JP6250367B2 - Feed forward active noise control apparatus and method - Google Patents

Description

  The present invention relates to a feedforward active noise control apparatus and a feedforward active noise control method that actively reduce noise waves by using an adaptive filter.

  For example, when operating a device such as a vehicle or a construction machine, noise waves are generally generated from an engine, a fan, a hydraulic pump, or the like provided in the device. This noise wave causes discomfort to an operator (driver) who operates the device in a room for operating the device. Therefore, a noise reduction device that reduces the noise wave may be provided in the room.

  This noise reduction device is generally roughly classified into a passive noise control device and an active noise control device. A passive noise control device (Passive Noise Control; PNC) attenuates a noise wave by absorbing the noise wave with a sound absorbing material or the like that effectively absorbs sound, and reduces the noise wave. The active noise control device (Active Noise Control; ANC) generates a sound wave that has the same amplitude and opposite phase as the noise wave, attenuates the noise wave by interfering with the noise wave, and reduces the noise wave. To do. The passive noise control device can effectively reduce noise waves having a relatively high frequency, while the active noise control device can effectively reduce noise waves having a relatively low frequency. For this reason, both are often used in combination. The active noise control device is roughly divided into two methods, a feedforward type and a feedback type. One of the feedforward type active noise control devices is a typical algorithm, Filtered-X LMS (minimum average). There is an apparatus to which a square (least mean square) method is applied (see, for example, Patent Document 1).

  In this feedforward type active noise control device, a reference signal representative of a noise wave source (a signal correlated with a noise wave) is generally detected or estimated in advance, and this reference signal and a control point at which the noise wave is to be reduced are detected. Based on the residual wave signal (error signal) of the residual noise wave of the noise wave and the noise cancellation wave, the coefficient of the adaptive filter is updated so as to generate the noise cancellation wave having the same amplitude and opposite phase as the noise wave. A reference signal is filtered by the adaptive filter, and a silencer is generated by driving a silencer source with the filtered signal. In the noise control apparatus to which the Filtered-X LMS method is applied, the LMS method is used to determine the filter coefficient of the adaptive filter, and the propagation path from the silencer source that generates the silencer to the control point ( A reference signal filtered by a filter (secondary system filter) having a transfer function (transfer characteristic) in the secondary system) is used.

Japanese Patent Laid-Open No. 5-289777

  By the way, when a noise wave generation state changes in the noise wave source, a frequency component or intensity included in the noise wave changes, and as a result, when the noise wave changes, a noise control apparatus to which the Filtered-X LMS method is applied. Follows the change of the noise wave by sequentially updating the coefficient of the adaptive filter by the LMS method so as to minimize the error signal (residual wave signal) which becomes a large value immediately after the change of the noise wave. I will do it. Since the noise wave is not effectively reduced during the following time (convergence time of the LMS method), it is desirable that the following time is short.

  This reduction in follow-up time is realized by, for example, using hardware with high processing capability for information processing of the LMS method, or by reducing the amount of information processing of the LMS method, such as shortening the filter length of the adaptive filter. it can. However, the hardware with high processing capability is expensive. Further, the reduction of the filter length may increase the reverberation (noise wave that could not be silenced by the silencer).

  The present invention has been made in view of the above-described circumstances, and its purpose is to increase the tracking time even when hardware with the same processing capability and the same filter length are used for information processing of the LMS method. To provide a feedforward active noise control device and a feedforward active noise control method that can be shortened.

As a result of various studies, the present inventor has found that the above object is achieved by the present invention described below. That is, the feedforward active noise control device according to one aspect of the present invention detects a noise wave generation state of a noise wave source, and generates a reference signal related to the noise wave based on the detected noise wave generation state. A reference signal generator, a silencer generator for generating a silencer corresponding to the control signal, and a residual wave signal by detecting a residual wave of the noise wave and the silencer remaining at a predetermined control point. Based on the residual wave signal, a filtered reference signal obtained by filtering the reference signal with a transfer characteristic from the silencer output point of the silencer to the control point, and the residual wave signal. A filter coefficient is determined so as to generate the muffler having the same amplitude as the wave and having an opposite phase, and the control signal is generated by filtering the reference signal with the calculated filter coefficient. And a data portion, the reference signal generator comprises a sine wave having a fundamental frequency of the noise wave on the basis of the noise wave generation state, one or more higher-order sine with the frequency of high-order with respect to the fundamental frequency The reference signal is generated by superimposing waves on each other, and the silencer control adaptive filter unit uses the fundamental period of the reference signal or the fundamental as the filtered reference signal and the residual wave signal for obtaining the filter coefficient. A filtered reference signal and a residual wave signal extracted so that a period that is an integral multiple of the period matches the filter length are used.

  In such a feedforward type active noise control apparatus, the fundamental period of the reference signal or an integer multiple of the fundamental period coincides with the filter length as the filtered reference signal and residual wave signal for obtaining the filter coefficient of the adaptive filter. The filtered reference signal and the residual wave signal taken out in (1) are used. If the filtered reference signal and the residual wave signal are extracted without considering the filter length, each of the extracted filtered reference signal and residual wave signal includes a waveform that is less than one period. It becomes discontinuous and includes not only the fundamental frequency and its integral multiple frequency components but also other frequency components. As a result, in the calculation of the LMS algorithm, the convergence calculation is executed including the other frequency components that are not originally required, and the convergence time becomes long. However, when each of the filtered reference signal and the residual wave signal is extracted as described above, each of the extracted filtered reference signal and residual wave signal does not include a waveform that is less than one cycle, and one cycle. Or only an integral multiple of the waveform is included, and a continuous waveform is obtained. As a result, the convergence calculation is not unnecessarily lengthened and is shortened. Therefore, such a feedforward type active noise control device can further reduce the follow-up time even when hardware having the same processing capability and the same filter length are used for information processing of the LMS method.

  According to another aspect, in the above-described feedforward active noise control device, the silencer control adaptive filter unit further includes, as the reference signal to be filtered, a fundamental period of the reference signal or an integer of the fundamental period. A reference signal extracted so that the double period matches the filter length is used.

  In such a feedforward type active noise control device, a reference signal extracted so that the fundamental period of the reference signal or an integer multiple of the fundamental period matches the filter length is used as the reference signal to be filtered by the adaptive filter. . As described above, the filter coefficient is obtained based on the filtered reference signal and the residual wave signal that are extracted so that the fundamental period of the reference signal or an integer multiple of the fundamental period matches the filter length. The filter coefficient obtained in this way and the reference signal extracted in consideration of the filter length have the same length. Therefore, such a feedforward type active noise control apparatus can accurately filter the reference signal extracted in consideration of the filter length with the filter coefficient thus obtained.

  According to another aspect, in the above-described feedforward active noise control device, the silencer control adaptive filter unit filters the reference signal with a transfer characteristic from the silencer output point of the silencer to the control point. And the reference signal, the filtered reference signal, and the residual signal, each of which has a filter length of a basic period of the reference signal or an integer multiple of the basic period A sampling unit that samples each of the noise wave and the filtered reference signal and the residual wave signal sampled by the sampling unit, respectively, and generates the muffler that has the same amplitude and opposite phase as the noise wave. The filter coefficient update unit for obtaining the filter coefficient and the filter coefficient obtained by the filter coefficient update unit Filtering said reference signal sampled at the ring portion, characterized in that it comprises an adaptive filter configured to generate the control signal.

