JP2014153571A - Active noise control device and control method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a sensorless active noise control device which attains high-speed control and increases followability and a noise reduction effect.SOLUTION: An active noise control device includes: a control device 11 of a power source; a control target area 16 to which a sound generated by the power source is transferred as a control target sound; a DSP 13 for performing signal processing with a drive control signal of the power source as an input signal; an amplifier 14 for amplifying an output signal from the DSP 13; and an actuator speaker 15 for outputting a control sound from the amplified output signal to the control target area 16. The DSP 13 comprises: a signal processing circuit 18 for estimating frequency characteristics of the control target sound from the input signal and calculating a control target frequency of order components; and an ADF/LMS algorithm 20 for performing ANC control by using an individual convergence coefficient for each frequency of order components of the calculated control target frequency. The actuator speaker 15 outputs the control target sound and control sound to the control target area 16 in response to the output signal from the ADF/LMS algorithm 20.

Description

  The present invention estimates a frequency characteristic of a control target sound from a drive control signal of a power source, narrows the control target to the estimated frequency to improve followability and noise reduction effect, and a sensor such as a microphone having a sound quality adjustment function The present invention relates to an active noise control device and a control method therefor.

  One of the noise countermeasure techniques for sound generated from a power source is an active noise control (ANC) technique. This ANC technique is a technique for reducing noise by causing the sound to be removed to interfere with wavelengths of the same amplitude and opposite phase.

  There is an active noise control apparatus provided with a Filtered-X LMS (Least Mean Square) algorithm in the existing ANC technology. As shown in FIG. 1, the active noise control apparatus includes a DSP (Digital Signal Processor) 1 and an output signal generated by the DSP 1 between the engine 2 side microphone 2a and the error microphone 3a on the control target region 3 side. (Control signal) It is composed of an amplifier 5 for amplifying y and an actuator / speaker 6 for generating a control sound z.

  Then, using the reference signal (input signal) x obtained from the engine 2 side microphone 2a, the adaptive filter 7 performs filtering by changing the filter coefficient Wi to generate the control signal y. The filter coefficient Wi of the adaptive filter 7 is automatically optimized using an LMS (Least Mean Square) algorithm 8. The error signal e obtained by the error microphone 3a in the control target area 3 is input to the LMS algorithm 8 after A / D conversion, and the error signal e is reduced by using the LMS algorithm 8 using the convergence coefficient μ. The coefficient Wi constituting the ADF 7 is updated by optimization.

  The control signal y filtered and output by the ADF 7 of the DSP 1 is D / A converted, becomes an analog waveform signal, is amplified by the amplifier 5, and the control sound z is output from the actuator speaker 6 to the control target region 3. . This control sound z interferes with the control target sound d from the engine noise, resulting in noise reduction.

  As described above, the ANC technique outputs the control sound z having the same amplitude and opposite phase from the actuator / speaker 6 with respect to the control target sound d from the engine noise, thereby reducing the noise in the control target region 3. .

In the ANC technique, as shown in FIG. 2, the error signal e detected from the control target sound d and the control sound z by the error microphone 3a in the control target area 3 is ANC feedback controlled by the DSP 1, and the error signal e becomes small. Thus, signal processing is performed by the LMS algorithm 8 using the convergence coefficient μ, and the filter coefficient Wi of the adaptive filter 7 is optimally changed and controlled by the generated filter signal W. In the ANC technology, the “Filtered-X” LMS algorithm 9 is controlled in consideration of a plant model G _M (z) obtained by modeling the transfer function G (z) of the error path from the speaker 6 to the microphone 3a. The input signal x is filtered by the adaptive filter 7 under the control of the “Filtered-X” LMS algorithm 9.

JP-A-8-286675 JP-A-6-259083 JP-A-3-27517

  Patent Documents 1 and 2 describe an active noise control device including a DSP adaptive filter. This noise control device includes an error microphone for observing an error signal in a control target region.

  However, error microphones are vulnerable to heat and water, are susceptible to thermal degradation, have great restrictions on installation locations, and have problems with cost and durability.

  Further, in the conventional active noise control device, the control target sound is controlled to be reduced in the entire frequency range, but the convergence target must be set to one for the entire band, and the control target sound is controlled. On the other hand, the frequency that exerts the noise reduction effect is limited by the largest peak value in the frequency domain, and the noise reduction effect is limited.

  Further, Patent Document 3 describes a microphone-less active silencer system that does not have a feedback loop. However, in a microphoneless active silencer system that does not have a feedback system, the control target sound in the control target region can be detected by a sensor. Although position estimation is a very important technique, no consideration is given to the position estimation technique of the control target region at this sensor position.

  The present invention has been made in consideration of the above-described circumstances. The frequency characteristic of the sound to be controlled is estimated based on the drive control signal from the power source, and the estimated frequency is narrowed down to the control object to speed up the control. It is an object of the present invention to provide a microphone-less active noise control device and a control method therefor that improve the followability and noise reduction effect.

