the control filter using the transfer characteristic of the sound from the speaker to the error microphone, and the output of the correction filter from the error signal. A subtractor for subtracting a signal, and an output signal of the subtractor is a control-off signal representing noise before control due to the interference, and the error signal is As a control-on signal representing noise after control due to interference, an effect measurement unit that measures a noise reduction effect at the control point based on a difference between the control-off signal and the control-on signal, and a noise control device comprising: Provided.

  According to the above aspect, a signal obtained by subtracting the output signal of the correction filter from the error signal indicating the residual noise detected by the error microphone is used as the control off signal, the error signal is used as the control on signal, and the control off signal and the control on signal are used. The noise reduction effect at the control point is measured based on the difference of That is, the noise reduction effect at the control point is measured based on the output signal of the correction filter, which is the difference between the signal obtained by subtracting the output signal of the correction filter from the error signal and the error signal.

  Therefore, a sound unrelated to the noise generated by the target noise source propagates to the control point, and the error signal indicating the residual noise detected by the error microphone contains a sound unrelated to the noise generated by the noise source. Even in this case, the effect of reducing the noise propagated from the noise source at the control point can be accurately measured based on only the output signal of the correction filter that is not related to the unrelated sound.

  In the above aspect, an adaptive state determination unit that determines whether or not the coefficient updater should update the control coefficient may be further provided.

  According to this aspect, it is possible to determine whether or not the coefficient updater should update the control coefficient. Therefore, when noise at the control point is deteriorated by causing the coefficient updater to update the control coefficient, the coefficient updater does not update the control coefficient, and the coefficient updater updates the control coefficient. The coefficient updater can be made to update the control coefficient only when the noise at the control point is reduced by executing the above.

  In the above aspect, the coefficient updater updates the control coefficient using a predetermined convergence constant, the effect measurement unit measures the difference between the control OFF signal and the control ON signal as the reduction effect, When it is determined in the determination process that the reduction effect has reached the target value, the control coefficient is optimal. If it is determined that the value has converged to a value, the updating of the control coefficient by the coefficient updater is stopped, the control coefficient is fixed to the optimum value, and the reduction effect does not reach the target value, As a control coefficient has not converged to an optimal value, a new value is added to the coefficient updater by adding a predetermined value to the convergence constant used by the coefficient updater at the time of measuring the reduction effect, and Yo The may be resumed updating of said control coefficient using a new convergence constant.

  According to this aspect, when it is considered that the difference between the control off signal and the control on signal has reached a predetermined target value and the control coefficient has converged to an optimum value, the control coefficient is set to It is possible to prevent the control coefficient from being unnecessarily updated by fixing it to the optimum value. On the other hand, in the case where the difference does not reach the predetermined target value and it is considered that the control coefficient has not converged to the optimum value, a new convergence constant larger than the measurement time point of the reduction effect is used. The control coefficient can be updated. Thus, according to this aspect, the control coefficient can be efficiently converged to the optimum value.

  In the above aspect, the effect measurement unit performs signal processing on each of the control off signal and the control on signal by using an A characteristic coefficient indicating an A characteristic simulating a human auditory characteristic, and after the signal processing, A difference between the control-off signal and the control-on signal after the signal processing may be measured as the reduction effect.

  According to this aspect, it is possible to measure the reduction effect in consideration of human auditory characteristics. Therefore, even if a sound unrelated to the noise generated by the target noise source is heard by a person at the control point, it is propagated from the noise source at the control point without being affected by the unrelated sound. The noise reduction effect can be accurately measured.

  In the above aspect, the effect measurement unit is a frequency analysis unit that calculates frequency characteristics of the control OFF signal and the control ON signal, and a difference between the control OFF signal and the control ON signal for each frequency in the frequency characteristics. And a frequency difference effect calculation unit that calculates the first difference as the index of the reduction effect.

  According to this aspect, the first difference, which is the difference between the control OFF signal and the control ON signal of each frequency in the frequency characteristics of the control OFF signal and the control ON signal, can be calculated as the index of the reduction effect. Therefore, it is possible to determine whether or not the reduction effect has reached a predetermined target value, in accordance with the number of first differences that have achieved a predetermined value according to the target value.

  In the above aspect, the effect measurement unit is a frequency analysis unit that calculates frequency characteristics of the control-off signal and the control-on signal, and using the frequency characteristics, each of the control-off signal and the control-on signal. An overall calculation unit that calculates an overall value in all frequency bands, and a second difference that is a difference between the overall value of the control-off signal and the overall value of the control-on signal, and an overall-value difference that is calculated as an index of the reduction effect. An effect calculation unit may be provided.

  According to this aspect, the second difference, which is the difference between the overall value of the control off signal and the overall value of the control on signal calculated using the frequency characteristics of the control off signal and the control on signal, is used as the index of the reduction effect. Can be calculated. Therefore, it is possible to determine whether or not the reduction effect has reached a predetermined target value, by determining whether or not the second difference has reached a predetermined value according to the target value. it can.

  In the above aspect, the effect measurement unit is a frequency analysis unit that calculates frequency characteristics of the control OFF signal and the control ON signal, and a difference between the control OFF signal and the control ON signal for each frequency in the frequency characteristics. A first difference that is a frequency difference effect calculation unit that calculates as an index of the reduction effect, and using the frequency characteristics, calculate an overall value in each frequency band of the control OFF signal and the control ON signal. And an overall value difference effect calculation unit that calculates a second difference that is a difference between the overall value of the control off signal and the overall value of the control on signal as an index of the reduction effect. Good.

  According to this aspect, the first difference, which is the difference between the control OFF signal and the control ON signal of each frequency in the frequency characteristics of the control OFF signal and the control ON signal, can be calculated as the index of the reduction effect. Therefore, it is possible to determine whether or not the reduction effect has reached a predetermined target value, in accordance with the number of first differences that have achieved a predetermined value according to the target value.

  Further, a second difference, which is a difference between the overall value of the control off signal and the overall value of the control on signal, calculated using the frequency characteristics of the control off signal and the control on signal, may be calculated as an index of the reduction effect. it can. Therefore, it is also possible to determine whether or not the reduction effect has reached a predetermined target value, by determining whether or not the second difference has reached a predetermined value according to the target value. You can

  In the above aspect, the effect measurement unit, by using the frequency characteristic, a band limiting unit that extracts a signal of a frequency within a predetermined evaluation target frequency band included in each of the control OFF signal and the control ON signal. Furthermore, the overall calculation unit, the band limiting unit calculates the overall value in each frequency band of the signal extracted from the control OFF signal and the control ON signal, the overall value difference effect calculation unit, The difference between the overall value of the signal extracted by the band limiting unit from the control off signal and the overall value of the signal extracted by the band limiting unit from the control on signal may be the second difference.

  According to this aspect, the difference between the overall value of the signal in the evaluation target frequency band included in the control-off signal and the overall value of the signal in the evaluation target frequency band included in the control-on signal is calculated as the second difference. Therefore, even if a signal outside the evaluation target frequency band is included in the control-off signal and the control-off signal due to occurrence of noise outside the target, the second difference is a predetermined value corresponding to the target value. By determining whether or not the value has been achieved, it is possible to accurately determine whether or not the reduction effect has achieved a predetermined target value, excluding the influence of signals outside the frequency band to be evaluated. You can

  In the above aspect, the coefficient updater updates the control coefficient using a predetermined convergence constant, and the effect measurement unit determines whether or not the reduction effect has reached a predetermined target value. In the determination process, the first difference of a majority of the frequencies calculated by the frequency difference effect calculation unit, which are included in the predetermined evaluation target frequency band, is a predetermined first value according to the target value. When one target value is achieved, it is determined that the reduction effect has achieved the target value, and it is determined that the control coefficient has converged to the optimum value, and the updating of the control coefficient by the coefficient updater is stopped. Then, the control coefficient is fixed to the optimum value, the frequency difference effect calculation unit calculates, the first difference of a majority of the frequencies included in the evaluation target frequency band is the first target value. Achievement If not, it is determined that the reduction effect does not reach the target value, the control coefficient is not converged to the optimum value, the coefficient updater was used at the time of calculating the first difference. A new convergence constant may be obtained by adding a predetermined value to the convergence constant, and the updating of the control coefficient by the coefficient updater using the new convergence constant may be restarted.

  According to this aspect, it is determined whether or not the reduction effect achieves the target value by achieving a predetermined first target value according to the target value among frequencies included in a predetermined evaluation target frequency band. It is possible to make an accurate determination based on the ratio of the frequencies corresponding to the first difference.

  Then, when it is considered that the reduction effect has reached a predetermined target value and the control coefficient has converged to the optimum value, the control coefficient is fixed to the optimum value and the control coefficient is unnecessarily updated. It is possible to avoid doing. On the other hand, when the reduction effect does not reach the predetermined target value and it is considered that the control coefficient does not converge to the optimum value, a new convergence constant larger than the first difference calculation time is used. It is possible to update the control coefficient. Thus, according to this aspect, the control coefficient can be efficiently converged to the optimum value.

  In the above aspect, the coefficient updater updates the control coefficient using a predetermined convergence constant, and the effect measurement unit determines whether or not the reduction effect has reached a predetermined target value. In the determination process, when the second difference has reached a predetermined second target value according to the target value, it is determined that the reduction effect has achieved the target value, the control Assuming that the coefficient has converged to the optimum value, the updating of the control coefficient by the coefficient updater is stopped to fix the control coefficient to the optimum value, and the second difference does not reach the second target value. In this case, it is determined that the reduction effect has not reached the target value, and the control coefficient has not converged to an optimum value, and the convergence used by the coefficient updater at the time of calculating the second difference. New with a constant value plus a predetermined value And Do convergence constant, a new convergence constant the by the coefficient updater may be resumed updating of said control coefficient used.

  According to this aspect, whether the reduction effect has reached the target value, depending on whether the second difference has reached a predetermined second target value according to the target value, It can be accurately determined.

  Then, when it is considered that the reduction effect has reached a predetermined target value and the control coefficient has converged to the optimum value, the control coefficient is fixed to the optimum value and the control coefficient is unnecessarily updated. It is possible to avoid doing. On the other hand, when the reduction effect does not reach the predetermined target value and it is considered that the control coefficient has not converged to the optimum value, a new convergence constant larger than the second difference calculation time point is used. It is possible to update the control coefficient. Thus, according to this aspect, the control coefficient can be efficiently converged to the optimum value.

