CN110889197B - Self-adaptive feedforward active noise reduction method based on neural network, computer readable storage medium and electronic equipment - Google Patents
Self-adaptive feedforward active noise reduction method based on neural network, computer readable storage medium and electronic equipment Download PDFInfo
- Publication number
- CN110889197B CN110889197B CN201911056457.3A CN201911056457A CN110889197B CN 110889197 B CN110889197 B CN 110889197B CN 201911056457 A CN201911056457 A CN 201911056457A CN 110889197 B CN110889197 B CN 110889197B
- Authority
- CN
- China
- Prior art keywords
- neural network
- network model
- noise reduction
- active noise
- adaptive feedforward
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R1/00—Details of transducers, loudspeakers or microphones
- H04R1/10—Earpieces; Attachments therefor ; Earphones; Monophonic headphones
- H04R1/1083—Reduction of ambient noise
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/10—Noise analysis or noise optimisation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Acoustics & Sound (AREA)
- Signal Processing (AREA)
- Soundproofing, Sound Blocking, And Sound Damping (AREA)
Abstract
The invention relates to a self-adaptive feedforward active noise reduction method based on a neural network, a computer readable storage medium and electronic equipment, which are used for improving the conditions of harmonic waves and intermodulation distortion generated by a link, wherein the method comprises the following steps: constructing a first neural network model, and taking the constructed first neural network model as a feedforward filter in an adaptive feedforward active noise reduction architecture based on the architecture; and/or constructing a second neural network model based on an adaptive feed-forward active noise reduction architecture with the second neural network model as a secondary channel estimate S' (z) in the architecture.
Description
Technical Field
The invention relates to earphone noise reduction, in particular to a self-adaptive feedforward active noise reduction method based on a neural network, a computer readable storage medium and electronic equipment.
Background
Referring to fig. 1, the basic principle of the adaptive feedforward active noise reduction architecture in the earphone is as follows: in the A bit, the reference microphone picks up an original noise signal in the environment, a signal (called as an inverted noise signal for short) inverted to the original noise signal is generated through a feedforward filter, then the inverted noise signal is output in the B bit through a loudspeaker, the original noise signal and the inverted noise signal are mutually offset in the B bit to generate a residual noise signal, and the self-adaptive feedforward active noise reduction is realized.
On the basis of self-adaptive feedforward active noise reduction, a C-bit error microphone is added, residual noise signals are collected by the C-bit error microphone, and after the residual noise signals are analyzed by an LMS algorithm, a weight coefficient is generated to adjust reverse noise signals output by a feedforward filter, so that the self-adaptive feedforward active noise reduction is realized.
In adaptive feedforward active noise reduction, a feedforward filter is generally implemented by an FIR (finite impulse response) filter or an IIR (infinite impulse response) filter, and the FIR is taken as an example, and the adaptive feedforward active noise reduction is implemented by the feedforward filterThe working process of the noise-reduced signal is shown in fig. 2, wherein x (n) is an original noise signal, P (z) is a transfer function of an original noise channel, the time delay of noise from A bit to B bit is represented, and x (n) outputs a target signal d (n) and W after passing through P (z) f (n) is a feedforward filter, y (n) is W f (n) the output inverse noise signal, S (z) is the secondary channel, i.e. W is input from acquisition x (n) f (n) to W f (n) outputting y (n) to a loudspeaker, pushing air through the loudspeaker, transmitting the air to a transfer function of the whole path sucked by an error microphone, representing time delay of noise through the path, superposing and canceling y (n) and d (n), sucking by the error microphone, obtaining residual noise signal e (n), S' (z) being an estimation of S (z), wherein the estimation is essentially a filter, and a weight coefficient iterative calculation formula is arranged in the estimation, and multiplying x (n) by the output weight of the formula, so as to correct x (n).
When in use, on one hand, e (n) is sucked by the error microphone as one input of the LMS algorithm, on the other hand, x (n) is corrected by S' (z) and then is used as the other input of the LMS algorithm, and the iteration formula output W of the LMS algorithm is utilized f Weight coefficient of (n) to obtain W f After the weight coefficient of (n), taking x (n) as W f An input signal of (n) is passed through W f The weight coefficient of (n) is x (n) so as to output y (n), thereby achieving the purpose of self-adaptive feedforward active noise reduction.