  Another object of the present invention is to estimate the frequency characteristics of the sound to be controlled from the drive control signal of the power source, to decompose the frequency peak for each estimated frequency to be controlled and adjust the sound volume to adjust the sound quality, It is an object of the present invention to provide an active noise control device and a control method thereof that can prevent auditory discomfort.

  In order to solve the above-described problem, an active noise control device according to the present invention is a drive source control device that drives and controls a power source, and a control target region in which a sound generated from the power source is transmitted as a control target sound. A DSP that performs signal processing using the drive control signal of the power source as an input signal, an amplifier that amplifies the output signal from the DSP, and an actuator speaker that outputs control sound from the amplified output signal to the control target region The DSP estimates a frequency characteristic of the control target sound from the input signal, and calculates a fundamental frequency and a control target frequency of an order component estimated from the fundamental frequency, and a calculated signal processing circuit An ADF / LMS algorithm that performs ANC control using an individual convergence coefficient for each frequency of the order component of the frequency to be controlled; It said actuator speaker receives an output signal from Gorizumu is characterized in that for outputting a control sound to the control region.

  Further, in order to solve the above-described problem, the active noise control method according to the present invention estimates the frequency characteristics of the control target sound from the drive control signal of the drive source, and the basic (primary) frequency and the order component of the control target sound. The control target frequency of the frequency is calculated by the signal processing circuit, the calculated order component control target frequency is distributed for each order and input to the ADF / LMS algorithm to perform ANC control, and the control output by this ANC control The signal is amplified and input to the actuator / speaker, and the control sound is output from the actuator / speaker to the control target sound in the control target region. The ADF / LMS algorithm uses each order from the order component control target frequency. This is a control method characterized by filtering with an adaptive filter every time.

  The active noise control device and the control method thereof according to the present invention estimate the fundamental frequency of the sound to be controlled from the drive control signal of the generation source, calculate the control target frequency of the order component of the fundamental frequency, and control the order component. ANC control is performed by the ADF / LMS algorithm by distributing the frequency for each order component, the frequency range of the frequency to be controlled is reduced, the ANC control is speeded up, and the followability is improved. The ADF / LMS algorithm is Since the control target frequency of the order component is filtered by setting the optimum convergence coefficient for each order component, the noise reduction effect is improved. Further, since the frequency to be controlled of each order component can be set to reduce or increase sound for each order by the ADF / LMS algorithm, the volume and sound quality can be controlled.

  Furthermore, the active noise control device and the control method thereof include a microphone-less active noise control device that eliminates the need for a microphone by predicting an error signal between a control target sound and a control sound by an amplitude phase estimation unit from a drive control signal of a power source. Therefore, it is possible to reduce the cost by eliminating the need for a microphone having low durability due to heat and temperature.

The block diagram which shows the existing active noise control apparatus provided with "Filtered-X LMS algorithm." Schematic of "Filtered-X LMS algorithm" with which the existing active noise control apparatus is equipped. The block diagram which shows 1st Embodiment of the active noise control apparatus which concerns on this invention. The flowchart explaining the operation example of the active noise control apparatus which concerns on this invention. (A) is a graph which shows the noise reduction effect of ANC control in the active noise control apparatus which concerns on this invention, (B) is a graph which shows the noise reduction effect of ANC control in the existing active noise control apparatus. The graph which compares and shows the trackability in the active noise control apparatus which concerns on this invention, and the existing active noise control apparatus. (A) is a graph showing an example of sound quality / volume control at the time of ANC control ON / OFF during steady running of a vehicle equipped with the active noise control device according to the present invention, and (B) is an ANC control ON / ON at the time of vehicle acceleration. The graph which shows the example of sound quality and volume control at the time of OFF. The block diagram which shows 2nd Embodiment of the active noise control apparatus which concerns on this invention.

  Embodiments of the present invention will be described below with reference to the accompanying drawings.

FIG. 3 is a configuration diagram illustrating the first embodiment of the active noise control device. This active noise control device 10 is a DSP that performs signal processing using an engine 12 that is driven and controlled by a power source control device 11 and an engine sensor signal that senses the state of the engine 12 as an input signal x as a drive control signal for the power source. and (Digital signal Processor) 13, an amplifier 14 for amplifying the output signal y from the DSP 13, mainly from the actuator speaker 15 for outputting by generating a control sound _{z n} from the control signal of the amplified output signal _{y n} Composed. Actuator speaker 15 is disposed in the control region 16, and outputs the control sound z _n as interference noise to the control target sound d _n is the engine noise from the engine 12 in this control region 16.

The DSP 13 receives an engine sensor signal (waveform signal, pulse signal) from the engine 12 that is a drive control signal for the power source, is converted into a digital signal by the A / D converter 17, and is converted into a basic frequency (basic) of the signal processing circuit 18. Period) is sent to the extraction unit 19. In fundamental period extraction unit 19, the basic control target sound d _n (primary) frequency x ₁ is extracted by the signal processing by the calculation, the basic control target sound d (n) (primary) frequency x ₁ is estimated. Arbitrary n-th order component frequency (x ₂ , x ₃ ,..., X _n ) of the fundamental frequency that is an arbitrary signal component is calculated from the estimated fundamental frequency x _1. The The basic (primary) frequency and the order component frequency signals x ₁ , x ₂ ,..., X _n are distributed and input to the ADF / LMS algorithm 20 (20 ₁ , 20 ₂ ,..., 20 _n ) for each order. The Depending on the type of engine, the order component n may be other than an integer, for example, a 0.5 order component, as in a V-twin engine.