  In the above aspect, the coefficient updater updates the control coefficient using a predetermined convergence constant, and the effect measurement unit determines whether or not the reduction effect has reached a predetermined target value. In the determination process, the first difference of the majority of the frequencies calculated by the frequency difference effect calculation unit, which are included in the predetermined evaluation target frequency band, is a predetermined first value according to the target value. When the target value is achieved, and the second difference has reached a predetermined second target value according to the target value, it is determined that the reduction effect has reached the target value, and the control is performed. Assuming that the coefficient has converged to the optimum value, the update of the control coefficient by the coefficient updater is stopped, the control coefficient is fixed to the optimum value, and the frequency difference effect calculation unit calculates the frequency band to be evaluated. Lap included in If the first difference in the frequency of the majority of the number does not reach the first target value, it is determined that the reduction effect does not reach the target value, the control coefficient converges to an optimum value. If not, a new convergence constant obtained by adding a predetermined value to the convergence constant used by the coefficient updater at the time of calculating the first difference is used, and the new convergence constant by the coefficient updater is used. If the second difference does not reach the second target value, it is determined that the reduction effect does not reach the target value, and the control coefficient becomes the optimum value. As a non-converged one, a new convergence constant obtained by adding a predetermined value to the convergence constant used by the coefficient updater at the time of calculating the second difference is set as the new convergence constant by the coefficient updater. The control coefficient using The update may be resumed.

  According to this aspect, it is determined whether or not the reduction effect achieves the target value by achieving a predetermined first target value according to the target value among frequencies included in a predetermined evaluation target frequency band. It is possible to make an accurate determination based on the ratio of the frequencies corresponding to the first difference. Further, whether or not the reduction effect has reached the target value is accurately determined according to whether or not the second difference has reached a predetermined second target value according to the target value. be able to.

  Then, when it is considered that the reduction effect has reached a predetermined target value and the control coefficient has converged to the optimum value, the control coefficient is fixed to the optimum value and the control coefficient is unnecessarily updated. It is possible to avoid doing. On the other hand, when the reduction effect does not reach the predetermined target value and it is considered that the control coefficient does not converge to the optimum value, a new difference larger than the calculation time point of the first difference or the second difference is added. The control coefficient can be updated using the convergence constant. Thus, according to this aspect, the control coefficient can be efficiently converged to the optimum value.

  In the above aspect, the effect measurement unit, in the determination process, among the frequencies in the predetermined noise increase band included in the evaluation target frequency band, calculated by the frequency difference effect calculation unit, of a predetermined number or more of frequencies. If the first difference exceeds a predetermined allowable value according to the target value, it may be determined that an abnormality has occurred in the control coefficient, and the update of the control coefficient by the coefficient updater may be stopped.

  According to this aspect, it is determined that an abnormality has occurred in the control coefficient by exceeding a predetermined allowable value according to the target value among frequencies within a predetermined noise increase band included in a predetermined evaluation target frequency band. It is possible to make an accurate determination based on the ratio of the frequencies corresponding to the first difference. When it is determined that the control coefficient is abnormal, the update of the control coefficient by the coefficient updater can be appropriately stopped.

  In the above aspect, the predetermined number may be one.

  According to this aspect, among the frequencies in the predetermined noise increase band included in the predetermined evaluation target frequency band, one frequency corresponding to the first difference exceeding the predetermined allowable value according to the target value is one. However, if it exists, it is possible to stop the update of the control coefficient by the coefficient updater on the assumption that an abnormality has occurred in the control coefficient.

  In the above aspect, a plurality of the error microphones may be provided, and the effect measurement unit may individually set each error microphone in advance for each of the plurality of the error microphones using the installation location of the error microphone as the control point. The determination process may be performed by using the individual target value set in step 1 as the target value.

  According to this aspect, it is possible to individually determine whether or not the noise reduction effect at the installation location of each error microphone achieves the individual target value set individually for each error microphone in advance.

  In the above aspect, the individual target value is associated with a priority order, and the effect measuring unit uses the individual target value associated with the highest priority order as the target value, When it is determined that the reduction effect has reached the target value, it may be determined that the reduction effect has reached the target value in the determination processing for all the control points.

  According to this aspect, even if it is not determined whether or not the noise reduction effect at each of one or more control points achieves the individual target value, the noise reduction effect corresponds to the highest priority. By determining that the attached individual target value is achieved, it can be easily determined that the noise reduction effect at all control points achieves the individual target value.

  In the above aspect, the adaptable state determination unit updates the control coefficient in the coefficient updater when a value obtained by averaging the instantaneous value level of the error signal in a predetermined period is within a predetermined threshold range. You may decide to make it.

  According to this aspect, even when the instantaneous value level of the error signal instantaneously exceeds the threshold range, a value obtained by averaging the instantaneous value levels of the error signal within a predetermined period falls within the predetermined threshold range. If so, the coefficient updater can be made to update the control coefficient.

  Each of the embodiments described below shows an example of a preferred specific example of the present disclosure. The components, the arrangement positions and connection forms of the components, the order of operations, and the like shown in the following embodiments are examples, and do not limit the present disclosure. The present disclosure is limited only by the claims.

  Therefore, among the constituent elements in the following embodiments, the constituent elements not described in the independent claims showing the highest concept of the present disclosure are not necessarily required to achieve the object of the present disclosure, but It is described as constituting the preferred form.

(Embodiment 1)
The configuration of the noise control device according to the first embodiment will be described. FIG. 1 is a configuration diagram of a noise control device 1000 according to the first embodiment.

  The noise control device 1000, like the conventional noise control device 1000b shown in FIG. 16, detects vibrations detected by the sensors 1a, 1b, 1c, 1d (FIGS. 15A and 15B) provided in the suspension portion of the automobile 100. Road noise caused by the vibration signal shown is reduced at a control point which is a place where the error microphones 2a, 2b, 2c, and 2d are installed.

  For convenience of explanation, FIG. 1 illustrates only the components used by the noise control device 1000 for the control for reducing the road noise in the front half of the automobile 100, similar to the conventional noise control device 1000b shown in FIG. .. However, in reality, the noise control device 1000 further includes the same components as those shown in FIG. 1 in the rear half of the automobile 100. The noise control device 1000, similarly to the noise control device 1000b, similarly performs control for reducing road noise in the front half of the vehicle 100 and the rear half of the vehicle 100. Therefore, in the following, only the control in which the noise control device 1000 shown in FIG. 1 reduces the road noise in the front half of the automobile 100 will be described in detail.

  Two sensors 1a, 1b, four control filters 20aa, 20ab, 20ba, 20bb and eight transfer characteristic correction filters 62aaa, 62aab, 62aba, 62abbb, 62baa, 62bab, 62bba, 62bbb, eight shown in FIG. The LMS calculators 61aaa, 61aaab, 61aba, 61abb, 61baa, 61bab, 61bba, 61bbb, two adders 30a, 30b, two speakers 3a, 3b, and two error microphones 2a, 2b are shown in FIG. It has the same configuration as shown in. That is, the noise control device 1000 is caused by the vibration signal indicating the vibration detected by the sensors 1a and 1b by the adaptive operation of updating the control coefficients of the control filters 20aa, 20ab, 20ba, and 20bb like the conventional noise control device 1000b. The road noise generated is reduced at the control point where the error microphones 2a and 2b are installed.

  Further, once the control coefficient converges to the optimum value, the noise control device 1000 thereafter performs a fixing operation of fixing the control coefficient to the optimum value. Hereinafter, a method for the noise control device 1000 to determine whether or not the control coefficient has converged to the optimum value will be described.

First, in the error microphone 2a, as a result of the road noise in the vehicle interior caused by the vibration signal indicating the vibration detected by the sensors 1a and 1b and the control sound reproduced by the speakers 3a and 3b, the error microphone 2a is installed at the installation location. An error signal indicating the residual noise remaining at a certain control point is output. Assuming that the signal indicating road noise at the installation location of the error microphone 2a is the signal N1, the signal reproduced by the speaker 3a is the signal y1, and the signal reproduced by the speaker 3b is the signal y2, the error signal e1 output by the error microphone 2a. Can be expressed by Equation 1.
e1 = N1 + C11 * y1 + C21 * y2 (Equation 1)
Here, C11 indicates the transfer characteristic of sound from the speaker 3a to the error microphone 2a.
C21 indicates a transfer characteristic of sound from the speaker 3b to the error microphone 2a.
* Indicates a convolution operation.

  On the other hand, the signal y1 is input to the subtractor 41a via the transfer characteristic correction filter 40aa. The transfer characteristic correction filter (correction filter) 40aa uses the same coefficient as the transfer characteristic correction filter 62aaa, which approximates the transfer characteristic C11 of the sound from the speaker 3a to the error microphone 2a, and convolves the signal y1 (signal processing, (3 signal processing) and outputs the signal after the convolution processing to the subtractor 41a. Similarly, the signal y2 is input to the subtractor 41a via the transfer characteristic correction filter 40ba.

The subtractor 41a subtracts the output signals of the transfer characteristic correction filter 40aa and the transfer characteristic correction filter 40ba from the error signal output by the error microphone 2a. Specifically, the subtractor 41a performs the calculation of Expression 2.
off1 = e1-C11 * y1-C21 * y2 (Equation 2)
Here, C11 indicates the transfer characteristic of sound from the speaker 3a to the error microphone 2a.
C21 indicates a transfer characteristic of sound from the speaker 3b to the error microphone 2a.
off1 indicates the output signal of the subtractor 41a.

By substituting the expression 1 into the expression 2, the output signal off1 of the subtractor 41a can be expressed by the expression 3.
off1 = N1 (Equation 3)

  Thus, the output signal off1 of the subtractor 41a is the same signal as the signal indicating the road noise at the setting location of the error microphone 2a, and the road noise at the setting location of the error microphone 2a and the output of the two speakers 3a, 3b. The signal represents the noise before control due to interference with the signal. On the other hand, the error signal e1 in Expression 1 is a signal on1 representing the noise after control due to the interference.

  That is, in the noise control device 1000, the signal off1 representing the noise before control due to the interference between the road noise at the setting location of the error microphone 2a and the output signals of the two speakers 3a and 3b, and the noise after the control due to the interference. The signal on1 and the signal represented are simultaneously calculated. The calculated signal off1 representing the noise before control and the calculated signal on1 representing the noise after control are input to the effect measurement unit 50a.

