CN111862925B - Adaptive active noise control system and method based on inertia learning - Google Patents
Adaptive active noise control system and method based on inertia learning Download PDFInfo
- Publication number
- CN111862925B CN111862925B CN202010629803.9A CN202010629803A CN111862925B CN 111862925 B CN111862925 B CN 111862925B CN 202010629803 A CN202010629803 A CN 202010629803A CN 111862925 B CN111862925 B CN 111862925B
- Authority
- CN
- China
- Prior art keywords
- control
- signal
- noise
- circuit
- input
- 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
- 238000000034 method Methods 0.000 title claims abstract description 31
- 230000003044 adaptive effect Effects 0.000 title claims description 17
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 54
- 230000008569 process Effects 0.000 claims abstract description 12
- 238000001228 spectrum Methods 0.000 claims description 28
- 238000013450 outlier detection Methods 0.000 claims description 18
- 239000000284 extract Substances 0.000 claims description 6
- 238000007621 cluster analysis Methods 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 5
- 238000001308 synthesis method Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 2
- 230000009467 reduction Effects 0.000 abstract description 9
- 238000013461 design Methods 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 230000008030 elimination Effects 0.000 description 5
- 238000003379 elimination reaction Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 230000000052 comparative effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000002592 echocardiography Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009323 psychological health Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 230000001743 silencing effect Effects 0.000 description 1
- 238000002945 steepest descent method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10K—SOUND-PRODUCING DEVICES; METHODS OR DEVICES FOR PROTECTING AGAINST, OR FOR DAMPING, NOISE OR OTHER ACOUSTIC WAVES IN GENERAL; ACOUSTICS NOT OTHERWISE PROVIDED FOR
- G10K11/00—Methods or devices for transmitting, conducting or directing sound in general; Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
- G10K11/16—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
- G10K11/175—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general using interference effects; Masking sound
- G10K11/178—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general using interference effects; Masking sound by electro-acoustically regenerating the original acoustic waves in anti-phase
- G10K11/1785—Methods, e.g. algorithms; Devices
- G10K11/17853—Methods, e.g. algorithms; Devices of the filter
- G10K11/17854—Methods, e.g. algorithms; Devices of the filter the filter being an adaptive filter
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24143—Distances to neighbourhood prototypes, e.g. restricted Coulomb energy networks [RCEN]
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10K—SOUND-PRODUCING DEVICES; METHODS OR DEVICES FOR PROTECTING AGAINST, OR FOR DAMPING, NOISE OR OTHER ACOUSTIC WAVES IN GENERAL; ACOUSTICS NOT OTHERWISE PROVIDED FOR
- G10K11/00—Methods or devices for transmitting, conducting or directing sound in general; Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
- G10K11/16—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
- G10K11/175—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general using interference effects; Masking sound
- G10K11/178—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general using interference effects; Masking sound by electro-acoustically regenerating the original acoustic waves in anti-phase
- G10K11/1781—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general using interference effects; Masking sound by electro-acoustically regenerating the original acoustic waves in anti-phase characterised by the analysis of input or output signals, e.g. frequency range, modes, transfer functions
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Soundproofing, Sound Blocking, And Sound Damping (AREA)
Abstract
The invention relates to a self-adaptive active noise control system and a method thereof based on inertia learning, wherein the system comprises a reference microphone, a controller, an actuator and an error microphone; the reference microphone and the error microphone are respectively arranged at a position close to the sound source and a control point, respectively collect a reference signal and a feedback signal to be controlled, and convert the collected signals into electric signals to be input to the input end of the controller; the controller processes the reference signal and the feedback signal to generate a control signal with the same amplitude and opposite phase to the noise signal to be controlled, and the control signal is input to the input end of the actuator; the actuator converts the control signal into control sound waves, and the control sound waves are overlapped with noise to be controlled at a control point to eliminate the noise. The invention has reasonable design, combines the k-NN method and the self-adaptive LMS algorithm, reduces the convergence time in the control process, accelerates the convergence speed, and effectively meets the requirement of rapid convergence of the noise reduction system under special working conditions.