The existing adaptive feedforward active noise reduction scheme has the following problems:
because of the feedforward filter W f The FIR filter or the IIR filter adopted in (n) belongs to a linear filter, so if a nonlinear link exists in a propagation path from an a-bit noise source to a B-bit speaker, for example, the original noise is too large, so that the reference microphone generates nonlinearity, or if a nonlinear link exists in a propagation path from the B-bit speaker to a C-bit error microphone, for example, the speaker is saturated, the noise reduction effect of the whole link can be obviously reduced because the FIR filter or the IIR filter cannot handle the harmonic wave and intermodulation distortion generated by nonlinearity.
Disclosure of Invention
The present invention is directed to an adaptive feed-forward active noise reduction method for improving the harmonic and intermodulation distortion generated by a link.
For this purpose, an adaptive feedforward active noise reduction method based on a neural network is provided, which includes:
constructing a first neural network model, and based on an adaptive feedforward active noise reduction architecture, using the constructed first neural network model as a feedforward filter in the architecture; and/or
A second neural network model is constructed based on an adaptive feedforward active noise reduction architecture, and the second neural network model is used as a secondary channel estimate S' (z) in the architecture.
As an embodiment, the building a first neural network model further includes:
s101, constructing a theoretical model of a first neural network model;
s102, acquiring S '(z) of a self-adaptive feedforward active noise reduction framework in the earphone, wherein S' (z) is the estimation of a secondary channel in the self-adaptive feedforward active noise reduction framework;
s103, collecting x from self-adaptive feedforward type active noise reduction framework 1 (n)、y 1 (n)、e 1 (n) wherein x 1 (n) is the historical raw noise signal, y 1 (n) is the historical inverse noise signal output by the feedforward filter, e 1 (n) is a historical residual noise signal;
s104, x is 1 (n) S' (z) corrected value sum e 1 (n) as input, calculating a weight coefficient of the first neural network model by using a BP algorithm;
s105, x is 1 (n), weight coefficient sum y of first neural network model 1 (n) training samples as theoretical models of the first neural network model, wherein x is taken as 1 (n) and the weight coefficient of the first neural network model are taken as input, y 1 (n) training the theoretical model using machine learning as an output.
And fitting the trained first neural network model by adopting a logistic regression algorithm.
The first neural network model takes x (n) and the current weight coefficient of the first neural network model as input, and outputs y (n) to a loudspeaker in the adaptive feedforward active noise reduction framework for broadcasting, wherein x (n) is a current original noise signal, and y (n) is a current inverse noise signal output by the feedforward filter.
As another embodiment, the building a second neural network model further includes:
s201, constructing a theoretical model of a second neural network model;
s202, collecting x from adaptive feedforward type active noise reduction framework 1 (n)、h 1 (n)、e 1 (n) wherein x 1 (n) is the historical original noise signal, h 1 (n) is x 1 (n) the value output after correction by S' (z), e 1 (n) is a historical residual noise signal;
s203, x is 1 (n) and e 1 (n) calculating the weight coefficient of the second neural network model by using BP algorithm as input;
s204, x is 1 (n), weight coefficient of the second neural network model and h 1 (n) training samples of a theoretical model of the second neural network model, wherein x is 1 (n) the weight coefficient of the second neural network model is taken as input, h is taken as 1 (n) training the theoretical model using machine learning as an output.
And fitting the trained second neural network model by adopting a logistic regression algorithm.
The second neural network model takes x (n) and the current weight coefficient of the second neural network model as input, h (n) is output to an adaptive algorithm in the adaptive feedforward active noise reduction framework, the weight coefficient is generated after analysis of the adaptive algorithm, the inverse noise signal output by the feedforward filter is adjusted, wherein x (n) is the current original noise signal, and h (n) is the value of x (n) corrected by the second neural network model.
There is also provided an electronic apparatus, wherein the electronic apparatus includes:
a controller; the method comprises the steps of,
a memory arranged to store computer executable instructions which, when executed, cause the controller to perform the method described above.
There is also provided a computer readable storage medium storing one or more programs which, when executed by a controller, implement the above-described method.