  Similarly, in the noise control device 1000, the signal off2 representing the noise before control due to the interference between the road noise at the setting location of the error microphone 2b and the output signals of the two speakers 3a and 3b, and the signal after the control due to the interference. The signal on2 representing noise is calculated at the same time. The calculated signal off2 representing the noise before control and the calculated signal on2 representing the noise after control are input to the effect measurement unit 50b.

  The effect measuring unit 50a uses a signal (control OFF signal) off1 indicating noise before control due to interference between road noise at the setting location of the error microphone 2a and output signals of the two speakers 3a and 3b and after control due to the interference. The road noise reduction effect at the setting location of the error microphone 2a is measured based on the signal representing noise (control-on signal) on1. The effect measurement unit 50b uses a signal (control OFF signal) off2 indicating noise before control due to interference between road noise at the setting location of the error microphone 2b and output signals of the two speakers 3a and 3b and after control due to the interference. The road noise reduction effect at the setting location of the error microphone 2b is measured based on the signal representing noise (control-on signal) on2.

  FIG. 2 is a diagram showing an example of the configuration of the effect measuring unit 50a. The effect measuring unit 50b has the same configuration as the effect measuring unit 50a. Therefore, only the configuration of the effect measurement unit 50a will be described below as a representative. As shown in FIG. 2, the effect measurement unit 50a includes two A characteristic filter units 51a and 51b, two frequency analysis units 52a and 52b, two overall calculation units 53a and 53b, and a frequency difference effect. The calculation unit 54a and the overall value difference effect calculation unit 54b are provided.

  A signal (hereinafter, noise pre-control signal) off1 representing noise before control due to interference between the road noise input to the effect measurement unit 50a and the output signals of the two speakers 3a and 3b, road noise, and two speakers. Signals (hereinafter, noise-controlled signals) on1 that represent noise after control due to interference with output signals of 3a and 3b are input to the A characteristic filter units 51a and 51b, respectively.

  The A characteristic filter unit 51a performs convolution processing (signal processing) on the input pre-noise control signal off1 by using a coefficient (A characteristic coefficient) indicating the A characteristic simulating human auditory characteristics. Similarly, the A characteristic filter unit 51b performs a convolution process on the input noise-controlled signal on1 by using a coefficient indicating the A characteristic that simulates the human auditory characteristic (hereinafter, the A characteristic coefficient).

  The frequency analysis unit 52a calculates a frequency characteristic of the pre-noise control signal off1 after the convolution processing by the A characteristic filter unit 51a by performing a predetermined frequency analysis processing such as FFT. The frequency analysis unit 52b calculates the frequency characteristic of the noise-controlled signal on1 after the convolution processing by the A characteristic filter unit 51b by performing a predetermined frequency analysis processing such as FFT.

  The frequency difference effect calculation unit 54a uses the pre-noise control signal off1 after the convolution processing by the A characteristic filter unit 51a and the A characteristic filter unit 51b for each frequency in the frequency characteristics calculated by the frequency analysis unit 52a and the frequency analysis unit 52b. The difference from the post-convolution noise control signal on1 (hereinafter referred to as the first difference) is calculated as an index of the road noise reduction effect at the setting location of the error microphone 2a.

  The overall calculation unit 53a uses the frequency characteristics of the noise control pre-signal off1 after the convolution processing by the A characteristic filter unit 51a calculated by the frequency analysis unit 52a, and uses the overall value of the noise control pre-signal off1 in all frequency bands. To calculate. Hereinafter, the overall value calculated by the overall calculation unit 53a will be referred to as the first overall value. The overall calculation unit 53b uses the frequency characteristic of the noise-controlled signal on1 after the convolution processing by the A-characteristic filter unit 51b, which is calculated by the frequency analysis unit 52b, and uses the overall value of the noise-controlled signal on1 in the entire frequency band. To calculate. Hereinafter, the overall value calculated by the overall calculation unit 53b will be referred to as the second overall value.

  The overall value difference effect calculation unit 54b calculates the difference (hereinafter, second difference) between the first overall value calculated by the overall calculation unit 53a and the second overall value calculated by the overall calculation unit 53b from the error microphone 2a. Calculated as an index of the road noise reduction effect at the setting location.

  FIG. 3 is a diagram showing an example of the noise reduction effect measured by the effect measurement unit 50a. In the section (a) of FIG. 3, the frequency characteristic of the noise control before signal off1 calculated by the frequency analysis unit 52a is shown by a solid line, and the frequency characteristic of the noise control after signal on1 calculated by the frequency analysis unit 52b is shown by a broken line. Has been done. In the section (b) of FIG. 3, the first for each frequency calculated by the frequency difference effect calculation unit 54a corresponding to the difference between the frequency characteristic indicated by the solid line and the frequency characteristic indicated by the broken line in the section (a) of FIG. Differences are shown.

  For example, in the examples shown in sections (a) and (b) of FIG. 3, road noise of frequencies f1 to f2 corresponding to the first difference below 0 dB is reduced at the setting location of the error microphone 2a. You can figure out that. Further, since there is no frequency corresponding to the first difference above 0 dB, it can be understood that the road noise of all frequency components has not increased.

  Further, to the right of the frequency characteristic in the section (a) of FIG. 3, the first overall value (for example, 85 dBA) calculated by the overall calculation unit 53a and the second overall value calculated by the overall calculation unit 53b ( For example, 80 dBA) is shown. Further, to the right of the frequency characteristic in the section (a) of FIG. 3, a second difference (for example, −) that is a difference between the first overall value and the second overall value calculated by the overall value difference effect calculation unit 54b. 5dBA) is shown. In the example shown in the section (a) of FIG. 3, since the second difference is −5 dBA, it can be seen that the road noise at the setting location of the error microphone 2a is reduced by 5 dBA.

  If it is desired to evaluate the road noise reduction effect without considering the human auditory characteristics, the effect measurement section 50a may not include the A characteristic filter sections 51a and 51b. In accordance therewith, the frequency analysis unit 52a calculates the frequency characteristic of the noise control pre-signal (control OFF signal) off1 input to the effect measurement unit 50a, and the frequency analysis unit 52b is input to the effect measurement unit 50a. The frequency characteristic of the post-noise control signal (control ON signal) on1 may be calculated.

  The effect measuring unit 50a further uses the first difference for each frequency calculated by the frequency difference effect calculating unit 54a and the second difference calculated by the overall value difference effect calculating unit 54b to detect the error microphone 2a. A determination process is performed to determine whether or not the road noise reduction effect at the set location has reached the target value.

Specifically, the effect measuring unit 50a determines whether or not the road noise reduction effect at the setting location of the error microphone 2a has reached the target value in the determination process as described in 1) and 2) below. To determine.
1) The first difference of a majority of frequencies included in a predetermined frequency band to be evaluated (for example, frequencies f1 to f2 in sections (a) and (b) of FIG. 3) is set in advance. When the first target value is achieved, it is determined that the reduction effect has achieved the target value. The effect measuring unit 50a may make the determination under more severe conditions. For example, when the effect measurement unit 50a achieves the first target value by the first difference of a predetermined number (for example, 70%) or more of a majority of frequencies included in the evaluation target frequency band. Alternatively, it may be determined that the reduction effect has reached the target value.
2) When the second difference reaches a preset second target value different from the first target value, it is determined that the reduction effect has reached the target value.

  The effect measuring unit 50b performs a determination process for determining whether or not the road noise reducing effect at the setting location of the error microphone 2b has reached the target value, similarly to the effect measuring unit 50a.

  As a result of the determination processing performed by the effect measurement unit 50a and the effect measurement unit 50b, it is determined that the road noise reduction effect at the setting locations of all the error microphones 2a, 2b installed in the vehicle has reached the target value. To do. In this case, the effect measurement unit 50a or the effect measurement unit 50b determines that the control coefficients of all the control filters 20aa, 20ab, 20ba, 20bb have converged to the optimum values, and stops the adaptive operation.

  Specifically, the effect measurement unit 50a or the effect measurement unit 50b uses four LMS calculators (coefficient updaters) 61aaa, 61aab, 61aba, 61abb, 61baa, 61bab, 61bba, 61bbb, and four control filters 20aa. , 20ab, 20ba, 20bb stop updating the control coefficients. Then, the effect measurement unit 50a or the effect measurement unit 50b fixes the control coefficient of each control filter 20aa, 20ab, 20ba, 20bb to each control coefficient when it is determined that the control filter 20aa, 20ab, 20ba, 20bb has converged to the optimum value.

  According to the above configuration, the pre-noise control signal off1 representing the road noise before control due to the interference between the road noise at the control point where the error microphones 2a and 2b are installed and the control sound reproduced by the speakers 3a and 3b. , Off2 and post-noise control signals on1 and on2 representing road noise before control due to the interference at the control point can be obtained at the same time.

  Further, the pre-noise control signal off1 is obtained by subtracting the output signals of the transfer characteristic correction filters 40aa and 40ba from the error signal indicating the residual noise detected by the error microphone 2a, and the error signal indicating the residual noise detected by the error microphone 2a. The noise reduction effect at the installation location of the error microphone 2a is measured based on the output signals of the transfer characteristic correction filters 40aa and 40ba that are the difference between the post-noise control signal on1.

  Therefore, a sound unrelated to the noise generated by the target noise source propagates to the control point, and the error signal indicating the residual noise detected by the error microphone 2a includes a sound unrelated to the noise generated by the noise source. Even if so, the noise reduction effect at the installation location of the error microphone 2a can be accurately measured based on only the output signals of the transfer characteristic correction filters 40aa and 40ba that are not related to the unrelated sound.

  Therefore, for example, the car maker does not need to drive the automobile 100 to be sold on the test course and determine the control coefficients of the control filters 20aa, 20ab, 20ba, and 20bb as described above. A general user can appropriately set the control coefficients of the respective control filters 20aa, 20ab, 20ba, 20bb while driving the automobile 100.

  Further, in the case of broadband noise such as road noise, once the control coefficient is obtained, the constant effect can be maintained without changing the control coefficient frequently. Therefore, each control filter 20aa, 20ab, 20ba, 20bb can be operated to measure the noise reduction effect at the installation location of the error microphone 2a.

