Active noise reduction device and active noise reduction program

Abstract

To provide an inexpensive active noise reduction system that can maintain control performance even if acoustic transmission characteristics change.SOLUTION: An active noise reduction system 11 includes: a plurality of noise microphones 13A-13E that generate a plurality of noise signals x on the basis of noise; and a controller 16 that controls canceling sound output devices 12A-12D on the basis of the plurality of noise signals x. The controller 16 selects a plurality of reference signals r' corresponding to noise and an error signal e corresponding to an error between the noise and canceling sound y from among the plurality of noise signals x, generates a control signal u from the plurality of reference signals r' using a plurality of control filters W, and adaptively updates the plurality of control filters W using a plurality of acoustic transmission characteristic filters C^A and C^B. The plurality of acoustic transmission characteristic filters C^A and C^B include an adaptive update filter C^B to be adaptively updated, and a non-adaptive update filter C^A to be updated on the basis of an update value of the adaptive update filter C^B.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an active noise reduction device and an active noise reduction program that reduce noise by interfering with the noise with a canceling sound that is in the opposite phase to the noise. [Background technology]

[0002] 2. Description of the Related Art Active noise reduction devices are known that reduce noise by interfering with a canceling sound that is in the opposite phase to the noise.

[0003] For example, Patent Document 1 discloses an active noise reduction device (noise cancellation device) that includes a speaker that outputs a cancellation sound, an acceleration sensor that generates a signal according to the noise, an error microphone that detects a composite sound of the noise and the cancellation sound and outputs a composite sound signal, and an adaptive signal processing unit that controls the speaker based on signals from the acceleration sensor and the error microphone. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 7-28474 Summary of the Invention [Problem to be solved by the invention]

[0005] In the above-described conventional technology, an acceleration sensor is used to generate a signal corresponding to noise, but acceleration sensors are generally relatively expensive components, so generating a signal corresponding to noise using an acceleration sensor may result in an increase in the cost of the active noise reduction device.

[0006] Furthermore, in the above-described conventional technology, an estimated value of the acoustic transfer characteristic from the speaker to the error microphone (see "Filter FXF") is used as a control parameter. However, the acoustic transfer characteristic may change due to various factors (for example, aging of the vehicle body, changes in the opening and closing state of the windows, and changes in the inclination of the seat). When the acoustic transfer characteristic changes in this way, an error occurs between the acoustic transfer characteristic and its estimated value, which may result in a decrease in the control performance (i.e., noise reduction performance) of the active noise reduction device or the generation of abnormal noise.

[0007] In view of the above background, an object of the present invention is to provide an inexpensive active noise reduction device that can maintain control performance even when acoustic transmission characteristics change. [Means for solving the problem]

[0008] In order to solve the above-mentioned problems, one aspect of the present invention is an active noise reduction device (11) comprising: a noise cancellation output device (12A-12D) that outputs a cancellation sound to cancel out a noise; a plurality of noise microphones (13A-13E) that generate a plurality of noise signals based on the noise; and a control device (16) that controls the noise cancellation output device based on the plurality of noise signals, wherein the control device acquires the plurality of noise signals output from the plurality of noise microphones, selects from the plurality of noise signals a plurality of reference signals corresponding to the noise and an error signal corresponding to an error between the noise and the cancellation sound, generates a control signal for the noise cancellation output device from the plurality of reference signals using a plurality of control filters (W), and adaptively updates the plurality of control filters using a plurality of acoustic transfer characteristic filters (C^), wherein the plurality of acoustic transfer characteristic filters include an adaptive update filter (C^B) that is adaptively updated, and a non-adaptive update filter (C^A) that is updated based on an update value of the adaptive update filter.

[0009] According to this aspect, both the reference signal and the error signal can be generated using multiple noise microphones. This eliminates the need to use expensive sensors such as acceleration sensors to generate reference signals, making it possible to provide an inexpensive active noise reduction device. Furthermore, by updating multiple acoustic transfer characteristic filters, the multiple acoustic transfer characteristic filters can track changes in the acoustic transfer characteristics. Therefore, even if the acoustic transfer characteristics change, the control performance of the active noise reduction device can be maintained and noise can be effectively reduced. Furthermore, by updating the non-adaptive update filter based on the update value of the adaptive update filter, the amount of calculation (computational load) of the control device can be reduced compared to when all acoustic transfer characteristic filters are adaptively updated. This allows the control device to be configured using a relatively inexpensive processor.

[0010] In the above aspect, a plurality of the adaptive update filters may be provided, and the control device may calculate estimated values ​​of the plurality of adaptive update filters by performing a predetermined averaging process on update values ​​of the plurality of adaptive update filters, and update the non-adaptive update filter based on the estimated values ​​of the plurality of adaptive update filters.

[0011] According to this aspect, even if there is variation in the update values ​​of the multiple adaptive update filters, it is possible to calculate the estimated values ​​of the multiple adaptive update filters with high accuracy, thereby improving the control performance of the active noise reduction device.

[0012] In the above aspect, a plurality of the adaptive update filters may be provided, and the control device may adaptively update the other adaptive update filters based on an update value of one of the adaptive update filters.

[0013] According to this aspect, it is possible to suppress variations in the update values ​​of the plurality of adaptive update filters, thereby improving the control performance of the active noise reduction device.

[0014] In the above aspect, the control device may classify the plurality of control filters and the adaptive update filter into a plurality of filter groups, and adaptively update the plurality of control filters and the adaptive update filter for each filter group in a predetermined update order.

[0015] According to this aspect, the update frequency of the control filters and adaptive update filters can be reduced compared to when all the control filters and adaptive update filters are adaptively updated each time, thereby further reducing the amount of calculation by the control device.

[0016] In the above aspect, when the amount of change in the adaptive update filter due to one adaptive update exceeds a predetermined threshold, the control device may adaptively update the adaptive update filter continuously for a predetermined time regardless of the update order.

[0017] According to this aspect, it is possible to suppress a decrease in the ability to follow changes in the acoustic transfer characteristics caused by reducing the update frequency of the adaptive update filter.