In this way, the basic period extraction unit 19 extracts the basic frequency x ₁ and calculates the order component frequencies x ₂ , x ₃ ,..., X _n from the basic frequency x ₁ . Fundamental frequency x ₁ can be calculated from the engine sensor signal such as an ignition pulse, also order component frequency _{x 2, x 3, ...,} x n can be readily determined by multiplying n of the fundamental frequency. In any case, the signals of the order component frequencies x ₁ , x ₂ ,..., X _n generate sine wave signals.

The ADF / LMS algorithm 20 (20 ₁ , 20 ₂ ,..., 20 _n ) includes an adaptive filter (ADF) 21 (21 ₁ , 21 ₂ ,..., 21 _n ) and an LMS algorithm 22 (22 ₁ , 22 ₂ ,. 22 _n ) and filters the input signals x ₁ , x ₂ ,..., X _n of the respective orders using the filter coefficients of the adaptive filter 21 (21 ₁ , 21 ₂ ,..., 21 _n ). processing the input signals _{x 1,} x 2 frequency components of each order of minutes, ..., _x 1 to _{x n,} x 2, ..., the output signal for each order component frequency corresponding to _{x n y 1, y 2,} ..., and it outputs the _{y n,} respectively. The adaptive filter 21 (21 ₁ , 21 ₂ ,..., 21 _n ) is an ADF (Adaptive Digital Filter), and uses an LMS (Least Mean Square) algorithm 22 (22 ₁ , 22 ₂ ,..., 22 _n ). filter coefficient _{w 1,} w 2, ..., slide into updated from time to time a _{w n.} Adaptive filter _{21 (21 1, 21 2,} ..., 21 n) in the input signal _{x 1,} x 2 of each order component frequency, ..., _{x n} is the filter coefficient and the product-sum operation is updated every moment filtering to signal processing, the processing result is the following number of the frequency of the output signal (control _{signal) y 1, y 2, ...} , are outputted individually as y _n.

The LMS algorithm 22 (22 ₁ , 22 ₂ ,..., 22 _n ) is one of known optimization methods, and an algorithm using instantaneous values of gradient vectors for reducing the amount of computation of the steepest descent method. It is. LMS algorithm _{22 (22 1, 22 2,} ..., 22 n) in the adaptive filter _{21 (21 1, 21 2,} ..., 21 n) in order to impart the optimum filter coefficient, an engine sensor amplitude phase estimation unit 30 signal frequency characteristics of the controlled object sound d _n is deduced from error signal e _n, the convergence of each next few minutes between the control sound z _n from the control target sound d _n and the actuator speaker 15 from engine noise The coefficients μ ₁ , μ ₂ ,..., Μ _n are individually set and sent to the LMS algorithm 22 (22 ₁ , 22 ₂ ,..., 22 _n ).

The LMS algorithm _{22 (22 1, 22 2,} ..., 22 n) in the filter signal _w 1 so that the error signal _{e n} is _minimized, w 2, ..., update the _{w n,} the adaptive filter 21 ₍₂₁ 1, 21 ₂ ,..., 21 _n ) and ANC control is performed by performing a convolution operation. In the adaptive filter 21 (21 ₁ , 21 ₂ ,..., 21 _n ) to which the filter signals w ₁ , w ₂ ,..., W _n are input, an optimum filter coefficient that changes from time to time is given to perform filtering signal processing. In this way, the ADF / LMS algorithm 20 (20 ₁ , 20 ₂ ,..., 20 _n ) uses the optimum filter coefficients in the adaptive filter 21 (21 ₁ , 21 ₂ ,..., 21 _n ). input signal _x 1 _{frequency, x 2, ...,} x _n filtering process to the input signals _{x 1,} x 2, ..., the output signal _y 1 of each order component frequency corresponding to _{x n, y 2, ...,} y _n is output as the control signal y.

Control signals (output signals) y ₁ , y ₂ ,..., Y _n generated by the ADF / LMS algorithm 20 (20 ₁ , 20 ₂ ,..., 20 _n ) are added by the adder 25 to be 1 One control signal y is output as an analog waveform signal from the DSP 13 by the D / A converter 26. The analog waveform signal (control signal y) output from the DSP 13 is amplified by the amplifier 14 and then output from the actuator speaker 15 to the control target region 16 as the control sound z _n . Control sound z _n are made to interfere as interference noise with respect to the control target sound d _n of the engine noise.