  FIG. 18 is a configuration diagram showing a modification of the noise control device 1000 according to the first embodiment. In such a case, the LMS calculators 61aaa to 61bbb and the transfer characteristic correction filters 62aaa to 62bbb may be removed from the noise control device 1000 (FIG. 1). Thereby, a noise control device 1002 having a simplified configuration as shown in FIG. 18 may be configured.

  That is, when the noise control device 1002 performs the control for reducing the road noise in the front half of the automobile 100, the two sensors 1a and 1b and the vibration signals output by the two sensors 1a and 1b are predetermined. Four control filters 20aa, 20ab, 20ba, 20bb that perform convolution processing using control coefficients, two adders 30a, 30b, two speakers 3a, 3b, and two error microphones 2a, 2b. And four transfer characteristic correction filters (correction filters) 40aa, 40ab, 40ba, 40bb, two subtractors 41a, 41b, and two effect measurement units 50a, 50b. ..

  FIG. 4 is a diagram showing another example of the noise reduction effect measured by the effect measurement unit 50a. Similar to the section (a) of FIG. 3, the section (a) of FIG. 4 shows a frequency characteristic of the pre-noise control signal off1 calculated by the frequency analysis unit 52a by a solid line, and the noise control calculated by the frequency analysis unit 52b. The frequency characteristic of the rear signal on1 is shown by a broken line. Similar to the section (b) of FIG. 3, the section (b) of FIG. 4 corresponds to the difference between the frequency characteristic indicated by the solid line and the frequency characteristic indicated by the broken line in the section (a) of FIG. The first difference for each frequency calculated by the unit 54a is shown.

  Further, in section (a) of FIG. 4, the frequency characteristic and the noise control of the noise control pre-signal off1 when the noise propagating to the error microphone 2a changes during the measurement of the road noise reduction effect by the effect measurement unit 50a. The frequency characteristic of the rear signal on1 is shown by the dotted line. For example, the noise propagated to the error microphone 2a changes when the traveling speed of the automobile 100 changes or when road conditions such as the road surface on which the automobile travels change. Further, the noise propagating to the error microphone 2a is caused by a passenger talking, music played by a car audio, voice guidance provided by a navigation system, or a large vehicle such as a truck passing each other. It also changes depending on the case.

  According to the above configuration, as shown by the dotted line portion in the section (a) of FIG. 4, even when the noise propagating to the error microphone 2a changes, the frequency characteristic of the pre-noise control signal off1 and the post-noise control signal A similar change appears in the frequency characteristic of the signal on1. Therefore, as shown in the section (b) of FIG. 4, the first difference of each frequency has the same characteristic as the section (b) of FIG.

  This can also be understood from the above-mentioned equation 1, equation 2 and equation 3. The reason is that in Expression 1, if the signal N1 indicating the road noise at the installation location of the error microphone 2a changes to the signal N1 ′, the Expression 1 in which the signal N1 ′ is substituted into the Expression 2 This is because off1 = N1 ′, which is the same formula as in 3, is obtained. That is, the noise control post-signal on1 which is the error signal e1 output by the error microphone 2a and the pre-noise control signal off1 shown in the equation 1 include the same signal N1 'indicating the changed noise. Therefore, the signal N1 'is canceled by calculating the difference between the post-noise control signal on1 and the pre-noise control signal off1.

  By the way, as shown in FIG. 4, when a sound having a frequency within the evaluation target frequency band (frequency f1 to f2) is generated as a sound irrelevant to the target noise, a particularly large problem occurs in the configuration described above. Does not occur. FIG. 5: is a figure which shows another example of the noise reduction effect measured in the effect measurement part 50a. However, as shown by the dotted line in section (a) of FIG. 5, a sound having a frequency outside the evaluation target frequency band is generated as a sound unrelated to the target noise, and the level of the unrelated sound is the evaluation target. It is assumed that the level is not sufficiently low with respect to the sound level of the frequency within the frequency band. In this case, as shown in section (b) of FIG. 5, the first difference is similar to the first difference shown in section (b) of FIGS. 3 and 4. However, the level of the irrelevant sound affects the first overall value and the second overall value, and the second difference, which is the difference between them, is different from the second difference when there is no irrelevant sound. There is.

  For example, in the example shown in section (a) of FIG. 5, the first overall value which is the overall value of the pre-noise control signal off1 is 87 dBA, which is 2 dBA higher than the example shown in section (a) of FIG. The second overall value, which is the overall value of the post-noise control signal on1, is 85 dBA, which is 5 dBA higher than the example shown in section (a) of FIG. As a result, the second difference, which is the difference between the first overall value and the second overall value, becomes −2dBA, and the noise reduction effect deteriorates by 3dBA compared to the example shown in section (a) of FIG. become.

  As described above, even if no problem occurs in the first difference shown in section (b) of FIG. 5, when a problem occurs in the second difference, setting of a second target value that is a target of the second difference, This will hinder the determination of the achievement of that goal.

  Therefore, as shown in FIG. 6, the configuration of the effect measurement unit 50a may be changed. FIG. 6 is a diagram showing an example of another configuration of the effect measurement unit 50a. That is, the effect measuring unit 50a may further include band limiting units 55a and 55b. Then, the band limiting unit 55a uses the frequency characteristic of the pre-noise control signal off1 calculated by the frequency analysis unit 52a to determine the frequency within the evaluation target frequency band (frequency f1 to f2) included in the pre-noise control signal off1. It is also possible to extract only the above signal and output the extracted signal to the overall calculation section 53a. Similarly, the band limiting unit 55b uses the frequency characteristic of the post-noise control signal on1 calculated by the frequency analysis unit 52b to evaluate the frequency band to be evaluated (frequency f1 to f2) included in the post-noise control signal on1. It is also possible to extract only the signal of the internal frequency and output the extracted signal to the overall calculation unit 53b.

  Then, the overall value difference effect calculation unit 54b calculates a second difference that is a difference between the first overall value calculated by the overall calculation unit 53a and the second overall value calculated by the overall calculation unit 53b. You may do so. Then, the second difference may be used as an index of the road noise reduction effect at the setting location of the error microphone 2a.

  In the first embodiment, the example in which the noise control device 1000 is applied to the automobile 100 has been described, but the present invention is not limited to this, and the noise control device 1000 may be applied to an aircraft, a train, or the like.

(Embodiment 2)
The configuration of the noise control device according to the second embodiment will be described.

  In the first embodiment, it has been described that the adaptive operation for updating the control coefficient and the measurement of the road noise reduction effect can be performed at the same time. However, for example, when the driver reproduces audio at a high volume, when a truck that is larger than the automobile 100 runs side by side, or when a noise that is greater than the road noise generated by the user driving the automobile 100 propagates, control is performed. This may adversely affect the adaptive behavior of updating the coefficients.

  In preparation for such a case, unlike the noise control device 1000 according to the first embodiment, the noise control device 1001 according to the second embodiment is adapted only when a predetermined condition that does not adversely affect the adaptive operation is satisfied. Take action. When the adaptive operation is stopped and the control coefficient is fixed, the control coefficient does not change even if a noise larger than the road noise propagates, so there is no need to make the configuration in this case different from that of the first embodiment. ..

  Hereinafter, the flow of the adaptive operation performed in the noise control device 1001 according to the second embodiment will be described. In the following description, eight LMS calculators 61aaa, 61aab, 61aba, 61abb, 61baa, 61bab, 61bba, 61bbb and eight transfer characteristic correction filters 62aaa, 62aab, 62aba, 62abb, 62baa, 62bab, 62bba. , 62bbb are collectively referred to as a coefficient updater 60. When the two effect measuring units 50a and 50b are collectively referred to, the effect measuring unit 50 will be described.

  FIG. 7 is a configuration diagram of the noise control device 1001 according to the second embodiment. As shown in FIG. 7, the noise control device 1001 includes an adaptive state determination unit 70 in addition to the configuration of the noise control device 1000 (FIG. 1) according to the first embodiment. The adaptable state determination unit 70 is configured by the CPU executing a program stored in advance in the ROM. The adaptable state determination unit 70 determines whether or not the coefficient updater 60 should update the control coefficient by determining whether or not the environment inside the vehicle satisfies a predetermined adaptation condition for executing the adaptive operation. judge.

  FIG. 8 is a flowchart showing the flow of the adaptive operation. As shown in FIG. 8, when the adaptive operation is started at a predetermined timing such as when the noise control device 1001 is powered on, the adaptable state determination unit 70 determines that the environment inside the vehicle is in the adaptive operation. It is determined whether or not the adaptive condition for implementation is satisfied (step S1). When it is determined in step S1 that the adaptive condition is satisfied (YES in step S1), the effect measurement unit 50 causes the coefficient updater 60 to perform the adaptive operation (step S2). The details of step S1 will be described later.

  Thereafter, the adaptable state determination unit 70 makes the same determination as in step S1 in parallel while the adaptive operation is being performed (step S3). When it is determined in step S3 that the adaptive condition is not satisfied (NO in step S3), the effect measurement unit 50 causes the coefficient updater 60 to suspend the implementation of the adaptive operation (step S4). After that, the processing from step S1 is performed again. The details of step S3 will be described later.

  On the other hand, if it is determined in step S3 that the adaptation condition is satisfied (YES in step S3), the effect measurement unit 50 starts the adaptive operation in step S2 while continuing the coefficient updater 60 to perform the adaptive operation. Then, it is determined whether or not a preset predetermined time (for example, 30 seconds) has elapsed (step S5). When it is determined in step S5 that the predetermined time has elapsed (YES in step S5), the effect measurement unit 50 ends the adaptive operation by the coefficient updater 60 and fixes the control coefficient to the control coefficient at the end point. The fixed coefficient operation is performed (step S6).

  Then, as described in the first embodiment, the effect measurement unit 50 performs the determination process of determining whether or not the road noise reduction effect at the control point, which is the setting location of each error microphone, achieves the target value. Perform (step S7). When the effect measuring unit 50 determines in step S7 that the road noise reduction effect does not reach the target value (NO in step S7), the process returns to step S1. On the other hand, when the effect measuring unit 50 determines in step S7 that the road noise reduction effect has reached the target value (YES in step S7), the fixed coefficient operation is continued (step S8). In addition, when the effect measuring unit 50 determines that an abnormality has occurred in the control coefficient during the execution of step S7, the effect measuring section 50 stops designing the control coefficient (step S9).