[0018] In the above aspect, the control device may adaptively update the plurality of control filters and the adaptive update filter for a predetermined period of time, and then suspend the adaptive updating of the plurality of control filters and the adaptive update filter, and while the adaptive updating of the plurality of control filters and the adaptive update filter is suspended, determine whether or not the noise control effect has decreased based on the error signal, and if it determines that the noise control effect has decreased, resume the adaptive updating of the plurality of control filters and the adaptive update filter.

[0019] According to this aspect, it is possible to further reduce the amount of calculations performed by the control device while suppressing a decrease in the ability to follow changes in the acoustic transfer characteristics.

[0020] In the above aspect, the control device may acquire buffer data in which the noise signals are accumulated in time series, and process the noise signals for each piece of buffer data.

[0021] According to this aspect, it is possible to further reduce the amount of calculation in a control device that processes buffered data (for example, a control device that is applied to a smart device such as a smartphone).

[0022] One aspect of the present invention is an active noise reduction program that causes a computer to execute the following steps: acquiring a plurality of noise signals output from a plurality of noise microphones (13A-13E); selecting, from the plurality of noise signals, a plurality of reference signals corresponding to the noise and an error signal corresponding to the error between the noise and a sound to be cancelled; generating a control signal for the sound to be cancelled from the plurality of reference signals using a plurality of control filters (W); and adaptively updating the plurality of control filters using a plurality of acoustic transfer characteristic filters (C^), wherein the plurality of acoustic transfer characteristic filters include an adaptive update filter (C^B) that is adaptively updated and a non-adaptive update filter (C^A) that is updated based on the update value of the adaptive update filter.

[0023] According to this aspect, both the reference signal and the error signal can be generated using multiple noise microphones. Therefore, there is no need to use an expensive sensor such as an acceleration sensor to generate the reference signal, and an inexpensive active noise reduction device can be provided. Furthermore, by updating multiple acoustic transfer characteristic filters, the multiple acoustic transfer characteristic filters can track changes in the acoustic transfer characteristics. Therefore, even if the acoustic transfer characteristics change, control performance can be maintained and noise can be effectively reduced. Furthermore, by updating the non-adaptive update filter based on the update value of the adaptive update filter, the amount of calculation (computational load) of a computer can be reduced compared to when all acoustic transfer characteristic filters are adaptively updated. Therefore, the control device can be configured using a relatively inexpensive processor. [Effects of the Invention]

[0024] According to the above aspect, it is possible to provide an inexpensive active noise reduction device that can maintain control performance even when the acoustic transmission characteristics change. [Brief explanation of the drawings]

[0025] [Figure 1] FIG. 1 is a schematic diagram showing a vehicle to which an active noise reduction device according to a first embodiment is applied; [Figure 2] FIG. 1 is a functional block diagram showing an active noise reduction device according to a first embodiment; [Figure 3] FIG. 10 is an explanatory diagram showing a second calculation method for a C^B estimated value according to the first embodiment; [Figure 4] FIG. 10 is an explanatory diagram showing a third calculation method for a C^B estimated value according to the first embodiment; [Figure 5] Graph showing noise reduction effect [Figure 6] FIG. 10 is a functional block diagram showing an active noise reduction device according to a second embodiment. [Figure 7] 10 is a table showing an update order table according to the second embodiment. [Figure 8] Flowchart showing change amount determination control according to the third embodiment [Figure 9] Graph showing the definition of delay time td according to the third embodiment [Figure 10] FIG. 10 is a schematic diagram showing a vehicle to which an active noise reduction device according to a fourth embodiment is applied. [Figure 11] Flowchart showing update interruption control according to the fourth embodiment [Figure 12] Graph showing the amount of calculation of the control device of the comparative example and the control devices of the first, second, and fourth embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0026] Hereinafter, embodiments of the present invention will be described with reference to the drawings. In this specification, the "^" (hat) next to various symbols indicates an identified value or an estimated value. While the "^" is placed above various symbols in figures and formulas, it is placed after the various symbols in the main text.

[0027] (First embodiment) First, a first embodiment of the present invention will be described with reference to FIGS.

[0028] <Active Noise Reduction Device 11> 1 is a schematic diagram showing a vehicle 1 to which an active noise reduction device 11 (hereinafter abbreviated as "noise reduction device 11") according to a first embodiment is applied. The noise reduction device 11 is an ANC (Active Noise Control Device) for reducing noise d generated in a passenger compartment 2 of the vehicle 1. More specifically, the noise reduction device 11 generates a canceling sound y that is in the opposite phase to the noise d, and reduces the noise d by causing the generated canceling sound y to interfere with the noise d.

[0029] For example, the noise d to be reduced by the noise reduction device 11 is road noise caused by wheel vibration due to force from the road surface. Note that the noise d to be reduced by the noise reduction device 11 may be noise other than the above-mentioned road noise (for example, drivetrain noise caused by vibration of the drive source 3 such as an internal combustion engine or an electric motor).

[0030] 1 and 2, noise reduction device 11 includes a plurality of speakers 12A-12D (an example of a noise cancellation output device) that output a cancellation sound y to cancel out noise d, a plurality of noise microphones 13A-13E that generate a plurality of noise signals x based on noise d, and a control device 16 that controls speakers 12A-12D based on the noise signals x.

[0031] <Speakers 12A-12D> Speakers 12A-12D of noise reduction device 11 are disposed at positions corresponding to a plurality of passenger seats 6A-6D provided in vehicle 1. For example, speakers 12A and 12B are disposed in the doors on the sides of front passenger seats 6A and 6B, and speakers 12C and 12D are disposed behind rear passenger seats 6C and 6D.

[0032] <Noise microphones 13A-13E> Noise microphones 13A-13E of noise reduction device 11 are installed at any location in vehicle 1. For example, noise microphones 13A-13D are arranged at positions corresponding to passenger seats 6A-6D. More specifically, noise microphones 13A-13D are arranged on headrests 7 of passenger seats 6A-6D. For example, noise microphone 13E is arranged near a noise source.

[0033] <Control device 16> The control device 16 of the noise reduction device 11 is configured by a computer having an arithmetic processing unit (a processor such as a CPU or an MPU) and a storage device (memory such as a ROM or a RAM). The control device 16 may be configured as a single piece of hardware, or may be configured as a unit consisting of multiple pieces of hardware.