The beneficial effects are that:
the invention is realized by changing the feedforward filter in the self-adaptive feedforward active noise reduction framework into the first neural network model, and the noise is better controlled by utilizing the nonlinear filter characteristic of the neural network, so that the harmonic wave and intermodulation distortion phenomenon caused by nonlinear links in the propagation path from the A-bit noise source to the B-bit loudspeaker are improved; similarly, S' (z) is changed to be realized by a second neural network model, so that the estimation of a secondary channel is more accurate, and the harmonic wave and intermodulation distortion phenomena caused by nonlinear links in the propagation path from the B-bit loudspeaker to the C-bit error microphone are improved.
The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present invention more readily apparent.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to designate like parts throughout the figures. In the drawings:
FIG. 1 is a schematic diagram of an adaptive feedforward noise reduction architecture in a conventional earphone;
FIG. 2 illustrates a digital noise reduction scheme implementation of an adaptive feedforward noise reduction architecture in a conventional earphone;
FIG. 3 is a schematic diagram showing the implementation of the digital noise reduction scheme when the feedforward filter is changed to the first neural network model in the present embodiment;
FIG. 4 shows a schematic structural diagram of a first neural network model;
FIG. 5 shows a schematic structural diagram of a neuron;
FIG. 6 shows a schematic diagram of a first neural network model in the presence of several input signals;
FIG. 7 is a schematic diagram showing the implementation of the digital noise reduction scheme when the secondary channel estimation S' (z) is changed to the second neural network model in the present embodiment;
FIG. 8 shows a schematic diagram of a digital noise reduction scheme implementation when both the feedforward filter and the secondary channel estimate S' (z) are implemented with a neural network in this embodiment;
fig. 9 shows a schematic structural diagram of the electronic device of the present invention;
fig. 10 shows a schematic structure of a computer-readable storage medium of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
For the nonlinear link existing in the propagation path from the A-bit noise source to the B-bit loudspeaker, for example, the nonlinear link is generated by the reference microphone due to the overlarge original noise, if the nonlinear filter can be used as the feedforward filter, the noise can be better controlled, and the neural network is a nonlinear controller, so that the feedforward filter can be realized by adopting the neural network, the harmonic wave and intermodulation distortion phenomenon caused by the nonlinear link existing in the propagation path from the A-bit noise source to the B-bit loudspeaker can be improved, and particularly, the feedforward filter is realized by a first neural network model on the basis of the adaptive feedforward active noise reduction architecture in the background technology, and the method specifically comprises the following steps:
s101, constructing a theoretical model of a first neural network model, and specifically:
assuming that the first neural network model is implemented by a 3×n×1 forward neural network as shown in fig. 4, that is, the input layer is 3 neurons, the hidden layer is N neurons, and the output layer is one neuron;
wherein w is ij,h Representing the weight value, w, of the ith input to the jth neuron of the hidden layer j,o Weights representing the neuron-to-output of j hidden layers;
the definition of the neurons is shown in FIG. 5, on the basis of which it is assumed that there are several input signals x i I=1, 2, …, N, the first neural network model becomes as shown in fig. 6, where w i The weight coefficient corresponding to the ith input is represented, θ represents the threshold value of the neuron, and can be understood as another constant input value; sigma represents accumulation, net represents an accumulated value obtained by multiplying each input value by a weight value, and y is an output value; f is a nonlinear function, such as Sigmoid function, or hyperbolic tangent function, as follows:
s102, acquiring S '(z) of an adaptive feedforward type active noise reduction framework in the earphone, wherein S' (z) is an estimation of a secondary channel in the adaptive feedforward type active noise reduction framework, and specifically:
in this step, S '(z) is still implemented by an FIR filter or an IIR filter temporarily, and here, taking the FIR filter as an example, the weight coefficient of each step of S' (z) is calculated by the following iterative calculation formula:
w 1 (n+1)＝w 1 (n)+λ 2 e(n)x(n)
s103, collecting x from self-adaptive feedforward type active noise reduction framework 1 (n)、y 1 (n)、e 1 (n) wherein x 1 (n) is the historical raw noise signal, y 1 (n) is the historical inverse noise signal output by the feedforward filter, e 1 (n) is the history residualNoise signals.