  Next, details of step S1 and step S3 will be described in detail. As shown in FIG. 7, information is input to the adaptable state determination unit 70 from the navigation system 81, the audio system 82, the tachometer (rotation speed) 83, and the speed meter 84. The output signals of the error microphones 2a and 2b are also input to the adaptable state determination unit 70.

  The information input from the audio system 82 to the adaptable state determination unit 70 includes, for example, switch information indicating whether or not the audio system 82 is activated and an audio signal. When the switch information input from the audio system 82 indicates that the audio system 82 is activated, the adaptable state determination unit 70 determines that the adaptation condition is not satisfied. In addition, the adaptable state determination unit 70 determines that the adaptation condition is not satisfied when the signal level of the audio signal input from the audio system 82 is equal to or higher than a predetermined threshold value.

  The information input from the navigation system 81 to the adaptable state determination unit 70 includes, for example, a voice guidance signal. The adaptable state determination unit 70 determines that the adaptation condition is not satisfied when the signal level of the voice guidance signal input from the navigation system 81 is equal to or higher than a predetermined threshold.

  The tachometer 83 inputs the rotational speed of the engine related to the road noise to the adaptable state determination unit 70. The adaptable state determination unit 70 satisfies the adaptation condition when the input engine speed is equal to or lower than a predetermined first speed (eg, 1000 rpm) or equal to or higher than a predetermined second speed (eg, 4000 rpm). It is determined not to. In addition, the traveling speed related to road noise is input from the speedometer 84 to the adaptable state determination unit 70. The adaptable state determination unit 70 satisfies the adaptation condition when the input traveling speed is equal to or lower than a predetermined first speed (for example, 40 km / h) or is equal to or higher than a predetermined second speed (for example, 130 km / h). It is determined not to.

  As described above, the reason why the adaptable state determination unit 70 makes the determination is that the road noise level is presumed to be smaller than that during normal traveling when the traveling speed is low or the engine speed is low. This is because it is considered that they have not reached. Further, when the traveling speed is considerably high or the engine speed is considerably high, it is estimated that the level of road noise is higher than that during normal traveling, and it is considered that the adaptive condition is exceeded.

  The signal input from the error microphones 2a and 2b to the adaptable state determination unit 70 is exactly the sound of the environment inside the vehicle, which is road noise during driving, passenger's conversation voice, reproduced sound by the audio system 82, and navigation system. 81 voice guidance, noise propagated from outside the vehicle (for example, noise of another vehicle traveling in parallel or passing each other), etc. are included. Therefore, the adaptable state determination unit 70 determines that the adaptation condition is not satisfied when the level of the signal input from the error microphones 2a and 2b is equal to or higher than the predetermined first threshold or equal to or lower than the second threshold.

  Next, a method in which the adaptable state determination unit 70 measures the level of the input signal will be described. FIG. 9A is a configuration diagram of the adaptable state determination unit 70. FIG. 9B is a diagram illustrating an example of the determination condition used by the adaptable state determination unit 70. As shown in FIG. 9A, the adaptable state determination unit 70 includes an instantaneous value level calculation unit 71, an averaging unit 72, and a threshold value determination unit 73.

  The instantaneous value level calculator 71 calculates the level (for example, -26 dB) at the moment when the output signal of the error microphone 2a is input.

  The averaging unit 72 averages the instantaneous value levels calculated by the instantaneous value level calculating unit 71 in a predetermined period. The predetermined period may be determined by time such as 1/10 seconds, or may be determined by the number of input instantaneous value levels such as a period during which 1000 instantaneous value levels are input. May be

  The threshold determination unit 73 determines whether the signal level (value) averaged by the averaging unit 72 is within a predetermined threshold range. FIG. 9B shows a graph showing a time-series change in the signal level averaged by the averaging unit 72, and a lower limit value THL1 and an upper limit value THL2 of the threshold range. The threshold value determination unit 73 determines that the adaptive condition is satisfied when the averaged signal level is equal to or higher than the lower limit value THL1 and equal to or lower than the upper limit value THL2.

  Therefore, as shown in the graph of FIG. 9B, when the signal level (value) averaged by the averaging unit 72 is input to the threshold determination unit 73, the averaged signal level remains unchanged until time t1. Since it is within the threshold range, the threshold determination unit 73 determines that the adaptation condition is satisfied. During the period from time t1 to time t2, the averaged signal level exceeds the upper limit value THL2, so the threshold value determination unit 73 determines that the adaptation condition is not satisfied. During the period from time t2 to time t3, since the averaged signal level is within the threshold range, the threshold determination unit 73 determines again that the adaptive condition is satisfied. During the period from time t3 to time t4, since the averaged signal level is less than the lower limit value THL1, it is determined that the adaptive condition is not satisfied.

  9A and 9B, an example in which the adaptable state determination unit 70 determines whether or not the adaptive condition is satisfied when the output signal of the error microphone 2a is input has been described. The same determination is made when the output signal of the error microphone 2b is input to the unit 70. As described above, the information used by the adaptable state determination unit 70 to determine whether or not the adaptive condition is satisfied is input from the audio system 82 and the speedometer 84 as well as the output signals of the error microphones 2a and 2b. It also includes information to be provided. The adaptable state determination unit 70 determines whether or not the adaptive condition is satisfied by using all the input information, and the environment inside the vehicle is determined only when it is determined that the adaptive condition is satisfied in all the determinations. It is determined that the adaptation condition for executing the adaptation operation is satisfied.

  According to the above configuration, the update of the control coefficient by the coefficient updater 60 is executed only when the adaptable state determination unit 70 determines that the environment inside the vehicle satisfies the adaptive condition for performing the adaptive operation. Therefore, the optimum control coefficient can be set more stably.

  However, when the adaptive operation is actually performed, as shown in FIGS. 3 and 4, the road noise does not increase, and only the reducing effect is hardly obtained. This is because, as shown in FIGS. 15A, 15B and 4, the locations where the sensors 1a, 1b, 1c, 1d, the error microphones 2a, 2b, 2c, 2d or the speakers 3a, 3b, 3c, 3d are installed. , Because there are practical limits. Hereinafter, the sensors 1a, 1b, 1c, and 1d are collectively referred to as the sensor 1 when collectively referred to. The error microphones 2a, 2b, 2c, and 2d are collectively referred to as the error microphone 2. The speakers 3a, 3b, 3c, and 3d are collectively referred to as the speaker 3.

  FIG. 10 is a diagram showing a distance D1 from the sensor 1 to the error microphone 2 and a distance D2 from the speaker 3 to the error microphone 2 in the noise control device 1001. For example, as shown in FIG. 10, the difference D1-D2 between the distance D1 from the sensor 1 that detects noise to the error microphone 2 and the distance D2 from the speaker 3 to the error microphone 2 is the signal processing in the noise control device 1001. It is not possible to secure a sufficient distance with respect to time. In this case, the causality rule in the noise control device 1001 is not satisfied.

When the signal processing time in the noise control device 1001 is T, the expression 4 must be satisfied at all frequencies in order for the causality condition to be satisfied.
T ≦ (D1-D2) / v (Formula 4)
Here, v indicates the speed of sound.

  However, as described above, if the distance D1-D2 is not sufficiently long, the causality condition (Equation 4) cannot be satisfied, particularly when processing a high frequency signal having a short wavelength. On the other hand, in consideration of the noise reduction effect, the noise reduction effect tends to be improved as the sensor 1 for detecting noise is closer to the control point where the error microphone 2 is installed. Therefore, if the sensor 1, the error microphone 2, and the speaker 3 are installed in consideration of the noise reduction effect, the distance D1-D2 becomes short, and it becomes difficult to satisfy the causal laws.

  Further, the characteristics of the speaker 3 also affect the causality condition. In particular, the speaker 3 has a large phase rotation at the low resonance frequency, and the delay (group delay) of signals near the low resonance frequency becomes large. Therefore, it becomes difficult to satisfy the causal laws when processing a signal near the low resonance frequency. That is, in the noise control device 1001, it is necessary to make the distance D1-D2 sufficiently long in order to correct the group delay of a signal having a low resonance frequency or lower.

  If the causality condition is not sufficiently satisfied, the noise reduction effect of the noise control device 1001 is as shown in FIG. 11, for example. FIG. 11 is a diagram showing another example of the noise reduction effect measured by the effect measurement unit 50. In the example shown in the sections (a) and (b) of FIG. 11, the road noises of the frequencies f1 to f3 are increased, but this is caused by the influence of the group delay near the low resonance frequency of the speaker 3. Often The reason why the road noise of the frequency f1 or lower does not increase is that the sound of the frequency f1 or lower cannot be reproduced due to the performance of the speaker 3.

  Further, the road noises of the frequencies f4 to f2 also increase, because the phase shift easily occurs due to the high frequency. It should be noted that the reason why the road noise of frequency f2 or higher does not increase is that the signal level of the road noise itself is low, and the control filter 20aa, 20ab, 20ba, 20bb convolves the control coefficient with the result that the signal level is further increased. Because it will decrease.

  As described above, the frequency band in which the expected noise reduction effect is obtained and the frequency band in which the unexpected noise increase occurs are mixed. This is a general control effect in the majority of noise control cases. Is. Therefore, achieving the desired noise reduction effect and suppressing the increase in noise is a problem in actually designing the control coefficient.

  The determination of the noise reduction effect, which is a key point in designing the control coefficient, will be described below with reference to FIG. FIG. 12 is an operation flow chart showing the flow of the control coefficient designing operation based on the result of determining the noise reduction effect in the effect measurement unit 50. The flow shown in FIG. 12 corresponds to step S7 in FIG.

  That is, it is assumed that the effect measurement unit 50 starts the determination process of step S7 that determines whether or not the road noise reduction effect at the control point that is the setting location of the error microphone 2 has reached the target value. In this case, as shown in FIG. 12, the A characteristic filter units 51a and 51b (FIG. 2) use the A characteristic coefficient for the noise control before signal off1 and the noise control signal on1 input to the effect measurement unit 50. A convolution process is performed (step P1). Next, the frequency analysis units 52a and 52b (FIG. 2) perform frequency analysis processing to calculate frequency characteristics of the pre-noise control signal off1 and the post-noise control signal on1 after the convolution processing in step P1 (step P2). ).

