Method, device, equipment and medium for determining a feedback active noise reduction headphone controller

Abstract

The invention belongs to the technical field of active noise control, and specifically relates to a method, device, equipment and medium for determining a controller of a feedback active noise reduction headphone. The present application obtains a secondary path data set of a target headphone, wherein the secondary path data set includes a plurality of primary data sets of different noise frequencies, each of the primary data sets includes a plurality of secondary data sets of different wearing methods, and each of the secondary data sets includes a plurality of secondary paths; constructs an initial controller of multiple control cores; determines the loss function of the initial controller according to the secondary path data set; optimizes each control core of the initial controller according to the loss function to obtain the most suitable controller for the target headphone. The present application uses the measured secondary path response to perform robust optimization of the feedback active noise control headphone, and gets rid of the problem of template design required in the traditional loop shaping method through the constraint design controller.

Description

Technical Field

[0001] The present invention belongs to the technical field of active noise control, and specifically relates to a method, device, equipment and medium for determining a feedback active noise reduction headphone controller. Background Art

[0002] Active noise cancellation (ANC) headphones have become one of the most popular hearing protectors due to their improved low-frequency noise reduction and listening experience in noisy environments. To achieve high noise reduction, a hybrid feedforward and feedback architecture is widely used in ANC headphones. This hybrid architecture uses a feedforward section to attenuate primary noise associated with a reference signal and a feedback section to eliminate residual noise not observed by the reference microphone. However, inappropriate feedback controller coefficients can interfere with the output of the feedforward section and even cause the ANC system to become unstable or diverge.

[0003] Many existing feedback controller design methods utilize a popular technique called loop shaping. The key concept of loop shaping is to adjust the open-loop response to a desired shape, thereby maintaining the closed-loop stability of the feedback system and achieving good noise reduction. However, shaping templates for the desired shape are not designed for specific headphones, making it difficult to ensure the stability of feedback ANC systems under extreme usage conditions. Furthermore, optimizing the template requires trial and error and relies heavily on personal experience. Summary of the Invention

[0004] In response to the above technical problems, the present invention proposes a method, device, equipment and medium for determining a controller of a feedback active noise reduction headphone. The present application obtains a secondary path data set of the target headphone, wherein the secondary path data set includes multiple primary data sets of different noise frequencies, each of the primary data sets includes multiple secondary data sets of different wearing methods, and each of the secondary data sets includes multiple secondary paths; constructs an initial controller with multiple control cores; determines the loss function of the initial controller according to the secondary path data set; optimizes each control core of the initial controller according to the loss function to obtain the most suitable controller for the target headphone. The present application uses the measured secondary path response to perform robust optimization of the feedback active noise control headphone. Through the constraint-designed controller, better noise reduction performance can be obtained, which gets rid of the problem of template design required in the traditional loop shaping method.

[0005] To solve the above technical problems, the technical solution adopted by the present invention includes four aspects.

[0006] In a first aspect, a method for determining a controller of a feedback active noise-cancelling headphone is provided, comprising: obtaining a secondary path data set of a target headphone, wherein the secondary path data set includes a plurality of primary data sets of different noise frequencies, each of the primary data sets includes a plurality of secondary data sets of different wearing methods, and each of the secondary data sets includes a plurality of secondary paths; constructing an initial controller with multiple control cores; determining a loss function of the initial controller based on the secondary path data set; and optimizing each control core of the initial controller based on the loss function to obtain an optimal controller for the target headphone.

[0007] In some embodiments, determining the loss function of the initial controller based on the secondary path data set includes: determining the sensitivity function of the target earphone at different noise frequencies based on each of the primary data sets; obtaining a preset weight function set, wherein the weight function set includes multiple weight functions, and each weight function corresponds to one primary data set; determining the objective function of the loss function based on the weight functions and the sensitivity functions corresponding to all primary data sets; determining the real eigenvalues and imaginary eigenvalues of each primary data set based on all secondary paths in each primary data set; determining a frequency threshold based on the real eigenvalues and the imaginary eigenvalues of all primary data sets; dividing all the primary data sets into linear dispersion data sets and circular dispersion data sets according to the noise frequency and the frequency threshold corresponding to each primary data set; determining a first constraint function of the loss function based on the sensitivity functions corresponding to all primary data sets; determining a second constraint function of the loss function based on the linear dispersion data set and the Nyquist point; determining a third constraint function of the loss function based on the circular dispersion data set and the Nyquist point; and determining the loss function based on the first constraint function, the second constraint function, the third constraint function, and the objective function.