Description
Technical Field
The invention belongs to the technical field of active noise control, and particularly relates to a self-adaptive active noise control system and a self-adaptive active noise control method based on inertia learning.
Background
The low-frequency line spectrum noise is used as a common noise in the current life and industrial production, has great influence on the normal life and work of people, and even affects the physical and psychological health of residents and workers when being exposed to a noise environment for a long time, so that the high-efficiency noise control technology is an important problem in the field of acoustic research for a long time.
In general, common noise reduction means mainly include two types of passive noise elimination and active noise elimination. Among them, passive sound attenuation is generally achieved by providing a sound attenuation material to block a sound wave propagation path or attenuating echoes and transmitted waves of noise by an acoustic cover layer. The working frequency of the passive noise elimination technology is mainly medium and high frequency, and particularly has a good effect on high-frequency noise, but has a limited effect on low-frequency noise. Thus, currently suppression of low frequency noise is often considered using active noise control techniques.
The active noise control technology utilizes the acoustic superposition principle, namely noise can be counteracted by generating a control sound wave with the same amplitude and opposite phases through a secondary sound source, and the purpose of noise suppression is achieved. Active noise control systems typically consist of a pick-up microphone that collects acoustic signals, a controller that processes the noise signals, and an actuator that emits a control sound wave. The common control system generally controls the output of the secondary sound source through an adaptive filtering algorithm, so that the system can have the functions of automatically detecting noise, automatically calculating filter parameters and eliminating noise, and the combination function of active noise elimination and active control is realized.
In an active noise control system, how to reasonably optimize a control algorithm is always the focus and hot spot of research in the field. The current least mean square (least mean square, LMS) algorithm and its modified algorithm are the most commonly applied methods in adaptive control processes. The algorithm is an optimized extension of the wiener filtering theory combined with the steepest descent method, no prior knowledge of the statistical characteristics of the target signal is needed, and the filter coefficient is updated by adding a proportion term of a negative mean square error gradient to the filter coefficient at the last moment. The algorithm has the advantages of low computational complexity and good convergence on stationary signals. The Morgan provides an Fx-LMS algorithm on the basis, takes secondary channel transfer functions generated by delay and error of electronic components into consideration, indicates that an actual error signal is not a simple combination of noise and filter output, filters according to the secondary channel transfer functions before updating filter parameters, achieves the purpose of eliminating the error, and is an adaptive algorithm with the best stability and the most wide application in the field of active noise control.
Generally, the convergence speed and control accuracy of a controller are two of the most important criteria for evaluating a control system. Although the expected value of the minimum mean square error algorithm can be converged to the wiener solution unbiasedly, in order to reduce the calculation complexity in the process of using gradient descent, the derivative of the square instantaneous value of the error signal is directly taken as the estimated value of the mean square error gradient, the optimal solution cannot be converged smoothly in the iterative process, residual errors are inevitably generated, and the convergence speed of the algorithm is influenced.
In a practical operating environment, many cases require a faster response speed of the system. For example, when the active noise control technology is used for submarine underwater sound elimination, whether the system can quickly converge determines the sound stealth performance of the ship, and the defect that the traditional minimum mean square error algorithm converges too slowly increases the probability of submarine exposure, so that the active noise reduction technology cannot be well utilized to achieve the sound stealth aim.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides a self-adaptive active noise control system and a self-adaptive active noise control method based on inertia learning, which solve the problem that the traditional minimum mean square error algorithm cannot meet the requirement of rapid convergence of a noise reduction system under a special working condition.
The invention solves the technical problems by adopting the following technical scheme:
an adaptive active noise control system based on inertia learning comprises a reference microphone, a controller, an actuator and an error microphone;
the reference microphone and the error microphone are respectively arranged at a position close to the sound source and a control point, respectively collect a reference signal and a feedback signal to be controlled, and convert the collected signals into electric signals to be input to the input end of the controller;
the controller processes the reference signal and the feedback signal to generate a control signal with the same amplitude and opposite phase to the noise signal to be controlled, and the control signal is input to the input end of the actuator;
the actuator converts the control signal into control sound waves, and the control sound waves are overlapped with noise to be controlled at a control point to eliminate the noise.
Further, the controller comprises an FFT circuit, an outlier detection circuit, a priori data set, a k-NN regressor circuit, an LMS filter circuit and a driving circuit;
the FFT circuit performs fast Fourier transform on the reference signal and inputs the calculated frequency spectrum into the outlier detection circuit;
the outlier detection circuit extracts outliers in the frequency spectrum through a dbscan algorithm to serve as frequency input k-NN regressor circuits of the line spectrum;
the frequency input k-NN regressor circuit searches sample points adjacent to the frequency in the prior data set and carries out weighted average to obtain an iteration initial value of the LMS algorithm;
the LMS filter circuit re-iterates the LMS module tap weight vector according to the error signal and the reference signal, and calculates a control signal;
the drive circuit outputs a control signal to the actuator.
Further, the controller is an integrated circuit chip.
An adaptive active noise control method based on inertia learning comprises the following steps:
step 1, sampling rate is Fs, and frequency f is respectively calculated i ＝1，2，……，f m x Establishing single frequency signals
Step 2, the iteration step length of the self-adaptive filtering algorithm is mu, the length of the FIR filter is l, and each f is obtained through iteration by using the LMS algorithm i Corresponding tap weight vector omega i To convergence, a priori data set w= (ω) 1 ，ω 2 ，…，ω max ) T ；
Step 3, transmitting the noise x (t) to be controlled to an input end of the FFT circuit, obtaining a frequency spectrum F (t) of the noise x (t) at the moment through fast Fourier transform, and transmitting the obtained frequency spectrum F (t) to an input end of the outlier detection circuit;
step 4, the outlier detection circuit performs cluster analysis on the frequency spectrum and extracts n outliers o in the frequency spectrum 1 ，o 2 ，…，o n Wherein o j Representing line spectrum signals constituting different frequencies with control noise, f j Representing the signal frequency, j=1, 2, …, n;
step 5, f j An input terminal of the input k-NN regressor circuit searches the prior-examination data set for two sample points W (f k )，W(f k+1 ) Let f k ＜f j ＜f k+1 And according to it andf j the distance of (2) is weighted averaged to obtain an estimation filter:
ω j ＝(f n -f k )ω k +(f k+1 -f n )ω k+1
according to the frequency synthesis method, the tap weight vector omega predict Expressed as:
step 6, the tap weight vector omega predict As an iteration initial value, the method inputs an LMS filter circuit, takes an input signal vector as follows:
X(t)＝[x(t)，x(t-1)，…，x(t-l+1)]
the output signal of the LMS filter circuit is:
step 7, outputting an output signal of the LMS filter circuit into a control sound wave by an actuator through a driving circuit, superposing the control sound wave with noise at a control point, and acquiring an error signal e (t) through an error microphone; the input LMS filter circuit iterates the tap weight vector omega, and the iteration formula is as follows:
ω′＝′+2μe(t)x(t)
where ω' is the new tap weight vector after the iteration.
Further, the amplitude and phase of the signal in the step 1 do not affect the tap weight vector after the filter converges.
Further, the outlier detection circuit in the step 4 performs cluster analysis and extraction on the frequency spectrum by adopting a dbscan algorithm.
The invention has the advantages and positive effects that:
1. the active noise control method combines a k-nearest neighbor inertia learning algorithm (k-NN algorithm) and a minimum mean square error algorithm (self-adaptive LMS algorithm), greatly reduces convergence time in the control process, and can effectively meet the requirement of rapid convergence of the noise reduction system under special working conditions.
2. The invention takes the LMS algorithm as the basis, calculates and obtains the optimal filter tap vector corresponding to each frequency line spectrum signal before the control starts, and avoids repeated calculation in the control process. The algorithm can converge from a better iteration initial value, has obvious control effect on line spectrum signals with higher energy in noise, does not influence the subsequent control effect of the LMS algorithm on other broadband noise, and greatly accelerates the convergence speed.
3. The algorithm used by the invention can be used simultaneously with optimization methods such as a classical Fx-LMS algorithm, a variable step length LMS algorithm and the like, and meets the requirements under different working conditions.
4. The invention utilizes the computing power provided by the modern high-performance computing chip, and can utilize the statistical characteristics of noise in the control process to provide basis for other signal processing means.
Drawings
FIG. 1 is a connection diagram of an adaptive active noise control system of the present invention;
FIG. 2 is a graph comparing simulation results of noise reduction performance of the present invention with that of a conventional LMS algorithm;
in the figure, a 1-reference microphone, a 2-FFT circuit, a 3-outlier detection circuit, a 4-prior data set and k-NN regressor circuit, a 5-LMS filter circuit, a 6-driving circuit, a 7-controller, an 8-actuator and a 9-error microphone.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
The design idea of the invention is as follows:
the wide use of the traditional minimum mean square error algorithm benefits from low calculation complexity besides the advantage of accurate convergence, and is suitable for meeting the real-time requirement of noise reduction under the condition of limited calculation power. With the continuous breakthrough of high-performance computing chips in recent years, new ideas are provided for the development of active noise control technology, so that some efficient machine learning algorithms can participate in the control process.
The nearest neighbor regression (k-Nearest Neighbors, k-NN) algorithm is a classical lazy learning algorithm. The algorithm requires a certain scale of training samples, but does not analyze the data set before receiving the data to be predicted, but finds k sample points adjacent to the point in the parameter space after taking the data, and takes a weighted average of the k sample points as the predicted value of the data point. The minimum mean square error algorithm is optimized by utilizing nearest neighbor regression, filter parameters corresponding to single-frequency signals with different frequencies can be calculated as prior data sets according to given conditions such as sampling frequency, filter tap number and the like before noise reduction begins, acoustic parameters are extracted according to collected noise in the noise reduction process, the proper filter parameters are calculated by utilizing the nearest neighbor algorithm and a frequency synthesis method, the parameters are used as iteration initial values of the minimum mean square error algorithm, the iteration process of the minimum mean square error algorithm can be greatly reduced, and therefore the effect of accelerating system convergence is achieved.
The invention filters noise based on a finite impulse response (Finite Impulse Response, FIR) filter, solves the FIR filter tap weight vector corresponding to the noise to be controlled by utilizing a nearest neighbor regression algorithm and combining a minimum mean square error algorithm, controls a secondary sound source to actively emit control sound waves with the same phase opposite to the amplitude of the noise to be controlled, and overlaps the noise to be controlled at a control point, thereby achieving the purpose of eliminating the noise.
Based on the design concept, the invention provides an adaptive active noise control system based on inertia learning, which comprises at least one group of reference microphones 1, a controller 7, an actuator 8 and an error microphone 9 as shown in fig. 1.
The reference microphone 1 is disposed near the sound source, the error microphone 9 is disposed at the control point, collects the reference signal and the feedback signal to be controlled, and converts the collected signals into electrical signals to be input to the input end of the controller 7.
The controller 7 generates a control signal having the same phase as the amplitude of the noise signal to be controlled and inputs the control signal to the input of the actuator 8.
The actuator 8 converts the control signal into control sound waves, and the control sound waves are overlapped with noise to be controlled at a control point to play a silencing effect.
The controller 7 comprises an FFT circuit 2, an outlier detection circuit 3, an a priori data set and k-NN regressor circuit 4, an LMS filter circuit 5 and a driving circuit 6.
The FFT circuit 2 performs fast Fourier transform on a reference signal with a certain length, and inputs the calculated frequency spectrum into the outlier detection circuit 3;
the outlier detection circuit 3 extracts outliers in the frequency spectrum through a dbscan algorithm to serve as frequency input k-NN regressor circuits 4 of the line spectrum;
the frequency input k-NN regressor circuit 4 searches sample points adjacent to the frequency in the prior data set and performs weighted average to obtain an iteration initial value of the LMS algorithm; an error signal is acquired by an error microphone 9 and output to the LMS filter circuit 5.
The LMS filter circuit 5 re-iterates the LMS module tap weight vector according to the error signal and the reference signal, and calculates a control signal.
The drive circuit 6 outputs a control signal to the actuator 8.
The controller 7 may be integrated into a hardware chip in practical use.
Based on the adaptive active noise control system based on the inertia learning, the invention also provides an adaptive active noise control method based on the inertia learning, which comprises the following steps:
step 1, sampling rate is Fs, and frequency f is respectively calculated i ＝1，2，……，f m x Establishing single frequency signalsWherein the amplitude and phase of the signal do not affect the tap weight vector after the filter converges.
Step 2, the iteration step length of the self-adaptive filtering algorithm is mu, the length of the FIR filter is l, and each f is obtained through iteration by using the LMS algorithm i Corresponding tap weight vector omega i To convergence, a priori data set w= (ω) 1 ，ω 2 ，…，ω max ) T 。
Step 3, transmitting the noise x (t) to be controlled to the input end of the FFT circuit 2, obtaining a frequency spectrum F (t) of the moment through fast Fourier transform, and transmitting the obtained frequency spectrum F (t) to the input end of the outlier detection circuit 3;
step 4, the outlier detection circuit 3 performs cluster analysis on the frequency spectrum by using a dbscan algorithm and extracts n outliers o therein 1 ，o 2 ，…，o n Wherein o j (j=1, 2, …, n) represents the line spectrum signals of different frequencies constituting the band control noise, f j Representing the signal frequency.
Step 5, f j The input terminal of the input k-NN regressor circuit 4 searches the first-check data set for two sample points W (f k )，W(f k+1 ) Let f k ＜f j ＜f k+1 And according to the sum f j Is weighted averaged to obtain an estimated filter
ω j ＝(f n -f k )ω k +(f k+1 -f n )ω k+1 #(1)
The tap weight vector omega can be obtained according to the frequency synthesis method predict Represented as
Step 6, the tap weight vector omega predict As an iteration initial value, the input signal vector is taken as an input signal vector to be input into the LMS filter circuit 5
X(t)＝[x(t),x(t-1)，…，x(t-l+1)]#(3)
The output signal of the LMS filter circuit 5 is
Step 7, an output signal of the LMS filter circuit 5 is output by the actuator 8 through the driving circuit 6 to be a control sound wave, the control sound wave is overlapped with noise at a control point, and an error signal e (t) is obtained through the error microphone 9; the input LMS filter circuit 5 iterates the tap weight vector omega, and the iteration formula is that
ω′＝ω+2μe(t)x(t)#(5)
Where ω' is the new tap weight vector after the iteration.
The following gives a comparative example of the present invention for noise control with the conventional LMS algorithm, and the simulation of the intelligent active noise reduction adaptive active noise control system is performed in the Simulink module using MATLAB, and the simulation result is divided into three parts from top to bottom as shown in fig. 2: the uppermost is a reference signal waveform diagram, the middle is an error signal waveform diagram for noise control by using a conventional LMS algorithm, and the lowermost is an error signal waveform diagram for noise control by using the present invention. The noise to be controlled is set as a single-frequency signal with the frequency of 600.2Hz, the sampling rate of the system is 4000Hz, and the control time is 2sec; the horizontal axis of the coordinates is a time axis, the vertical axis is the amplitude of the error signal, and the smaller the vibration amplitude is, the more obvious the control effect is. Compared with the traditional algorithm, the new algorithm has obvious advantages in the aspects of convergence speed and control precision, and is more suitable for research and application in special acoustic environments such as underwater acoustic countermeasure and the like.
It should be emphasized that the examples described herein are illustrative rather than limiting, and therefore the invention includes, but is not limited to, the examples described in the detailed description, as other embodiments derived from the technical solutions of the invention by a person skilled in the art are equally within the scope of the invention.
Claims (4)
1. A control method of an adaptive active noise control system based on inertia learning is characterized by comprising the following steps: the control system comprises a reference microphone, a controller, an actuator and an error microphone; the reference microphone and the error microphone are respectively arranged at a position close to the sound source and a control point, respectively collect a reference signal and a feedback signal to be controlled, and convert the collected signals into electric signals to be input to the input end of the controller; the controller processes the reference signal and the feedback signal to generate a control signal with the same amplitude and opposite phase to the noise signal to be controlled, and the control signal is input to the input end of the actuator; the actuator converts the control signal into control sound waves, and the control sound waves are overlapped with noise to be controlled at a control point to eliminate the noise; the controller comprises an FFT circuit, an outlier detection circuit, a priori data set, a k-NN regressor circuit, an LMS filter circuit and a driving circuit; the FFT circuit performs fast Fourier transform on the reference signal and inputs the calculated frequency spectrum into the outlier detection circuit; the outlier detection circuit extracts outliers in the frequency spectrum through a dbscan algorithm to serve as frequency input k-NN regressor circuits of the line spectrum; the frequency input k-NN regressor circuit searches sample points adjacent to the frequency in the prior data set and carries out weighted average to obtain an iteration initial value of the LMS algorithm; the LMS filter circuit re-iterates the LMS module tap weight vector according to the error signal and the reference signal, and calculates a control signal; the driving circuit outputs a control signal to the actuator; the control method comprises the following steps:
step 1, sampling rate is Fs, and frequency f is respectively calculated i ＝1，2，......，f max Establishing single frequency signals
Step 2, taking the iteration step length of the adaptive filtering algorithm as mu, the length of the FIR filter as l, and respectively and iteratively solving the tap weight vector omega corresponding to each fi by using the LMS algorithm i To convergence, a priori data set w= (ω) 1 ，ω 2 ，...，ω max ) T ；
Step 3, transmitting the noise x (t) to be controlled to an input end of the FFT circuit, obtaining a frequency spectrum F (t) of the noise x (t) at the moment through fast Fourier transform, and transmitting the obtained frequency spectrum F (t) to an input end of the outlier detection circuit;
step 4, the outlier detection circuit performs cluster analysis on the frequency spectrum and extracts n outliers o in the frequency spectrum 1 ，o 2 ，...，o n Wherein o j Representing line spectrum signals constituting different frequencies with control noise, f j Which is indicative of the frequency of the signal, j=1, 2,. -%, n;
step 5,Will f j An input terminal of the input k-NN regressor circuit searches the prior-examination data set for two sample points W (f k )，W(f k+1 ) Let f k ＜f j ＜f k+1 And according to the sum f j The distance of (2) is weighted averaged to obtain an estimation filter:
ω j ＝(f n -f k )ω k +(f k+1 -f n )ω k+1
according to the frequency synthesis method, the tap weight vector omega predict Expressed as:
step 6, the tap weight vector omega predict As an iteration initial value, the method inputs an LMS filter circuit, takes an input signal vector as follows:
X(t)＝[x(t)，x(t-1)，…，x(t-l+1)]
the output signal of the LMS filter circuit is:
step 7, outputting an output signal of the LMS filter circuit into a control sound wave by an actuator through a driving circuit, superposing the control sound wave with noise at a control point, and acquiring an error signal e (t) through an error microphone; the input LMS filter circuit iterates the tap weight vector omega, and the iteration formula is as follows:
ω′＝ω+2μe(t)x(t)
where ω' is the new tap weight vector after the iteration.
2. The method for controlling an adaptive active noise control system based on lazy learning of claim 1, wherein: the controller is an integrated circuit chip.
3. The method for controlling an adaptive active noise control system based on lazy learning of claim 1, wherein: the amplitude and the phase of the signal in the step 1 do not affect the tap weight vector after the filter converges.
4. The method for controlling an adaptive active noise control system based on lazy learning of claim 1, wherein: and 4, the outlier detection circuit performs cluster analysis and extraction on the frequency spectrum by adopting a dbscan algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010629803.9A CN111862925B (en) | 2020-07-03 | 2020-07-03 | Adaptive active noise control system and method based on inertia learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010629803.9A CN111862925B (en) | 2020-07-03 | 2020-07-03 | Adaptive active noise control system and method based on inertia learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111862925A CN111862925A (en) | 2020-10-30 |
| CN111862925B true CN111862925B (en) | 2024-04-12 |
Family
ID=73151926
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010629803.9A Active CN111862925B (en) | 2020-07-03 | 2020-07-03 | Adaptive active noise control system and method based on inertia learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111862925B (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102098258A (en) * | 2009-12-14 | 2011-06-15 | 中兴通讯股份有限公司 | Method for removing narrow-band interference and self-adapting filter |
| CN104038181A (en) * | 2014-06-05 | 2014-09-10 | 北京航空航天大学 | Self-adapting filter construction method based on NLMS algorithm |
| CN105103568A (en) * | 2012-09-24 | 2015-11-25 | Actiwave公司 | Control and protection of loudspeakers |
| CN105261354A (en) * | 2015-09-09 | 2016-01-20 | 东南大学 | Adaptive active noise control system for active noise reduction and controlling method thereof |
| CN105913836A (en) * | 2016-04-07 | 2016-08-31 | 虞安波 | Fixed-point real time noise reduction method based on DSP |
| KR20180031332A (en) * | 2016-09-20 | 2018-03-28 | 전북대학교산학협력단 | METHOD AND SYSTEM FOR k-NN QUERY PROCESSING BASED ON GARBLED CIRCUIT |
| CN108761530A (en) * | 2018-05-22 | 2018-11-06 | 闽南师范大学 | A kind of seismic signal spectral factorization method |
| JP2019015953A (en) * | 2017-07-04 | 2019-01-31 | 鹿島建設株式会社 | Active noise control device and active noise control method |
| CN109547920A (en) * | 2018-09-30 | 2019-03-29 | 西安电子科技大学 | WiFi indoor locating system and method based on adaptive noise cancellation |
| CN109714023A (en) * | 2018-12-28 | 2019-05-03 | 歌尔股份有限公司 | Adaptive filter method, sef-adapting filter and noise control system |
| CN110866135A (en) * | 2019-11-12 | 2020-03-06 | 重庆邮电大学 | Response length hiding-based k-NN image retrieval method and system |
| CN111312275A (en) * | 2020-02-13 | 2020-06-19 | 大连理工大学 | Online sound source separation enhancement system based on sub-band decomposition |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2133866B1 (en) * | 2008-06-13 | 2016-02-17 | Harman Becker Automotive Systems GmbH | Adaptive noise control system |
| US8583429B2 (en) * | 2011-02-01 | 2013-11-12 | Wevoice Inc. | System and method for single-channel speech noise reduction |
| JP2016061868A (en) * | 2014-09-17 | 2016-04-25 | ソニー株式会社 | Noise suppression device, noise suppression method and program |
-
2020
- 2020-07-03 CN CN202010629803.9A patent/CN111862925B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102098258A (en) * | 2009-12-14 | 2011-06-15 | 中兴通讯股份有限公司 | Method for removing narrow-band interference and self-adapting filter |
| CN105103568A (en) * | 2012-09-24 | 2015-11-25 | Actiwave公司 | Control and protection of loudspeakers |
| CN104038181A (en) * | 2014-06-05 | 2014-09-10 | 北京航空航天大学 | Self-adapting filter construction method based on NLMS algorithm |