  When Step P2 is performed, the frequency difference effect calculation unit 54a (FIG. 2) causes the pre-noise control signal off1 and the A characteristic after the convolution processing by the A characteristic filter unit 51a for each frequency in the frequency characteristic calculated in Step P2. The first difference, which is the difference from the noise-controlled signal on1 after the convolution processing by the filter unit 51b, is calculated (step P4).

  On the other hand, when step P2 is performed, the overall calculation units 53a and 53b (FIG. 2) calculate the first overall value and the second overall value, respectively (step P3). The effect measuring unit 50a is assumed to have a configuration including band limiting units 55a and 55b as shown in FIG. In this case, in step P3, the overall calculation section 53a may calculate the overall value in the entire frequency band of the signal extracted by the band limiting section 55a as the first overall value. Similarly, the overall calculating unit 53b may calculate the overall value in the entire frequency band of the signal extracted by the band limiting unit 55b as the second overall value. Next, the overall value difference effect calculation unit 54b calculates the second difference that is the difference between the first overall value calculated in step P3 and the second overall value (step P5).

  The effect measurement unit 50 determines whether or not the second difference calculated in Step P5 has reached the preset second target value (Step P6). For example, assume that the second target value is set to -3dBA. In this case, the effect measurement unit 50 determines that the second difference has reached the second target value when the second difference is less than the second target value.

  On the other hand, the effect measurement unit 50 uses the first difference of each frequency calculated in step P4 to determine the majority of the frequencies included in the predetermined effect expected band (FIG. 11) within the predetermined evaluation target band. It is determined whether or not the first difference has reached a preset first target value (step P7). For example, assume that the first target value is set to 5 dB. In this case, when the first difference of the majority frequency among the frequencies included in the expected effect band (FIG. 11) is larger than the first target value, the effect measurement unit 50 determines that the first difference of the majority frequency is It is determined that the first target value has been achieved.

  In addition, in step P7, the effect measurement unit 50 may perform the determination under more severe conditions. For example, whether or not the effect measuring unit 50 has achieved the first target value by the first difference of a predetermined number (for example, 80%) or more of a majority of frequencies included in the expected effect band. Alternatively, it may be determined.

  In addition, the effect measurement unit 50 uses the first difference of each frequency calculated in step P4 to determine the majority of the frequencies included in the predetermined noise increase band (FIG. 11) within the predetermined evaluation target band. It is determined whether the first difference exceeds a preset allowable value (step P8). For example, assume that the allowable value is set to 2 dB. In this case, when the first difference of the majority frequency among the frequencies included in the noise increase band (FIG. 11) is larger than the allowable value, the effect measuring unit 50 determines that the first difference of the majority frequency is the allowable value. It is determined that it exceeds.

  In addition, in step P8, the effect measurement unit 50 may perform the determination under more severe conditions. For example, the effect measurement unit 50 determines whether or not the first difference of a predetermined number (for example, 30%) of less than a majority of the frequencies included in the noise increase band exceeds the allowable value. You may do it. In addition, the effect measuring unit 50 determines whether or not the first difference of one or more frequencies among the frequencies included in the noise increase band exceeds the allowable value, and makes a determination under more severe conditions. You may do it. Alternatively, in step P8, the effect measurement unit 50 may make the determination under a gentler condition. For example, the effect measurement unit 50 determines whether the first difference of a predetermined number (for example, 70%) or more of a majority of frequencies included in the noise increase band exceeds the allowable value. You may do it.

  Then, the effect measurement unit 50 determines in step P6 that the second difference does not reach the second target value (NO in step P6), or (OR), and in step P7, the first difference of the majority frequency. Is determined not to have reached the first target value (NO in step P7). In this case, it is assumed that the effect measurement unit 50 further (AND2) determines in step P8 that the first difference of the majority frequency does not exceed the allowable value (NO in step P8). In this case, the effect measurement unit 50 determines that the road noise reduction effect at the control point does not reach the target value (NO in step S7). In this case, the effect measuring unit 50 determines that the control coefficient has not converged to the optimum value and continues the adaptive operation to continue the design of the control coefficient (step P9 (corresponding to NO in step S7 of FIG. 8)). ).

  Further, the effect measurement unit 50 determines that the second difference has reached the second target value in Step P6 (YES in Step P6), and (AND1), and the first difference of the majority frequency in Step P7. Is determined to have achieved the first target value (YES in step P7). In this case, it is assumed that the effect measurement unit 50 further determines (AND1) that the first difference of the majority frequency does not exceed the allowable value in step P8 (NO in step P8). In this case, the effect measurement unit 50 determines that the road noise reduction effect at the control point has reached the target value (corresponding to YES in step S7). In this case, the effect measuring unit 50 determines that the control coefficient has converged to the optimum value, normally completes the design of the control coefficient, and fixes the control coefficient to the optimum value (step P10 (corresponding to step S8 in FIG. 8). )).

  Further, it is assumed that the effect measurement unit 50 determines in step P8 that the first difference of the majority frequency exceeds the allowable value (YES in step P8). In this case, it is assumed that the increase in noise is at a level that cannot be ignored. Therefore, the effect measuring unit 50 determines that an abnormality has occurred in the control coefficient during the execution of step S7, and forcibly cancels the coefficient design (step P11 (corresponding to step S9 in FIG. 8)).

  According to the above configuration, it is possible to realize a realistic and practical design of the control coefficient even in the case where the increase in noise as shown in FIG. 11 occurs. Further, of the frequencies included in the expected effect band, the ratio of frequencies at which the first difference reaches the first target value, and of the frequencies included in the noise increase band, the ratio of frequencies at which the first difference reaches the allowable value. By grasping the above, it is possible to suppress an increase in undesired noise while realizing a desired noise reduction effect. This makes it possible to properly balance the design of control coefficients. As a result, in any case, the user can be provided with the optimal control effect at that time.

  Note that, for example, when the noise control device 1001 is applied to an automobile, the characteristics of road noise are often different between the front seat (driver's seat and front passenger seat) and the rear seat. Therefore, for each control point that is the setting location of each error microphone 2 installed in the vehicle, it is determined whether or not the road noise reduction effect at each control point has reached the target value (step S7). When performing, the same target value may be used. However, instead of this, an individual target value that is individually set in advance for each error microphone 2 may be used. Then, a value corresponding to each individual target value may be individually set as the first target value, the second target value, and the allowable value.

  In this case, the noise reduction effect is optimized for each seat. Particularly, it is assumed that the noise control device 1001 is applied to a space having a large number of seats such as an airplane and various types of seats such as window seats and aisle seats. In such a case, the individual target value corresponding to each seat in which the error microphone 2 is set is individually set, and the first target value, the second target value, and the allowable value corresponding to the individual target value are set. Alternatively, they may be set individually.

  For example, in FIG. 7, the noise reduction effect of the error microphone 2a installed near the head of the driver's seat is measured by the effect measurement unit 50a, and the noise reduction effect of the error microphone 2b installed near the head of the passenger seat. Is measured by the effect measuring unit 50b. In this case, the target values used by the effect measuring unit 50a and the effect measuring unit 50b in the determination process of step S7 may be the individual target value Ka and the individual target value Kb, respectively. In response to this, the first target value, the second target value, and the allowable value used by the effect measurement unit 50a in steps P7, P6, and P8 include the first individual target value K1a corresponding to the individual target value Ka, and the first target value K1a. (2) The individual target value K2a and the individual allowable value K3a may be set. Similarly, the effect measuring unit 50b uses the first target value, the second target value, and the allowable value used in steps P7, P6, and P8 as the first individual target value K1b and the second individual target value K1b according to the individual target value Kb. The target value K2b and the individual allowable value K3b may be set. Note that the expected effect band and the noise increase band shown in FIG. 11 are not limited to these, and may be set individually so as to be associated with each error microphone 2.

  As shown in FIG. 7, in the noise control device 1001 as a whole, noises at the installation locations of the error microphones 2a and 2b are simultaneously controlled. Therefore, not only the control filters 20aa and 20ba control noise at the installation location of the error microphone 2a, but similarly, only the control filters 20ab and 20bb control noise at the installation location of the error microphone 2b. Do not mean.

  That is, the control as a whole acts so that the noises at the installation locations of the error microphones 2a and 2b are collectively optimized. Therefore, as described above, when the target value or the like is individually set for each error microphone 2, if the value is set to deviate significantly from the other target values or the like, noise at the installation location of the error microphones 2a and 2b is reduced. Is not optimized, and it is possible to fall into a state where the control coefficient design is not completed indefinitely.

  For example, assume that the second individual target value K2a corresponding to the error microphone 2a is set to 3dBA and the second individual target value K2b corresponding to the error microphone 2b is set to 4dBA. In this case, when the second target value is not individually set for each error microphone 2, assuming that the noise reduction effects of the error microphones 2a and 2b both settle within the range of 3.0 to 3.5 dBA, If the individual target value K2b is set to 4dBA, this may hinder the design of the control coefficient from being completed.

  Therefore, in order to avoid such a situation, in the case of a control configuration in which a plurality of control points are collectively controlled, the control points are prioritized within the control configuration unit, and individual control points with a high priority are assigned. When the target value is reached, the control coefficient design may be completed. For example, if the second individual target value K2a is set to 3dBA and the priority is set to the highest, even if the second individual target value K2b is set to 4dBA, regardless of the noise reduction effect at the installation location of the error microphone 2b, The design of the control coefficient can be completed when the noise reduction effect at the installation location of the error microphone 2a becomes 3 dBA or more. When the control coefficient design is completed, the control coefficient at the time of completion may be the final control coefficient.

  On the other hand, when the noise is increasing, it is considered that it is not preferable to exceed the allowable value at any control point. Therefore, at one control point among all the control points, the reduction effect is less than the allowable value. If it exceeds, the coefficient design may be stopped. When the adaptive operation is stopped, the control coefficient with the best effect up to the stop may be used as the final control coefficient.

  Note that, for example, in the case of the automobile 100, it can be assumed that the seats (driver's seat and passenger seat) in the front of the vehicle body and the rear seats are all “in the control unit”, but when controlling noise in a large space such as an aircraft. Does not need to collectively control the noise in the control constituent unit as "in the control constituent unit" by grouping seats that are separated from each other by a predetermined distance or more. For example, the control structural unit may be constructed such that the adjacent seats are “within the control structural unit”.

  From the above explanation, the noise reduction effect is measured, and based on the result, whether the control coefficient design has been completed or whether it is necessary to continue the control coefficient design, and the noise level at a specific frequency It was possible to show the overall flow of the control coefficient design operation, which is whether the control coefficient design should be stopped accordingly.