[0034] The control device 16 has, as functional components, a signal selection unit 21, n (n≧2) feedback cancellation units 22, n control signal output units 23, an adder 25, and an estimated value calculation unit 26.

[0035] <Signal selection unit 21> Signal selection unit 21 of control device 16 is connected to noise microphones 13A-13E and acquires noise signals x output from noise microphones 13A-13E. Signal selection unit 21 selects, from noise signals x, n reference signals r' (r1', . . . , rn') corresponding to noise d itself and error signal e corresponding to the error between noise d and cancellation sound y. Signal selection unit 21 outputs the selected reference signals r' to feedback elimination unit 22 and also outputs the selected error signal e to control signal output unit 23.

[0036] The signal selection unit 21 may select an error signal e and a reference signal r' from the noise signal x based on the positions of the speakers 12A-12D to be controlled. For example, when controlling the speaker 12A corresponding to the passenger seat 6A, the signal selection unit 21 may select the noise signal x output from the noise microphone 13A corresponding to the passenger seat 6A as the error signal e, and may select the noise signals x output from the noise microphones 13B-13E other than the noise microphone 13A as the reference signal r'. On the other hand, when controlling the speaker 12B corresponding to the passenger seat 6B, the signal selection unit 21 may select the noise signal x output from the noise microphone 13B corresponding to the passenger seat 6B as the error signal e, and may select the noise signals x output from the noise microphones 13A, 13C-13E other than the noise microphone 13B as the reference signal r'.

[0037] As described above, noise signal x output from noise microphone 13A is selected as error signal e in the control of speaker 12A, and is selected as reference signal r' in the control of speaker 12B. Because the control of speakers 12A and 12B is executed simultaneously, noise signal x output from noise microphone 13A is used simultaneously as error signal e and reference signal r' (the same applies to noise signals x output from noise microphones 13B-13E).

[0038] Hereinafter, noise microphones 13A-13E that generate reference signal r' will be referred to as "reference microphones 13r," and noise microphones 13A-13E that generate error signal e will be referred to as "error microphones 13e." As is clear from the above description, noise microphones 13A-13E will be used simultaneously as reference microphones 13r and error microphones 13e. Symbol C in FIG. 2 indicates the transfer characteristic (transfer characteristic of the secondary path) of cancellation sound y from speakers 12A-12D to error microphone 13e, and symbol C in FIG. H indicates the transfer characteristics of the cancellation sound y from the speakers 12A-12D to the reference microphone 13r.

[0039] <Howling elimination section 22> The feedback canceller 22 of the control device 16 includes a feedback filter 31 , a polarity inverter 32 , and an adder 33 .

[0040] The howling filter unit 31 is a howling filter C^ H (C^1 H , ···, C^n H ) is composed of the feedback filter C^ H is the transfer characteristic C of the cancellation y from the loudspeakers 12A-12D to the reference microphone 13r. H This is a filter that corresponds to the estimated value of the feedback filter C^. H For the filter, an FIR filter (finite impulse response filter) or an SAN filter (adaptive notch filter) may be used.

[0041] The feedback filter unit 31 generates a feedback signal yh (y1h, ..., ynh) by filtering the control signal u (details of which will be described later) output from the adder 25. The feedback signal yh is a signal corresponding to a component of the cancellation sound y (more specifically, a component of the cancellation sound y that leaks from the speakers 12A-12D to the reference microphone 13r). The feedback filter unit 31 outputs the generated feedback signal yh to the polarity inversion unit 32.

[0042] The polarity inversion unit 32 inverts the polarity of the howling signal yh output from the howling filter unit 31. The polarity inversion unit 32 outputs the howling signal yh with the inverted polarity to the adder 33.

[0043] The adder 33 generates a corrected reference signal r (r1, ..., rn) by adding the reference signal r' output from the signal selection unit 21 and the howling signal yh output from the polarity inversion unit 32. The corrected reference signal r is expressed by the following equation (1). Note that * in the following equation (1) indicates a convolution operation.

number

[0044] <Control signal output unit 23> The control signal output units 23 of the control device 16 correspond to the corrected reference signals r (r1, ..., rn), respectively. The control signal output unit 23 includes one control signal output unit 23A (a C^ fixed-type control signal output unit) and K (K=n-1) control signal output units 23B (C^ learning-type control signal output units). In other embodiments, the control signal output unit 23 may include multiple control signal output units 23A.

[0045] <Control signal output unit 23A> The control signal output unit 23A includes a control filter unit 36, a secondary path filter unit 37, and a control update unit 38.

[0046] The control filter unit 36 ​​is configured with a control filter W1. The control filter W1 may be an FIR filter or a SAN filter. The control filter unit 36 ​​generates a control signal component u1' by performing filtering on the corrected reference signal r1. The control filter unit 36 ​​outputs the generated control signal component u1' to the adder 25.

[0047] The secondary path filter unit 37 is configured with a secondary path filter C^A (an example of an acoustic transfer characteristic filter and a non-adaptive update filter). The secondary path filter C^A is a filter corresponding to an estimated value of the transfer characteristic C of the cancellation sound y from the speakers 12A-12D to the error microphone 13e. An FIR filter is used for the secondary path filter C^A. However, in other embodiments, a SAN filter may be used for the secondary path filter C^A. The secondary path filter unit 37 performs filtering on the corrected reference signal r1 and outputs the filtered corrected reference signal r1 to the control update unit 38.

[0048] The control update unit 38 adaptively updates the control filter W1 using an adaptive algorithm such as an LMS (Least Mean Square) algorithm. More specifically, the control update unit 38 updates the control filter W1 so that the error signal e output from the signal selection unit 21 is minimized.

[0049] <Control signal output unit 23B> The control signal output unit 23B has a control signal generation unit 41, a first cancellation estimation signal generation unit 42, a noise estimation signal generation unit 43, a second cancellation estimation signal generation unit 44, a control filter update unit 45, and a virtual error signal generation unit 46.

[0050] The control signal generation unit 41 is configured with a control filter W (W2, ..., Wn). The control filter W may be an FIR filter or a SAN filter. The control signal generation unit 41 generates control signal components u' (u2', ..., un') by performing filter processing on the corrected reference signal r (r2, ..., rn). The control signal generation unit 41 outputs the generated control signal components u' to the adder 25.