Furthermore, the active noise control apparatus 10, an engine sensor signal from the engine 12 to obtain the error (error) signal e _n, is input to the amplitude phase estimation unit 30 through the A / D converter 29, the amplitude and phase the error signal e _n is obtained which corresponds to the difference between the control sound d _n from the control target sound d _n and the actuator speaker 15 of engine noise in estimating section 30. The error signal e _n is equivalent to the error signal e obtained in error microphone 3a of existing active noise control apparatus (see FIGS. 1 and 2).

The active noise control apparatus 10 of the first embodiment, provided with an amplitude phase estimation unit 30 to the DSP 13, the frequency of the control target sound d _n by amplitude phase estimation unit 30 for inputting an engine sensor signal for sensing the state of the engine 12 characteristics to estimate, and obtains the error signal e _n. This estimation enabled microphoneless noise control without using a microphone without using an error microphone.

Also obtains an error signal e _n corresponding to the signal e from the existing error microphone amplitude phase estimation unit 30. Convergence coefficient mu ₁ to the error signal _{e n} from the amplitude phase estimation unit _{30, μ 2, ..., μ} n is granted LMS algorithm _{22 (22 1, 22 2,} ..., 22 n) are input to, the LMS algorithm _{22 (22 1, 22 2,} ..., 22 n) convergence coefficient mu ₁ so that the error signal _{e n} becomes smaller, mu _2, ..., to signal processing using a mu _n, the adaptive filter 21 ₍₂₁ 1, 21 _2, ..., optimal filter signal _w 1 which changes every moment with _{21 n), w 2, ...} , w n are updated respectively.

On the other hand, the DSP 13 is provided with a sound quality adjustment circuit 33 (33 ₁ , 33 ₂ ,..., 33 _n ) that adjusts the amplitude of each order component frequency and adjusts the sound quality. Tone control circuit _{33 (33 1, 33 2,} ..., 33 n) , each frequency input signal _x 1 of the next few _minutes, x 2 ..., amplification value _a 1 to _{x n,} a 2 _..., an amplifier which imparts _{a n 34 (34 1, 34 2,} ..., 34 n) and the amplification value _{a 1, a 2 ..., a} n frequency input signals of the next few minutes is assigned _{x 1,} x 2 ..., amplitude phase estimation of _{x n} adder _{35 (35 1, 35 2,} ..., 35 n) to be added to the error signal _{e n} from the part 30 constituted by a.

Tone control circuit _{33 (33 1, 33 2,} ..., 33 n) , the input signal _x 1 of the next few minutes frequency amplification value a is _applied, x 2 ..., _{x n} is the adder 35 ₍₃₅ 1, 35 _2, ..., 35 _n) in is added to the error signal e _n from the amplitude phase estimation unit 30, the frequency of each order component can be controlled to a desired amplitude error signal e _n is reduced.

In the DSP 13, the fundamental (primary) frequency is extracted from the frequency characteristic of the control target sound by the basic period extraction unit 19 to which the engine sensor signal is input, the frequency peak is decomposed, and the frequency of each order component is set as the control target frequency. Then, the volume control circuit 33 (33 ₁ , 33 ₂ ,..., 33 _n ) is guided for each frequency signal x ₁ , x ₂ ..., X _{n of} each order component to reduce or increase the sound for each frequency to be controlled. amplification value such sounds _{a 1, a} 2 _..., it is possible to impart _{a n.} Thereby, the volume adjustment of the control target frequency can be performed for each control target frequency or each control target order, and the sound quality can be adjusted.

[Operation of First Embodiment]
FIG. 4 is a flowchart for explaining the operation of the active noise control device 10.

In this active noise control device 10, a DSP (Digital Signal Processor) 13 acquires an engine sensor signal (pulse signal) in step 1 using wiring from the engine 12, and the acquired engine sensor signal is the basic of the DSP 13. A basic frequency component is calculated by the period extraction unit 19 (step 2). Control target sound d _n is, n multiples of the fundamental frequency, for example, it is known to occur in this fundamental frequency and the order component frequency. In other words, the control target sound _{d n,} the frequency _{x 1,} x 2 ... of order components, has been found to occur with _{x n.}

Therefore, control should order component frequency of the control target sound d _n is generated as the input signal sine wave (Sin wave) in step 3. The input signal Sin wave of each order component ADF · LMS algorithm _{20 (20 1, 20 2,} ..., 20 n) the adaptive filter _{21 (21 1, 21 2, ...} , 21 n) by the input, the filtering process (Step 4).

