Active noise reduction device

Abstract

To maintain maximum noise reduction performance while ensuring control stability. [Solution] The active noise reduction device 1 comprises a sound cancellation output device 12 that outputs a sound cancellation y to cancel out noise d, an error microphone 13 that generates an error signal e based on noise d and sound cancellation y, and a control device 15 that controls the sound cancellation output device 12 based on the error signal e. The control device 15 includes at least one adaptively updatable filter W, C^, H^, and changes the amount by which the values ​​of the filters W, C^, H^ before adaptive updating are reflected in the adaptively updated values ​​of the filters W, C^, H^ based on a predetermined signal.

Description

[Technical Field]

[0001] This invention relates to an active noise reduction device. [Background technology]

[0002] Conventionally, active noise reduction devices are known that reduce noise by interfering with the noise with a canceling sound that is in the opposite phase to the noise. An active noise reduction device comprises a canceling sound output device (e.g., a speaker) that outputs a canceling sound to cancel out the noise, an error microphone that generates an error signal based on the noise and the canceling sound, and a control device that controls the canceling sound output device based on the error signal (see Patent Documents 1 and 2). [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 7-28474 [Patent Document 2] Japanese Patent Publication No. 2010-70023 [Overview of the initiative] [Problems that the invention aims to solve]

[0004] In active noise reduction devices like those described above, there is a risk of control instability if disturbances are introduced into the noise being controlled (for example, if wind blows on the error microphone). Therefore, there is a need to develop an active noise reduction device that can maintain maximum noise reduction performance while ensuring control stability.

[0005] In view of the above background, one aspect of the present invention aims to provide an active noise reduction device that can maintain maximum noise reduction performance while ensuring control stability. [Means for solving the problem]

[0006] To solve the above problems, one aspect of the present invention provides an active noise reduction device comprising: a noise cancellation output device that outputs a cancellation sound to cancel out noise; an error microphone that generates an error signal based on the noise and the cancellation sound; and a control device that controls the noise cancellation output device based on the error signal, wherein the control device includes at least one adaptively updatable filter and changes the amount by which the value of the filter before adaptive updating is reflected in the value of the filter after adaptive updating based on a predetermined signal. [Effects of the Invention]

[0007] According to the above embodiments, it is possible to provide an active noise reduction device that can maintain maximum noise reduction performance while ensuring control stability. [Brief explanation of the drawing]

[0008] [Figure 1] A schematic diagram showing a vehicle to which an active noise reduction device according to the first embodiment is applied. [Figure 2] Block diagram showing an active noise reduction device according to the first embodiment. [Figure 3] Graph showing the relationship between the norm of the control signal and the forgetting coefficient in the first embodiment. [Figure 4] Block diagram showing an active noise reduction device according to the second embodiment. [Figure 5] Graph showing the relationship between the norm signal and the forgetting coefficient in the second embodiment. [Figure 6] Block diagram showing an active noise reduction device according to the third embodiment. [Figure 7] Graph showing the relationship between the norm signal and the forgetting coefficient in the third embodiment. [Figure 8] Block diagram showing an active noise reduction device according to the fourth embodiment. [Figure 9] Graph showing the relationship between the reference signal norm and the forgetting coefficient in the fourth embodiment. [Modes for carrying out the invention]

[0009] <<First Embodiment>> Hereinafter, an active noise reduction device 1 according to a first embodiment (hereinafter abbreviated as "noise reduction device 1") will be described with reference to FIGS. 1 to 3. In this specification, the "^" (hat) appended to various reference signs indicates an identified value or an estimated value. In drawings and mathematical formulas, "^" is placed above various reference signs, while in the main text, it is placed after various reference signs.

[0010] <Vehicle 3> FIG. 1 is a schematic diagram showing a vehicle 3 to which the noise reduction device 1 is applied. The vehicle 3 includes a vehicle body 5, a plurality of wheels (not shown) arranged below the vehicle body 5, and a plurality of suspensions 6 arranged between the vehicle body 5 and the plurality of wheels. A vehicle compartment 8 is formed inside the vehicle body 5. A plurality of occupant seats 9 are installed in the vehicle compartment 8. Each occupant seat 9 includes a seat cushion 9A, a seat back 9B arranged rearward and above the seat cushion 9A, and a headrest 9C fixed to an upper end of the seat back 9B.

[0011] <Noise reduction device 1> Referring to FIG. 1, the noise reduction device 1 is an ANC device (Active Noise Control Device) for reducing noise d generated in the vehicle compartment 8 of the vehicle 3. More specifically, the noise reduction device 1 generates a cancellation sound y having a phase opposite to that of the noise d, and reduces the noise d by causing the generated cancellation sound y to interfere with the noise d.

[0012] The noise reduction device 1 includes a plurality of acceleration sensors 11 (an example of a reference signal generation device) that generate a reference signal r corresponding to the noise d, a plurality of speakers 12 (an example of a cancellation sound output device) that output a cancellation sound y for canceling the noise d, a plurality of error microphones 13 that generate an error signal e based on the noise d and the cancellation sound y, and a control device 15 that controls the plurality of speakers 12 based on the reference signal r and the error signal e.

[0013] <Acceleration sensor 11> Referring to Figure 1, each acceleration sensor 11 is installed on the corresponding suspension 6. The acceleration sensor 11 detects the acceleration of the suspension 6 in response to the noise d and generates a reference signal r corresponding to the acceleration of the suspension 6. The acceleration sensor 11 has a normal operating range (a range in which the acceleration of the suspension 6 can be detected normally).

[0014] <Speaker 12> Referring to Figure 1, each speaker 12 is installed in a location in the vehicle 3 other than the passenger seat 9 (for example, the side door of the passenger seat 9 or the space behind the passenger seat 9). In other embodiments, the speaker 12 may be installed on the passenger seat 9 (for example, the headrest 9C of the passenger seat 9). The speaker 12 has a normal operating range (a range in which it can output the canceling sound y normally).

[0015] <Error Microphone 13> Referring to Figure 1, each error microphone 13 is installed on the headrest 9C of the passenger seat 9. In other embodiments, the error microphones 13 may be installed in locations other than the passenger seats 9 in the vehicle 3 (for example, on the ceiling above the passenger seats 9).

[0016] <Control device 15> Referring to Figure 2, the control device 15 is composed of a computer having an arithmetic processing unit (processor such as a CPU or MPU) and a storage device (memory such as ROM or RAM). The control device 15 may be configured as a single piece of hardware or as a unit consisting of multiple pieces of hardware. The control device 15 has as functional components a control signal generation unit 16, a sound field learning unit 17, and a forgetting coefficient setting unit 18.

[0017] <Control signal generation unit 16> The control signal generation unit 16 of the control device 15 includes a control filter unit 24, a reference signal correction unit 25, and a control update unit 26.

[0018] The control filter unit 24 is composed of a control filter W. The control filter W is composed of an FIR filter (finite impulse response filter). In other embodiments, the control filter W may be composed of a SAN filter (single frequency adaptive notch filter) or the like. The control filter unit 24 generates a control signal u for controlling the speaker 12 by filtering the reference signal r with the control filter W. The control filter unit 24 outputs the generated control signal u to the speaker 12, the sound field learning unit 17, and the forgetting coefficient setting unit 18. When the control signal u is input to the speaker 12, the speaker 12 generates a cancellation sound y corresponding to the magnitude of the control signal u.

[0019] The reference signal correction unit 25 is composed of a secondary path filter C^. The secondary path filter C^ is a filter that shows an estimated value of the transfer function C of the secondary path from the speaker 12 to the error microphone 13. The secondary path filter C^ is composed of an FIR filter. In other embodiments, the secondary path filter C^ may be composed of a SAN filter or the like. The reference signal correction unit 25 corrects the reference signal r by applying filtering to the reference signal r using the secondary path filter C^. The reference signal correction unit 25 outputs the corrected reference signal r to the control update unit 26.

[0020] The control update unit 26 adaptively updates the control filter W using an adaptive algorithm such as the LMS algorithm (Least Mean Square Algorithm). More specifically, the control update unit 26 adaptively updates the control filter W according to equation (1) below so that the error signal e output from the error microphone 13 is minimized.

number

[0021] <Sound Field Learning Section 17> The sound field learning unit 17 of the control device 15 includes a sound cancellation estimation signal generation unit 31, a secondary path update unit 32, a noise estimation signal generation unit 33, a primary path update unit 34, a sound cancellation estimation signal inversion unit 35, a noise estimation signal inversion unit 36, and a virtual error signal generation unit 37.

[0022] The negation sound estimation signal generation unit 31, like the reference signal correction unit 25, is configured with a second-order path filter C^. The negation sound estimation signal generation unit 31 applies filtering to the control signal u using the second-order path filter C^ to generate a negation sound estimation signal y^ that indicates the estimated value of the negation sound y. The negation sound estimation signal generation unit 31 outputs the generated negation sound estimation signal y^ to the negation sound estimation signal inversion unit 35.

[0023] The secondary path update unit 32 adaptively updates the secondary path filter C^ of the cancellation sound estimation signal generation unit 31 using an adaptive algorithm such as the LMS algorithm. More specifically, the secondary path update unit 32 updates the virtual error signal e output from the virtual error signal generation unit 37. vThe quadratic path filter C^ is adaptively updated by equation (2) below so that (details below) is minimized.

number

[0024] Furthermore, when the secondary path update unit 32 updates the secondary path filter C^ of the sound cancellation estimation signal generation unit 31 as described above, the updated secondary path filter C^ is output to the reference signal correction unit 25, and the secondary path filter C^ of the reference signal correction unit 25 is updated.

[0025] The noise estimation signal generation unit 33 is composed of a primary path filter H^. The primary path filter H^ is a filter that shows an estimated value of the transfer function H of the primary path from the noise source to the error microphone 13. The primary path filter H^ is composed of, for example, an FIR filter. In other embodiments, the primary path filter H^ may be composed of a SAN filter or the like. The noise estimation signal generation unit 33 generates a noise estimation signal d^ that shows an estimated value of noise d by applying filtering to the reference signal r using the primary path filter H^. The noise estimation signal generation unit 33 outputs the generated noise estimation signal d^ to the noise estimation signal inversion unit 36.

[0026] The primary path update unit 34 adaptively updates the primary path filter H^ using an adaptive algorithm such as the LMS algorithm. More specifically, the primary path update unit 34 updates the virtual error signal e output from the virtual error signal generation unit 37. vThe first-order path filter H^ is adaptively updated by equation (3) below so that (details below) is minimized.

number

[0027] The sound cancellation estimation signal inversion unit 35 inverts the polarity of the sound cancellation estimation signal y^ output from the sound cancellation estimation signal generation unit 31. The sound cancellation estimation signal inversion unit 35 outputs the sound cancellation estimation signal y^ with inverted polarity to the virtual error signal generation unit 37.

[0028] The noise estimation signal inversion unit 36 ​​inverts the polarity of the noise estimation signal d^ output from the noise estimation signal generation unit 33. The noise estimation signal inversion unit 36 ​​outputs the noise estimation signal d^ with inverted polarity to the virtual error signal generation unit 37.

[0029] The virtual error signal generation unit 37 adds the error signal e output from the error microphone 13, the cancellation estimation signal y^ that has passed through the cancellation estimation signal inversion unit 35, and the noise estimation signal d^ that has passed through the noise estimation signal inversion unit 36 ​​to generate the virtual error signal e v The virtual error signal generation unit 37 generates the virtual error signal e. v This is output to the secondary route update unit 32 and the primary route update unit 34.

[0030] <Forgetting Factor Setting Unit 18> The forgetting coefficient setting unit 18 of the control device 15 receives a control signal u from the control filter unit 24. The forgetting coefficient setting unit 18 sets the forgetting coefficient η based on the control signal u. The method of setting the forgetting coefficient η by the forgetting coefficient setting unit 18 will be described later.

[0031] The forgetting coefficient setting unit 18 outputs a forgetting coefficient η to the control update unit 26, the secondary path update unit 32, and the primary path update unit 34. The control update unit 26, the secondary path update unit 32, and the primary path update unit 34 respectively adaptively update the control filter W, the secondary path filter C^, and the primary path filter H^ using the forgetting coefficient η output from the forgetting coefficient setting unit 18 (see equations (1) to (3) above). Hereinafter, the control filter W, the secondary path filter C^, and the primary path filter H^ will be collectively referred to as "adaptive update filters W, C^, and H^".

[0032] <Method for setting the forgetting coefficient η using the forgetting coefficient setting unit 18> The forgetting coefficient setting unit 18 calculates the norm ||u|| of the control signal u (an example of a predetermined signal) based on the control signal u. More specifically, the forgetting coefficient setting unit 18 calculates the norm ||u|| of the control signal u by the following equation (4).

number

[0033] Referring to Figure 3, the forgetting coefficient setting unit 18 sets the forgetting coefficient η based on the norm ||u|| of the control signal u. More specifically, when the norm ||u|| of the control signal u is in a first range R1 (the range in which the control signal u is within the normal operating range of the speaker 12), the forgetting coefficient setting unit 18 keeps the forgetting coefficient η at its maximum value (e.g., 1). When the norm ||u|| of the control signal u is in a second range R2 that exceeds the first range R1 (the range in which the control signal u is beyond the normal operating range of the speaker 12), the forgetting coefficient setting unit 18 continuously decreases the forgetting coefficient η as the norm ||u|| of the control signal u increases. As a result, when the norm ||u|| of the control signal u is in the second range R2, the larger the norm ||u|| of the control signal u becomes, the less the current values ​​of the adaptive update filters W, C^, and H^ are reflected in the updated values ​​of the adaptive update filters W, C^, and H^ (ηW(t), ηC^(t), ηH^(t)) (see equations (1) to (3) above). When the norm ||u|| of the control signal u is in the third range R3, which is greater than the second range R2, the forgetting coefficient setting unit 18 keeps the forgetting coefficient η at its minimum value (for example, 0).

[0034] The norm ||u|| of the control signal u is an example of an indicator of control stability. The relationship between such indicators of control stability and the forgetting coefficient η is stored internally in the control device 15 in table format.

[0035] <Effects> In the noise reduction device 1 described above, it is also possible to set the secondary path filter C^ to a fixed value. However, if the secondary path filter C^ is set to a fixed value, the difference between the transfer function C of the secondary path and the secondary path filter C^ will increase when the transfer function C of the secondary path changes. As a result, there is a risk that the noise reduction device 1 may not be able to sufficiently reduce noise d, or that noise d may be amplified. Therefore, the control device 15 adaptively updates the secondary path filter C^. This allows the secondary path filter C^ to change in response to changes in the transfer function C of the secondary path. As a result, the noise reduction device 1 can sufficiently reduce noise d and suppress the amplification of noise d.

[0036] On the other hand, even if the control device 15 adaptively updates the secondary path filter C^, the control may become unstable if disturbances are introduced into the noise d being controlled (for example, if wind blows on the error microphone 13). Therefore, the control device 15 changes the amount by which the current values ​​of the adaptive update filters W, C^, and H^ are reflected in the update values ​​of the adaptive update filters W, C^, and H^ based on an index indicating control stability (norm of the control signal u||u||). This makes it possible to maintain control stability even if disturbances are introduced into the noise d being controlled.

[0037] Furthermore, the control device 15 can determine that the control is unstable when an indicator of control stability falls below a pre-set threshold, and then attenuate the canceling sound y based on the forgetting coefficient η. However, if such control is adopted, the control device 15 will set the threshold to a small value in order to reliably suppress the amplification of noise d. As a result, there will be more situations in which the control is determined to be unstable (situations in which the canceling sound y is attenuated), and there is a risk that the noise reduction performance will not be able to be maximized. Also, if the above-mentioned control is adopted, the control device 15 will set the forgetting coefficient η to a small value in order to reliably suppress the amplification of noise d. As a result, if the control is determined to be unstable, there is a risk that the noise reduction performance will rapidly decrease, and it will take time to recover the noise reduction performance. Therefore, the control device 15 continuously changes the amount by which the current value of the control filter W is reflected in the update value of the control filter W by continuously changing the forgetting coefficient η. This suppresses a rapid decrease in noise reduction performance and allows the noise reduction performance to be maintained at its maximum.

[0038] Furthermore, the control device 15 continuously decreases the forgetting coefficient η as the norm ||u|| of the control signal u increases, provided that the norm ||u|| of the control signal u is within a certain range. This allows the control filter W to be gradually attenuated as the control signal u increases, provided that the norm ||u|| of the control signal u is within a certain range. Therefore, it is possible to ensure control stability by suppressing excessive growth of the control signal u, while maintaining noise reduction performance by suppressing abrupt decreases in the control signal u. In other words, it is possible to balance the requirement of improving control stability early on and the requirement of maintaining maximum noise reduction performance (two requirements that are in a trade-off relationship).

[0039] Incidentally, conventionally, compressors have been used to prevent the output of audio equipment from exceeding the maximum volume of the speaker. It is conceivable that this technology could be applied to the noise reduction device 1, and by reducing the output gain with a compressor, the generation of distorted sound caused by the control signal u exceeding the normal operating range of the speaker 12 could be suppressed. However, since the control device 15 performs adaptive control, even if the output gain is reduced with a compressor, the control device 15 will increase the coefficient of the control filter W in order to adapt to the change in gain. As a result, there is a risk that the over-enhancement of the control signal u cannot be suppressed, and the generation of distorted sound cannot be suppressed. In contrast, when the control signal u is in a range that exceeds the normal operating range of the speaker 12, the control device 15 continuously decreases the forgetting coefficient η as the norm ||u|| of the control signal u increases. This makes it possible to effectively suppress the generation of distorted sound caused by the control signal u exceeding the normal operating range of the speaker 12.

[0040] <<Second Embodiment>> Next, with reference to Figures 4 and 5, the active noise reduction device 41 according to the second embodiment (hereinafter abbreviated as "noise reduction device 41") will be described. Note that, except for the method of setting the forgetting coefficient η by the forgetting coefficient setting unit 18, the explanation is the same as that of the noise reduction device 1 according to the first embodiment, and therefore will be omitted.

[0041] <Method for setting the forgetting coefficient η using the forgetting coefficient setting unit 18> Referring to Figure 4, the forgetting coefficient setting unit 18 receives a control signal u from the control filter unit 24 and a reference signal r from the acceleration sensor 11. Based on the control signal u and the reference signal r, the forgetting coefficient setting unit 18 calculates a norm signal N1 (an example of a predetermined signal). More specifically, the forgetting coefficient setting unit 18 calculates the norm signal N1 by the following equation (5).

number

[0042] Referring to Figure 5, the forgetting coefficient setting unit 18 sets the forgetting coefficient η based on the norm signal N1. More specifically, when the norm signal N1 is in the first range R1, the forgetting coefficient setting unit 18 keeps the forgetting coefficient η at its maximum value (e.g., 1). When the norm signal N1 is in the second range R2, which is beyond the first range R1, the forgetting coefficient setting unit 18 continuously decreases the forgetting coefficient η as the norm signal N1 increases. As a result, when the norm signal N1 is in the second range R2, the amount by which the current values ​​of the adaptive update filters W, C^, and H^ are reflected in the update values ​​of the adaptive update filters W, C^, and H^ (ηW(t), ηC^(t), ηH^(t)) continuously decreases as the norm signal N1 increases (see equations (1) to (3) above). When the norm signal N1 is in a third range R3 that exceeds the second range R2, the forgetting coefficient setting unit 18 keeps the forgetting coefficient η at its minimum value (for example, 0).

[0043] <Effects> When vehicle 3 is traveling on a rough road surface, the vibration input from the road surface increases, and the required control output (sound cancellation y) also increases. In the first embodiment, the control device 15 sets the forgetting coefficient η based only on the control signal u, so even in situations where a large control output is required, the control signal u is limited, and there is a risk that the noise reduction performance cannot be fully demonstrated. In contrast, the control device 15 of the second embodiment sets the forgetting coefficient η based not only on the control signal u but also on the reference signal r. More specifically, the control device 15 of the second embodiment calculates the norm signal N1 by dividing the norm ||u|| of the control signal u by the norm ||r|| of the reference signal r, and sets the forgetting coefficient η based on the calculated norm signal N1. Therefore, when the reference signal r increases due to an increase in vibration input from the road surface, the norm signal N1 decreases, and the forgetting coefficient η increases. This prevents the control signal u from becoming unnecessarily small, and the noise reduction performance can be maintained to the maximum extent.

[0044] <<Third Embodiment>> Next, with reference to Figures 6 and 7, the active noise reduction device 51 according to the third embodiment (hereinafter abbreviated as "noise reduction device 51") will be described. Note that, except for the method of setting the forgetting coefficient η by the forgetting coefficient setting unit 18, the explanation is the same as that of the noise reduction device 1 according to the first embodiment, and therefore will be omitted.

[0045] <Method for setting the forgetting coefficient η using the forgetting coefficient setting unit 18> Referring to Figure 6, the forgetting coefficient setting unit 18 receives the coefficient of the control filter W from the control filter unit 24, the coefficient of the secondary path filter C^ from the sound cancellation estimation signal generation unit 31, and the coefficient of the primary path filter H^ from the noise estimation signal generation unit 33. Based on the coefficient of the control filter W, the coefficient of the secondary path filter C^, and the coefficient of the primary path filter H^, the forgetting coefficient setting unit 18 calculates a norm signal N2 (an example of a predetermined signal). More specifically, the forgetting coefficient setting unit 18 calculates the norm signal N2 by the following equation (6).

number

[0046] Referring to FIG. 7, the forgetting coefficient setting unit 18 sets the forgetting coefficient η based on the norm signal N2. More specifically, when the norm signal N2 is in the first range R1, the forgetting coefficient setting unit 18 maintains the forgetting coefficient η at a maximum value (e.g., 1). When the norm signal N2 is in the second range R2 that exceeds the first range R1, the forgetting coefficient setting unit 18 continuously decreases the forgetting coefficient η as the norm signal N2 increases. Accordingly, when the norm signal N2 is in the second range R2, as the norm signal N2 increases, the reflection amounts (ηW(t), ηC^(t), ηH^(t)) of the current values of the adaptive update filters W, C^, and H^ with respect to the update values of the adaptive update filters W, C^, and H^ continuously decrease (see the above formulas (1) to (3)). When the norm signal N2 is in the third range R3 that exceeds the second range R2, the forgetting coefficient setting unit 18 maintains the forgetting coefficient η at a minimum value (e.g., 0).

[0047] <Effect> When the canceling sound y and the noise d completely cancel each other out, ||C^×W|| (characteristics of the canceling sound y at the control point) and ||H^|| (characteristics of the noise d at the control point) become equal. On the other hand, when the canceling sound y is unnecessarily large relative to the noise d, ||C^×W|| becomes larger than ||H^||. Therefore, the control device 15 calculates the norm signal N2 by dividing the product of the norm ||W|| of the control filter W and the norm ||C^|| of the second-order path filter C^ by the norm ||H^|| of the first-order path filter H^, and sets the forgetting coefficient η based on the calculated norm signal N2. As a result, when the canceling sound y is unnecessarily large relative to the noise d, the norm signal N2 becomes large and the forgetting coefficient η becomes small. This makes it possible to suppress the canceling sound y from becoming unnecessarily large relative to the noise d, and improves the stability of the control.

[0048] <<Fourth Embodiment>> Next, with reference to Figures 8 and 9, the active noise reduction device 61 according to the fourth embodiment (hereinafter abbreviated as "noise reduction device 61") will be described. Note that, except for the method of setting the forgetting coefficient η by the forgetting coefficient setting unit 18, the explanation is the same as that of the noise reduction device 1 according to the first embodiment, and therefore will be omitted.

[0049] <Method for setting the forgetting coefficient η using the forgetting coefficient setting unit 18> Referring to Figure 8, the forgetting coefficient setting unit 18 receives a reference signal r from the acceleration sensor 11. Based on the reference signal r, the forgetting coefficient setting unit 18 calculates the norm ||r|| of the reference signal r (an example of a predetermined signal). More specifically, the forgetting coefficient setting unit 18 calculates the norm ||r|| of the reference signal r by the following equation (7).

number

[0050] Referring to Figure 9, the forgetting coefficient setting unit 18 sets the forgetting coefficient η based on the norm ||r|| of the reference signal r. More specifically, when the norm ||r|| of the reference signal r is in a first range R1 (the range in which the reference signal r falls within the normal operating range of the acceleration sensor 11), the forgetting coefficient setting unit 18 keeps the forgetting coefficient η at its maximum value (e.g., 1). When the norm ||r|| of the reference signal r is in a second range R2 that exceeds the first range R1 (the range in which the reference signal r exceeds the normal operating range of the acceleration sensor 11), the forgetting coefficient setting unit 18 continuously decreases the forgetting coefficient η as the norm ||r|| of the reference signal r increases. As a result, when the norm ||r|| of the reference signal r is in the second range R2, the larger the norm ||r|| of the reference signal r becomes, the less the current values ​​of the adaptive update filters W, C^, and H^ are reflected in the updated values ​​of the adaptive update filters W, C^, and H^ (ηW(t), ηC^(t), ηH^(t)) (see equations (1) to (3) above). When the norm ||r|| of the reference signal r is in the third range R3, which is beyond the second range R2, the forgetting coefficient setting unit 18 keeps the forgetting coefficient η at its minimum value (for example, 0).

[0051] <Effects> When the reference signal r is in a range that exceeds the normal operating range of the acceleration sensor 11, the control device 15 continuously decreases the forgetting coefficient η as the norm ||r|| of the reference signal r increases. This prevents the generation of abnormal noise when the vibration input from the road surface increases and the reference signal r exceeds the normal operating range of the acceleration sensor 11.

[0052] <Variation> In the first to fourth embodiments described above, the control device 15 adaptively updates the adaptive update filters W, C^, and H^ using equations (1) to (3) above. However, equations (1) to (3) above are merely examples of equations used for the adaptive update of the adaptive update filters W, C^, and H^. Therefore, in other embodiments, the control device 15 may adaptively update the adaptive update filters W, C^, and H^ using equations other than those (1) to (3) above. For example, the control device 15 may adaptively update the adaptive update filters W, C^, and H^ using equations (8) to (10) below.

number

number

number

number

number

number

number

[0053] In the first to fourth embodiments described above, the control device 15 modifies the amount of current value reflection in the updated value for all adaptive update filters W, C^, and H^ based on the forgetting coefficient η (see equations (1) to (3) above). In other embodiments, the control device 15 may modify the amount of current value reflection in the updated value for only some of the adaptive update filters W, C^, and H^ based on the forgetting coefficient η. For example, the control device 15 may modify the amount of current value reflection in the updated value for only the control filter W based on the forgetting coefficient η. Alternatively, the control device 15 may modify the amount of current value reflection in the updated value for only the control filter W and the secondary path filter C^ based on the forgetting coefficient η.

[0054] In the first to fourth embodiments described above, the control device 15 sets the forgetting coefficient η for all adaptive update filters W, C^, and H^ using the same table. In other embodiments, the control device 15 may set the forgetting coefficient η for each adaptive update filter W, C^, and H^ using different tables.

[0055] In the first to fourth embodiments described above, the control device 15 uses the current values ​​(W(t), C^(t), H^(t)) of the adaptive update filters W, C^, and H^ as the values ​​before adaptive update of the adaptive update filters W, C^, and H^. In other embodiments, the control device 15 may use the past values ​​(for example, W(t-1), C^(t-1), H^(t-1)) of the adaptive update filters W, C^, and H^ as the values ​​before adaptive update of the adaptive update filters W, C^, and H^, or it may use both the current and past values ​​of the adaptive update filters W, C^, and H^.

[0056] In the first to fourth embodiments described above, the control device 15 includes multiple adaptive update filters W, C^, and H^. In other embodiments, the control device 15 may include only one adaptive update filter.

[0057] In the first to fourth embodiments described above, the noise reduction devices 1, 41, 51, and 61 are applied to the passenger compartment 8 of the vehicle 3. In other embodiments, the noise reduction devices 1, 41, 51, and 61 may be applied to the interior space of a moving object other than the vehicle 3 (e.g., a ship or an aircraft), or to the interior space of a fixed object (e.g., a house).

[0058] This concludes the description of specific embodiments, but the present invention is not limited to the above embodiments or modifications and can be broadly modified and implemented.

[0059] <Summary of Embodiments> The active noise reduction devices 1, 41, 51, and 61 each include a noise cancellation output device 12 that outputs a noise cancellation sound y to cancel out noise d, an error microphone 13 that generates an error signal e based on the noise d and the noise cancellation sound y, and a control device 15 that controls the noise cancellation output device 12 based on the error signal e. The control device 15 includes at least one adaptively updatable filter W, C^, H^, and modifies the amount by which the adaptively updated values ​​of the filter W, C^, H^ are reflected in the adaptively updated values ​​of the filter W, C^, H^ based on predetermined signals ||u||, N1, N2, and ||r||.

[0060] According to this embodiment, by changing the reflection amount based on predetermined signals ||u||, N1, N2, and ||r||, it is possible to provide active noise reduction devices 1, 41, 51, and 61 that can maintain maximum noise reduction performance while ensuring control stability.

[0061] The at least one filter W, C^, H^ includes a control filter W that generates a control signal u input to the sound cancellation output device 12, the predetermined signal ||u|| is generated based on the control signal u, and the control device 15 reduces the amount ηW(t) that reflects the value of the control filter W before adaptive update to the value of the control filter W after adaptive update, as the predetermined signal ||u|| becomes larger.

[0062] According to this embodiment, the control filter W can be attenuated as the control signal u increases. Therefore, while ensuring control stability by suppressing excessive increases in the control signal u, noise reduction performance can be maintained by suppressing abrupt decreases in the control signal u. In other words, it is possible to satisfy in a good balance the requirement to improve control stability early on and the requirement to maintain maximum noise reduction performance (two requirements that are in a trade-off relationship).

[0063] The active noise reduction device 61 further comprises a reference signal generator 11 that generates a reference signal r corresponding to the noise d, the predetermined signal ||r|| is generated based on the reference signal r, and the control device 15 reduces the reflection amount as the predetermined signal ||r|| increases in the range R2 where the reference signal r exceeds the normal operating range of the reference signal generator 11.

[0064] According to this embodiment, when the vibration input from the road surface increases and the reference signal r exceeds the normal operating range of the reference signal generator 11, the generation of abnormal noise can be avoided.

[0065] The predetermined signal N1 is generated based on the control signal u input to the sound cancellation output device 12 and the reference signal r corresponding to the noise d.

[0066] According to this embodiment, by generating a predetermined signal N1 based not only on the control signal u but also on the reference signal r, the reflection amount can be set with high precision. This makes it possible to maintain maximum noise reduction performance while increasing the stability of the control.

[0067] The predetermined signal N1 is generated by dividing a first calculated value ||u|| calculated based on the control signal u by a second calculated value ||r|| calculated based on the reference signal r, and the control device 15 reduces the reflection amount as the predetermined signal N1 becomes larger.

[0068] According to this embodiment, when the reference signal r increases due to an increase in vibration input from the road surface, a predetermined signal N1 can be reduced. This prevents the reflection amount from becoming unnecessarily small, and allows the noise reduction performance to be maintained to the maximum extent.

[0069] The at least one filter W, C^, H^ includes a control filter W that generates a control signal u input to the sound cancellation output device 12, a second-order path filter C^ that shows an estimated value of the transfer function C from the sound cancellation output device 12 to the error microphone 13, and a first-order path filter H^ that shows an estimated value of the transfer function H from the noise source to the error microphone 13, and the predetermined signal N2 is generated based on the control filter W, the first-order path filter H^, and the second-order path filter C^.

[0070] According to this embodiment, the reflection amount can be set with high precision by generating a predetermined signal N2 based on the control filter W, the secondary path filter C^, and the primary path filter H^. This makes it possible to maintain maximum noise reduction performance while increasing the stability of the control.

[0071] The predetermined signal N2 is generated by dividing the product of a first intermediate value ||W|| calculated based on the control filter W and a second intermediate value ||C^|| calculated based on the secondary path filter C^ by a third intermediate value ||H^|| calculated based on the primary path filter H^, and the control device 15 reduces the reflection amount as the predetermined signal N2 becomes larger.

[0072] According to this embodiment, when the canceling sound y becomes unnecessarily large compared to the noise d, a predetermined signal N2 becomes larger and the amount of reflection decreases. This makes it possible to suppress the canceling sound y from becoming unnecessarily large and improve the stability of the control.

[0073] The at least one filter W, C^, H^ includes a control filter W that generates a control signal u input to the sound cancellation output device 12, and the control device 15 changes the amount ηW(t) that reflects the value of the control filter W before adaptive update to the value of the control filter W after adaptive update, based on the predetermined signals ||u||, N1, N2, ||r||.

[0074] According to this embodiment, by changing the reflection amount ηW(t) based on predetermined signals ||u||, N1, N2, and ||r||, it is possible to provide active noise reduction devices 1, 41, 51, and 61 that can maintain maximum noise reduction performance while ensuring control stability.

[0075] The at least one filter W, C^, H^ further includes a second-order path filter C^ that shows an estimated value of the transfer function C from the sound cancellation output device 12 to the error microphone 13, and the control device 15 changes the amount of reflection ηC^(t) of the value of the second-order path filter C^ before adaptive update to the value of the second-order path filter C^ after adaptive update, based on the predetermined signals ||u||, N1, N2, ||r||.