S104, x is 1 (n) S' (z) corrected value sum e 1 (n) as input, calculating the weight coefficient of the first neural network model using BP (back propagation) algorithm, specifically:
according to BP (Back propagation) back propagation algorithm, an iterative calculation formula for each weight of the output layer in the first neural network model is calculated as follows, wherein f' o (Net o ) Is f o (Net o ) Is the derivative of:
the iterative calculation formula of each weight of the hidden layer is as follows:
the two formulas above are either written collectively:
w node，o (n+1)＝w node，o (n)+λ 1 δ node (n)x node，h (n)
wherein lambda is 1 Is the convergence coefficient, delta nod∈ (n) is calculated as follows:
s105, x is 1 (n), weight coefficient sum y of first neural network model 1 (n) training samples as theoretical models of the first neural network model, wherein x is taken as 1 (n) and the weight coefficient of the first neural network model are taken as input, y 1 (n) training a theoretical model by using a machine learning mode as output, and fitting a first neural network model by using a Logistic Regression (LR) algorithm until a model meeting the requirements is obtained, so as to realize the construction of the first neural network model;
s106, using the constructed first neural networkThe complex model is used as a feedforward filter W in a self-adaptive feedforward active noise reduction architecture f (n) using, for example, the first neural network model with x (n) being the current raw noise signal and the current weighting coefficients of the first neural network model as inputs, outputting y (n) being the current inverse noise signal output by the feedforward filter, to the loudspeakers in the adaptive feedforward active noise reduction architecture.
For the existence of a nonlinear link in the propagation path from the B-bit loudspeaker to the C-bit error microphone, such as that the loudspeaker is saturated, if the estimation of the secondary channel adopts nonlinear estimation, the estimation of the secondary channel is more accurate, further, considering that S (z) has nonlinearity, so S '(z) can be estimated by using a second neural network model, and parameters of the second neural network model are iteratively solved according to a BP (back propagation) algorithm, specifically, see fig. 7, the S' (z) is changed to be implemented by the second neural network model on the basis of the adaptive feedforward active noise reduction architecture of the background technology, and the method comprises the following steps:
s201, constructing a theoretical model of a second neural network model, wherein the theoretical model of the second neural network model is constructed by referring to the theoretical model of the first neural network model, and the theoretical model is not repeated here.
S202, collecting x from adaptive feedforward type active noise reduction framework 1 (n)、h 1 (n)、e 1 (n) wherein x 1 (n) is the historical original noise signal, h 1 (n) is x 1 (n) the value output after correction by S' (z), e 1 (n) is a historical residual noise signal;
s203, x is 1 (n) and e 1 (n) as input, calculating weight coefficients of the second neural network model using BP (back propagation) algorithm, specifically:
according to BP (Back propagation) back propagation algorithm, the iterative calculation formula for calculating each weight in the second neural network model is as follows:
w node，o (n+1)＝w node，o (n)+λ 1 δ node (n)x node，h (n)
s204, x is 1 (n), weight coefficient of the second neural network model and h 1 (n) training samples of a theoretical model of the second neural network model, wherein x is 1 (n) the weight coefficient of the second neural network model is taken as input, h is taken as 1 (n) training a theoretical model by using a machine learning mode as output, and fitting a second neural network model by using a Logistic Regression (LR) algorithm until a model meeting the requirements is obtained, so as to realize the construction of the second neural network model;
s205, using the constructed second neural network model as S '(z) in the adaptive feedforward type active noise reduction framework, wherein S' (z) is an estimation of a secondary channel in the adaptive feedforward type active noise reduction framework, for example, using x (n) and e (n) as inputs, calculating a current weight coefficient of the second neural network model by using a BP algorithm, using the current weight coefficients of the x (n) and the second neural network model as inputs, outputting h (n) to an LMS algorithm in the adaptive feedforward type active noise reduction framework, generating a weight coefficient after analysis by the LMS algorithm, and adjusting an inverse noise signal output by the feedforward filter, wherein x (n) is a current original noise signal, h (n) is a value of the x (n) corrected by the second neural network model, and e (n) is a current residual noise signal.
For the case where there is a nonlinear link in the propagation path from the a-bit noise source to the B-bit speaker and in the propagation path from the B-bit speaker to the C-bit error microphone, both the feedforward filter and the secondary channel estimate S '(z) may be implemented with a neural network, specifically, as shown in fig. 8, using the above-described trained first neural network model as the feedforward filter and the above-described trained second neural network model as the secondary channel estimate S' (z).