| CN105261354A (en) * | 2015-09-09 | 2016-01-20 | 东南大学 | Adaptive active noise control system for active noise reduction and controlling method thereof |
| CN105913836A (en) * | 2016-04-07 | 2016-08-31 | 虞安波 | Fixed-point real time noise reduction method based on DSP |
| KR20180031332A (en) * | 2016-09-20 | 2018-03-28 | 전북대학교산학협력단 | METHOD AND SYSTEM FOR k-NN QUERY PROCESSING BASED ON GARBLED CIRCUIT |
| JP2019015953A (en) * | 2017-07-04 | 2019-01-31 | 鹿島建設株式会社 | Active noise control device and active noise control method |
| CN108761530A (en) * | 2018-05-22 | 2018-11-06 | 闽南师范大学 | A kind of seismic signal spectral factorization method |
| CN109547920A (en) * | 2018-09-30 | 2019-03-29 | 西安电子科技大学 | WiFi indoor locating system and method based on adaptive noise cancellation |
| CN109714023A (en) * | 2018-12-28 | 2019-05-03 | 歌尔股份有限公司 | Adaptive filter method, sef-adapting filter and noise control system |
| CN110866135A (en) * | 2019-11-12 | 2020-03-06 | 重庆邮电大学 | Response length hiding-based k-NN image retrieval method and system |
| CN111312275A (en) * | 2020-02-13 | 2020-06-19 | 大连理工大学 | Online sound source separation enhancement system based on sub-band decomposition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111862925A (en) | 2020-10-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108428445B (en) | Self-adaptive active noise reduction method of error-free microphone | |
| CN109741727B (en) | Active noise reduction earphone based on active noise control algorithm, noise reduction method and storage medium | |
| CN109817209B (en) | Intelligent voice interaction system based on double-microphone array | |
| CN106875938B (en) | Improved nonlinear self-adaptive voice endpoint detection method | |
| US10037755B2 (en) | Method and system for active noise reduction | |
| CN108600894A (en) | A kind of earphone adaptive active noise control system and method | |
| Kim et al. | Active control of multi-tonal noise with reference generator based on on-line frequency estimation | |
| CN112053676A (en) | Nonlinear adaptive active noise reduction system and noise reduction method thereof | |
| Shi et al. | Active noise control in the new century: The role and prospect of signal processing | |
| Shen et al. | A momentum two-gradient direction algorithm with variable step size applied to solve practical output constraint issue for active noise control | |
| CN112270915A (en) | Active noise reduction method for indoor space | |
| Albu | The constrained stability least mean square algorithm for active noise control | |
| CN111862925B (en) | Adaptive active noise control system and method based on inertia learning | |
| Jabeen et al. | FxRLS algorithms based active control of impulsive noise with online secondary path modeling | |
| TWI459381B (en) | Speech enhancement method | |
| CN113078884B (en) | Adaptive algorithm adding nonlinear fitting | |
| CN109658947A (en) | A kind of active noise controlling method of synchronous modeling and control | |
| CN113851104B (en) | Feedback type active noise control system and method containing secondary channel online identification | |
| CN111193497A (en) | Secondary channel modeling method based on EMFNL filter | |
| Huang et al. | Adaptive active noise suppression using multiple model switching strategy | |
| CN113223552B (en) | Speech enhancement method, device, apparatus, storage medium, and program | |
| Kajikawa et al. | Active noise control without a secondary path model by using a frequency-domain simultaneous perturbation method with variable perturbation | |
| CN114333753A (en) | Microphone array-based wind turbine pipeline ANC system reference signal construction method | |
| JP3765567B2 (en) | Sound collection device, sound collection method, sound collection program, and recording medium | |
| TWI579833B (en) | Signal processing device and signal processing method |
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 |