[0051] The first cancellation estimation signal generating unit 42 includes a secondary path filter unit 51 and a secondary path updating unit 52.

[0052] The secondary path filter unit 51 includes a secondary path filter C^B (C^B1, ..., C^B K : an example of an acoustic transfer characteristic filter and an adaptive update filter). The secondary path filter C^B is a filter corresponding to an estimated value of the transfer characteristic C of the cancellation sound y from the speakers 12A-12D to the error microphone 13e. An FIR filter is used for the secondary path filter C^B. However, in other embodiments, a SAN filter may be used for the secondary path filter C^B.

[0053] The secondary path filter unit 51 generates a canceling noise estimation signal y^1 by filtering the control signal component u'. The secondary path filter unit 51 outputs the generated canceling noise estimation signal y^1 to the virtual error signal generation unit 46.

[0054] The secondary path update unit 52 adaptively updates the coefficients of the secondary path filter C^B using an adaptive algorithm such as an LMS algorithm. More specifically, the secondary path update unit 52 updates the coefficients of the secondary path filter C^B so that the virtual error signal e1 (described in detail later) output from the virtual error signal generation unit 46 is minimized.

[0055] The noise estimation signal generator 43 includes a primary path filter unit 54 and a primary path update unit 55 .

[0056] The primary path filter unit 54 is a primary path filter H^ (H^1, ..., H^ K ) The primary path filter H^ is a filter corresponding to an estimated value of the transfer characteristic of the noise d from the noise source to the error microphone 13e. An FIR filter is used as the primary path filter H^. However, in other embodiments, a SAN filter may be used as the primary path filter H^.

[0057] The primary path filter unit 54 generates a noise estimation signal d^ by filtering the corrected reference signal r. The primary path filter unit 54 outputs the generated noise estimation signal d^ to the virtual error signal generator 46.

[0058] The primary path update unit 55 adaptively updates the coefficients of the primary path filter H^ using an adaptive algorithm such as an LMS algorithm. More specifically, the primary path update unit 55 updates the coefficients of the primary path filter H^ so that the virtual error signal e1 output from the virtual error signal generation unit 46 is minimized.

[0059] The second cancellation estimation signal generation unit 44 is configured with the secondary path filter C^B, similar to the first cancellation estimation signal generation unit 42. When the coefficients of the secondary path filter C^B are updated in the first cancellation estimation signal generation unit 42, the updated coefficients of the secondary path filter C^B are output to the second cancellation estimation signal generation unit 44, and the coefficients of the secondary path filter C^B are updated in the second cancellation estimation signal generation unit 44. In other words, the coefficients of the secondary path filter C^B set in the second cancellation estimation signal generation unit 44 are not fixed values ​​but are values ​​that are successively updated based on the signal from the first cancellation estimation signal generation unit 42.

[0060] The second cancellation estimation signal generation unit 44 generates a cancellation estimation signal y^2 by performing a filter process on the corrected reference signal r. The second cancellation estimation signal generation unit 44 outputs the generated cancellation estimation signal y^2 to the control filter update unit 45.

[0061] The control filter update unit 45 includes a control filter unit 57 and a control update unit 58 .

[0062] The control filter unit 57 is configured by control filters W (W2, ..., Wn) similarly to the control signal generation unit 41. The control filter unit 57 generates a canceling noise estimation signal y^ by performing a filter process on the canceling noise estimation signal y^2 output from the second canceling noise estimation signal generation unit 44. The control filter unit 57 outputs the generated canceling noise estimation signal y^ to the virtual error signal generation unit 46.

[0063] The control update unit 58 uses an adaptive algorithm such as an LMS algorithm to update the coefficients of the control filter W. More specifically, the control update unit 58 updates the coefficients of the control filter W so that the virtual error signal e2 (details of which will be described later) output from the virtual error signal generation unit 46 is minimized.

[0064] When the coefficients of the control filter W are updated in this manner in the control filter update unit 45, the updated coefficients of the control filter W are output to the control signal generation unit 41, and the coefficients of the control filter W are updated in the control signal generation unit 41. In other words, the coefficients of the control filter W set in the control signal generation unit 41 are not fixed values, but are values ​​that are successively updated based on the signal from the control filter update unit 45.

[0065] The virtual error signal generator 46 includes a first polarity inverter 61 , a second polarity inverter 62 , a first adder 63 , and a second adder 64 .

[0066] The first polarity inversion unit 61 inverts the polarity of the cancellation estimation signal y^1 output from the first cancellation estimation signal generation unit 42. The second polarity inversion unit 62 inverts the polarity of the noise estimation signal d^ output from the noise estimation signal generation unit 43.

[0067] The first adder 63 generates a virtual error signal e1 by adding together the error signal e, the canceling estimation signal y^1 that has passed through the first polarity inversion unit 61, and the noise estimation signal d^ that has passed through the second polarity inversion unit 62. The first adder 63 outputs the generated virtual error signal e1 to the first canceling estimation signal generation unit 42 and the noise estimation signal generation unit 43.

[0068] The second adder 64 generates a virtual error signal e2 by adding the noise estimation signal d^ output from the noise estimation signal generation unit 43 and the cancellation estimation signal y^ output from the control filter update unit 45. The second adder 64 outputs the generated virtual error signal e2 to the control filter update unit 45.

[0069] <Adder 25> The adder 25 of the control device 16 generates a control signal u for the speakers 12A-12D by adding together the control signal components u' (u1', ..., un') output from the n control signal output units 23. The adder 25 outputs the generated control signal u to the speakers 12A-12D and the feedback elimination unit 22. As a result, the speakers 12A-12D output a cancellation sound y corresponding to the control signal u.

[0070] <Estimated value calculation unit 26> The estimated value calculation unit 26 of the control device 16 calculates an estimated value of the secondary path filter C^B (hereinafter referred to as the "C^B estimated value") based on the updated value of the secondary path filter C^B, and updates the secondary path filter C^A based on the calculated C^B estimated value. For example, the estimated value calculation unit 26 updates the secondary path filter C^A by copying the calculated C^B estimated value to the secondary path filter C^A. Hereinafter, the method for calculating the C^B estimated value by the estimated value calculation unit 26 will be described.