[0008] In some embodiments, determining the second constraint function of the loss function based on the linearly dispersed data set and the Nyquist point includes: constructing a rectangular model based on each of the secondary data sets in the linearly dispersed data set; and determining the second constraint function based on the rectangular model and the Nyquist point.

[0009] In some embodiments, constructing a rectangular model based on each of the secondary data sets in the linearly dispersed data set includes: determining the long side slope and the wide side slope of the rectangular model, as well as the wide side intercept between the two wide sides according to the two secondary paths with the farthest Euclidean distance in the secondary data set; determining the long side intercept between the two long sides in the rectangular model according to the two secondary paths with the farthest Euclidean distance in the wide side direction projection of the secondary data set; and determining the rectangular model according to the long side slope, the long side intercept, the wide side slope, and the wide side intercept.

[0010] In some embodiments, the third constraint function of the loss function is determined based on the circular scattered data set and the Nyquist point, including: constructing a single circle model based on each of the first-level data sets in the circular scattered data set; and determining the third constraint function based on the single circle model and the Nyquist point.

[0011] In some embodiments, the wearing mode includes: a normal wearing mode; constructing a single circle model based on each of the first-level data sets in the circular scattered data sets includes: determining the origin position of the single circle model based on the second-level data set corresponding to the normal wearing mode in the first-level data set; determining the farthest secondary path farthest from the origin position in the second-level data sets corresponding to the remaining wearing modes, wherein the remaining wearing modes are other wearing modes except the normal wearing mode; determining the radius of the single circle model based on the distance between the farthest secondary path and the origin position; determining the single circle model based on the radius and the origin position.

[0012] In some embodiments, the initial controller is optimized according to the loss function to obtain the target earphone Optimal The controller includes: performing trial and error on the core parameters of each control core in the initial controller through a genetic algorithm; determining that the core parameters that minimize the loss function are optimal parameters; when the core parameters of each control core in the initial controller are all the optimal parameters, determining that the initial controller is the optimal controller for the target headset.

[0013] In second aspect, the present application provides a determination device for a feedback active noise reduction headphone controller, comprising: a first acquisition module, used to obtain a secondary path data set of a target headphone, wherein the secondary path data set includes multiple primary data sets of different noise frequencies, each of the primary data sets includes multiple secondary data sets of different wearing methods, and each of the secondary data sets includes multiple secondary paths; a first execution module, used to construct an initial controller of multiple control cores; a first determination module, used to determine the loss function of the initial controller based on the secondary path data set; a second execution module, used to optimize each control core of the initial controller based on the loss function to obtain the optimal controller of the target headphone.

[0014] In a third aspect, the present application proposes an electronic device comprising: a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the method described in any one of the first aspects is executed.

[0015] In a fourth aspect, the present application proposes a storage medium, which stores a computer program that can be executed by one or more processors, and the computer program can be used to implement any method described in the first aspect.

[0016] The beneficial effects created by the present invention are as follows: This application obtains a secondary path dataset of a target headset, wherein the secondary path dataset includes multiple primary datasets of different noise frequencies, each of the primary datasets includes multiple secondary datasets of different wearing styles, and each of the secondary datasets includes multiple secondary paths; constructs an initial controller with multiple control cores; determines a loss function of the initial controller based on the secondary path dataset; and optimizes each control core of the initial controller based on the loss function to obtain the optimal controller for the target headset. This application utilizes the measured secondary path response to perform robust optimization of feedback-type active noise control headsets. By constraining the designed controller, better noise reduction performance can be achieved, eliminating the problem of template design required in traditional loop shaping methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The scope of the present disclosure may be better understood by reading the following detailed description of exemplary embodiments in conjunction with the accompanying drawings, which include:

[0018] Figure 1 An overall flow chart of a method for determining a feedback active noise reduction headphone controller provided in an embodiment of the present application;

[0019] Figure 2 A schematic diagram of the structure of an ANC headset provided in an embodiment of the present application;

[0020] Figure 3A graph showing the variation of λ2 / λ1 with frequency provided in an embodiment of the present application;

[0021] Figure 4 A schematic diagram of a rectangular model provided in an embodiment of the present application;

[0022] Figure 5 A schematic diagram of a single circle model provided in an embodiment of the present application;

[0023] Figure 6 This is a structural block diagram of a determination device for a feedback active noise reduction headphone controller provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0025] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0026] If similar descriptions of "first\second\third" appear in the application documents, the following explanation will be added. In the following description, the terms "first\second\third" are only used to distinguish similar objects and do not represent a specific order for the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0028] Example 1:

[0029] Many existing feedback controller design methods utilize a popular technique called loop shaping. The key concept of loop shaping is to adjust the open-loop response to a desired shape, thereby maintaining the closed-loop stability and noise reduction effectiveness of the feedback system. While effective, these initial open-loop response shaping template designs struggle to ensure the stability of feedback ANC systems under extreme conditions. Furthermore, optimizing the template requires trial and error and relies heavily on personal experience.

[0030] In view of the problems existing in the existing technology, such as Figure 1 As shown, the present application provides a method for determining a feedback active noise-canceling headphone controller, the method being applied to an electronic device, which may be a server, a mobile terminal, a computer, a cloud platform, etc. The functions implemented by the device data processing provided in the embodiments of the present application can be implemented by a processor of the electronic device calling a program code, wherein the program code can be stored in a computer storage medium. The method for determining a feedback active noise-canceling headphone controller includes:

[0031] Step S1: Obtain a secondary path dataset of the target earphone, wherein the secondary path dataset includes multiple primary datasets of different noise frequencies, each of the primary datasets includes multiple secondary datasets of different wearing methods, and each of the secondary datasets includes multiple secondary paths.

[0032] When obtaining the secondary path data set, pink noise will be played through the speaker of the headset, and then the reception result of the headset feedback microphone will be obtained, and then the secondary path can be obtained. Since pink noise includes noises of various frequencies, the secondary path can be divided into secondary paths under different noise frequencies according to the noise frequency. There are generally three wearing methods, one is normal wearing, one is pressing wearing, and the other is loose wearing. Under different wearing methods, the secondary paths generated are different, and the present application also obtains multiple secondary paths under the same wearing method, so a data set corresponding to different wearing methods at the same noise frequency is formed, that is, a secondary data set. Each secondary data set includes the same number of secondary paths, and multiple secondary data sets at the same noise frequency form a primary data set. When the headset is fed back, as Figure 2 As shown in the figure, G(z) represents the secondary path, which is the transfer function from the speaker input to the feedback microphone output, and C(z) represents the controller, which can drive the speaker to generate the anti-noise signal y(n).

[0033] Step S2: Construct an initial controller of multiple control cores.

[0034] The controller is the core of improving the noise reduction effect of headphones. Therefore, before optimizing the controller parameters, it is necessary to establish an initial controller in advance. The initial controller may include multiple control cores. Therefore, in some embodiments, the controller established in this application uses a cascaded biquad IIR filter, which has the same structure as commercial ANC chips.

[0035] So the transfer function of the initial controller is:

[0036]

[0037] in Among them F s is the sampling frequency, f m is the center frequency, and the gain value g m (dB) determines the height or depth of the peaks and valleys, the quality factor Q m Determines the bandwidth of the peak or valley. Therefore, the kernel parameters of each control kernel are the parameter group [g m ,f m ,Q m ]. Then all parameters of the initial controller are determined by the parameter vector of length 3M+1:

[0038] α=[g1,f1,Q1,...,g m ,f m ,Q m ,...g M ,f M ,Q M ,A], where M is the number of control cores of the initial controller and m is the mth control core.

[0039] Step S3: Determine the loss function of the initial controller according to the secondary path data set.

[0040] The construction of the loss function is a key part of controller optimization, but the loss function needs to be established based on actual data. Therefore, in some embodiments, step S3 "determining the loss function of the initial controller based on the secondary path data set" includes:

[0041] Step S31: determining the sensitivity function of the target earphone at different noise frequencies according to each of the first-level data sets.

[0042] like Figure 2 As shown, the error signal e(n) measured by the feedback microphone placed near the speaker represents the acoustic superposition of the main noise d(n) and the anti-noise signal y(n) and can be expressed as follows:

[0043] E(z)=D(z)-E(z)C(z)G(z).。

[0044] Where E(z) and D(z) are the z-transforms of e(n) and d(n), respectively. Therefore, the closed-loop transfer function from primary noise d(n) to error signal e(n) can be obtained from the above equation as:

[0045]

[0046] This transfer function is also called the sensitivity function, and its modulus represents the amount of noise attenuation. G(z)C(z) represents the open-loop transfer function. If G(z) is relatively flat and has no phase shift, the controller gain can be set large enough to satisfy |S(z)| << 1, which is ideal noise reduction performance. However, due to circuit delays and sound wave propagation time, the secondary path G(z) inevitably has a certain delay, resulting in a phase shift that increases with frequency. When the G(z)C(z) phase shift approaches 180°, the negative feedback described above becomes positive feedback, causing the feedback control system to become unstable. In addition to the inherent delay, the amplitude and phase deviation of G(z)C(z) caused by changes in the secondary path is another key factor affecting the stability of the feedback ANC system. Therefore, in this application, it is necessary to obtain the sensitivity function at different noise frequencies. Since noise reduction only needs to ensure the best noise reduction effect under normal wearing conditions, the construction of the sensitivity function at different frequencies can be determined by only one secondary path under normal wearing conditions.

[0047] Step S32: obtaining a preset weight function set, wherein the weight function set includes a plurality of weight functions, and each weight function corresponds to one of the first-level data sets.

[0048] Step S33: determining the objective function of the loss function according to the weight functions and the sensitivity functions corresponding to all the primary data sets.

[0049] For the loss function, the objective function is very important, and the objective function is an important part of the loss function. Due to the different spectral characteristics of external noise, different weights can be given to the sensitivity functions under different noise frequencies, so a weight function set is preset in this application. The weight function set includes multiple weight functions, each weight function corresponds to a noise frequency. Therefore, the objective function obtained based on the weight function and sensitivity function under all noise frequencies is shown as follows:

[0050] Where S(jω) is the sensitivity function and W1(ω) is the weight function.

[0051] A loss function consisting solely of an objective function is flawed, so in this application, a penalty mechanism needs to be added to the loss function. Therefore, in steps S37, S38, and S39, a constraint function is determined as a penalty function for the loss function.

[0052] Step S34: determining the real eigenvalue and the imaginary eigenvalue of each of the first-level data sets according to all the secondary paths in each of the first-level data sets.

[0053] Step S35: determining a frequency threshold according to the real eigenvalues and the imaginary eigenvalues of all the primary data sets.

[0054] Step S36: According to the noise frequency and the frequency threshold corresponding to each of the first-level data sets, all the first-level data sets are divided into linear dispersion data sets and circular dispersion data sets.

[0055] Because the existing feedback controller design method is still not accurate enough in dividing the range of secondary path changes, this defect will affect the noise reduction performance of the feedback system. For example, the multi-circle model algorithm tends to repeatedly surround the discrete points of the secondary path under the same frequency, resulting in computational redundancy, and this shortcoming is particularly obvious when the secondary path data is linearly distributed in the complex plane. The single-circle model will have a large area of blank area surrounded, and the division of the uncertain area is also too broad. Therefore, in this application, different processing methods will be adopted for the secondary paths under different noise frequencies based on the distribution performance of the secondary paths under different noise frequencies.

[0056] Therefore, in some embodiments, in order to effectively characterize the distribution of data points, an obvious idea is to transform a coordinate system so that the projections of the data points in the new coordinate system are as separated as possible, that is, maximum separability. Therefore, this application changes the coordinate system of the secondary paths in each primary data set. After the coordinate system is changed, we find that the dispersion degree of the secondary paths in some primary data sets is circularly distributed, while the fractional degree of the secondary paths in some primary data sets is linearly distributed.

[0057] Therefore, this application determines the real eigenvalue and imaginary eigenvalue of each primary data set based on the secondary paths in the primary data set, and the ratio of the real eigenvalue to the imaginary eigenvalue can represent the degree of dispersion of the secondary paths in each primary data set in the changed coordinate system. When λ2 / λ1→1, the sample points (multiple secondary paths) are distributed in a circular shape in the new coordinate system. When λ2 / λ1→0, the sample points are distributed in a linear shape in the new coordinate system, where λ1 is the real eigenvalue and λ2 is the imaginary eigenvalue.

[0058] like Figure 3 As shown, Figure 3 The curve in the figure shows the value change of λ2(ω) / λ1(ω) under different noise frequencies, so according to Figure 3We set the positioning frequency threshold for the noise frequency corresponding to λ2(ω) / λ1(ω)=0.5. The primary dataset corresponding to the noise frequencies corresponding to the portion with λ2(ω) / λ1(ω)>=0.5 is set as a circular dispersion dataset, because the secondary paths in these datasets have a circular dispersion shape in the new coordinate system and can be processed using a single circle model. Similarly, the primary dataset corresponding to the noise frequencies corresponding to the portion with λ2(ω) / λ1(ω)<0.5 is set as a linear dispersion dataset, because the secondary paths in these datasets have a linear dispersion shape in the new coordinate system and can be processed using a rectangular model. We can then use different methods to determine the constraint function based on the degree of dispersion of different primary datasets.

[0059] Step S37: Determine the first constraint function of the loss function according to the sensitivity functions corresponding to all the first-level data sets.

[0060] The sensitivity function can form the objective function together with the weight function, but we also need to use the sensitivity function to form the constraint function. Therefore, the first constraint function in this application is as follows:

[0061] Where β represents the noise amplification factor, Π l represents a linearly dispersed data set, Π d Represents a circularly scattered dataset.

[0062] Step S38: determining a second constraint function of the loss function according to the linearly dispersed data set and the Nyquist point.

[0063] This application divides the secondary path data set into a linear dispersion data set and a circular dispersion data set through a frequency threshold, and uses different processing methods for data sets with different dispersion degrees to determine the constraint function of the loss function.

[0064] Therefore, in some embodiments, step S38 of “determining a second constraint function of the loss function according to the linearly dispersed data set and the Nyquist point” includes:

[0065] Step S381: constructing a rectangular model based on each of the secondary data sets in the linearly dispersed data sets.

[0066] In some embodiments, the step S381 of “constructing a rectangular model based on each of the secondary data sets in the linearly dispersed data set” includes:

[0067] Step S3811: Determine the long side slope and the broad side slope of the rectangular model, as well as the broad side intercept between the two broad sides, based on the two secondary paths with the longest Euclidean distance in the secondary data set.

[0068] Step S3812: Determine the long side intercept between the two long sides of the rectangular model according to the two secondary paths with the longest Euclidean distance in the wide side projection of the secondary data set.

[0069] Step S3813: Determine the rectangular model according to the long side slope, the long side intercept, the broad side slope and the broad side intercept.

[0070] Since the secondary paths in a linearly dispersed dataset are linearly distributed, the secondary paths in each secondary dataset will form an accumulation within a certain range. Therefore, we need to construct a rectangular model for each secondary dataset and then use the rectangular model to determine the second constraint function. Since the rectangular model is expressed in the form of a rectangle, there are many ways to determine the rectangle. Among them, this application uses a fixed point plus slope method to determine the rectangular model of the secondary path under each wearing method.

[0071] Therefore, for each secondary data set, we need to first obtain the two points with the longest Euclidean distance in the secondary data set. These two points are named G x1 and G x2 , and then based on these two points we can get the slope k of the long side and the slope of the wide side of the rectangle and the broadside intercept b of the equation of the line on which the two broadsides lie x1 ,b x2 Then find the two points with the longest Euclidean distance after projection in the broadside direction in the secondary data set, which are G x3 and G x4 Based on these two points, we can get the long side intercept b of the straight line equation where the two long sides are located. x3 ,b x4 Finally, the rectangle of the secondary data set is obtained according to the long side intercept, long side slope, wide side intercept and wide side slope. The result is as follows Figure 4 As shown, Figure 4 The three rectangles in the middle represent the corresponding rectangular models for different wearing methods, among which the rectangular model in the lower right corner is the effect after partial enlargement.

[0072] Step S382: Determine the second constraint function according to the rectangular model and the Nyquist point.

[0073] After obtaining the rectangular model in step S381, the second constraint function needs to be derived based on the rectangular model. According to the Nyquist stability criterion, the model constructed from the data of a stable feedback system should not contain the Nyquist point, that is, the rectangular model should not contain the point (-1,0). Therefore, based on the Nyquist point and the rectangular model, the second constraint function can be derived as follows:

[0074] (k(ω)-b x3 (ω))*(k(ω)-bx4 (ω))<0 or where π l Represents a linearly scattered dataset.

[0075] Step S39: determining a third constraint function of the loss function according to the circularly scattered data set and the Nyquist point.

[0076] Similarly, after frequency thresholding, we found that at certain noise frequencies, the dispersion shape of the secondary path is circular, so we divide these primary data sets with lower dispersion into circular dispersion data sets. For circular dispersion data sets, we use the method of establishing a single circle model to determine the constraint function of the loss function.

[0077] Therefore, in some embodiments, step S39 of “determining a third constraint function of the loss function based on the circularly dispersed data set and the Nyquist point” includes:

[0078] Step S391: constructing a single circle model according to each of the first-level data sets in the circular scattered data set.

[0079] In some embodiments, step S391 of “constructing a single circle model based on each of the first-level data sets in the circular scattered data set” includes:

[0080] Step S3911: Determine the origin position of the single circle model according to the secondary data set corresponding to the normal wearing mode in the primary data set.

[0081] Step S3922: determining the farthest secondary path farthest from the origin position in the secondary data sets corresponding to the remaining wearing modes, where the remaining wearing modes are wearing modes other than the normal wearing mode.

[0082] Step S3933: Determine the radius of the single circle model according to the distance between the farthest secondary path and the origin position.

[0083] Step S3934: Determine the single circle model according to the radius and the origin position.

[0084] The single circle model is a model that wraps all the secondary paths in a certain primary data set through a circular model, and the blank area in the circle needs to be reduced as much as possible. Since many wearing methods are based on the normal wearing method, such as loose wearing and tight wearing, which are the results of the normal wearing method extending in two directions, the secondary data set corresponding to the normal wearing method is selected in this application as the origin of the circle, where the origin is determined based on the distribution of secondary paths in the secondary data set of the normal wearing method. After determining the origin, in order to ensure that all secondary paths under the noise frequency can be wrapped by the circular model, it is also necessary to determine a radius. The radius is the distance of the secondary path farthest from the origin among all secondary paths. This allows the single circle model to achieve the effect of minimizing the blank area while wrapping all the secondary paths. The final single circle model is as follows: Figure 5 shown.

[0085] Step S392: Determine the third constraint function according to the single circle model and the Nyquist point.

[0086] Similarly, according to the Nyquist stability criterion, the model constructed from the data in a stable feedback system should not include the Nyquist point, that is, the single-circle model should not include the (-1,0) point. Therefore, the third constraint function determined in this application based on the single-circle model and the Nyquist point is as follows:

[0087] Where γ represents the coefficient, G0(jω) is the origin position, B(ω) is the radius, and C(jω) is the feedback control output of the controller.

[0088] Step S310: determining the loss function according to the first constraint function, the second constraint function, the third constraint function and the objective function.

[0089] After determining the objective function, the first constraint function, the second constraint function, and the third constraint function, the loss function can be determined. The determined loss function is as follows:

[0090] In this formula, the second constraint function and the third constraint function are expressed by P.

[0091] Where ε(·) is the penalty function, defined as:

[0092]

[0093] p i (ω) is the penalty value of the three-rectangle model under three headphone wearing methods, defined as:

[0094]

[0095] Where G left (jω)*C(jω) is the leftmost point of the four vertices of the rectangle, which is the real(G left (jω)*C(jω))<-1, so |G left (jω1)*C(jω1)|>1, according to the penalty function ε(·), a larger penalty value will be generated. Using this method to obtain the penalty value makes the penalty value change more reasonable.

[0096] Step S4: Optimizing each control core of the initial controller according to the loss function to obtain the optimal controller for the target headset.

[0097] In some embodiments, step S4 of "optimizing each control core of the initial controller according to the loss function to obtain the optimal controller for the target headset" includes:

[0098] Step S41: performing trial and error on the core parameters of each control core in the initial controller by using a genetic algorithm.

[0099] Step S42: Determine the kernel parameters that minimize the loss function as the optimal parameters.

[0100] Step S43: When the core parameters of each control core in the initial controller are all the optimal parameters, the initial controller is determined to be the optimal controller for the target headset.

[0101] In order to obtain the most suitable controller, the present application needs to make the parameters of each control core in the initial controller the optimal parameters. Therefore, the present application uses the trial and error method to make the parameter group of each controller core parameter [g m ,f m ,Q m ] changes and records the value of the loss function. When the value of the loss function is minimum, the value in the parameter group at this time is determined to be the optimal parameter of the control core. Then the above steps are applied to the next control core until the parameters of all control cores are optimal parameters. Then, the optimal controller can be obtained. Of course, since the existing technology generally manually adjusts the parameters in the parameter group by observing the power spectrum of the error signal, but this method is inefficient, the present application adopts a genetic algorithm for automatic trial and error to achieve automatic optimization of the controller.

[0102] This application eliminates the template design issues inherent in traditional loop shaping methods. Based on the principle of maximum separability, data points at different frequencies are divided into two distribution shapes: circular and linear. We then propose a three-rectangle model to more accurately delineate the uncertainty range of the linearly distributed secondary path, while using the traditional single-circle model to estimate the uncertainty range of the circularly distributed secondary path. By employing a constrained feedback controller, we achieve even better noise reduction performance.

[0103] Example 2:

[0104] Based on the foregoing embodiments, an embodiment of the present application provides a device for determining a feedback active noise reduction headphone controller. The modules included in the device, and the units included in each module, can be implemented by a processor in a computer device; of course, they can also be implemented by a specific logic circuit; in the implementation process, the processor can be a central processing unit (CPU), a microprocessor (MPU), a digital signal processor (DSP), or a field programmable gate array (FPGA), etc.

[0105] like Figure 6 As shown, a determination device for a feedback active noise reduction headphone controller includes: a first acquisition module 1, a first execution module 2, a first determination module 3 and a second execution module 4.

[0106] The first acquisition module 1 is used to obtain a secondary path dataset of the target earphone, wherein the secondary path dataset includes multiple primary datasets of different noise frequencies, each of the primary datasets includes multiple secondary datasets of different wearing styles, and each of the secondary datasets includes multiple secondary paths. The first execution module 2 is used to construct an initial controller with multiple control cores. The first determination module 3 is used to determine the loss function of the initial controller based on the secondary path dataset. The second execution module 4 is used to optimize each control core of the initial controller based on the loss function to obtain the optimal controller for the target earphone.

[0107] Each module in the determination device of the above-mentioned feedback active noise reduction headphone controller can be implemented in whole or in part by software, hardware, or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the device in hardware form, or can be stored in the memory of the processing device in software form, so that the processor can call and execute the operations corresponding to the above modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical function division. In actual implementation, there may be other division methods.

[0108] Example 3:

[0109] A third aspect provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of a method for determining a feedback active noise reduction headphone controller are implemented.

[0110] Example 4:

[0111] A fourth aspect provides a storage medium, which stores a computer program that can be executed by one or more processors. The computer program can be used to implement the steps of any one of the methods for determining a feedback active noise reduction headphone controller in the first aspect.

[0112] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0113] It should be understood that "one embodiment" or "an embodiment" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. The above-mentioned serial numbers of the embodiments of the present application are for description only and do not represent the advantages and disadvantages of the embodiments.

[0114] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0115] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0116] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed across multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0117] In addition, all functional units in the embodiments of the present application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the above-mentioned integrated units can be implemented in the form of hardware or in the form of hardware plus software functional units.

[0118] Those skilled in the art will understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROMs), magnetic disks, optical disks, and other media that can store program codes.

[0119] Alternatively, if the above-mentioned integrated unit of the present application is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a controller to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks or optical disks.

[0120] The above is merely an embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for determining a feedback active noise reduction headphone controller, characterized in that: include: Obtaining a secondary path dataset of the target headset, wherein the secondary path dataset includes multiple primary datasets of different noise frequencies, each of the primary datasets includes multiple secondary datasets of different wearing styles at the same noise frequency, and each of the secondary datasets includes multiple secondary paths; Build an initial controller for multiple control cores; determining a loss function of the initial controller based on the secondary path data set; Each control core of the initial controller is optimized according to the loss function to obtain an optimal controller for the target headset.

2. The method according to claim 1, characterized in that The determining the loss function of the initial controller according to the secondary path data set includes: Determining the sensitivity function of the target earphone at different noise frequencies according to each of the primary data sets; Obtaining a preset weight function set, wherein the weight function set includes a plurality of weight functions, and each weight function corresponds to one of the primary data sets; Determine the objective function of the loss function according to the weight function and the sensitivity function corresponding to all primary data sets; Determine the real eigenvalue and the imaginary eigenvalue of each of the first-level data sets according to all secondary paths in each of the first-level data sets; Determine a frequency threshold according to the real eigenvalues and the imaginary eigenvalues of all the primary data sets; According to the noise frequency and the frequency threshold corresponding to each of the primary data sets, all the primary data sets are divided into linear dispersion data sets and circular dispersion data sets; Determine a first constraint function of the loss function according to the sensitivity functions corresponding to all primary data sets; Determine a second constraint function of the loss function according to the linearly dispersed data set and the Nyquist point; Determine a third constraint function of the loss function according to the circularly scattered data set and the Nyquist point; The loss function is determined according to the first constraint function, the second constraint function, the third constraint function and the objective function.

3. The method according to claim 2, characterized in that The second constraint function of the loss function is determined according to the linearly dispersed data set and the Nyquist point, comprising: constructing a rectangular model based on each of the secondary data sets in the linearly dispersed data sets; The second constraint function is determined according to the rectangular model and the Nyquist point.

4. The method according to claim 3, characterized in that The step of constructing a rectangular model based on each of the secondary data sets in the linearly dispersed data sets includes: Determine the long side slope and the broad side slope of the rectangular model, and the broad side intercept between the two broad sides according to the two secondary paths with the farthest Euclidean distance in the secondary data set; Determine the long side intercept between two long sides of the rectangular model according to the two secondary paths with the longest Euclidean distance in the broadside projection of the secondary data set; The rectangular model is determined according to the long side slope, the long side intercept, the broad side slope, and the broad side intercept.

5. The method according to claim 2, characterized in that The third constraint function of determining the loss function according to the circularly dispersed data set and the Nyquist point includes: constructing a single circle model according to each of the first-level data sets in the circular scattered data sets; The third constraint function is determined according to the single circle model and the Nyquist point.

6. The method according to claim 5, characterized in that The wearing mode includes: a normal wearing mode; and constructing a single circle model according to each of the first-level data sets in the circular scattered data sets includes: Determine the origin position of the single circle model according to the secondary data set corresponding to the normal wearing mode in the primary data set; Determining the farthest secondary path farthest from the origin position in the secondary data sets corresponding to the remaining wearing modes, wherein the remaining wearing modes are wearing modes other than the normal wearing mode; Determining the radius of the single circle model according to the distance between the farthest secondary path and the origin position; The single circle model is determined according to the radius and the origin position.

7. The method according to claim 1, characterized in that Optimizing the initial controller according to the loss function to obtain the optimal controller for the target headset includes: Performing trial and error on the core parameters of each control core in the initial controller by using a genetic algorithm; Determining the kernel parameters that minimize the loss function as optimal parameters; When the core parameters of each control core in the initial controller are all the optimal parameters, the initial controller is determined to be the optimal controller for the target headset.

8. A device for determining a feedback active noise reduction headphone controller, characterized in that: include: A first acquisition module is configured to acquire a secondary path dataset of a target headset, wherein the secondary path dataset includes a plurality of primary datasets of different noise frequencies, each of the primary datasets includes a plurality of secondary datasets of different wearing styles at the same noise frequency, and each of the secondary datasets includes a plurality of secondary paths; A first execution module is used to construct an initial controller of multiple control cores; a first determining module, configured to determine a loss function of the initial controller based on the secondary path data set; The second execution module is configured to optimize each control core of the initial controller according to the loss function to obtain an optimal controller for the target headset.

9. An electronic device, characterized in that: include: A memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the method according to any one of claims 1 to 7 is performed.

10. A storage medium, characterized in that: The computer program stored in the storage medium can be executed by one or more processors, and the computer program can be used to implement the method according to any one of claims 1 to 7.

Images