It should be noted that:
the method according to the present embodiment can be implemented by being transferred to a program step and a device that can be stored in a computer storage medium, and being called and executed by a controller.
The algorithms and displays presented herein are not inherently related to any particular computer, virtual machine, or other apparatus. Various general purpose devices may also be used with the teachings herein. The required structure for the construction of such devices is apparent from the description above. In addition, the present invention is not directed to any particular programming language. It will be appreciated that the teachings of the present invention described herein may be implemented in a variety of programming languages, and the above description of specific languages is provided for disclosure of enablement and best mode of the present invention.
In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be construed as reflecting the intention that: i.e., the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.
Those skilled in the art will appreciate that the modules in the apparatus of the embodiments may be adaptively changed and disposed in one or more apparatuses different from the embodiments. The modules or units or components of the embodiments may be combined into one module or unit or component and, furthermore, they may be divided into a plurality of sub-modules or sub-units or sub-components. Any combination of all features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or units of any method or apparatus so disclosed, may be used in combination, except insofar as at least some of such features and/or processes or units are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings), may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.
Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments.
Various component embodiments of the invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that some or all of the functions of some or all of the components in an apparatus for detecting the wearing state of an electronic device according to an embodiment of the present invention may be implemented in practice using a microprocessor or Digital Signal Processor (DSP). The present invention can also be implemented as an apparatus or device program (e.g., a computer program and a computer program product) for performing a portion or all of the methods described herein. Such a program embodying the present invention may be stored on a computer readable medium, or may have the form of one or more signals. Such signals may be downloaded from an internet website, provided on a carrier signal, or provided in any other form.
For example, fig. 9 shows a schematic structural diagram of an electronic device according to an embodiment of the present invention. The electronic device conventionally comprises a processor 61 and a memory 62 arranged to store computer executable instructions (program code). The memory 62 may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read only memory), an EPROM, a hard disk, or a ROM. Memory 62 has a storage space 63 storing program code 64 for performing any of the method steps in the embodiments. For example, the memory space 63 for the program code may include individual program code 64 for implementing the various steps in the above method, respectively. The program code can be read from or written to one or more computer program products. These computer program products comprise a program code carrier such as a hard disk, a Compact Disc (CD), a memory card or a floppy disk. Such a computer program product is typically a computer readable storage medium as described for example in fig. 10. The computer readable storage medium may have memory segments, memory spaces, etc. arranged similarly to the memory 62 in the electronic device of fig. 6. The program code may be compressed, for example, in a suitable form. Typically, the memory unit stores program code 71 for performing the method steps according to the invention, i.e. program code readable by a processor such as 61, which when run by an electronic device causes the electronic device to perform the steps in the method described above.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The use of the words first, second, third, etc. do not denote any order. These words may be interpreted as names.
Claims (5)
1. The adaptive feedforward active noise reduction method based on the neural network is characterized by comprising the following steps of:
constructing a first neural network model based on an adaptive feedforward active noise reduction architecture, using the constructed first neural network model as a feedforward filter in the architecture, the constructing the first neural network model further comprising: s101, constructing a theoretical model of a first neural network model; s102, acquiring S '(z) of a self-adaptive feedforward active noise reduction framework in the earphone, wherein S' (z) is the estimation of a secondary channel in the self-adaptive feedforward active noise reduction framework; s103, collecting x from self-adaptive feedforward type active noise reduction framework 1 (n)、y 1 (n)、e 1 (n) wherein x 1 (n) is the historical raw noise signal, y 1 (n) is the historical inverse noise signal output by the feedforward filter, e 1 (n) is a historical residual noise signal; s104, with x 1 (n) value h after S' (z) correction 1 (n) and e 1 (n) as input, calculating a weight coefficient of the first neural network model by using a BP algorithm; s105, in x 1 (n), weight coefficient sum y of first neural network model 1 (n) training samples as theoretical models of the first neural network model, wherein x is taken as 1 (n) and the weight coefficient of the first neural network model are taken as input, y 1 (n) training the theoretical model by means of machine learning as output, and y being output 1 (n) broadcasting to speakers in the adaptive feed-forward active noise reduction architecture;
constructing a second neural network model based on an adaptive feedforward active noise reduction architecture, using the second neural network model as a secondary channel estimate S' (z) in the architecture, the constructing the second neural network model further comprising: s201, constructing a theoretical model of a second neural network model; s202, collecting x from adaptive feedforward type active noise reduction framework 1 (n)、h 1 (n)、e 1 (n) wherein x 1 (n) is the historical original noise signal, h 1 (n) is x 1 (n) the value output after correction by S' (z), e 1 (n) is a historical residual noise signal; s203, in x 1 (n) and e 1 (n) calculating the weight coefficient of the second neural network model by using BP algorithm as input; s204, with x 1 (n), second neural network modelWeight coefficient sum h of (2) 1 (n) training samples of a theoretical model of the second neural network model, wherein x is 1 (n) the weight coefficient of the second neural network model is taken as input, h is taken as 1 (n) training the theoretical model using machine learning as an output.