  Such a feedforward type active noise control device uses a relatively simple method of adjusting a sampling unit to convert a reference signal, a filtered reference signal, and a residual wave signal to a fundamental period of the reference signal or an integer of the fundamental period. Double the period can match the filter length.

  According to another aspect, in the feedforward active noise control device described above, the noise wave source is an engine, the noise wave generation state is an engine rotation speed, and the reference signal generation unit is A rotational speed of the engine is detected, and a sine wave having a frequency corresponding to the detected rotational speed as a fundamental frequency and one or more high-order sine waves having a higher order frequency with respect to the fundamental frequency are superimposed on each other. The reference signal is generated, and the sampling unit of the silencer control adaptive filter unit changes the sampling rate in accordance with the rotational speed of the engine, whereby the reference signal, the filtered reference signal, and the residual wave signal respectively Are respectively sampled so that the fundamental period of the reference signal or an integral multiple of the fundamental period matches the filter length. And features.

  Such a feedforward type active noise control device uses a relatively simple method of changing the sampling rate in accordance with the rotational speed of the engine, and converts each of the reference signal, the filtered reference signal, and the residual wave signal to the reference signal. The fundamental period or an integral multiple of the fundamental period can be made to match the filter length.

  According to another aspect, in the feedforward active noise control apparatus described above, the sampling unit of the silencer control adaptive filter unit is configured so that the engine is divided by a filter length equal to one rotation divided equally by a filter length. The sampling rate is changed so that the reference signal, the filtered reference signal, and the residual wave signal are sampled each time the rotation is performed.

  Such a feedforward type active noise control device has the reference signal, the filtered reference signal, and the residual wave signal each time the engine rotates by a filter length equally divided by one filter length. Each of the reference signal, the filtered reference signal, and the residual wave signal is set to a fundamental period of the reference signal or an integral multiple of the fundamental period by a relatively simple method of changing the sampling rate to sample each. The period can be made to match the filter length.

According to another aspect of the present invention, a feedforward active noise control method detects a noise wave generation state of a noise wave source, and generates a reference signal related to the noise wave based on the detected noise wave generation state. A reference signal generation step to generate, a silencer generation step to generate a silencer corresponding to the control signal, and a residual by detecting a residual wave that remains between the noise wave and the silencer at a predetermined control point A residual wave detection step of outputting a wave signal, a filtered reference signal obtained by filtering the reference signal with a transfer characteristic from the silencer output point of the silencer to the control point, and the residual wave signal, A filter coefficient is obtained so as to generate the muffler having the same amplitude and opposite phase as the noise wave, and the control signal is generated by filtering the reference signal with the obtained filter coefficient. And a filter process, the reference signal generating step includes a sine wave having a fundamental frequency of the noise wave on the basis of the noise wave generation state, one or more higher-order sine with the frequency of high-order with respect to the fundamental frequency The reference signal is generated by superimposing waves on each other, and the silencer control adaptive filter step uses the fundamental period of the reference signal or the fundamental as the filtered reference signal and the residual wave signal for obtaining the filter coefficient. A filtered reference signal and a residual wave signal extracted so that a period that is an integral multiple of the period matches the filter length are used.

  In such a feedforward type active noise control method, the fundamental period of the reference signal or an integer multiple of the fundamental period matches the filter length as the filtered reference signal and residual wave signal for obtaining the filter coefficient of the adaptive filter. The filtered reference signal and the residual wave signal taken out in (1) are used. For this reason, each of the extracted filtered reference signal and residual wave signal does not include a waveform that is less than one period, and includes only a waveform of one period or an integral multiple thereof, thereby forming a continuous waveform. As a result, the convergence calculation is not unnecessarily lengthened and is shortened. Therefore, such a feedforward type active noise control device can further reduce the follow-up time even when hardware having the same processing capability and the same filter length are used for information processing of the LMS method.

  The feedforward type active noise control apparatus and the feedforward type active noise control method according to the present invention further increase the follow-up time even when hardware having the same processing capability and the same filter length are used for information processing of the LMS method. Can be shortened.

It is a side view showing a hydraulic excavator in an embodiment. FIG. 2 is a diagram for explaining arrangement positions of a sound deadening output unit and a residual wave input unit in a cab chamber of the excavator shown in FIG. 1. It is a figure which shows the structure of the feedforward type | mold active noise control apparatus in embodiment. It is a flowchart which shows the identification operation | movement of the secondary system filter part in the feedforward type active noise control apparatus of embodiment. It is a flowchart which shows operation | movement of the feedforward type active noise control apparatus in embodiment. It is a flowchart which shows the calculation operation | movement of the reference signal in the feedforward type active noise control apparatus of embodiment. It is a figure for demonstrating the follow-up time (convergence time of a LMS method) in the feedforward type active noise control apparatus of an Example. It is a figure for demonstrating the tracking time (convergence time of a LMS method) in the feedforward type active noise control apparatus of a comparative example.

  Hereinafter, an embodiment according to the present invention will be described with reference to the drawings. In addition, the structure which attached | subjected the same code | symbol in each figure shows that it is the same structure, The description is abbreviate | omitted suitably. Further, in this specification, when referring generically, it is indicated by a reference symbol without a suffix, and when referring to an individual configuration, it is indicated by a reference symbol with a suffix.

  The feedforward-type active noise control device in the embodiment, when operating a device such as a vehicle or a construction machine (construction machine), for example, generates noise waves generated from an engine, a fan, a hydraulic pump, and the like provided in the device. This is a device for reducing in the operation room for operating the. The feedforward active noise control device in the embodiment may be mounted on a vehicle such as a train or an automobile, but in the present embodiment, as an example, the feedforward type active noise control device is mounted on a hydraulic excavator that is one of construction machines. explain.

  FIG. 1 is a side view showing a hydraulic excavator in the embodiment. FIG. 2 is a diagram for explaining the arrangement positions of the sound deadening output unit and the residual wave input unit in the cab chamber of the excavator shown in FIG.

  A hydraulic excavator HS equipped with the feedforward type active noise control device ANC according to the embodiment includes a lower traveling body 101 having a left traveling crawler 101L and a right traveling crawler, and an upper frame provided on the lower traveling body 101 so as to be able to turn. 102, a machine room 103 provided on the upper frame 102, in which an engine (not shown) and the like are disposed, a work attachment 104 provided on the upper frame 102 so as to be able to undulate, and a work attachment 104. An operation member 106 such as a lever for operation, a cab chamber (operation chamber) 105 provided with the operation member 106, and a feedforward type active noise control device ANC described later are provided. In the cab chamber 105, the silencer output unit 21 and the residual wave input unit 31 of the feedforward type active noise control device ANC are respectively arranged at predetermined positions shown in FIG. 2, as will be described later.

The work attachment 4 includes a boom 141, an arm 142 connected to the distal end portion of the boom 141, and a bucket 143 that is swingably attached to the distal end portion of the arm 142. The boom 141 is raised and lowered with respect to the upper frame 102 by the expansion and contraction operation of the boom cylinder 144a. Arm 14 2 is swung relative to the boom 141 by expansion and contraction of the arm cylinder 144b. The bucket 143 swings with respect to the arm 142 by the expansion / contraction operation of the bucket cylinder 144c. Each of the boom cylinder 144a, the arm cylinder 144b, and the bucket cylinder 144c is a hydraulic hydraulic cylinder.

  Next, the structure of the feedforward type active noise control apparatus in the embodiment will be described. FIG. 3 is a diagram illustrating a configuration of the feedforward active noise control apparatus according to the embodiment.

The feed-forward type active noise control device generates a sound wave having the same amplitude and opposite phase as the noise wave whose size is to be reduced, and the noise wave is attenuated by interfering the sound wave with the noise wave. It is to reduce. More specifically, the feedforward active noise control apparatus according to the present embodiment generates a reference wave related to a noise wave based on a noise wave generation state of a noise wave source, and a predetermined control point for reducing the noise wave. The residual wave (residual wave) between the noise wave and the silenced sound wave at the center is detected, and the detected residual wave and the reference wave at the control point (reference wave after passing through the secondary system, filtered reference wave) And generating a control signal by filtering the reference wave while adaptively changing a filter characteristic so as to generate the muffler having the same amplitude and the opposite phase as the noise wave. It is an apparatus that generates the silencer according to a signal. Thereby, the feedforward type active noise control device can reduce the noise wave at the control point by causing the noise wave to interfere with the noise wave. The feedforward-type active noise control device according to the present embodiment generates a reference wave electrical signal (reference signal) and generates a sine wave having a fundamental frequency of the noise wave based on the noise wave generation state. And one or a plurality of higher-order sine waves having higher-order frequencies with respect to the fundamental frequency are superimposed on each other, and when the filter characteristic is adaptively changed, An electric signal of the filtered reference wave (filtered reference signal) and an electric signal of the residual wave (residual wave signal) so that a fundamental period or a period that is an integral multiple of the fundamental period matches the filter length. To take out. As a result, the feedforward type active noise control apparatus can adaptively change the filter characteristics in just one period or a plurality of periods, so that the tracking time (LMS method convergence time) can be shortened.

  For example, as shown in FIG. 3, the feedforward active noise control device ANC according to this embodiment includes a reference signal generation unit 1, a silencer generation unit 2, a residual wave detection unit 3, and a silencer. And a control adaptive filter unit 4.

The reference signal generation unit 1 is a device that detects a noise wave generation state of the noise wave source NS that generates a noise wave, and generates a reference signal related to the noise wave based on the detected noise wave generation state. More specifically, in the present embodiment, the reference signal generation unit 1 detects a noise wave generation state of the noise wave source NS, and based on the detected noise wave generation state, a sine wave having a fundamental frequency of the noise wave, and a basic The reference signal is generated by superimposing one or more higher-order sine waves having higher-order frequencies with respect to frequencies. The reference signal generation unit 1 is connected to the silencer control adaptive filter unit 4 and outputs the reference signal to the silencer control adaptive filter unit 4. The noise wave source NS may be any object that generates a noise wave. In this specification, the noise wave includes not only a sound wave but also a vibration wave. In the present embodiment, for example, the noise wave source NS is an engine mounted on the hydraulic excavator HS. The noise wave generation state is a state of the noise wave source NS that causes the generation of the noise wave. When the noise wave generation state changes, the frequency component and intensity included in the noise wave change, and the noise wave also changes. This noise wave generation state is a parameter that changes the profile of the noise wave.

  In the present embodiment, such a reference signal generation unit 1 includes, for example, a noise wave generation state detection unit 11, an analog / digital conversion unit (hereinafter abbreviated as “ADC unit”) 12, a fundamental frequency detection unit 13, and the like. And a reference signal synthesis unit 14.

The noise wave generation state detection unit 11 is a device that detects the noise wave generation state of the noise wave source NS. For the noise wave generation state detection unit 11, an appropriate device is selected according to the noise wave generation state to be detected. The noise wave generation state detection unit 11 is connected to the ADC unit 12 and outputs the detected noise wave generation state signal (noise wave generation state signal) to the ADC unit 12. More specifically, in the present embodiment, the noise wave source NS is, for example, an engine of a hydraulic excavator HS, and the noise wave is generated by the rotation of the engine. Therefore, the noise wave generation state detection unit 11 generates the noise wave. This is a device that detects the rotational speed of the engine (the rotational speed per unit time, usually simply called the rotational speed) as a state, and outputs an engine rotational signal representing the detected rotational speed. Such a noise wave generation state detection unit 11 includes, for example, a rotary encoder. The ADC unit 12 is a circuit that converts the noise wave generation state signal input from the noise wave generation state detection unit 11 from an analog signal to a digital signal. The ADC unit 12 is connected to the fundamental frequency detection unit 13, and outputs the converted noise wave generation state signal of the digital signal to the fundamental frequency detection unit 13. The fundamental frequency detection unit 13 is a circuit that detects the fundamental frequency of the noise wave based on the noise wave generation state signal input from the ADC unit 12. The fundamental frequency detector 13 is connected to the reference signal synthesizer 14 and outputs the detected fundamental frequency to the reference signal synthesizer 14. The reference signal synthesis unit 14 is a circuit that generates a reference signal based on the fundamental frequency input from the fundamental frequency detection unit 13. More specifically, in the present embodiment, the reference signal synthesis unit 14 generates a sine wave having a fundamental frequency (a sine wave of a fundamental wave component), and uses this fundamental frequency to obtain a higher-order frequency relative to the fundamental frequency. One or a plurality of higher-order sine waves (sine waves of higher-order wave components) are generated, and the sine waves of the fundamental wave components and the sine waves of one or more higher-order wave components are aligned to overlap each other. To synthesize. As a result, the reference signal synthesis unit 14 generates a reference signal superimposed with the sine wave of the fundamental wave component and the sine wave of one or more higher-order wave components. The reference signal synthesis unit 14 outputs the generated reference signal to the silencer control adaptive filter unit 4.

  The silencer generating unit 2 is an apparatus that is connected to the silencer control adaptive filter unit 4 and generates and outputs a silencer corresponding to the control signal input from the silencer control adaptive filter unit 4. In the present embodiment, the silencer generation unit 2 includes, for example, a silencer output unit 21, a drive unit 22, and a digital / analog conversion unit (hereinafter abbreviated as “DAC unit”) 23.

  The DAC unit 23 is connected to a later-described second filter unit 41-2 in the silencer control adaptive filter unit 4, and receives a control signal input from the second filter unit 41-2 of the silencer control adaptive filter unit 4 as a digital signal. Is a circuit for converting from analog signals to analog signals. The DAC unit 23 is connected to the drive unit 22, and outputs the converted analog signal control signal to the drive unit 22. The drive unit 22 is a circuit that is connected to the silencer output unit 21 and drives the silencer output unit 21 by a control signal input from the DAC unit 23, and includes, for example, a power amplifier. The silencer output unit 21 is a device that is driven by the drive unit 22 to generate and output a silencer corresponding to the control signal, and includes, for example, a speaker. The silencer output unit 21 is disposed at a suitable position in the cab chamber 105 so that the silencer can be radiated to the space in the cab chamber 105 toward the control point CP where it is desired to reduce the noise wave. Preferably, the sound deadening output unit 21 is disposed in the vicinity of the control point CP. The control point CP is appropriately set in advance according to the place where the noise wave is desired to be reduced. For example, in the case of the hydraulic excavator HS described above, the control point CP is a location where the head of the operator who operates the hydraulic excavator HS is supposed to be located. .

  The residual wave detection unit 3 is a device that detects a residual wave (residual wave) remaining after the interference between the noise wave and the silencer at the control point CP and outputs a residual wave signal. The residual wave detection unit 3 is connected to the silencer control adaptive filter unit 4 and outputs a residual wave signal to the silencer control adaptive filter unit 4. The residual wave detection unit 3 includes, for example, a residual wave input unit 31 and an ADC unit 32 in the present embodiment.

  The residual wave input unit 31 is a device that outputs a residual wave signal by detecting the residual wave of the non-erasing residual wave at the control point CP, and includes a microphone, for example. The residual wave (error wave) is transmitted to the noise wave propagating through the primary system (propagation path of the primary system) from the noise wave source NS to the control point CP at the control point CP. This is a noise wave that is caused to interfere with the extinguished sound wave that has propagated through the secondary system (secondary system propagation path) from the unit 21 to the control point CP, and as a result remains due to the interference of the extinguishing sound. The residual wave input unit 31 is arranged at an appropriate position in the cab chamber 105 so as to pick up an unerased residual wave of the control point CP. Preferably, the residual wave input unit 31 is disposed in the vicinity of the control point CP. The residual wave input unit 31 is connected to the ADC unit 32 and outputs a residual wave signal to the ADC unit 32. The ADC unit 32 is a circuit that converts the residual wave signal input from the residual wave input unit 31 from an analog signal to a digital signal. The ADC unit 32 is connected to a later-described third filter unit 41-3 in the silencer control adaptive filter unit 4, and outputs the converted residual signal of the digital signal to the silencer control adaptive filter unit 4.

  The silencer control adaptive filter unit 4 obtains a filter coefficient based on the filtered reference signal and the residual wave signal so as to generate a silenced wave having the same amplitude and opposite phase as the noise wave, and has the obtained filter coefficient. A circuit that generates the control signal by filtering a reference signal with an adaptive filter unit. The filtered reference signal is referred to by a transfer characteristic (secondary system transfer characteristic (transfer function)) from the silencer output point of the silencer output from the silencer output unit 21 of the silencer generation unit 2 to the control point CP. A signal obtained by filtering the signal. In this embodiment, the silencer control adaptive filter unit 4 uses the fundamental period of the reference signal or an integer multiple of the fundamental period to match the filter length as the filtered reference signal and residual wave signal for obtaining the filter coefficient. The filtered reference signal and the residual wave signal thus extracted are used. Furthermore, in this embodiment, the silencer control adaptive filter unit 4 uses, as the reference signal to be filtered, a reference signal extracted so that the basic period of the reference signal or an integer multiple of the basic period matches the filter length. .

  In the present embodiment, for example, the silencer control adaptive filter unit 4 includes the first to third filter units 41-1 to 41-3, the secondary filter unit 42, and the first to third sampling rates. Conversion units (hereinafter abbreviated as “SRC units”) 43-1 to 43-3, a filter coefficient update unit 44, and an adaptive filter unit 45 are provided.

  The first to third filter units 41-1 to 41-3 are filter circuits for removing so-called noise components unnecessary for the feedforward active noise control device ANC. The first filter unit 41-1 is connected to the reference signal synthesis unit 14 of the reference signal generation unit 1 and removes a noise component from the output of the reference signal synthesis unit 14. The first filter unit 41-1 is connected to each of the first SRC unit 43-1 and the secondary system filter unit 42, and outputs the output (reference signal) of the reference signal synthesis unit 14 from which the noise component has been removed to the first SRC unit 43-1. And the secondary filter unit 42 respectively. The second filter unit 41-2 is connected to the adaptive filter unit 45 and removes a noise component from the output of the adaptive filter unit. The second filter unit 41-2 is connected to the DAC unit 23 of the silencer generation unit 2, and outputs the output (control signal) of the adaptive filter unit 45 from which the noise component has been removed to the DAC unit 23 of the silencer generation unit 2. . The third filter unit 41-3 is connected to the ADC unit 32 of the residual wave detection unit 3 and removes noise components from the output of the ADC unit 32. The 3rd filter part 41-3 is connected to the 3rd SRC part 43-3, and outputs the output (residual wave signal) of the ADC part 32 which removed the noise component to the 3rd SRC part 43-3. Each of the first to third filter units 41-1 to 41-3 includes a low-pass filter, a high-pass filter, or a band-pass filter as appropriate.

  The secondary filter unit 42 filters the reference signal input from the reference signal synthesis unit 14 of the reference signal generation unit 1 via the first filter unit 41-1 with the transfer characteristic of the secondary system by filtering. A circuit for generating a reference signal. The transfer characteristic (transfer function) of the secondary system is obtained in advance as described later, and is set in the secondary filter unit 42. The secondary filter unit 42 is connected to the second SRC unit 43-2, and outputs the generated filtered reference signal to the second SRC unit 43-2.

  The first SRC unit 43-1 changes the sampling rate based on the noise wave generation state, thereby the reference input from the reference signal synthesis unit 14 of the reference signal generation unit 1 via the first filter unit 41-1. It is a circuit that samples a signal so that the basic period of the reference signal or an integer multiple of the basic period matches the filter length of the adaptive filter unit 45. The first SRC unit 43-1 is connected to the adaptive filter unit 45 and outputs the sampled reference signal to the adaptive filter unit 45. The second SRC unit 43-2 changes the sampling rate based on the noise wave generation state to change the filtered reference signal input from the secondary system filter unit 42 into the fundamental period of the reference signal or the fundamental period. This is a circuit that performs sampling so that the integer multiple period matches the filter length of the adaptive filter unit 45. The second SRC unit 43-2 is connected to the filter coefficient update unit 44 and outputs the sampled filtered reference signal to the filter coefficient update unit 44. The third SRC unit 43-3 changes the sampling rate based on the noise wave generation state, and thereby the residual input from the ADC unit 32 of the residual wave detection unit 3 via the third filter unit 41-3. This is a circuit that samples a wave signal so that a fundamental period of the reference signal or a period that is an integral multiple of the fundamental period matches the filter length of the adaptive filter unit 45. The third SRC unit 43-3 is connected to the filter coefficient updating unit 44 and outputs the sampled residual wave signal to the filter coefficient updating unit 44.

  In the present embodiment, as described above, since the engine speed is used as the noise wave generation state, each of the first to third SRC units 43-1 to 43-3 is detected by the reference signal generation unit 1. The sampling rate is changed according to the engine speed.

  The first to second SRC units 43-1 to 43-3 are configured such that the reference signal, the filtered reference signal, and the residual wave signal have a basic period of the reference signal or a period that is an integral multiple of the basic period matches the filter length. Corresponds to an example of a sampling unit for sampling each of the above.

  The filter coefficient updating unit 44 adjusts the noise wave based on the filtered reference signal and the residual wave signal which are adjusted and matched to match the filter length by the second and third SRC units 43-2 and 43-3, respectively. This is a circuit for obtaining a filter coefficient of the adaptive filter unit 45 so as to generate a silencer having the same amplitude and an opposite phase. For the calculation of the filter coefficient, a known adaptive algorithm, for example, an LMS (Least-mean-square) algorithm (least mean square algorithm) is used. The filter coefficient updating unit 44 is connected to the adaptive filter unit 45, outputs the obtained filter coefficient to the adaptive filter unit 45, and sets the filter coefficient in the adaptive filter unit 45.

  The adaptive filter unit 45 generates a control signal by filtering the reference signal that has been adjusted and matched with the filter length by the first SRC unit 43-1 with the filter coefficient obtained by the filter coefficient updating unit 44. Circuit. The adaptive filter unit 45 includes, for example, a FIR (finite impulse response) filter (finite impulse response filter), an IIR (infinite impulse response) filter (infinite impulse response filter), and the like. The adaptive filter unit 45 is connected to the DAC unit 23 of the silencer generating unit 2 via the second filter unit 41-2, and the generated control signal is sent to the silencer generating unit 2 via the second filter unit 41-2. Are output to the DAC section 23 of

Here, for example, in the case of the FIR filter, if the input signal is x (n), the output signal is y (n), and the filter coefficient of the i item is h _i (i = 0 to N), the FIR filter The input / output relationship is expressed by the following equation 1 consisting of (N + 1) terms. The filter length is the number of terms in the following expression 1 representing the above input / output relationship, and is (N + 1).
y (n) = h ₀ x (n) + h ₁ x (n−1) +... + h _N x (n−N) (1)

Filter coefficient updating unit 44, when the adaptive filter 45 is an FIR filter is determined by the LMS algorithm based on the filter coefficient h _i in said filtered reference signal and the residual wave signal. In this LMS algorithm, the vector of the j-th filter coefficient in the iterative convergence calculation is set as H _j (H _j = (h _{0, j} , h _{1, j} ,..., H _{N, j} ; h _{i, j} is filter coefficients of i item in the j-th)), the step size and mu, the j-th residual wave signal and e _j, and, when a vector containing the j-th reference signal and x _j, table by the following equation 2 Is done. The step size μ is a parameter for controlling the amount of change of the filter coefficient in the repeated convergence calculation.
H _{j + 1} = H _j + μe _j x _j (2)

  Next, the operation of the feedforward active noise control device ANC in this embodiment will be described. FIG. 4 is a flowchart illustrating the identification operation of the secondary filter unit in the feedforward active noise control device of the embodiment. FIG. 5 is a flowchart showing the operation of the feedforward active noise control apparatus in the embodiment. FIG. 6 is a flowchart illustrating a reference signal calculation operation in the feedforward active noise control apparatus according to the embodiment.

First, the identification operation of the secondary filter unit 42 will be described. In this identification operation, a predetermined identification wave is output from the silencer output unit 21 that is a secondary system wave source, and the square error between the identification wave and the identification wave propagated through the secondary system is minimized. This is done by determining the filter coefficients by the LMS algorithm. More specifically, in FIG. 4 , white noise is radiated from the silencer output unit 21 toward the control point CP, and the white noise propagated through the secondary system is transmitted by the residual wave input unit 31 in the vicinity of the control point CP. Received (S1). The filter coefficients are sequentially updated by the LMS algorithm based on the white noise radiated from the muffler output unit 21 and the white noise received by the residual wave input unit 31 (white noise after secondary system propagation). The transfer characteristic of the secondary system is identified (S2). As a result, the filter coefficient of the secondary filter unit 42 is obtained in advance and set in the secondary filter unit 42.

  After performing the identification operation of the secondary system filter unit 42 as described above, the feedforward active noise control device in the embodiment reduces noise waves by operating as follows.

In FIG. 5 , first, the reference signal generation unit 1 changes the noise wave generation state (engine rotation state) by the noise wave generation state detection unit 11 (rotary encoder) from the noise wave source NS (engine in the above example, the same applies hereinafter). The fundamental frequency is detected by the fundamental frequency detection unit 13 based on a noise wave generation state signal (a pulse signal described later) via the ADC unit 12, and is referenced by the reference signal synthesis unit 14 based on the detected fundamental frequency. Generate a signal. In the present embodiment, the silencer control adaptive filter unit 4 obtains the rotational position of the engine based on the noise wave generation state (engine rotational state) detected by the reference signal generation unit 1 (S11).

  This process S11 will be described more specifically with reference to FIG. In the above example, the ADC unit 12 samples the output of the rotary encoder at a predetermined sampling frequency, for example, 4 kHz, converts the sampling result into a digital signal, and outputs the digital signal to the basic frequency detection unit 13 to detect the basic frequency. The unit 13 samples the output of the ADC unit 12 at a sampling period of 4 kHz so as to obtain one pulse signal by one rotation of the engine (shaft) (S21).

  Next, the fundamental frequency detector 13 determines the presence / absence of the pulse signal by determining whether the output of the ADC 12 is equal to or greater than a preset threshold Thp (S22). More specifically, when the output of the ADC unit 12 is less than the threshold value Thp, the basic frequency detection unit 13 determines that there is no pulse signal (No), and executes a later-described process S25 while detecting the basic frequency. When the output of the ADC unit 12 is equal to or greater than the threshold Thp, the unit 13 determines that the pulse signal is present (Yes), and executes the next process S23.

  In this process S23, the fundamental frequency detector 13 detects the fundamental frequency f of the noise wave from the interval between pulse signals that are temporally adjacent to each other, and stores the detected fundamental frequency f. More specifically, the fundamental frequency detection unit 13 obtains the number of revolutions (rotation speed) of the engine per second from the time interval from the determination that there is a pulse signal to the next determination that there is a pulse signal, and this is determined as noise. This is stored as the fundamental frequency f of the wave. The time interval is determined based on the number of samples (sampling number) and the divided sampling frequency (sampling frequency of the basic frequency detection unit 13) between the determination of the presence of a pulse signal and the next determination of the presence of the pulse signal. Desired.

Next, the reference signal synthesis unit 14 generates a reference signal as follows. First, in process S24, the reference signal synthesizing unit 14 obtains an engine rotation angle displacement amount Δa for each sampling. This rotation angle displacement amount Δa is an angle at which the engine rotates from sampling until the next sampling (an angle at which the engine rotates during one sampling interval). If the sampling frequency is fs, the fundamental frequency f Is expressed by the following equation 3.
Δa = 2π / fs / f (3)
In the present embodiment, the sampling frequency is 4 kHz (fs = 4 kHz), and Equation 3 is Δa = 2π / 4000 / f.

An engine noise wave is generated by an explosion inside the engine. The noise wave includes a fundamental wave component of the fundamental frequency, a harmonic component thereof, that is, a higher-order wave having a higher frequency than the fundamental frequency. Ingredients included. For this reason, the reference signal synthesizing unit 14 obtains and stores the phase fluctuation amount of the sine wave in a predetermined harmonic component (high-order wave component). phase deviation Delta] p _n sine wave at n harmonic components (n order wave component) is represented by the rotation angle displacement amount Δa of the fundamental wave component by the following equation 4. The higher-order wave component included in the reference wave (reference signal) is appropriately set according to the noise wave reduction target engine for which the noise wave is desired to be reduced. For example, by measuring each noise wave in a plurality of engines of the same model number and obtaining a frequency spectrum of each noise, a higher order component that is predominantly included in the noise wave is included in the reference wave (reference signal). Set as a component. For example, in the present embodiment, as the higher-order wave components included in the reference wave (reference signal), the fourth harmonic (fourth-order wave component), the sixth harmonic (the sixth-order wave component), the seventh harmonic (the seventh-order wave component), 8 Overtones (8th order wave components), 9th overtones (9th order wave components) and 10th overtones (10th order wave components) are obtained.
Δp _n = n × Δa (4)

Next, in the process S25, first, the reference signal synthesis unit 14 updates the current engine rotation angle a by the following equation 5, and similarly updates the current phase in the sine wave of each overtone by the following equation 6. .
a = a + Δa (5)
p _n = p _n + Δp _n (6)
_{(A ← a + Δa; p} n ← p n + Δp n)

Next, the reference signal synthesis unit 14 generates a reference signal superimposed with the sine wave of the fundamental wave component and the sine wave of one or more higher-order wave components by using these. This reference signal x is expressed by the following equation 7 in this example.
x = A ₁ × sin (a)
+ A ₄ × sin (p ₄ )
+ A ₆ × sin (p ₆ )
+ A ₇ × sin (p ₇ )
+ A ₈ × sin (p ₈ )
+ A ₉ × sin (p ₉ )
+ A ₁₀ × sin (p ₁₀ ) (7)
_An is the amplitude of the nth-order wave component (where n = 1 is the fundamental wave component). Since the reference signal x is filtered by the adaptive filter unit 45, these amplitudes _An may be the same arbitrary numerical values, or may be arbitrarily different arbitrary numerical values. For example, these amplitudes _An are all set to 0.5 or 1. Further, for example, from the viewpoint of shortening the convergence time of the LMS and the viewpoint of converging the filter coefficient with an appropriate value, these amplitudes _An are set when the higher-order wave component included in the reference wave (reference signal) is set. You may set so that it may become each component ratio in the frequency spectrum of the measured noise wave.

Then, the first to third SRC units 43-1 to 43-3 in the silencer control adaptive filter unit 4 obtain the current rotation position Q from the current engine rotation angle a and the filter length (N + 1) by the following equation (8). . The calculation of the current rotational position Q of the engine may be performed by each of the SRC units 43-1 to 43-3, or one of the first to third SRC units 43-1 to 43-3. Any one of them, for example, the first SRC unit 43-1 may be representatively implemented. As will be described later, this current rotational position Q is used to sample each signal so that the fundamental period of the signal or an integer multiple of the fundamental period matches the filter length of the adaptive filter unit 45. This information is necessary for each SRC unit 43-1 to 43-3.
Q = int ((a / 2π) × (N + 1)) (8)
Here, int (B) is an operator for obtaining an integer value not exceeding B. The filter length is related to the cutoff characteristic in the filter characteristics of the adaptive filter unit 45, the convergence time of the LMS, and the like. If the filter length is increased, the adaptive filter unit 45 has a steeper filter characteristic, and the convergence time of the LMS. Becomes longer. From these viewpoints, the filter length may be set as appropriate, but is set to 180, for example, in the present embodiment. Therefore, in this embodiment, one rotation of the engine is equally divided into 180, and the position of the engine is monitored at every 4 kHz interruption.

  Returning to FIG. 5, in step S <b> 12, the reference signal x generated by the reference signal generation unit 1 as described above is input to the first filter unit 41-1 of the silencer control adaptive filter unit 4, and the first filter The noise component is removed by the unit 41-1 and output to the first SRC unit 43-1 and the secondary filter unit 42, respectively. On the other hand, the residual wave signal e detected and generated by the residual wave detection unit 3 is input to the third filter unit 41-3 of the silencer control adaptive filter unit 4, and noise is generated by the third filter unit 41-3. The component is removed and output to the third SRC unit 43-3.

  Next, the secondary system filter unit 42 filters the reference signal x with the secondary system transfer characteristic identified as described above to generate the filtered reference signal r. That is, the secondary filter unit 42 performs a convolution operation (convolution) of a vector (reference signal vector) X representing the reference signal x and a vector (secondary system transfer characteristic vector) C representing the transfer characteristic of the secondary system. Thus, a vector representing the filtered reference signal r (filtered reference signal vector R) is obtained (S13).

  As described above in the processing S11 described with reference to FIG. 6, in this embodiment, one rotation of the engine is equally divided by the filter length of 180, and the engine position is monitored every 4 kHz interruption. Accordingly, each time the engine rotates by a fraction of 180 of the filter length, the values of the reference signal x, the filtered reference signal r, and the residual signal e are updated, that is, the engine is divided by a fraction of the filter length. If the sampling rates of the first to third SRC units 43-1 to 43-3 are changed so that the reference signal x, the filtered reference signal r, and the residual wave signal e are sampled each time the signal is rotated, Each of the reference signal x, the filtered reference signal r, and the residual wave signal e can be sampled such that the fundamental period of the signal or an integer multiple of the fundamental period matches the filter length of the adaptive filter unit 45. For this reason, after the processing S13, first, the first to third SRC units 43-1 to 43-3 have the rotational position Q (Qt) obtained by the current interruption determined by the previous interruption. It is determined whether or not the same as Qb) (S14). This determination may be performed individually by each of the SRC units 43-1 to 43-3, or may be performed by, for example, the first SRC unit 43-1 as a representative.

  If the result of this determination is the same (Qt = Qb, Yes), the process S15 is performed, and if different (Qt ≠ Qb, No), the process S16 is performed.

  In the process S15, when the above is the same (Qt = Qb, Yes), the engine speed is low, and after the interrupt process is performed until the next interrupt process is performed (during the interrupt process execution interval). , During the interruption interval), it can be determined that the engine has not yet rotated by an angle equal to 180 (the int function is used in Equation 8), so the first to third SRC units 43-1 to 43- 3 requires that the sampling rate be slower than the current rate. Therefore, each of the first to third SRC units 43-1 to 43-3 performs processing as follows.

  The first SRC unit 43-1 includes the reference signal x (xt) in the current interrupt process, the reference signal having the same rotational position Qb as the rotational position Qt obtained in the previous interrupt process and the current interrupt process. An average value (average reference signal) x ′ for all of x (xb) is obtained, and the average reference signal x ′ is updated. For example, if there are five reference signals xb xb1, xb2, xb3, xb4, and xb5 having the same rotational position Qb as the rotational position Qt obtained in the current interrupt process, the average reference signal x ′ is (xt + xb1 + xb2 + xb3 + xb4 + xb5) / 6 is required.

  Similarly, the second SRC unit 43-2 includes the filtered reference signal r (rt) in the current interrupt process, and includes the same rotation position Qb as the rotation position Qt obtained in the previous interrupt process and obtained in the current interrupt process. An average value (average filtered reference signal) r ′ for all of the filtered reference signals r (rb) having is obtained, and the average filtered reference signal r ′ is updated.

  Similarly, the third SRC unit 43-3 includes the residual wave signal e (et) in the current interrupt process and includes the same rotational position as the rotational position Qt obtained in the previous interrupt process and in the current interrupt process. An average value (average residual wave signal) e ′ for all residual wave signals e (eb) having Qb is obtained, and the average residual wave signal e ′ is updated.

  Then, the adaptive filter unit 45 does not operate and outputs the control signal y as it was in the previous interrupt process, and the process S17 described later is performed. This slows down the sampling rate.

  On the other hand, in process S16, when the above-mentioned different cases (Qt ≠ Qb, No), it can be determined that the engine has rotated by an angle equal to 180, so the first to third SRC units 43-1 are updated to update the control signal y. Each of ˜43-3 performs processing as follows, whereby the adaptive filter unit 45 and the filter coefficient update unit 44 perform processing as follows.

  First, the first SRC unit 43-1 determines the average reference signal x ′ at each value at that time, and outputs the determined average reference signal x ′ to the adaptive filter unit 45. On the other hand, the second and third SRC units 43-2 and 43-3 determine the average filtered reference signal r ′ and the average residual wave signal e ′, respectively, at the respective values at that time, and determine the determined average filtered reference signal r. 'And the average residual wave signal e' are output to the filter coefficient updating unit 44. When the filter coefficient updating unit 44 receives the input of the determined average filtered reference signal r ′ and the average residual wave signal e ′, each vector of the determined average filtered reference signal r ′ and the average residual wave signal e ′. A vector H of filter coefficients h is obtained by an LMS algorithm using R ′ and E ′. Then, the filter coefficient updating unit 44 obtains an average value of each component in the obtained filter coefficient vector H, obtains a filter coefficient vector H from which a direct current component has been removed by subtracting the average value from each component, and obtains the obtained value. The filtered filter coefficient vector H is output to the adaptive filter unit 45. Then, upon receiving the determined average reference signal x ′ and the filter coefficient vector H, the adaptive filter unit 45 convolves the vector X ′ of the determined average reference signal x ′ and the filter coefficient vector H (convolution). Thus, a vector (control signal vector Y) representing the control signal y is obtained, and the obtained control signal y is output to the second filter unit 41-2.