  On the other hand, for example, when the noise control device 1001 is applied to an aircraft, a seat in front of the engine (first class or business class), a seat next to the engine (part of business class or economy class), and a seat behind the engine (economy). The noise level and the noise frequency characteristic are remarkably different between the two types. Further, since the number of seats in the aircraft is 100 to 200, or more, it is usual that the optimum noise reduction effect is different for each seat. Therefore, as described above, it is conceivable to install the error microphone 2 for each seat and individually set the first target value, the second target value, and the allowable value corresponding to each error microphone 2. However, other than that, it is preferable that the operating condition of the adaptive operation for updating the control coefficient is also set individually for each error microphone 2.

Specifically, the operating condition is the convergence constant μ of the LMS calculators 61aaa, 61aaab, 61aba, 61abb, 61baa, 61bab, 61bba, 61bbb. Hereinafter, the LMS calculators 61aaa, 61aab, 61aba, 61abb, 61baa, 61bab, 61bba, 61bbb are collectively referred to as the LMS calculator 61. As described in Patent Document 1 and the like, in the LMS calculator 61, the control coefficient is updated according to the following Expression 5.
W (n + 1) = W (n) -μ · er · r Equation (5)
Here, W (n) indicates the control coefficient of the control filter before update (for example, the control filter 20aa in FIG. 7), and W (n + 1) indicates the control coefficient of the control filter after update.
e indicates an error signal (for example, the output signal of the error microphone 2a in FIG. 7).
r indicates a reference signal (for example, the output signal of the transfer characteristic correction filter 62aaa in FIG. 7).
μ represents a convergence constant (step size parameter).
-Indicates multiplication.

  That is, the convergence constant μ is a value for adjusting the convergence speed and the degree of convergence. When the convergence constant μ increases, the speed at which the control coefficient converges to the optimum value (hereinafter, convergence speed) increases, but the risk that the control coefficient update operation diverges increases. On the contrary, when the convergence constant μ becomes small, the control coefficient update operation is performed stably, but the convergence speed becomes slow and there is a problem that it takes time until the noise reduction effect is sufficiently obtained. ..

  Therefore, it is important to set an appropriate convergence constant μ. However, when a number of seats have different noise characteristics and noise levels as in an aircraft, it is considered that the optimum convergence constant μ for each seat also differs. Since it takes a great deal of time to check the optimum value of the convergence constant μ in advance, it is desirable that the noise control device 1001 automatically derives the optimum value of the convergence constant μ. Therefore, a method of deriving the optimum value of the convergence constant μ will be described below.

  13A and 13B are operation flowcharts showing the flow of the design operation of the control coefficient in the entire noise control device 1001. The operation flow shown in FIGS. 13A and 13B includes the same steps as those shown in FIGS. 8 and 12. Hereinafter, detailed description of the same step will be omitted, and a method of deriving an optimum value of the convergence constant μ will be mainly described.

  As shown in FIG. 13A, the effect measurement unit 50 sets a predetermined initial value to the convergence constant μ used by the LMS calculator 61 before executing step S1 (step S0). The convergence constant μ is a decimal number of 0 or more and 1 or less. For example, the initial value of the convergence constant μ is set to a value close to 0 in consideration of the stability of adaptive operation. However, the initial value of the convergence constant μ is not limited to this, and may be 0. In step S0, when the initial value is set for the convergence constant μ, the processes of step S1 and subsequent steps are performed.

  Thereafter, as shown in FIG. 13B, the effect measurement unit 50 determines that the second difference does not reach the second target value in step P6 (NO in step P6), or (OR), and in step P7, It is assumed that it is determined that the first difference of the majority of frequencies does not reach the first target value (NO in step P7). In this case, it is assumed that the effect measurement unit 50 further (AND2) determines in step P8 that the first difference of the majority frequency does not exceed the allowable value (NO in step P8). As a result, it is assumed that the effect measurement unit 50 determines that the road noise reduction effect at the control point does not reach the target value (NO in step S7).

  In this case, the effect measuring unit 50 predetermines the convergence constant μ at the time of calculating the first difference in step P4 or at the time of calculating the second difference in step P5, assuming that the control coefficient has not converged to the optimum value. A new convergence constant μ + Δ is obtained by adding the predetermined value Δ. Then, the effect measurement unit 50 restarts the update of the control coefficient using the new convergence constant μ + Δ by the coefficient updater 60. As a result, the effect measurement unit 50 continues the adaptive operation (step S79). After that, the processing from step S1 is performed.

  Therefore, after the execution of steps S1 to S6, the convergence constant μ increases by a predetermined value Δ each time the control coefficient design process from step P1 to step S79 is repeated. During that time, since the noise reduction effect is also measured, the convergence constant μ is adjusted to the convergence constant μ that provides the optimum noise reduction effect.

  In addition, the effect measurement unit 50 determines that the second difference has reached the second target value in Step P6 (YES in Step P6), and (AND1), and the first difference of the majority frequency in Step P7. Is determined to have achieved the first target value (YES in step P7). In this case, it is assumed that the effect measurement unit 50 further determines (AND1) that the first difference of the majority frequency does not exceed the allowable value in step P8 (NO in step P8). As a result, it is assumed that the effect measurement unit 50 determines that the road noise reduction effect at the control point has reached the target value (corresponding to YES in step S7). In this case, the effect measuring unit 50 determines that the control coefficient has converged to the optimum value, normally completes the design of the control coefficient, and fixes the control coefficient to the control coefficient at the time of completion (the last control coefficient) (step). S81 (corresponding to step S8 in FIG. 8)).

  In addition, when the effect measuring unit 50 determines in step P8 that the first difference of the majority frequency exceeds the allowable value (YES in step P8), the increase in noise is at a level that cannot be ignored. Since it is assumed, it is determined that an abnormality has occurred in the control coefficient during the execution of step S7. In this case, the design operation of the control coefficient is forcibly stopped, and the control coefficient is fixed to the optimum value when it is determined that the abnormality has converged to the optimum value in step S81 before it is determined that the abnormality has occurred (step S81). P91 (corresponding to step S9 in FIG. 8)).

  As described above, in the noise control device 1001, adaptive operation when the convergence constant μ is the initial value, measurement of the road noise reduction effect by fixing the control coefficient, the convergence constant μ is updated to a new convergence constant μ + Δ, and new. The adaptive operation using a different convergence constant μ + Δ is repeated. Accordingly, even if the noise control device 1001 is applied to a large space including a large number of seats such as an aircraft, the convergence constant μ can be automatically adjusted to the optimum convergence constant μ. As a result, it is possible to quickly realize the optimum noise reduction effect in each seat.

  In the above embodiment, the example in which the noise control device 1001 is applied to the automobile 100 or the aircraft is shown, but the application destination of the noise control device 1001 is not limited to this.

(Modified embodiment)
Although the embodiment of the present disclosure has been described above, the embodiment of the present disclosure is not limited to the above embodiment, and may be, for example, a modified embodiment described below.

  In the noise control device 1001 of the second embodiment, step P8 and step P11 may be omitted. In this case, the effect measurement unit 50 determines in step P6 that the second difference has not reached the second target value (NO in step P6), or (OR), and in step P7, the first of the majority frequencies is determined. It is assumed that it is determined that the one difference does not reach the first target value (NO in step P7). In this case, the effect measurement unit 50 may determine that the road noise reduction effect at the control point does not reach the target value (NO in step S7). Further, the effect measurement unit 50 determines that the second difference has reached the second target value in Step P6 (YES in Step P6), and (AND1), and the first difference of the majority frequency in Step P7. Is determined to have achieved the first target value (YES in step P7). In this case, the effect measurement unit 50 may determine that the road noise reduction effect at the control point has reached the target value (corresponding to YES in step S7).

  Furthermore, in the noise control device 1001 of the second embodiment, step P7 may be omitted. In this case, when the effect measuring unit 50 determines in step P6 that the second difference does not reach the second target value (NO in step P6), the road noise reduction effect at the control point achieves the target value. It may be determined that it has not been performed (NO in step S7). When the effect measuring unit 50 determines that the second difference has reached the second target value in step P6 (YES in step P6), the road noise reduction effect at the control point has reached the target value. May be determined (corresponding to YES in step S7).

  Alternatively, in the noise control device 1001 of the second embodiment, step P6 may be omitted. In this case, when the effect measurement unit 50 determines in step P7 that the first difference of the majority frequency does not reach the first target value (NO in step P7), the road noise reduction effect at the control point is obtained. May be determined to have not reached the target value (corresponding to NO in step S7). In addition, when the effect measuring unit 50 determines in step P7 that the first difference of the majority frequency has reached the first target value (YES in step P7), the road noise reduction effect at the control point is reduced. It may be determined that the target value has been reached (YES in step S7).

  Further, the sensors 1, 1a, 1b, 1c, 1d described above may be microphones that detect noise generated at the installation location and output a noise signal indicating the detected noise.