2. The method of claim 1, wherein fitting is performed on the trained first neural network model using a logistic regression algorithm.
3. The method of claim 1, wherein fitting is performed on the trained second neural network model using a logistic regression algorithm.
4. A computer readable storage medium, wherein the computer readable storage medium stores one or more programs which, when executed by a controller, implement the method of any of claims 1-3.
5. An electronic device, wherein the electronic device comprises: a controller; and a memory arranged to store computer executable instructions that, when executed, cause the controller to perform the method of any of claims 1-3.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911056457.3A CN110889197B (en) | 2019-10-31 | 2019-10-31 | Self-adaptive feedforward active noise reduction method based on neural network, computer readable storage medium and electronic equipment |
| CN202310204371.0A CN116305886A (en) | 2019-10-31 | 2019-10-31 | Self-adaptive feedforward active noise reduction method based on neural network filter, computer readable storage medium and electronic equipment |
| CN202310204390.3A CN116362014A (en) | 2019-10-31 | 2019-10-31 | Noise reduction method for constructing secondary channel estimation by using neural network, computer readable storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911056457.3A CN110889197B (en) | 2019-10-31 | 2019-10-31 | Self-adaptive feedforward active noise reduction method based on neural network, computer readable storage medium and electronic equipment |
Related Child Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310204390.3A Division CN116362014A (en) | 2019-10-31 | 2019-10-31 | Noise reduction method for constructing secondary channel estimation by using neural network, computer readable storage medium and electronic equipment |
| CN202310204371.0A Division CN116305886A (en) | 2019-10-31 | 2019-10-31 | Self-adaptive feedforward active noise reduction method based on neural network filter, computer readable storage medium and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110889197A CN110889197A (en) | 2020-03-17 |
| CN110889197B true CN110889197B (en) | 2023-04-21 |
Family
ID=69746747
Family Applications (3)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310204390.3A Pending CN116362014A (en) | 2019-10-31 | 2019-10-31 | Noise reduction method for constructing secondary channel estimation by using neural network, computer readable storage medium and electronic equipment |
| CN201911056457.3A Active CN110889197B (en) | 2019-10-31 | 2019-10-31 | Self-adaptive feedforward active noise reduction method based on neural network, computer readable storage medium and electronic equipment |
| CN202310204371.0A Pending CN116305886A (en) | 2019-10-31 | 2019-10-31 | Self-adaptive feedforward active noise reduction method based on neural network filter, computer readable storage medium and electronic equipment |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310204390.3A Pending CN116362014A (en) | 2019-10-31 | 2019-10-31 | Noise reduction method for constructing secondary channel estimation by using neural network, computer readable storage medium and electronic equipment |
Family Applications After (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310204371.0A Pending CN116305886A (en) | 2019-10-31 | 2019-10-31 | Self-adaptive feedforward active noise reduction method based on neural network filter, computer readable storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (3) | CN116362014A (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111800687B (en) * | 2020-03-24 | 2022-04-12 | 深圳市豪恩声学股份有限公司 | Active noise reduction method and device, electronic equipment and storage medium |
| CN113115196B (en) * | 2021-04-22 | 2022-03-29 | 东莞市声强电子有限公司 | Intelligent test method of noise reduction earphone |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101589628A (en) * | 2007-01-25 | 2009-11-25 | 沃福森微电子股份有限公司 | Ambient noise reduction |
| CN104952458A (en) * | 2015-06-09 | 2015-09-30 | 广州广电运通金融电子股份有限公司 | Noise suppression method, device and system |