  Then, in order to carry out the next preparation, the first SRC 43-1 passes the reference signal x input thereto as it is, and updates the average reference signal x 'using the reference signal x as a new average reference signal x'. Similarly, the second SRC 43-2 passes the filtered reference signal r inputted thereto as it is, and updates the average filtered reference signal r using the filtered reference signal r as a new average filtered reference signal r. Similarly, the third SRC 43-1 passes the residual wave signal e input thereto as it is, and updates the average residual wave signal e ′ using the residual wave signal e as a new average residual wave signal e ′. . Then, the next process S17 is performed.

  In process S17, the control signal y generated as described above by the adaptive filter unit 45 is input to the second filter unit 41-2 of the silencer control adaptive filter unit 4, and the second filter unit 41-2 generates a noise component. Are removed and output to the DAC unit 23 of the silencer generating unit 2. And the silencer production | generation part 2 produces | generates a silencer based on this control signal y, and radiates | emits a silencer toward the control point CP.

  By such processing, after the sampling rate conversion, the information at the time of one rotation of the engine sampled at 4 kHz, that is, the basic period exactly matches the filter length 180 in the adaptive filter unit 45 having 180 taps.

As described above, the feedforward active noise control apparatus ANC in the present embodiment, as filtered reference signal r and the residual wave signal e obtains the filter coefficient h _i of the adaptive filter section 45, the basic cycle of the reference signal x or the A filtered reference signal r and a residual wave signal e extracted so that a period that is an integral multiple of the fundamental period matches the filter length are used. If the filtered reference signal r and the residual wave signal e are extracted without considering the filter length, the extracted filtered reference signal r and the residual wave signal e are less than one period. The waveform is included and becomes discontinuous, and not only the fundamental frequency and its integral multiple frequency components but also other frequency components are included. As a result, in the calculation of the LMS algorithm, the convergence calculation is executed including the other frequency components that are not originally required, and the convergence time becomes long. However, when the filtered reference signal r, the residual wave signal e, and the reference signal x are extracted as described above, each of the extracted filtered reference signal r and residual wave signal e is less than one period. A waveform is not included, and only a waveform of one period or an integral multiple thereof is included, resulting in a continuous waveform. As a result, the convergence calculation is not unnecessarily lengthened and is shortened. Therefore, the feedforward type active noise control device ANC and the feedforward type active noise control method implemented in this embodiment are the cases where the same processing capacity hardware and the same filter length are used for information processing of the LMS method. Even if it exists, the tracking time (convergence time of the LMS method) can be further shortened.

In the feedforward active noise control device ANC according to the present embodiment, the reference signal x to be filtered by the adaptive filter unit 45 is such that the basic period of the reference signal x or an integer multiple of the basic period matches the filter length. The reference signal x taken out is used. As described above, the filter coefficient h _i is obtained based on the filtered reference signal r and the residual wave signal e that are extracted so that the fundamental period of the reference signal x or an integer multiple of the fundamental period matches the filter length. because it is, and is thus determined filter coefficients h _i and the reference signal x a filter length was taken in consideration, the same length. Therefore, such a feedforward active noise control apparatus ANC is the thus determined filter coefficients h _i, a reference signal taken out by considering the filter length x, can be accurately filtering.

  Further, the feedforward active noise control device ANC according to the present embodiment changes the sampling rate of the first to third SRC units 43-1 to 43-3 as an example of the sampling unit according to the rotational speed of the engine. In a relatively simple method, the reference signal x, the filtered reference signal r, and the residual wave signal e are set to the filter length of the adaptive filter unit 45 in the fundamental period of the reference signal x or an integer multiple of the fundamental period in substantially real time. Can match. In this embodiment, each time the engine rotates by a fraction of the filter length, the sampling rate is changed so that the reference signal x, the filtered reference signal r, and the residual wave signal e are sampled. With a simple method, the reference signal x, the filtered reference signal r, and the residual wave signal e can be set such that the fundamental period of the reference signal x or an integer multiple of the fundamental period matches the filter length.

  Next, each result of an Example and a comparative example is demonstrated. FIG. 7 is a diagram for explaining the tracking time (convergence time of the LMS method) in the feedforward active noise control device of the embodiment. FIG. 8 is a diagram for explaining the follow-up time (convergence time of the LMS method) in the feedforward active noise control device of the comparative example. FIGS. 7A and 8A show temporal changes in the frequency of the original sound, which is a noise wave, where each horizontal axis is a sample number, that is, time, and each vertical axis is frequency. FIGS. 7B and 8B show temporal changes in the power of the sound wave on the left side of the control point CP. Each horizontal axis is a sample number, that is, time, and each vertical axis is power. . 7 (C) and 8 (C) show the time change of the power of the sound wave on the right side of the control point CP. Each horizontal axis is a sample number, that is, time, and each vertical axis is power. It is. 7 and 8, a solid line indicates a case where active noise control is not performed (without ANC), and a broken line indicates a case where active noise control is performed (with ANC). In FIGS. 7A, 7B, and 7C, with and without ANC cannot be performed simultaneously with one construction machine, solid lines are related to each other, and broken lines are related to each other. Similarly, in FIGS. 8A, 8 </ b> B, and 8 </ b> C, with ANC and without ANC cannot be performed simultaneously by one construction machine, solid lines are related to each other, and broken lines are related to each other. .

  In the embodiment, the feedforward type active noise control device ANC shown in FIG. 3 is mounted on the construction machine shown in FIG. In the comparative example, the conventional feedforward active noise control device obtained by removing the first to third SRC units 43-1 to 43-3 from the feedforward active noise control device ANC shown in FIG. It is what is installed. And when the construction machine was relieved, the effect of active noise control was measured.

  In the embodiment, when the feedforward type active noise control device ANC is not operated (no ANC), the time change of the frequency in the original sound (noise wave) is as shown in the solid line of FIG. The frequency once increases from about 17.5 Hz in steady state to about 18.3 Hz, then decreases to about 17.5 Hz in steady state, and once lower from about 17.5 Hz to about 16.2 Hz in this steady state. After that, the profile rose up to about 17.5 Hz in a steady state and returned. With respect to the original sound whose frequency changes with time, the sound wave power on the left side of the control point CP changes with time as shown by the solid line in FIG. 7B, and the sound wave power on the right side of the control point CP becomes The time changed as indicated by the solid line in FIG. On the other hand, when the feedforward type active noise control device ANC is operated (with ANC), the time change of the frequency in the original sound (noise wave) is the same construction machine (engine), so FIG. As indicated by the broken line, the profile is the same as the profile indicated by the solid line in FIG. With respect to the original sound whose frequency changes with time, the sound wave power on the left side of the control point CP changes with time as shown by the broken line in FIG. 7B, and the sound wave power on the right side of the control point CP becomes As shown by the broken line in FIG. The follow-up time (the convergence time of the LMS method) in the feed-forward type active noise control device ANC is that the sound wave power at the control point CP has increased rapidly as a result of the loss of noise waves being lost due to changes in the frequency of the original sound. This is the time from the time point to the time point when the noise wave is reduced by active noise control, and in the case of the example, as shown in FIGS. 7B and 7C, time T1.