On the other hand, an engine sensor signal, the control region 16, in order to obtain an error signal between the control sound z _n from the control target sound d _n and the actuator speaker 15 from the engine noise, amplitude phase estimation of DSP13 Input to the unit 30. The amplitude phase estimation unit 30 acquires an engine sensor signal, estimates the amplitude phase of each order component frequency, and acquires an error signal in the control target region 16 (step 5). Is the error signal _{e n} was the acquired amplitude phase estimation unit 30, tone control circuit ₃₃ in step _{6 (33 1, 33 2,} ..., 33 n) in the amplification value _{a 1, a 2 ..., a} n each order component is added to the frequency of the input signal LMS algorithm _{22 (22 1, 22 2,} ..., 22 n) is sent to, the LMS algorithm _{22 (22 1, 22 2,} ..., 22 n) is the error signal _{e n} by It decreases, by using the signal for adjusting the sound volume and sound quality, the adaptive filter _{21 (21 1, 21 2,} ..., 21 n) filter coefficients _{w 1,} w 2 of, ..., changing the _{w n,} the optimum value ANC control is performed so that

The amplitude phase estimation of the DSP 13 is performed by the adaptive filter 21. Adaptive filter 21 is seeking the filter coefficients so that the error (error) signal e _n is the smallest. The optimization of the filter coefficient by the adaptive filter 21 is guaranteed that the amplitude phase estimation is optimally performed by the LMS algorithm 22 using the steepest descent method.

In this way, DSP 13 is, Sin wave input signal _x 1 of the frequency of each order _component, x 2 ..., each adapted as _{x n} filter _{21 1,} 21 2, ..., are input one by one to 21 _n, Each of the adaptive filters 21 ₁ , 21 ₂ ,..., 21 _n has a convergence coefficient μ ₁ , μ ₂ ,..., Μ _n and an amplification value a ₁ , a _{2 ,.} LMS algorithm _{22 1,} 22 2 using, ..., optimized filter coefficients wi are the 22 _n are filtering the input signal _{x 1,} x 2 of the frequency of each order component, ..., each order corresponding to _{x n} output signal _{y 1,} y 2 components, ..., the control signal y is generated from _{y n} (step 7).

Output signals (control signals) y ₁ , y ₂ ,..., Y _n from the adaptive filters 21 ₁ , 21 ₂ ,..., 21 _{n of} the ADF / LMS algorithm 20 generated in step 7 are added by an adder 25. The added control signal y is output to the D / A converter 26 as an output signal (step 8).

  In the D / A converter 26, the control signal is converted into an analog waveform signal in step 9.

Converted analog waveform signal is amplified by the amplifier 14 (step 10), is output from the actuator speaker 15 to the control target region 16 as the control sound z _n (step 11).

Control sound z _n outputted from the actuator speaker 15 is a control interference sound control target sound d _n from the engine noise is adjusted so that the desired sound in the control region 16.

Further, the ADF · LMS algorithm 20 of DSP 13, the ADF · LMS algorithm _{20 1,} 20 2, ..., 20 adaptive filter ₂₁ 1 of _n, 21 2, ..., 21 _n is the convergence coefficient most suitable for each order mu _1, mu _2, ..., mu LMS algorithm _{22 1 n using,} 22 2, ..., since the filter coefficients are set by the filter signal produced by 22 _n, the control region 16, the control target sound d The noise control effect of _n can be improved.

  Note that the left and right waveform example diagrams attached to FIG. 4 exemplify signal waveform examples respectively acquired in steps 1 to 8 as an example.

The active noise control apparatus 10 of the present embodiment, the frequency characteristics of the controlled object sound d _n inferred from an engine sensor signal, and a control target frequency, respectively to decompose the frequency peak as a basic (primary) frequency or order components Frequency . A convergence coefficient can be individually set for each frequency to be controlled, and ANC control can be performed by the DSP 13 for each frequency to be controlled.

  For this reason, the ANC control using the DSP 13 in the active noise control device 10 is performed as shown by the solid line A in FIG. As shown in the OFF state, each order component frequency peak is reduced, and a high noise reduction effect is obtained over a wide frequency range.

On the other hand, FIG. 5B shows the noise reduction effect obtained by the existing active noise control apparatus shown in FIG. As shown in FIG. 5B, in the existing active noise control device, since one convergence coefficient μ must be set in the entire frequency band of the DSP 13, as indicated by the solid line C even when the ANC is ON, as shown by the ANC is OFF dashed D, narrow frequency to exert a noise reduction effect to the control target sound d _n, is limited.

Furthermore, the active noise control apparatus 10 of the present embodiment, to estimate the frequency characteristics of the amplitude and phase of the control target sound d _n from the engine sensor signal, and extracted with fundamental frequency extraction unit 19 of the DSP 13, the basic (primary) frequency Ya The frequency of the nth order was obtained, and the frequency peak was decomposed to obtain the control target frequency. As a result, the convergence coefficient μ _n can be set for each component, and the ANC control by the ADF / LMS algorithm 20 of the DSP 13 can be speeded up as indicated by symbol E in FIG. In FIG. 6, symbol F indicates ANC control by the existing DSP 13, and it takes about 1.3 seconds to reduce the noise level to a certain level by the existing ANC control F. The DSP 13 of this embodiment is shown in FIG. In the ANC control of the ADF / LMS algorithm 20 according to the above, the speed can be increased by about 5 times as high as about 0.25 sec, and high followability can be obtained.

Furthermore, the active noise control apparatus 10 of the present embodiment, controlling the amplitude and frequency characteristics of the phase estimated from an engine sensor signal of the engine 12, order component frequency estimated from the fundamental frequency and the fundamental frequency of the control target sound d _n As the target frequency, it is possible to set sound reduction and sound increase for each control target frequency.