[0071] <Calculation method 1 of C^B estimated value> First, the estimated value calculation unit 26 acquires the updated values of K secondary path filters C^B (in this embodiment, FIR filters) from the secondary path filter unit 51 of the control signal output unit 23B. Here, the updated value of the k-th secondary path filter C^Bk is represented by the following equation (2). Note that M in the following equation (2) indicates the number of coefficients of the secondary path filter C^B.

Equation

[0072] The estimated value calculation unit 26 calculates the C^B estimated value by performing an averaging process on the updated values of the K secondary path filters C^B. For example, the estimated value calculation unit 26 calculates the m-th (m = 1,..., M) coefficient C^Bm of the C^B estimated value by the following equation (3). Note that a in the following equation (3) k indicates a weighted coefficient.

Equation

[0073] For example, the estimated value calculation unit 26 may fix the weighted coefficient a k to 1 / K. As a result, the m-th coefficient C^Bm of the C^B estimated value becomes a simple average value of the coefficients of the K secondary path filters C^B. Alternatively, the estimated value calculation unit 26 determines the weighted coefficient a according to the control channel (the error microphone 13e to be controlled)k It may be changed. As a result, the m-th coefficient C^Bm of the C^B estimated value becomes a weighted average value of the coefficients of K secondary path filters C^B.

[0074] <Calculation Method 2 of C^B Estimated Value> Referring to FIG. 3, first, the estimated value calculation unit 26 acquires update values of K secondary path filters C^B (in this embodiment, FIR filters) from the secondary path filter unit 51 of the control signal output unit 23B. In FIG. 3, for simplicity of explanation, the number of secondary path filters C^B is set to 2.

[0075] Next, the estimated value calculation unit 26 performs FFT (Fast Fourier Transform) on the K secondary path filters C^B. As a result, the estimated value calculation unit 26 calculates the frequency characteristics of the K secondary path filters C^B.

[0076] Next, the estimated value calculation unit 26 performs an averaging process on the frequency characteristics of the secondary path filter C^B to calculate the frequency characteristic C^B of the C^B estimated value. f For example, the estimated value calculation unit 26 calculates the frequency characteristic C^B of the C^B estimated value by the following formula (4). f In the following formula (4), a k、f represents a weighted coefficient.

Equation

[0077] Here, when the positions of the reference microphone 13r and the error microphone 13e are determined, the relationship between the reference microphone 13r and the error microphone 13e is also determined. Therefore, for each frequency band of the error microphone 13e, a reference microphone 13r with high calculation accuracy of the C^B estimated value can be determined. Therefore, for each frequency band of the error microphone 13e, the weighted coefficient a of the secondary path filter C^B corresponding to the reference microphone 13r with high calculation accuracy of the C^B estimated value k、fIf [[ID=]] is set large, the C^B estimated value can be accurately calculated in all frequency bands of the error microphone 13e. For example, in FIG. 3, in the frequency bands fa and fb, the weighted coefficient a corresponding to the secondary path filter C^B1 k、f is set large, and in the frequency band fc, the weighted coefficient a corresponding to the secondary path filter C^B2 k、f is set large.

[0078] Finally, the estimated value calculation unit 26 performs an IFFT (inverse fast Fourier transform) on the frequency characteristics C^B of the C^B estimated value f . As a result, the estimated value calculation unit 26 calculates a C^B estimated value corresponding to the FIR filter.

[0079] <Calculation method 3 of C^B estimated value> Referring to FIG. 4, the secondary path update unit 52 corresponding to the secondary path filter C^B1 adaptively updates the secondary path filter C^B1. Next, the secondary path update unit 52 corresponding to the secondary path filter C^B2 adaptively updates the secondary path filter C^B2 based on the updated value of the secondary path filter C^B1.

[0080] In this way, the secondary path update unit 52 repeatedly performs a process of adaptively updating the secondary path filter C^B k-1 based on the updated value (the updated value of the previous secondary path filter C^B). At that time, the secondary path update unit 52 calculates the updated value of the secondary path filter C^B k by the following equation (5). Note that t in the following equation (5) represents the discrete time, μ in the following equation (5) represents the step size parameter (a parameter for adjusting the update amount of the secondary path filter C^B k ), and * in the following equation (5) represents the convolution operation. k

Equation

[0081] In this way, when the secondary path filter C^B k is adaptively updated, the secondary path filter C^Bk The updated value of the secondary path filter C^Bk is output to the estimate calculation unit 26. The estimate calculation unit 26 calculates the C^B estimated value based on the updated value of the secondary path filter C^Bk. For example, the estimate calculation unit 26 may set the updated value of the secondary path filter C^Bk itself as the C^B estimated value. Alternatively, the estimate calculation unit 26 may calculate the C^B estimated value by performing an averaging process on the updated values ​​of the K secondary path filters C^B, including the updated value of the secondary path filter C^Bk, as in the first and second calculation methods of the C^B estimated value.

[0082] <Effects of the first embodiment> As described above, the control device 16 acquires multiple noise signals x output from the multiple noise microphones 13A-13E, selects from the multiple noise signals x multiple reference signals r' corresponding to the noise d and an error signal e corresponding to the error between the noise d and the cancellation sound y, uses multiple control filters W to generate control signals u for the speakers 12A-12D from the multiple reference signals r', and adaptively updates the multiple control filters W using multiple secondary path filters C^A and C^B. In other words, the active noise reduction program causes the control device 16 (computer) to execute the above-described processing. This allows both the reference signal r' and the error signal e to be generated by the multiple noise microphones 13A-13E. This eliminates the need to use an expensive sensor such as an acceleration sensor to generate the reference signal r', making it possible to provide an inexpensive noise reduction device 11.

[0083] In this embodiment, since the multiple noise microphones 13A-13E include only one error microphone 13e, the transfer characteristic C of the cancellation y from the speakers 12A-12D to the error microphone 13e is also determined to be one. Therefore, the values ​​of the secondary path filters CA and CA, which correspond to the estimated values ​​of the transfer characteristic C of the cancellation y from the speakers 12A-12D to the error microphone 13e, should theoretically be the same.

[0084] Therefore, the control device 16 adaptively updates the secondary path filter C^B and updates the secondary path filter C^A using this adaptively updated secondary path filter C^B. In other words, the control device 16 sets some control channels to the C^ learning type and uses the C^ learned in this C^ learning type control channel for other control channels. This makes it possible to reduce the amount of calculation (computational load) of the control device 16 while enabling the secondary path filters C^A and C^B to be updated.