  On the other hand, in the comparative example, when the feedforward active noise control device of the comparative example is not operated (no ANC), the time change of the frequency in the original sound (noise wave) is as shown by the solid line in FIG. During relief, the frequency once increases from about 17.5 Hz in steady state to about 18.1 Hz, then decreases to about 17.5 Hz in steady state, and from about 17.5 Hz in this steady state to about 16.1 Hz. The profile was once lowered to 8 Hz and then returned to a steady state of about 17.5 Hz. With respect to the original sound whose frequency changes with time, the sound wave power on the left side of the control point CP changes with time as shown by the solid line in FIG. 8B, and the sound wave power on the right side of the control point CP becomes The time changed as indicated by the solid line in FIG. On the other hand, when the feedforward type active noise control device of the comparative example is operated (with ANC), the time change of the frequency in the original sound (noise wave) is the same construction machine (engine). As indicated by the broken line in A), the profile was the same as the profile indicated by the solid line in FIG. With respect to the original sound whose frequency changes with time, the sound wave power on the left side of the control point CP changes with time as shown by the broken line in FIG. 8B, and the sound wave power on the right side of the control point CP is As shown by the broken line in FIG. In the case of the comparative example, the follow-up time (the convergence time of the LMS method) in the feedforward type active noise control device was time T2, as shown in FIGS. 8B and 8C.

  As can be seen by comparing the time T1 shown in FIG. 7 and the time T2 shown in FIG. 8, the follow-up time T1 in the feedforward active noise control device ANC of the embodiment is the follow-up time in the feedforward active noise control device of the comparative example. It is shorter than time T2, and the follow-up time is improved.

  In the above-described embodiment, the reference signal generation unit 1 includes the rotary encoder that detects the rotational state of the engine, and the fundamental frequency of the noise wave is obtained from the output thereof. However, the present invention is not limited to this. For example, the reference signal generation unit 1 includes a microphone that detects engine sound, and obtains the fundamental frequency of the noise wave by obtaining a spectrum of the noise wave by performing Fourier transform (for example, fast Fourier transform) on the output of the microphone. Also good.

  In the above-described embodiment, the feedforward active noise control device ANC is composed of one set of the reference signal generation unit 1, the silencer generation unit 2, and the residual wave detection unit 3. It may be configured. Also in this case, according to the conventional theory, the feedforward type active noise control device ANC can be configured in the same manner as in the above-described embodiment.

  In order to express the present invention, the present invention has been properly and fully described through the embodiments with reference to the drawings. However, those skilled in the art can easily change and / or improve the above-described embodiments. It should be recognized that this is possible. Therefore, unless the modifications or improvements implemented by those skilled in the art are at a level that departs from the scope of the claims recited in the claims, the modifications or improvements are not covered by the claims. To be construed as inclusive.

ANC feedforward type active noise control device NS noise source HS hydraulic excavator 1 reference signal generator 2 silencer generator 3 residual wave detector 4 silencer adaptive filter unit 43-1 first SRC unit 43-2 second SRC unit 43 -3 Third SRC part

Claims (6)

A reference signal generation unit that detects a noise wave generation state of a noise wave source and generates a reference signal related to the noise wave based on the detected noise wave generation state;
A silencer generator for generating a silencer in response to the control signal;
A residual wave detector for detecting a residual wave of the noise wave and the silencer remaining at a predetermined control point and outputting a residual wave signal;
Based on the filtered reference signal obtained by filtering the reference signal with the transfer characteristic from the silencer output point of the silencer to the control point, and the residual wave signal, the noise wave has the same amplitude and the opposite phase. A filter coefficient is obtained so as to generate a silencer, and a silencer control adaptive filter unit that generates the control signal by filtering the reference signal with the obtained filter coefficient,
The reference signal generation unit superimposes a sine wave having a fundamental frequency of the noise wave and one or more higher-order sine waves having a higher order frequency with respect to the fundamental frequency based on the noise wave generation state. Generating the reference signal by:
The silencer control adaptive filter unit extracts the filtered reference signal and the residual wave signal for obtaining the filter coefficient so that a fundamental period of the reference signal or an integer multiple of the fundamental period matches a filter length. A feedforward active noise control device using a filtered reference signal and a residual wave signal. The silencer control adaptive filter unit further uses, as the reference signal to be filtered, a reference signal extracted so that a basic period of the reference signal or an integer multiple of the basic period matches a filter length. The feedforward active noise control device according to claim 1, wherein The silencer control adaptive filter unit is
A secondary filter unit that filters the reference signal with a transfer characteristic from the silencer output point of the silencer to the control point and outputs a filtered reference signal;
A sampling unit that samples each of the reference signal, the filtered reference signal, and the residual wave signal such that a fundamental period of the reference signal or an integer multiple of the fundamental period matches a filter length;
A filter coefficient update unit for obtaining a filter coefficient so as to generate the muffler having the same amplitude and opposite phase as the noise wave based on the filtered reference signal and the residual wave signal sampled by the sampling unit,
The adaptive filter unit that generates the control signal by filtering the reference signal sampled by the sampling unit with the filter coefficient obtained by the filter coefficient update unit. Feed forward active noise control device. The noise wave source is an engine;
The noise wave generation state is the rotational speed of the engine,
The reference signal generation unit detects a rotation speed of the engine, and a sine wave having a frequency corresponding to the detected rotation speed as a fundamental frequency, and one or a plurality of higher orders having a higher order frequency with respect to the fundamental frequency The sine wave is superimposed on each other to generate the reference signal,
The sampling unit of the silencer control adaptive filter unit changes the sampling rate in accordance with the rotation speed of the engine, so that each of the reference signal, the filtered reference signal, and the residual wave signal is converted into the basic of the reference signal. The feedforward active noise control device according to claim 3, wherein sampling is performed so that a cycle or a cycle that is an integral multiple of the basic cycle matches a filter length. The sampling unit of the silencer control adaptive filter unit includes the reference signal, the filtered reference signal, and the residual wave signal each time the engine rotates by a filter length equal to one rotation divided by a filter length. The feedforward active noise control device according to claim 4, wherein the sampling rate is changed so that each is sampled. A reference signal generation step of detecting a noise wave generation state of the noise wave source and generating a reference signal related to the noise wave based on the detected noise wave generation state;
A silencer generating step for generating a silencer according to the control signal;
A residual wave detection step of detecting a residual residual wave of the noise wave and the silencer at a predetermined control point and outputting a residual wave signal;
Based on the filtered reference signal obtained by filtering the reference signal with the transfer characteristic from the silencer output point of the silencer to the control point, and the residual wave signal, the noise wave has the same amplitude and the opposite phase. A filter coefficient so as to generate a silencer, and a silencer control adaptive filter step of generating the control signal by filtering the reference signal with the obtained filter coefficient,
The reference signal generating step superimposes a sine wave having a fundamental frequency of the noise wave and one or more higher-order sine waves having a higher-order frequency with respect to the fundamental frequency on the basis of the noise wave generation state. Generating the reference signal by:
The silencer control adaptive filtering step extracts the filtered reference signal and the residual wave signal for obtaining the filter coefficient so that a fundamental period of the reference signal or an integer multiple of the fundamental period matches a filter length. A feedforward active noise control method using a filtered reference signal and a residual wave signal.