In this active noise control device path 10, the sound quality adjustment circuit 33 (33 ₁ , 33 ₂ ,..., 33 _n ) reduces or increases sound for each control target frequency that is a fundamental (primary) frequency and an n-th order frequency. Volume and sound quality can be adjusted.

FIG. 7A shows a case where sound quality adjustment control is performed using the active noise control device 10 for a vehicle at a constant speed of 50 km / hour, for example. This sound quality adjustment control, LMS algorithm _{22 1,} 22 2 of the DSP 13, ..., to 22 _n, as set in Zoon written sound reduction for each control target frequency is possible, tone control circuit 33 ₍₃₃ 1, 33 _2, ..., an amplifier _{34 1,} 34 2 33 _n), ..., amplification value at 34 _{n a 1,} a 2, ..., set the _{a n,} to create an algorithm that is suitable for sound reduction and Zoon Yes. 7A shows a state (solid line G) in which sound quality control is performed at a constant speed with the ADF / LMS algorithm 20 (20 ₁ , 20 ₂ ,..., 20 _n ) and sound quality control is not performed. It is a graph which compares and shows a frequency characteristic with a state (broken line H). As shown by the solid line G, Doing tone control, control target sound d _n is over a wide frequency range, so that the required sound pressure of the target, the sound control target sound d _n and sound reduction or Zoon The pressure level can be controlled to be substantially constant.

  FIG. 7B shows an example of sound quality control (sound quality evaluation test) during acceleration of the scooter vehicle. By performing this sound quality control according to the engine speed, at the time of acceleration, the engine noise adjustment can be connected with a smooth and pleasant continuous sound as shown by the solid line I with little fluctuation in the sound pressure as the engine speed increases. The volume and sound quality are controlled. Further, as shown by the broken line J, when the sound quality control is not performed, the sound pressure fluctuation increases as the engine speed increases, and a roaring sound is generated, and an audible discomfort sound is generated. By performing sound quality control, it is possible to prevent amplitude fluctuation during acceleration, connect with a pleasant smooth continuous sound, and prevent a sense of incongruity.

Furthermore, the control target sound d acquired at the position of the error microphone 3a (see FIG. 1) of the existing active noise control device is predicted to be a linearly transformed signal of the reference signal x. Then, the error signal e can be obtained by taking the difference from the signal obtained by convolving the transfer function G (z) from the control target signal d to the adaptive filter output. That is, the signal of the control target sound d _n in FIG. 3 can be expressed as the sensor input signal obtained by convoluting the transfer function to the signal of each order component of the engine sensor signal. Focusing on the sine wave that is, one frequency component, the transfer function is represented by the amplitude ratio and phase difference between the sound signals of the control target sound d _n of the engine sensor signal and the control target region 16. (For convenience, the transfer function of the control target signal d _n engine sensor signals x and error sensor position is referred to as a transfer function H (z).) Therefore, the sine wave signal and the transfer function H that is generated from the engine sensor signal The sensorless active noise control device 10 is obtained by using the error signal created using (z).

When the engine sensor signal is input and the control target sound signal is output, the transfer function H (z) can be uniquely determined by a linear output relationship between the input and output. Transfer function H (z) by having a prepared value of the transfer function data consisting of the impulse response, it is possible to estimate the sound signal of the control target sound d _n from the engine sensor signal.

  Therefore, it is possible to obtain a sensorless active noise control device that does not use a physical sensor such as a microphone by estimating the frequency characteristics of the control target sound of the engine noise.

The sound quality control in the DSP 13 of the active noise control device 10 is such that the amplitude of the high and long wave noise component synchronized with the engine rotation does not fluctuate over time, and the low-order noise order component in the adaptive sound quality control during vehicle acceleration. The present inventors have found that small is one index of comfortable pleasant sound (see CDROM P895-P896, Proceedings of the Acoustical Society of Japan 2010 Spring Conference). The amplification value a ₁ = a ₂ = 0 of the fundamental (primary) frequency or the secondary frequency component, and the amplification value of the third or higher order component a ₃ = a ₄ = a ₅ =... = A _n = constant. It is possible to control the sound quality. Therefore, as the control target signal, amplification value a ₃ tertiary above order _component, ..., by obtaining a _n, it is possible to perform a more preferred tone control.

[Second Embodiment]
FIG. 8 is a configuration diagram illustrating a second embodiment of the active noise control device.

  The active noise control device 10A shown in the second embodiment pays attention to the fact that not only the frequency that requires noise reduction is determined by the engine sensor signal, but also sound that does not depend on the engine sensor signal exists. The same components and functions as those of the active noise control device 10 shown in the first embodiment are denoted by the same reference numerals, and redundant description is omitted or simplified.

  In the active noise control device 10A shown in the second embodiment, in addition to the active noise control device 10 of the first embodiment, sound generated at a constant frequency is also added to noise reduction measures. The purpose is also to control.

The active noise control apparatus 10A adds an ADF / LMS algorithm 40 for controlling a constant frequency sound to the DSP 13A described in the first embodiment to the DSP 13A, and the ADF / LMS algorithm 40 is an adaptive filter for filtering the constant frequency sound. (ADF) 43 and LMS algorithm 44 are combined. LMS algorithm 44 is inputted error signal convergence coefficient μa to e _n from the amplitude phase estimation unit 30 sends the filtered signal wa that the signal processing so that the error signal e _n is reduced to the adaptive filter 43, the adaptive filter 43 Controls filter coefficients. The adaptive filter 43 performs filtering of a constant frequency sound with this filter coefficient. Filtered output signal by the adaptive filter 43 (control signal) _{y a} are added by the adder 25, is converted into an analog waveform signal is a control signal _y m _(y + _{y a)} by the D / A converter 26. Analog waveform (control) signal is amplified by subsequently amplifier 14, is output as the control sound z _m to the control region 16.

[Operation of Second Embodiment]
The active noise control device 10A shown in the second embodiment has the same operation and function as the active noise control device 10 shown in the first embodiment, and also does not depend on the engine sensor signal, for example, the path There is also a sound with a constant frequency that is strengthened by the length.

For this reason, the active noise control device 10A uses not only the frequency that needs noise reduction but also the constant frequency that does not depend on the engine sensor signal as the target sound for noise reduction measures. The active noise control system output signal obtained by filtering the signal of a constant frequency sound with an adaptive filter (ADF) 43 of the ADF · LMS algorithm 40 which is additionally provided 10A (control signal) _{y a} are added by the adder 25, the control signal y _m (y + y _a ) is output, converted into an analog waveform (control) signal, amplified, and output from the actuator speaker 15 as a control sound z _m . The control sound z _m are output to the control target sound d _m of engine noise, for example as an interference sound, including noise reduction measures to the path constant frequency sound to be intensified by the length.

The active noise control apparatus 10A shown in a second embodiment, the sound generated constant frequency to the control target sound d _m from the engine noise is also intended, the control target frequency is constant, the control only the phase and amplitude from the engine sensor signal estimates calculated as the target sound d _m, the noise reduction measures to certain frequency sound is also being implemented as an object.

[Others]
In the embodiment of the present invention, the engine sensor signal from the engine is used as an input signal, and the signal processing is performed by DSP ANC technology to reduce noise. However, the engine sensor signal of the engine is an ignition pulse signal as the input signal. A drive control signal of the power source such as (ignition signal) or throttle opening signal of the throttle valve may be used as the input signal. Moreover, you may include the generation | occurrence | production sound of the fixed frequency which exists in the environment which is not influenced by the drive control signal of a power source as a control object sound.

  Furthermore, in the embodiment, an example of an active noise control device intended for noise reduction control from an internal combustion engine as a power source has been shown. However, when a predetermined drive control input signal is given to the power source, the reproducibility is improved. Any power generation source that outputs sound exhibiting certain acoustic characteristics may be used. For example, not only a power generation source of an engine (internal combustion engine) but also a rotary electric machine such as an electric motor or a hybrid vehicle in which an internal combustion engine and a rotary electric machine are combined may be used as a power generation source.

In addition, the present invention aims to control sound, but it can also be applied to vibration suppression control of an object. The engine noise d _n of the controlled object to the engine vibration, similarly to the noise control by replacing the actuator speaker 15 to the vibrator, it is possible to vibration control. Therefore, vibration reduction control for a structure that exhibits reproducible vibration characteristics when a predetermined drive control input signal is applied may be an object of the present invention.

  In addition, the control target area may include not only the vehicle compartment, the engine room, and the motor room, but also intake system piping, exhaust system piping, and engine peripheral devices.

10, 10A Active noise control device 11 Power source control device 12 Engine 13, 13A DSP
14 Amplifier 15 Actuator / speaker 16 Control target region 17 A / D converter 18 Signal processing circuit 19 Basic period (frequency) extraction unit 20 (20 ₁ , 20 ₂ ,..., 20 _n ) ADF / LMS algorithm 21 (21 ₁ , 21 ₂ ,..., 21 _n ) Adaptive filter (ADF)
22 (22 ₁ , 22 ₂ ,..., 22 _n ) LMS algorithm 25 Adder 26 D / A converter 29 A / D converter 30 Amplitude phase estimation unit 33 (33 ₁ , 33 ₂ ,..., 33 _n ) Sound quality adjustment Circuit 34 (34 ₁ , 34 ₂ ,..., 34 _n ) Amplifier 35 (35 ₁ , 35 ₂ ,..., 35 _n ) Adder 40 ADF / LMS algorithm 43 Adaptive filter (ADF)
44 LMS algorithm

Claims (14)

A drive source control device for driving and controlling the power source;
A control target region in which sound generated from the power source is transmitted as control target sound;
A DSP that performs signal processing using the drive control signal of the power source as an input signal;
An amplifier for amplifying the output signal from the DSP;
An actuator speaker that outputs a control sound from the amplified output signal to the control target region;
The DSP estimates a frequency characteristic of the control target sound from the input signal, and calculates a control target frequency of an order component estimated from the fundamental frequency and the fundamental frequency; and
An ADF / LMS algorithm that performs ANC control using an individual convergence coefficient for each frequency of the calculated order component of the frequency to be controlled;
In response to an output signal from the ADF / LMS algorithm, the actuator / speaker outputs a control sound to the control target area. The signal processing circuit includes a fundamental period calculation unit that estimates a frequency characteristic of the control target sound from a drive control signal of the power source, estimates a fundamental frequency thereof, and calculates an order component frequency.
The active noise control device according to claim 1, wherein a control target frequency is distributed from the fundamental (primary) frequency and the order component frequency calculated by the fundamental period calculation unit and is input to the ADF / LMS algorithm. The ADF / LMS algorithm includes an ADF / LMS algorithm for each order that causes the control target frequency to be input for each order component,
3. The active noise control apparatus according to claim 1, wherein the ADF / LMS algorithm for each order is ANC controlled for each order component of the frequency to be controlled. The ADF / LMS algorithm is composed of an ADF / LMS algorithm for each order in combination with an optimal filter and an LMS algorithm,
The active noise control apparatus according to claim 3, wherein each LMS algorithm uses an individual convergence coefficient, adjusts a convergence environment, and optimizes a filter coefficient of each adaptive filter. The DSP inputs a drive control signal from the power source and obtains the basic (primary) frequency of the control target sound and the control target frequency of the order component obtained by the signal processing circuit for the ADF / LMS for each order component. While letting the algorithm input
2. The active noise control apparatus according to claim 1, wherein an output signal for each order component output from the ADF / LMS algorithm of the order component is added by an adder, and the added output signal is sent to an amplifier. The DSP includes an amplitude phase estimation unit that estimates and generates an error signal between a control target sound from the power source and a control sound from the actuator / speaker from a drive control signal of the power source,
The active noise control apparatus according to claim 1, wherein the error signal generated by the amplitude phase estimation unit is multiplied by a convergence coefficient and sent to the ADF / LMS algorithm. The DSP includes, for each order component, a sound quality adjustment circuit that individually guides the control target frequency of each order component from the control target frequency of the basic (primary) frequency and the order component frequency processed by the signal processing circuit,
The sound quality adjustment circuit for each order component individually adds an amplification value for adjusting the volume and sound quality for each control target frequency of each order component, and is added to the error signal from the amplitude phase estimation unit by an adder. Item 2. The active noise control device according to Item 1. The DSP includes an additional ADF / LMS algorithm for ANC controlling a signal of a constant frequency sound other than the order component of the fundamental frequency of the sound to be controlled,
The active noise control according to any one of claims 1, 6 and 7, wherein the additional ADF / LMS algorithm is multiplied by a convergence coefficient to an error signal from the amplitude phase estimation unit to optimize a filter coefficient of an adaptive filter. apparatus. The frequency characteristic of the control target sound is estimated from the drive control signal of the drive source, and the control target frequency of the basic (primary) frequency and the order component frequency of the control target sound is calculated by the signal processing circuit.
ANC control is performed by distributing the calculated order component control target frequency for each order and inputting it to the ADF / LMS algorithm.
The control signal output by this ANC control is amplified and input to the actuator / speaker,
A control sound is output from the actuator / speaker to the control target sound in the control target region,
In the active noise control method, the ADF / LMS algorithm performs filtering with an adaptive filter for each order from the order component control target frequency. The ADF / LMS algorithm includes an ADF / LMS algorithm for each order component distributed from each order component control target frequency,
The active noise control method according to claim 9, wherein the filter coefficients of the adaptive filter are individually optimized by the LMS algorithm in which the ADF / LMS algorithm for each order component has a convergence coefficient. 11. The active noise control method according to claim 10, wherein output signals of the respective order components are added from the ADF / LMS algorithm of the order components, and are sent as control signals to the actuator / speaker. An error signal between the control target sound from the power source and the control sound from the actuator speaker is estimated from the drive control signal of the power source by the amplitude phase estimation unit of the DSP,
The active noise control method according to claim 9 or 10, wherein an LMS algorithm obtains a filter coefficient of an adaptive filter that minimizes the error signal by using an error signal from the amplitude phase estimation unit. The basic (primary) frequency and the order component control target frequency of the control target sound are distributed from the signal processing circuit and input to the sound quality adjustment circuit for each order,
The sound quality adjustment circuit for the orders is added to the error signal from the amplitude phase estimation unit using an amplification value for adjusting the volume and sound quality for each order component of the order component control target frequency. The active noise control method described. ANC control is performed with an additional ADF / LMS algorithm for a signal of a constant frequency sound other than the order component of the fundamental frequency of the control target sound,
The active noise control method according to claim 9 or 13, wherein a filter coefficient is optimized and controlled by an adaptive filter for the constant frequency sound by multiplying an error signal from the amplitude phase estimation unit by a convergence coefficient.

Images