[0085] On the other hand, in an actual space, the frequency characteristics of the reference signal r' generated by the reference microphone 13r have peaks (areas with high sound pressure levels) and valleys (areas with low sound pressure levels) depending on the position in the vehicle interior 2 where the reference microphone 13r is disposed. In particular, in the valleys of the frequency characteristics of the reference signal r', the update accuracy of the secondary path filter C^B is likely to decrease as the sound pressure level decreases. If the secondary path filter C^A is updated based only on the secondary path filter C^B with such reduced update accuracy, this may lead to a decrease in the control performance of the noise reduction device 11.

[0086] Therefore, the control device 16 updates the secondary path filter C^A based on multiple secondary path filters C^B that have been adaptively updated. In other words, the control device 16 updates the secondary path filter C^A using reference signals r' from multiple reference microphones 13r arranged at different positions. This makes it possible to compensate for the valleys in the frequency characteristics of a certain reference signal r' with the peaks in the frequency characteristics of other reference signals r'. This prevents the secondary path filter C^A from being updated based only on the secondary path filter C^B with reduced update accuracy, thereby improving the control performance of the noise reduction device 11.

[0087] Fig. 5 is a graph showing the effect of reducing noise d. As shown in Fig. 5, when noise reduction device 11 is turned on, noise d can be reduced more effectively than when noise reduction device 11 is turned off.

[0088] (Second embodiment) Next, a second embodiment of the present invention will be described with reference to Figures 6 and 7. Note that descriptions that overlap with the first embodiment of the present invention will be omitted as appropriate.

[0089] <Active Noise Reduction Device 71> 6 is a functional block diagram showing an active noise reduction device 71 (hereinafter abbreviated as "noise reduction device 71") according to the second embodiment. Note that, in the noise reduction device 71 according to the second embodiment, the components other than the update determination unit 75 of the control device 73 are the same as those of the noise reduction device 11 according to the first embodiment, and therefore description thereof will be omitted.

[0090] <Update Decision Unit 75> The update determination unit 75 of the control device 73 classifies adaptively updatable filters (in this embodiment, the control filter W, the primary path filter H^, and the secondary path filter CB: hereinafter collectively referred to as "updatable filters") into P (P≧2) filter groups and determines the update order of the filter groups. A method for determining the update order of the filter groups by the update determination unit 75 will be described below.

[0091] 7, the update determination unit 75 stores an update order table T1. The update order table T1 is a table that defines the relationship between the filter group numbers (1, 2, ..., P) and the updatable filters included in each filter group.

[0092] 6, the update determination unit 75 calculates a count value clk of the clock signal cs by counting the clock signal cs output from the clock 76 at predetermined time intervals. The update determination unit 75 determines a filter group (hereinafter referred to as an "update filter group") that adaptively updates an updatable filter based on the count value clk of the clock signal cs. For example, the update determination unit 75 determines the update filter group using the following equation (6). Note that p in the following equation (6) represents the number of the update filter group, and % in the following equation (6) represents a remainder operation.

number

[0093] For example, if the number P of filter groups is 20 and the count value clk of the clock 76 is 41, the number p of the update filter group is calculated by the above formula (6) as p=41%20+1=2. Therefore, the update determination unit 75 determines the filter group with number 2 as the update filter group.

[0094] The update determination unit 75 transmits a flag value fv (1 or 0) according to the update filter group to the update units of the updatable filters (in this embodiment, the control update unit 38, the secondary path update unit 52, the primary path update unit 55, and the control update unit 58: hereinafter collectively referred to as the "adaptive update units"). More specifically, the update determination unit 75 transmits a flag value fv of 1 to the adaptive update units of the updatable filters included in the update filter group. On the other hand, the update determination unit 75 transmits a flag value fv of 0 to the adaptive update units of the updatable filters not included in the update filter group.

[0095] When the adaptive update unit of the updatable filter receives a flag value fv of 1 from the update determination unit 75, it adaptively updates the corresponding updatable filter. On the other hand, when the adaptive update unit of the updatable filter receives a flag value fv of 0 from the update determination unit 75, it waits without performing adaptive updating of the corresponding updatable filter.

[0096] 7, for example, when the filter group numbered 2 is the update filter group, the update determination unit 75 transmits a flag value fv of 1 to the primary path update unit 55 of the primary path filter H^1 and the secondary path update unit 52 of the secondary path filter C^B1. Therefore, the primary path update unit 55 adaptively updates the primary path filter H^1, and the secondary path update unit 52 adaptively updates the secondary path filter C^B1. Meanwhile, the update determination unit 75 transmits a flag value fv of 0 to adaptive update units other than the above two adaptive update units. Therefore, the adaptive update units other than the above two adaptive update units wait without performing adaptive updates on the corresponding updatable filters.

[0097] <Effects of the second embodiment> The control device 73 classifies the updatable filters into a plurality of filter groups and adaptively updates the updatable filters for each filter group in a predetermined update order. This reduces the update frequency of the updatable filters compared to when all the updatable filters are adaptively updated every time. This further reduces the amount of calculation by the control device 73.

[0098] <Modification of the second embodiment> In the second embodiment, the control filter W, the primary path filter H^, and the secondary path filter C^B are set as updatable filters. H The howling filter C^ is adaptively updated. H In this case, the update determination unit 75 may include the howling filter C^ in addition to the control filter W, the primary path filter H^, and the secondary path filter C^B. H It is recommended to set the update order table T1 by including the filter group.

[0099] (Third embodiment) Next, a third embodiment of the present invention will be described with reference to Figures 8 and 9. Note that, since the third embodiment is the same as the second embodiment except for the change amount determination control executed by the update determination unit 75 of the control device 73, the description thereof will be omitted.

[0100] When adaptive updating of the updatable filters is performed for each filter group as in the second embodiment described above, the learning frequency of the acoustic transfer characteristics (for example, the transfer characteristics of the cancellation sound y and the noise d from the speakers 12A-12D to the error microphone 13e) decreases. However, even if the learning frequency of the acoustic transfer characteristics decreases in this way, the updatable filters are actually updated at intervals of about several ms. For example, when the sampling frequency is 5 kHz and the number of filter groups P is 20, the update interval of the updatable filters included in one filter group is 4 ms. Therefore, even if the learning frequency of the acoustic transfer characteristics decreases, it is considered that the impact on the control effect of the noise reduction device 71 (i.e., the effect of reducing the noise d) is small.

[0101] However, if the learning frequency of the acoustic transfer characteristics decreases, the ability to follow changes in the acoustic transfer characteristics will still decrease. Therefore, the update determination unit 75 suppresses the decrease in the ability to follow changes in the acoustic transfer characteristics by executing the following change amount determination control.

[0102] <Change amount determination control> 8, when the primary path filter H^ and the secondary path filter CB (hereinafter collectively referred to as "acoustic learning filters") are adaptively updated (step ST1), the update determination unit 75 acquires the phase change amount ΔP and the gain change amount ΔG of the acoustic learning filter due to one adaptive update (one sample) (step ST2). For example, the update determination unit 75 acquires the delay time td (see FIG. 9) of the impulse response of the acoustic learning filter (an FIR filter in this embodiment) as the phase change amount ΔP of the acoustic learning filter. Furthermore, the update determination unit 75 acquires the change amount in the sum of squares of the coefficients of the acoustic learning filter as the gain change amount ΔG of the acoustic learning filter.

[0103] Next, the update determination unit 75 determines whether or not at least one of the following conditions 1 and 2 is met (step ST3). <Condition 1> The phase change amount ΔP of the acoustic learning filter due to one adaptive update exceeds a predetermined phase threshold. <Condition 2> The gain change amount ΔG of the acoustic learning filter due to one adaptive update exceeds a predetermined gain threshold.

[0104] If at least one of the above conditions 1 and 2 is satisfied (step ST3: Yes), the update determination unit 75 estimates that the acoustic transfer characteristics have changed significantly. Therefore, the update determination unit 75 adaptively updates the acoustic learning filter continuously for a predetermined time (a predetermined number of samples) regardless of the update order of the filter groups (step ST4).

[0105] On the other hand, if neither of the above conditions 1 nor 2 is satisfied (step ST3: No), the update determination unit 75 estimates that the acoustic transfer characteristics have not changed significantly, and therefore adaptively updates the acoustic learning filter based on the update order of the filter groups (step ST5).

[0106] <Effects of the third embodiment> When the change amount of the acoustic learning filter (phase change amount ΔP or gain change amount ΔG) due to one adaptive update becomes equal to or greater than a predetermined threshold, the control device 73 adaptively updates the acoustic learning filter continuously for a predetermined time, regardless of the update order of the filter groups. This makes it possible to prevent a decrease in the ability to follow changes in the acoustic transfer characteristics due to a decrease in the update frequency of the acoustic learning filter.

[0107] (Fourth embodiment) Next, a fourth embodiment of the present invention will be described with reference to Figures 10 and 11. Note that descriptions that overlap with the second embodiment of the present invention will be omitted as appropriate.

[0108] <Active Noise Reduction Device 81> 10 is a schematic diagram showing a vehicle 1 to which an active noise reduction device 81 according to a fourth embodiment (hereinafter abbreviated as "noise reduction device 81") is applied. Note that, of the noise reduction device 81 according to the fourth embodiment, the components other than the control device 83 are the same as those of the noise reduction device 71 according to the second embodiment, and therefore description thereof will be omitted.

[0109] <Control device 83> 10, a control device 83 of a noise reduction device 81 is provided in a smart device 17 (an example of a mobile terminal) that can be carried outside the vehicle 1. More specifically, the control device 83 is realized by an active noise reduction program (active noise reduction application) that runs on an OS of the smart device 17. The smart device 17 is configured by, for example, a smartphone.

[0110] The control device 83 is configured to acquire buffer data in which the noise signal x is accumulated in time series, and process each piece of buffer data of the noise signal x. That is, the control device 83 employs a smart device type signal processing method.

[0111] The control device 83 is connected to an interface 18 provided in the vehicle 1, and is connected to the speakers 12A-12D and the noise microphones 13A-13E via the interface 18. The interface 18 may be a wired interface such as USB, or a wireless interface such as Bluetooth (registered trademark).

[0112] The components of the control device 83 are the same as those of the control device 73 according to the second embodiment, and therefore a description thereof will be omitted. The update interruption control executed by the update determination unit 75 of the control device 83 will be described below.

[0113] <Update interruption control> Referring to FIG. 11, when the updatable filter is adaptively updated (step ST11), the update determination unit 75 determines whether the elapsed time since the adaptive update of the updatable filter started (hereinafter referred to as the "update time of the updatable filter") has exceeded a predetermined reference time (step ST12).

[0114] If the update time is equal to or shorter than the reference time (step ST12: No), it is considered that the update time of the updatable filter is insufficient (it is considered that the learning of the acoustic transfer characteristics is insufficient). Therefore, the process returns to step ST11, and the updatable filter is adaptively updated again.

[0115] On the other hand, if the update time exceeds the reference time (step ST12: Yes), it is considered that the update time of the updatable filter is sufficient (the learning of the acoustic transfer characteristics is considered sufficient), and therefore the update determination unit 75 suspends the adaptive update of the updatable filter (step ST13).

[0116] Next, the update determination unit 75 calculates the current sound pressure evaluation amount J(n) based on the following equation (7) (step ST14): In the following equation (7), J(n-1) represents the previous sound pressure evaluation amount, β represents an average coefficient, e(n) represents the current error signal e, and L represents the number of error signals e.

number

[0117] Next, the update determination unit 75 determines whether the control effect of the noise reduction device 81 (i.e., the effect of reducing the noise d) has decreased based on the current sound pressure evaluation quantity J(n) (step ST15). For example, the update determination unit 75 determines that the control effect of the noise reduction device 81 has decreased if at least one of the following conditions A and B is met. On the other hand, the update determination unit 75 determines that the control effect of the noise reduction device 81 has not decreased if neither of the following conditions A nor B is met. <Condition A> The current sound pressure evaluation value J(n) is greater than a predetermined first threshold value. <Condition B> The difference between the current sound pressure evaluation amount J(n) and the sound pressure evaluation amount J(n-ΔN) from a certain time ago is greater than a predetermined second threshold value.