1, 1a, 1b, 1c, 1d sensor (noise detector)
2, 2a, 2b Error microphone 3, 3a, 3b, 3c, 3d Speakers 20aa, 20ab, 20ba, 20bb Control filter 40aa, 40ba Transfer characteristic correction filter (correction filter)
41a, 41b Subtractor 50, 50a, 50b Effect measurement unit 52a, 52b Frequency analysis unit 53a, 53b Overall calculation unit 54a Frequency difference effect calculation unit 54b Overall value difference effect calculation unit 60 Coefficient updater 61, 61aaa, 61aaab, 61aba, 61abb, 61baa, 61bab, 61bba, 61bbb LMS calculator (coefficient updater)
62aaa, 62aab, 62aba, 62abb, 62baa, 62bab, 62bba, 62bbb Transfer characteristic correction filter (coefficient updater)
70 Adaptable state determination unit 1000, 1001 Noise control device C11, C12, C21, C22 Transfer characteristic Ka, Kb Individual target value e1 Error signal off1, off2 Noise control pre-signal (control off signal)
on1, on2 Noise control signal (control on signal)
μ convergence constant

Claims (19)

A noise detector for detecting noise generated by a noise source,
A control filter that performs signal processing using a predetermined control coefficient for a noise signal indicating noise detected by the noise detector,
A speaker that reproduces the output signal of the control filter as a control sound,
An error microphone installed at a control point where interference between noise propagating from the noise source and control sound reproduced by the speaker occurs, and detecting residual noise remaining at the control point due to the interference,
A transfer characteristic correction filter that performs signal processing of the noise signal using transfer characteristics of sound from the speaker to the error microphone,
Using an error signal indicating the residual noise detected by the error microphone and the output signal of the transfer characteristic correction filter, a coefficient updater for updating the control coefficient so as to minimize the error signal,
A correction filter that performs signal processing on the output signal of the control filter using the transfer characteristic of sound from the speaker to the error microphone,
A subtractor that subtracts the output signal of the correction filter from the error signal,
The output signal of the subtractor is a control-off signal representing noise before control due to the interference, the error signal is a control-on signal representing noise after control due to the interference, and the control-off signal and the control-on signal Based on the difference with, and an effect measurement unit that measures the noise reduction effect at the control point,
A noise control device. Further comprising an adaptable state determination unit that determines whether or not the coefficient updater should update the control coefficient,
The noise control device according to claim 1. The coefficient updater updates the control coefficient using a predetermined convergence constant,
The effect measurement unit,
The difference between the control-off signal and the control-on signal is measured as the reduction effect,
Perform a determination process to determine whether the reduction effect has reached a predetermined target value,
In the determination process,
When it is determined that the reduction effect has reached the target value, it is determined that the control coefficient has converged to an optimum value, and the updating of the control coefficient by the coefficient updater is stopped to set the control coefficient to the optimum value. Fixed to
When it is determined that the reduction effect does not reach the target value, it is determined that the control coefficient has not converged to an optimum value, and the convergence constant used by the coefficient updater at the time of measuring the reduction effect is used. A new convergence constant is obtained by adding a predetermined value, and the update of the control coefficient by the coefficient updater using the new convergence constant is restarted.
The noise control device according to claim 1. The effect measurement unit,
For each of the control off signal and the control on signal, signal processing is performed using an A characteristic coefficient indicating an A characteristic simulating human auditory characteristics, and the control off signal after the signal processing and the signal after the signal processing are performed. Measuring the difference with the control ON signal as the reduction effect,
The noise control device according to claim 3. The effect measurement unit,
A frequency analysis unit for calculating frequency characteristics of the control-off signal and the control-on signal;
For each frequency in the frequency characteristic, a first difference that is a difference between the control-off signal and the control-on signal, a frequency difference effect calculation unit that calculates as an index of the reduction effect,
The noise control device according to claim 1 or 2, further comprising: The effect measurement unit,
A frequency analysis unit for calculating frequency characteristics of the control-off signal and the control-on signal;
An overall calculator that calculates an overall value in each frequency band of the control-off signal and the control-on signal using the frequency characteristic,
A second difference, which is the difference between the overall value of the control off signal and the overall value of the control on signal, an overall value difference effect calculation unit that calculates as an index of the reduction effect,
The noise control device according to claim 1 or 2, further comprising: The effect measurement unit,
A frequency analysis unit for calculating frequency characteristics of the control-off signal and the control-on signal;
For each frequency in the frequency characteristic, a first difference that is a difference between the control-off signal and the control-on signal, a frequency difference effect calculation unit that calculates as an index of the reduction effect,
An overall calculator that calculates an overall value in each frequency band of the control-off signal and the control-on signal using the frequency characteristic,
A second difference, which is the difference between the overall value of the control off signal and the overall value of the control on signal, an overall value difference effect calculation unit that calculates as an index of the reduction effect,
The noise control device according to claim 1 or 2, further comprising: The effect measurement unit,
Using the frequency characteristic, further comprising a band limiting unit for extracting a signal of a frequency within a predetermined evaluation target frequency band included in each of the control OFF signal and the control ON signal,
The overall calculation unit calculates an overall value in each frequency band of the signals extracted from the control off signal and the control on signal by the band limiting unit,
The overall value difference effect calculation unit determines the difference between the overall value of the signal extracted from the control-off signal by the band limiting unit and the overall value of the signal extracted from the control-on signal by the band limiting unit. Two differences,
The noise control device according to claim 6 or 7. The coefficient updater updates the control coefficient using a predetermined convergence constant,
The effect measurement unit,
Perform a determination process to determine whether the reduction effect has reached a predetermined target value,
In the determination process,
The first difference of a majority of the frequencies included in the predetermined evaluation target frequency band calculated by the frequency difference effect calculation unit has reached a predetermined first target value according to the target value. In this case, it is determined that the reduction effect has reached the target value, and assuming that the control coefficient has converged to the optimum value, the update of the control coefficient by the coefficient updater is stopped and the control coefficient is optimized. Fixed to the value,
If the first difference of the majority of the frequencies included in the evaluation target frequency band calculated by the frequency difference effect calculation unit does not reach the first target value, the reduction effect is the target value. It is determined that the control coefficient has not converged to an optimum value, and a value obtained by adding a predetermined value to the convergence constant used by the coefficient updater at the time of calculating the first difference is determined as As a new convergence constant, the coefficient updater restarts updating the control coefficient using the new convergence constant,
The noise control device according to claim 5. The coefficient updater updates the control coefficient using a predetermined convergence constant,
The effect measurement unit,
Perform a determination process to determine whether the reduction effect has reached a predetermined target value,
In the determination process,
When the second difference has reached a predetermined second target value according to the target value, it is determined that the reduction effect has reached the target value, and the control coefficient has converged to an optimum value. As, the update of the control coefficient by the coefficient updater is stopped to fix the control coefficient to the optimum value,
When the second difference does not reach the second target value, it is determined that the reduction effect does not reach the target value, and the control coefficient does not converge to an optimum value, and the second A new value is added to the convergence constant used by the coefficient updater at the time of calculation of the difference as a new convergence constant, and the update of the control coefficient by the coefficient updater using the new convergence constant is restarted. Let
The noise control device according to claim 6. The coefficient updater updates the control coefficient using a predetermined convergence constant,
The effect measurement unit,
Perform a determination process to determine whether the reduction effect has reached a predetermined target value,
In the determination process,
Calculated by the frequency difference effect calculation unit, the first difference of a majority of the frequencies included in the predetermined evaluation target frequency band achieves a predetermined first target value according to the target value, and, When the second difference has reached a predetermined second target value according to the target value, it is determined that the reduction effect has reached the target value, and the control coefficient has converged to an optimum value. As, the update of the control coefficient by the coefficient updater is stopped to fix the control coefficient to the optimum value,
If the first difference of the majority of the frequencies included in the evaluation target frequency band calculated by the frequency difference effect calculation unit does not reach the first target value, the reduction effect is the target value. It is determined that the control coefficient has not converged to an optimum value, and a value obtained by adding a predetermined value to the convergence constant used by the coefficient updater at the time of calculating the first difference is determined as As a new convergence constant, the coefficient updater restarts the update of the control coefficient using the new convergence constant,
When the second difference does not reach the second target value, it is determined that the reduction effect does not reach the target value, and the control coefficient does not converge to an optimum value, and the second A new value is added to the convergence constant that was used by the coefficient updater at the time of calculating the difference, and is set as a new convergence constant, and the update of the control coefficient by the coefficient updater using the new convergence constant is restarted. Let
The noise control device according to claim 7. The effect measurement unit,
In the determination process, the frequency difference effect calculation unit calculates, among the frequencies in the predetermined noise increase band included in the evaluation target frequency band, the first difference of a predetermined number or more of the frequencies, the target value. If the control coefficient exceeds a predetermined allowable value, it is determined that an abnormality has occurred in the control coefficient, and the update of the control coefficient by the coefficient updater is stopped.
The noise control device according to claim 9.   The noise control device according to claim 12, wherein the predetermined number is one. Equipped with a plurality of the error microphone,
The effect measurement unit,
For each of the plurality of error microphones, the location where each error microphone is installed is used as the control point, and the individual target value set individually for each error microphone is used as the target value, and the determination process is performed. ,
The noise control device according to claim 3, 4, 9, 10, 12, or 13. Priority is associated with the individual target value,
The effect measurement unit,
In the determination process of using the individual target value associated with the highest priority as the target value, when it is determined that the reduction effect has achieved the target value, the above for all the control points In the determination process, it is determined that the reduction effect has reached the target value,
The noise control device according to claim 14. The adaptable state determination unit,
When the value obtained by averaging the instantaneous value level of the error signal in a predetermined period is within a predetermined threshold range, it is determined that the coefficient updater should update the control coefficient,
The noise control device according to claim 2. The computer of the noise control device
The noise generated by the noise source is detected using a sensor,
The first signal processing is performed using a predetermined control coefficient for the noise signal indicating the noise detected by the sensor,
The signal after the first signal processing is reproduced by a speaker as a control sound,
Using the error microphone installed at the control point where interference between the noise propagated from the noise source and the control sound reproduced by the speaker occurs, the residual noise remaining at the control point due to the interference is detected,
Performing second signal processing on the noise signal using the transfer characteristics of the sound from the speaker to the error microphone,
Using the error signal indicating the residual noise detected by the error microphone and the signal after the second signal processing, updating the control coefficient so as to minimize the error signal,
Performing a third signal processing on the signal after the first signal processing using the transfer characteristics of the sound from the speaker to the error microphone,
Subtract the signal after the third signal processing from the error signal,
The signal after the subtraction is a control off signal representing noise before control due to the interference, the error signal is a control on signal representing noise after control due to the interference, and the control off signal and the control on signal Based on the difference of, to measure the noise reduction effect at the control point,
Noise control method.   A program for causing a computer to execute the noise control method according to claim 17. A noise detector for detecting noise generated by a noise source,
A control filter that performs signal processing using a predetermined control coefficient for a noise signal indicating noise detected by the noise detector,
A speaker that reproduces the output signal of the control filter as a control sound,
An error microphone installed at a control point where interference between noise propagating from the noise source and control sound reproduced by the speaker occurs, and detecting residual noise remaining at the control point due to the interference,
A correction filter that performs signal processing on the output signal of the control filter using the transfer characteristic of sound from the speaker to the error microphone,
A subtractor that subtracts the output signal of the correction filter from the error signal,
The output signal of the subtractor is a control-off signal representing noise before control due to the interference, the error signal is a control-on signal representing noise after control due to the interference, and the control-off signal and the control-on signal Based on the difference with, and an effect measurement unit that measures the noise reduction effect at the control point,
A noise control device.

Images