[0076] According to this embodiment, by changing the reflection amount ηC^(t) based on predetermined signals ||u||, N1, N2, and ||r||, it is possible to provide active noise reduction devices 1, 41, 51, and 61 that can maintain maximum noise reduction performance while ensuring control stability.

[0077] The at least one filter W, C^, H^ further includes a first-order path filter H^ that shows an estimated value of the transfer function H from the noise source to the error microphone 13, and the control device 15 changes the amount of reflection ηH^(t) of the value of the first-order path filter H^ before adaptive update to the value of the first-order path filter H^ after adaptive update, based on the predetermined signals ||u||, N1, N2, ||r||.

[0078] According to this embodiment, by changing the reflection amount ηH^(t) based on predetermined signals ||u||, N1, N2, and ||r||, it is possible to provide active noise reduction devices 1, 41, 51, and 61 that can maintain maximum noise reduction performance while ensuring control stability.

[0079] The control device 15 changes a predetermined coefficient η based on the predetermined signals ||u||, N1, N2, and ||r||, and calculates the reflection amount by multiplying the predetermined coefficient η by the value of the filter before the adaptive update.

[0080] According to this embodiment, the amount reflected can be calculated by a simple method. [Explanation of symbols]

[0081] <<First Embodiment>> 1: Active noise reduction device 11: Accelerometer (an example of a reference signal generator) 12: Speaker (an example of a sound-canceling output device) 13: Error Microphone 15: Control device W: Control filter (an example of a filter) C^: Quadratic path filter (an example of a filter) H^: Primary path filter (an example of a filter) <<Second Embodiment>> 41: Active noise reduction device <<Third Embodiment>> 51: Active noise reduction device <<Fourth Embodiment>> 61: Active noise reduction device

Claims

1. A noise cancellation output device that outputs noise cancellation sound to cancel out noise, An error microphone that generates an error signal based on the aforementioned noise and the aforementioned cancellation sound, An active noise reduction device comprising a control device that controls the noise cancellation output device based on the error signal, The control device is Includes at least one adaptively updatable filter, An active noise reduction device that changes the amount by which the value of the filter before adaptation is reflected in the value of the filter after adaptation and update, based on a predetermined signal.

2. The at least one filter includes a control filter that generates a control signal input to the sound cancellation output device, The predetermined signal is generated based on the control signal, The control device, according to claim 1, reduces the amount by which the value of the control filter before adaptive update is reflected in the value of the control filter after adaptive update as the predetermined signal becomes larger.

3. The device further comprises a reference signal generating device that generates a reference signal corresponding to the noise, The predetermined signal is generated based on the reference signal, The control device, in a range where the reference signal exceeds the normal operating range of the reference signal generator, reduces the amount of reflection as the predetermined signal becomes larger, according to claim 1.

4. The active noise reduction device according to claim 1, wherein the predetermined signal is generated based on a control signal input to the noise cancellation output device and a reference signal corresponding to the noise.

5. The predetermined signal is generated by dividing a first calculated value calculated based on the control signal by a second calculated value calculated based on the reference signal. The control device is an active noise reduction device according to claim 4, wherein the amount of reflection decreases as the predetermined signal becomes larger.

6. The at least one filter is, A control filter that generates a control signal input to the sound-canceling output device, A quadratic path filter showing the estimated value of the transfer function from the sound cancellation output device to the error microphone, The system includes a first-order path filter that shows an estimated value of the transfer function from the noise source to the error microphone, The active noise reduction device according to claim 1, wherein the predetermined signal is generated based on the control filter, the primary path filter, and the secondary path filter.

7. The predetermined signal is generated by dividing the product of a first intermediate value calculated based on the control filter and a second intermediate value calculated based on the secondary path filter by a third intermediate value calculated based on the primary path filter. The control device is an active noise reduction device according to claim 6, wherein the amount of reflection decreases as the predetermined signal becomes larger.

8. The at least one filter includes a control filter that generates a control signal input to the sound cancellation output device, The active noise reduction device according to any one of claims 1 to 7, wherein the control device changes the amount by which the value of the control filter before adaptive update is reflected in the value of the control filter after adaptive update based on the predetermined signal.

9. The at least one filter further includes a quadratic path filter that shows an estimated value of the transfer function from the sound cancellation output device to the error microphone, The active noise reduction device according to claim 8, wherein the control device changes the amount by which the value of the secondary path filter before adaptive update is reflected in the value of the secondary path filter after adaptive update based on the predetermined signal.

10. The at least one filter further includes a first-order path filter that shows an estimate of the transfer function from the noise source to the error microphone, The active noise reduction device according to claim 9, wherein the control device changes the amount by which the value of the primary path filter before adaptive update is reflected in the value of the primary path filter after adaptive update based on the predetermined signal.

11. The control device is A predetermined coefficient is changed based on the predetermined signal. An active noise reduction device according to any one of claims 1 to 7, wherein the amount of reflection is calculated by multiplying the predetermined coefficient by the value of the filter before its adaptation update.

Images