[0118] If it is determined that the control effect of the noise reduction device 81 has not decreased (step ST15: No), the update determination unit 75 suspends the adaptive update of the updatable filter, returns to step ST14, and re-calculates the current sound pressure evaluation quantity J(n).

[0119] On the other hand, if it is determined that the control effect of the noise reduction device 81 has decreased (step ST15: Yes), the update determination unit 75 resumes adaptive updating of the updatable filter (step ST16) and ends the update interruption control.

[0120] <Effects of the Fourth Embodiment> FIG. 12 is a graph showing the amount of calculation of a control device (not shown) of a comparative example and the control devices 16, 73, and 83 of the first, second, and fourth embodiments.

[0121] The control device of the comparative example adaptively updates all of the secondary path filters C^. In contrast, the control device 16 of the first embodiment adaptively updates only some of the secondary path filters C^ (i.e., the secondary path filters C^B). Therefore, the control device 16 of the first embodiment has a reduced amount of calculation compared to the control device of the comparative example.

[0122] Furthermore, the control device 16 of the first embodiment adaptively updates all updatable filters each time. In contrast, the control device 73 of the second embodiment adaptively updates only the updatable filters included in the update filter group. Therefore, the control device 73 of the second embodiment requires a smaller amount of calculation than the control device 16 of the first embodiment.

[0123] Furthermore, the control device 73 of the second embodiment constantly performs adaptive updating of the updatable filter. In contrast, the control device 83 of the fourth embodiment provides a period during which the adaptive updating of the updatable filter is temporarily suspended. Therefore, when compared in units of buffer data, the control device 83 of the fourth embodiment has a reduced amount of calculation compared to the control device 73 of the second embodiment. In other words, by adopting the configuration of the fourth embodiment, the amount of calculation of the control device 83 can be significantly reduced in a smart device-type signal processing system.

[0124] <Modification of the Fourth Embodiment> In the above-described fourth embodiment, the control device 83 is provided in a smart device 17 (an example of a mobile terminal) that can be carried outside the vehicle 1. On the other hand, in a modification of the fourth embodiment, the control device 83 may be provided in an in-vehicle system (not shown) installed in the vehicle 1. More specifically, the control device 83 may be realized by an active noise reduction program (active noise reduction application) executed on the OS of the in-vehicle system.

[0125] Although the description of the specific embodiment has been completed above, the present invention is not limited to the above embodiment and its modifications, and can be modified in a wide range of ways. [Explanation of symbols]

[0126] 11: Active noise reduction device 12A-12D: Speaker (an example of a noise canceling output device) 13A-13E: Noise microphone 16: Control device 71: Active noise reduction device 73: Control device 81: Active noise reduction device 83: Control device C^A: Secondary path filter (an example of an acoustic transfer characteristic filter and a non-adaptive update filter) C^B: Secondary path filter (an example of an acoustic transfer characteristic filter and adaptive update filter) W: Control filter d: Noise e: error signal r´ :Reference signal u: control signal x: noise signal y :Cancellation sound ΔG: Gain change amount ΔP: Phase change amount

Claims

1. a noise canceling output device that outputs a noise canceling sound to cancel out the noise; a plurality of noise microphones for generating a plurality of noise signals based on the noise; a control device that controls the noise canceling output device based on the plurality of noise signals, The control device acquiring the plurality of noise signals output from the plurality of noise microphones; selecting, from the plurality of noise signals, a plurality of reference signals corresponding to the noise and an error signal corresponding to an error between the noise and the cancellation sound; generating a control signal for the cancellation output device from the plurality of reference signals using a plurality of control filters; adaptively updating the plurality of control filters using a plurality of acoustic transfer characteristic filters; the plurality of acoustic transfer characteristic filters include adaptive update filters that are adaptively updated and non-adaptive update filters that are updated based on update values ​​of the adaptive update filters; The control device is an active noise reduction device that adaptively updates the adaptive update filter based on the error signal.

2. a plurality of the adaptive update filters are provided; 2. The active noise reduction device according to claim 1, wherein the control device calculates estimated values ​​of the plurality of adaptive update filters by performing a predetermined averaging process on update values ​​of the plurality of adaptive update filters, and updates the non-adaptive update filter based on the estimated values ​​of the plurality of adaptive update filters.

3. a plurality of the adaptive update filters are provided; 3. The active noise reduction device according to claim 1, wherein the control device adaptively updates one of the adaptive update filters based on an update value of the other adaptive update filter.

4. The active noise reduction device according to any one of claims 1 to 3, wherein the control device classifies the plurality of control filters and the adaptive update filter into a plurality of filter groups, and adaptively updates the plurality of control filters and the adaptive update filter for each filter group in a predetermined update order.

5. 5. The active noise reduction device according to claim 4, wherein when an amount of change in the adaptive update filter due to a single adaptive update exceeds a predetermined threshold, the control device adaptively updates the adaptive update filter continuously for a predetermined time regardless of the update order.

6. The control device After adaptively updating the plurality of control filters and the adaptive update filter for a predetermined period of time, the adaptive update of the plurality of control filters and the adaptive update filter is interrupted; determining whether or not the noise control effect is reduced based on the error signal while the adaptive updating of the plurality of control filters and the adaptive update filter is suspended; 6. An active noise reduction device according to claim 1, wherein adaptive updating of the plurality of control filters and the adaptive update filter is resumed when it is determined that the noise control effect has decreased.

7. 7. The active noise reduction device according to claim 6, wherein the control device acquires buffer data in which the noise signals are accumulated in time series, and processes the noise signals for each buffer data.

8. 1. An active noise reduction program comprising: acquiring a plurality of noise signals output from a plurality of noise microphones; selecting, from the plurality of noise signals, a plurality of reference signals corresponding to noise and an error signal corresponding to an error between the noise and a cancellation sound; generating a control signal for the cancellation from the plurality of reference signals using a plurality of control filters; and adaptively updating the plurality of control filters using a plurality of acoustic transfer characteristic filters; the plurality of acoustic transfer characteristic filters include adaptive update filters that are adaptively updated and non-adaptive update filters that are updated based on update values ​​of the adaptive update filters; The computer adaptively updates the adaptive update filter based on the error signal.

Images