| CN107240391A (en) * | 2017-06-30 | 2017-10-10 | 邢优胜 | A kind of active noise controlling method based on fuzzy neural network, system and panzer helmet of driver |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7593535B2 (en) * | 2006-08-01 | 2009-09-22 | Dts, Inc. | Neural network filtering techniques for compensating linear and non-linear distortion of an audio transducer |
| EP2645362A1 (en) * | 2012-03-26 | 2013-10-02 | Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. | Apparatus and method for improving the perceived quality of sound reproduction by combining active noise cancellation and perceptual noise compensation |
-
2019
- 2019-10-31 CN CN202310204390.3A patent/CN116362014A/en active Pending
- 2019-10-31 CN CN201911056457.3A patent/CN110889197B/en active Active
- 2019-10-31 CN CN202310204371.0A patent/CN116305886A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101589628A (en) * | 2007-01-25 | 2009-11-25 | 沃福森微电子股份有限公司 | Ambient noise reduction |
| CN104952458A (en) * | 2015-06-09 | 2015-09-30 | 广州广电运通金融电子股份有限公司 | Noise suppression method, device and system |
| CN107240391A (en) * | 2017-06-30 | 2017-10-10 | 邢优胜 | A kind of active noise controlling method based on fuzzy neural network, system and panzer helmet of driver |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110889197A (en) | 2020-03-17 |
| CN116362014A (en) | 2023-06-30 |
| CN116305886A (en) | 2023-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111091805B (en) | Feedback type noise reduction method based on neural network | |
| Lu et al. | Adaptive Volterra filter with continuous lp-norm using a logarithmic cost for nonlinear active noise control | |
| Akhtar et al. | Improving robustness of filtered-x least mean p-power algorithm for active attenuation of standard symmetric-α-stable impulsive noise | |
| CN110889197B (en) | Self-adaptive feedforward active noise reduction method based on neural network, computer readable storage medium and electronic equipment | |
| Wang et al. | Stochastic analysis of FXLMS-based internal model control feedback active noise control systems | |
| Gholami-Boroujeny et al. | Non-linear active noise cancellation using a bacterial foraging optimisation algorithm | |
| CN105321524A (en) | Digital feed-forward adaptive hybrid active noise control method and device | |
| CN110992921B (en) | Method for dynamically selecting reference microphone in feedforward noise reduction system, electronic device and computer readable storage medium | |
| CN106531145A (en) | Recursion active noise control method based on M estimator | |
| US6381272B1 (en) | Multi-channel adaptive filtering | |
| Tan et al. | Nonlinear active noise control using diagonal-channel LMS and RLS bilinear filters | |
| Zhao et al. | Nonlinear adaptive filter-based simplified bilinear model for multichannel active control of nonlinear noise processes | |
| CN105976806A (en) | Active noise control method based on maximum entropy | |
| CN105913836A (en) | Fixed-point real time noise reduction method based on DSP | |
| Kukde et al. | Incremental learning based adaptive filter for nonlinear distributed active noise control system | |
| Yin et al. | Selective fixed-filter active noise control based on frequency response matching in headphones | |
| CN111193497A (en) | Secondary channel modeling method based on EMFNL filter | |
| JP4705554B2 (en) | Echo canceling apparatus, method thereof, program thereof, and recording medium thereof | |
| Diana et al. | Hybrid metaheuristic method of ABC kernel filtering for nonlinear acoustic echo cancellation | |
| Albu | Leading element dichotomous coordinate descent exponential recursive least squares algorithm for multichannel active noise control | |
| CN113889068A (en) | Active noise elimination method and device, electronic equipment and storage medium | |
| US9848261B2 (en) | Apparatus for removing howling and method for removing howling | |
| Thaitangam et al. | Speech enhancement using adaptive filter with bat algorithm | |
| CN113506556B (en) | Active noise control method, device, storage medium and computer equipment | |
| JP4255888B2 (en) | Signal separation method and apparatus, signal separation program, and recording medium recording the program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |