Multi-channel adaptive active wind noise reduction car seat and noise reduction method thereof

Abstract

The application discloses a multi-channel adaptive active wind noise reduction automobile seat and a noise reduction method thereof. The wind noise reduction automobile seat comprises a multi-channel adaptive active wind noise reduction system. The multi-channel adaptive active wind noise reduction system comprises a reference microphone, a loudspeaker group, an error microphone group and a multi-channel controller. The reference microphone is arranged near a noise source and in a wind environment, and is used for collecting a wind noise signal as a reference signal. The loudspeaker group is arranged in or near a headrest of the automobile seat. The error microphone group is arranged in or near the headrest of the automobile seat, and is used for collecting a residual noise signal in a noise reduction area as an error signal. The input end of the multi-channel controller is connected with the reference microphone and the error microphone group respectively. The output end of the multi-channel controller is connected with the loudspeaker group. The multi-channel controller can inhibit wind noise through an adaptive filter and a decoupling filter, and can significantly improve the active noise reduction performance of an in-vehicle wind noise environment.

Description

Technical Field

[0001] This invention relates to the field of automotive noise reduction technology, and in particular to a multi-channel adaptive active wind noise reduction car seat and its noise reduction method. Background Technology

[0002] With the development of the automotive industry and smart cockpit technology, noise control within the cockpit has become a key issue in improving the driving experience and comfort. Active noise control technology has an effective ability to suppress low-frequency noise and shows significant potential in optimizing the in-vehicle acoustic environment. Compared to single-channel solutions, multi-channel active noise control systems, by arranging multiple speakers and error microphones around the seat, can create a larger quiet area near the ear. However, multi-channel active noise control for automotive seating scenarios still faces a series of practical challenges.

[0003] In multi-channel active noise control systems, centralized control, while exhibiting superior global noise reduction and system stability, suffers from high computational complexity, making it difficult to meet the real-time requirements of in-vehicle systems. In contrast, distributed control structures offer higher computational efficiency and are easier to implement on embedded platforms. However, the secondary channel coupling phenomenon prevalent in distributed systems not only affects control accuracy but can also lead to overall performance degradation due to local instability, particularly pronounced in the confined space and complex reflection environments of a cockpit.

[0004] To address the secondary channel coupling problem, one paper proposed an active noise reduction system with an inverse filter structure to overcome the limitations of traditional multi-channel active systems in terms of crosstalk (Bouchard M. and Feng Y., Inverse structure for active noise control and combined active noise control / sound reproduction systems, IEEE Transactions on Speech and Audio Processing, 2001). However, such methods often introduce long time delays and are limited by the length of the inverse filter, resulting in high crosstalk removal residuals, which restricts their application in practical environments.

[0005] On the other hand, to address the problem of sound source movement, existing research has proposed a hybrid system combining adaptive filters and temporal convolutional recurrent networks (DRNs) to decouple the nonlinear components in the secondary loudspeaker through neural networks (Chen D., Cheng L., Yao D., et al., A secondary path decoupled active noise control algorithm based on deep learning, IEEE Signal Processing Letters, 2022). Similar work utilizes dual-gated DRNNs to model the inverse response of the secondary path, allowing the controller to track only changes in the primary path (Park J., Choi JH, Kim Y., et al., HAD-ANC: A hybrid system comprising an adaptive filter and deep neural networks for active noise control, INTERSPEECH 2023, 2023). While these methods improve system adaptability, they currently mainly focus on single-channel applications and have not fully utilized the potential of multi-channel structures in terms of coverage and spatial equalization. Summary of the Invention

[0006] This application provides a multi-channel adaptive active wind noise reduction car seat and its noise reduction method, which can effectively reduce modeling errors and crosstalk interference, overcome the limitations of headrest speaker layout, track noise source changes in real time, and significantly improve the active noise reduction performance of the in-vehicle wind noise environment.

[0007] The first aspect of this application provides a multi-channel adaptive active wind noise reduction car seat, including a multi-channel adaptive active wind noise reduction system. The multi-channel adaptive active wind noise reduction system includes a reference microphone, a speaker group, an error microphone group, and a multi-channel controller. The reference microphone is located near the noise source and in a windy environment to collect wind noise signals as reference signals. The speaker group is located in or near the headrest of the car seat. The error microphone group is located in or near the headrest of the car seat to collect residual noise signals in the noise reduction area as error signals. The input terminal of the multi-channel controller is connected to the reference microphone and the error microphone group, respectively. The output terminal of the multi-channel controller is connected to the speaker group. The multi-channel controller suppresses wind noise through an adaptive filter and a decoupling filter.

[0008] In one possible implementation, the multi-channel controller includes: An initialization module is used to initialize the internal parameters and system state variables of the multi-channel controller, and to measure the secondary channel response from each speaker in the speaker group to the corresponding error microphone. The decoupling filter, based on the IIR filter, performs inverse modeling of the secondary channel response from each of the loudspeakers to the corresponding error microphone, and is used to generate a predicted inverse wind noise signal. An adaptive filter, based on a variable step-size adaptive method for FIR filters, dynamically adjusts the coefficients of the FIR filter according to the reference signal and the error signal; The output module is used to input the output signal of the adaptive filter into the decoupling filter to obtain a control signal output to the corresponding speaker, so as to drive the speaker to generate a secondary sound field to cancel wind noise.

[0009] In one possible implementation, the coefficients of the multi-channel controller specifically include: The IIR filter has pole coefficients. , is the result of calculation for the decoupling filter. Each channel coefficient K The number of channels, and the coefficient for each channel includes... L Modeling coefficients; The FIR filter has the value obtained from the adaptive filter calculation. K Each channel coefficient includes [number] channel coefficients, and each channel coefficient includes [number] channel coefficients. J Zero-point coefficients; in K These are the given user coefficients, i.e., the IIR and FIR filters are used in the anc decoupling.

[0010] In one possible implementation, the reference microphone is located on the window frame outside the car window, or at the junction of the window glass and the car body, or on the inner surface of the car body adjacent to the window glass, for directly or indirectly collecting wind noise signals transmitted into the car body through the window structure as the reference signal.

[0011] In one possible implementation, the error microphone group and the speaker group are arranged symmetrically on the left and right sides of the headrest; The speaker assembly is built into the forward-curving structure on both sides of the headrest, and its sound radiation direction is towards the left and right ear areas of the occupant. The error microphone is positioned in front of the corresponding speaker and closer to the occupant's ear than the corresponding speaker, and is used to collect residual noise signals near the occupant's corresponding ear.

[0012] A second aspect of this application provides a multi-channel adaptive active wind noise reduction method, which reduces wind noise using a multi-channel adaptive active wind noise reduction car seat as described above. The multi-channel adaptive active wind noise reduction method includes the following steps: S10. The multi-channel controller issues a command to acquire wind noise, the reference microphone picks up the wind noise and feeds it back to the multi-channel controller, and the multi-channel controller generates a predicted inverse wind noise signal through the adaptive filter. S20. The multi-channel controller sends the predicted anti-phase wind noise signal to the decoupling filter, then sends the filtered anti-phase wind noise signal to the speaker group, and feeds back the real-time wind noise error signal collected by the error microphone to the adaptive filter. The anti-phase wind noise emitted by the speaker group reaches the error microphone group through the secondary channel to suppress the wind noise at the error microphone.

[0013] In one possible implementation, the filtering method of the decoupling filter includes: The secondary channel decoupling IIR filtering is specifically as follows: Step 1: Generate inverse model estimation signal. In the initial stage of system operation, the initialization module inputs identification signal to the speaker group. And use the error microphone group to collect the response. ; Secondary channel transfer function matrix from the speaker group to the error microphone group Each element is subscript Indicates the secondary speaker serial number. Indicates the error microphone serial number; Step 2, inverse model estimation and iterative update, based on the identified... Calculate its approximate inverse model This makes its components ;Will The model is an autoregressive model, and the coefficients are updated in real time using the recursive least squares method.

[0014] In one possible implementation, the inverse model estimation and iterative update in step 2 specifically includes: At each sampling time : Step 2.1, calculate the system error, the inverse filter output is calculated as follows: ,in For order is L The modeling coefficient vector and the inverse model estimation error signal are: ; Step 2.2, for the coefficients The update is performed using the recursive least squares method, and the update formula is as follows: ; ; ; in, For the driving signal vector, for The equivalent inverse matrix of the autocorrelation matrix of the time-driven signal vector is obtained through iterative updating of the formula. It is the identity matrix of the corresponding dimension; Step 2.3 involves applying the modeling coefficients obtained in steps 2.1 and 2.2. As coefficients of the secondary channel inverse model, the output signal of the adaptive filter serves as the input of the decoupling filter and is processed by the secondary channel inverse model. Filtering is performed to obtain decoupled, independent signals for each channel, which are then fed back to the corresponding loudspeakers.

[0015] In one possible implementation, the filtering method of the adaptive filter includes: Step 1: In each control cycle, residual noise signals are collected using the error microphone group. ,in For the current moment, Set the channel number; initialize the adaptive filters for each channel. J order coefficient vector And set the initial step size. Step size variation parameters and minimum step size and maximum step size ; Step 2, the reference signal Through coefficient The adaptive filter generates the secondary control signal. The secondary control signal After being filtered by the decoupling filter, the signal is fed back to the power amplifier to drive the speaker group to generate a secondary sound field to cancel out wind noise. Step 3: Calculate the instantaneous power of the error signal. ,in It is a smoothing factor; Dynamically adjust the step size based on error power. ,in Forgetting factor of step size This is the gain coefficient; Limit the step size ; Step 4, using variable step size Update adaptive filter coefficients ,in for J 1st order vector The pre-filtered signal is the reference signal for the current period and a past period. The vector formed, where the past time period refers to the time from time n to time n-L+1; Step 5, update the filter coefficients For the filtering calculation at the next time step, repeat steps 2 to 4 to reduce the residual noise signal. The mean square value gradually decreases (generally it will be greater than 0), and the system tends to converge and stabilize.

[0016] Beneficial Effects: Compared with existing technologies, the multi-channel adaptive active wind noise reduction car seat and its noise reduction method provided in this application construct a multi-channel active noise control system by deploying speaker groups and error microphone groups in the head area of ​​the seat. The multi-channel controller consists of an adaptive filter and a decoupling filter connected in series. The former uses a variable step-size tracking strategy to estimate the primary channel response in real time, while the latter decouples the primary channel based on the inverse model of the secondary channel. This effectively reduces modeling errors and crosstalk interference, overcomes the limitations of headrest speaker layout, and tracks changes in noise sources in real time, thereby significantly improving the active noise reduction performance of the in-vehicle wind noise environment.

[0017] These and other objects, features and advantages of the present invention will become fully apparent from the following detailed description. Attached Figure Description

[0018] Figure 1a This illustration shows a structural schematic diagram of a multi-channel adaptive active wind noise reduction car seat according to a preferred embodiment of this application. Figure 1b A schematic diagram of the structure of a reference microphone installed in the window frame of a vehicle is shown.

[0019] Figure 2 This illustration shows a schematic diagram of the overall system architecture of a multi-channel adaptive active wind noise reduction car seat according to a preferred embodiment of this application.

[0020] Figure 3 The diagram illustrates the principle of an adaptive filter in a multi-channel adaptive active wind noise reduction car seat system according to a preferred embodiment of this application.

[0021] Figure 4 The diagram illustrates the principle of a decoupling filter for a multi-channel adaptive active wind noise reduction car seat system according to a preferred embodiment of this application.

[0022] Figure 5The diagram illustrates a filtering method for a decoupling filter in a multi-channel adaptive active wind noise reduction car seat system according to a preferred embodiment of this application.

[0023] Figure 6 The diagram illustrates a filtering method for an adaptive filter in a multi-channel adaptive active wind noise reduction car seat system according to a preferred embodiment of this application.

[0024] Figure 7 This diagram illustrates a time-domain comparison of the noise reduction performance of a multi-channel adaptive active wind noise reduction car seat and its noise reduction method according to a preferred embodiment of this application.

[0025] Figure 8 This diagram illustrates the frequency domain noise reduction performance of a multi-channel adaptive active wind noise reduction car seat and its noise reduction method according to a preferred embodiment of this application. Detailed Implementation

[0026] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0027] Those skilled in the art should understand that, in the disclosure of this specification, the terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the above terms should not be construed as limiting the present invention.

[0028] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. It will also be understood that terms, such as those defined in commonly used dictionaries, shall be interpreted as having a meaning consistent with their meaning in the context of the relevant art and shall not be interpreted as having an idealized or overly formal meaning unless expressly so defined herein.

[0029] It is understood that the term "a" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0030] Despite significant progress made by neural networks in acoustic channel modeling, a key but often overlooked problem remains in traditional distributed filtering algorithms: the inverse model of the secondary path typically relies on a linear approximation of the finite impulse response, which introduces substantial modeling errors in environments with significant reverberation, such as car cabins. Furthermore, acoustic crosstalk between channels further weakens the overall control performance. Therefore, achieving efficient, accurate, and crosstalk-suppressive inverse modeling of the secondary path in multi-channel active noise control systems integrated into car seats remains a critical technological bottleneck that urgently needs to be overcome in this field.

[0031] refer to Figure 1a and Figure 1b This invention discloses a multi-channel adaptive active wind noise reduction car seat, the core of which lies in the integration of a multi-channel adaptive active wind noise reduction system. This system mainly includes: a reference microphone 2, a speaker group 3, an error microphone group 4, and a multi-channel controller 1.

[0032] The multi-channel controller 1 is the core of the entire system, and it integrates an adaptive filter and a decoupling filter. The input terminals of the multi-channel controller 1 are connected to the reference microphone 2 and the error microphone group 4, respectively, for receiving the reference signal and the error signal; its output terminal is connected to the speaker group 3 for outputting the calculated inverse noise-reduced signal.

[0033] The reference microphone 2 is used to collect wind noise signals as a reference signal. Its key placement location is near the noise source and along the wind noise transmission path. It can be placed on the outer window frame of the vehicle window, at the junction of the window glass and the vehicle body sheet metal, or on the inner surface of the passenger compartment adjacent to the window glass. These locations can directly or indirectly, and more sensitively, capture wind noise signals transmitted into the passenger compartment via window structure vibrations or airflow through gaps, providing the system with advanced and accurate reference input.

[0034] The speaker group 3 and the error microphone group 4 are arranged in a coordinated manner in the noise reduction target area—that is, near the occupant's head. Specifically, they are integrated into or near the headrest 5 of the car seat.

[0035] according to Figure 1a and Figure 1b In the illustrated embodiment, the system employs a dual-channel configuration, comprising two speakers, two error microphones, one reference microphone, and a multi-channel controller. The speaker group 3 is integrated within the headrest 5, specifically located within a slightly forward-curved structure on the left and right sides of the headrest 5, with its sound radiation directed towards the left and right ear areas of the occupant.

[0036] The error microphone group 4 and the speaker group 3 are arranged symmetrically on the left and right sides of the headrest 5. Each error microphone 4 is located in front of the corresponding speaker 3 and is positioned closer to the occupant's ear than the speaker 3.

[0037] Figure 2 This is a schematic diagram of the overall system architecture of a multi-channel adaptive active wind noise reduction car seat according to a preferred embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the principle of the adaptive filter in one embodiment of the present invention; Figure 4 This is a schematic diagram of the decoupling filter described in one embodiment of the present invention.

[0038] like Figure 2 As shown in the figure express n The reference signal received by the constant reference microphone is transmitted through two paths: On the one hand, the reference signal passes through the primary channel between the noise source and the error microphone group. Then, a noise signal to be eliminated is generated at the error microphone group; On the other hand, the reference signal passes through an adaptive filter in the multi-channel controller. and decoupling filter The filter ultimately drives the speaker assembly to emit secondary sound, which then passes through the secondary channel between the speaker assembly and the error microphone assembly. This generates a canceling sound at the error microphone.

[0039] After the primary noise and the cancelling sound are superimposed, the residual noise signal collected by the error microphone group is: .

[0040] Figure 3 The adaptive filter is further illustrated. A schematic diagram illustrating the principle. The adaptive filter uses a reference signal. Input, output secondary control signal Meanwhile, adaptive filter Based on the error signal received by the corresponding error microphone The filter is implemented using a variable step size adaptive method based on a finite-length unit impulse response (FIR) filter, which can dynamically update its filter coefficients to minimize residual noise energy.

[0041] Figure 4 The decoupling filter is further illustrated. A schematic diagram illustrating the principle. The decoupling filter... Receive secondary control signals from the adaptive filter. and output the decoupled signal. This filter employs a secondary channel inverse model identification method based on infinite impulse response (IIR) filtering, which generates an identification signal. Acquire secondary channel response Calculate the inverse model estimation error signal and iteratively update the modeling coefficients. This enables effective compensation for acoustic coupling between multiple channels.

[0042] pass Figures 2 to 4 The combined explanation clearly illustrates the complete control process of this application, from noise acquisition, adaptive phase inversion generation, multi-channel decoupling processing to secondary sound field control, as follows: S10. The multi-channel controller issues a command to acquire wind noise, the reference microphone picks up the wind noise and feeds it back to the multi-channel controller, and the multi-channel controller generates a predicted inverse wind noise signal through the adaptive filter. S20. The multi-channel controller sends the predicted anti-phase wind noise signal to the decoupling filter, and then sends the filtered anti-phase wind noise signal to the loudspeaker group. The real-time wind noise error signal collected by the error microphone is fed back to the adaptive filter. The anti-phase wind noise emitted by the loudspeaker group reaches the error microphone through the secondary channel to suppress the wind noise at the error microphone group.

[0043] Figure 5 This is a schematic diagram of a decoupling filter method for a multi-channel adaptive active wind noise reduction automotive seat system according to an embodiment of the present invention. The decoupling filter method includes: The secondary channel decoupling IIR filtering is specifically as follows: Step 1: Generate inverse model estimation signal. In the initial stage of system operation, the initialization module inputs identification signal to the speaker group. And use the error microphone group to collect the response. ; Secondary channel transfer function matrix from the speaker group to the error microphone group Each element is subscript Indicates the secondary speaker serial number. Indicates the error microphone serial number; Step 2, inverse model estimation and iterative update, based on the identified... Calculate its approximate inverse model This makes its components subscript Indicates the secondary speaker serial number. Indicate the error microphone serial number; The model is an autoregressive model, and the coefficients are updated in real time using the recursive least squares method.

[0044] Step 2, inverse model estimation and iterative update, specifically includes: At each sampling time : Step 2.1, calculate the system error; the output signal of the inverse filter is calculated as follows: ,in For order is L The modeling coefficient vector has the following error: ; Step 2.2, for the coefficients The update is performed using the recursive least squares method, and the update formula is as follows: ; ; ; in, For the driving signal vector, for The equivalent inverse matrix of the autocorrelation matrix of the time-driven signal vector is obtained through iterative updating of the formula. It is the identity matrix of the corresponding dimension.

[0045] Step 2.3 involves applying the modeling coefficients obtained in steps 2.1 and 2.2. As coefficients of the secondary channel inverse model, the output signal of the adaptive filter serves as the input of the decoupling filter and is processed by the secondary channel inverse model. Filtering is performed to obtain decoupled, independent signals for each channel, which are then fed back to the corresponding loudspeakers.

[0046] Figure 6 This is a schematic diagram of an adaptive filter method for a multi-channel adaptive active wind noise reduction automotive seat system according to an embodiment of the present invention. The adaptive filter method includes: Step 1: In each control cycle, residual noise signals are collected using the error microphone group. ,in For the current moment, Set the channel number; initialize the adaptive filters for each channel. J order coefficient vector And set the initial step size. Step size variation parameters and minimum step size and maximum step size .

[0047] Step 2, the reference signal Through coefficient The adaptive filter generates the secondary control signal. The secondary control signal After being filtered by the decoupling filter, the signal is fed back to the power amplifier to drive the speaker array to generate a secondary sound field to cancel out wind noise.

[0048] Step 3: Calculate the instantaneous power of the error signal. ,in It is a smoothing factor; Dynamically adjust the step size based on error power. ,in Forgetting factor of step size This is the gain coefficient; Limit the step size .

[0049] Step 4, using variable step size Update adaptive filter coefficients ,in for J 1st order vector The pre-filtered signal is the reference signal for the current period and a past period. The vector formed.

[0050] Step 5, update the filter coefficients For the filtering calculation at the next time step, repeat steps 2 to 4 to reduce the residual noise signal. As the mean square value gradually decreases, the system tends to converge and stabilize.

[0051] Figure 7 This is a time-domain comparison diagram of the noise reduction performance of a multi-channel adaptive active wind noise reduction car seat and its noise reduction method according to an embodiment of the present invention. It shows the comparison results of the residual error changing over time between the system of the present invention and a system using the traditional FxLMS algorithm under the same wind noise input conditions. The red dashed line in the figure represents the residual error convergence process of the traditional FxLMS algorithm, while the blue solid line represents the residual error convergence process of the system proposed in this invention.

[0052] It can be observed that the traditional FxLMS algorithm takes about 15 seconds to converge to a residual error level of -39dB; in contrast, the system proposed in this invention shows a significant advantage in convergence performance, reaching the same residual error energy in about 6 seconds, with a significant improvement in convergence speed.

[0053] This result demonstrates that the method described in this invention, while maintaining the same noise reduction depth, has a faster system response and stabilization speed, further enhancing its real-time noise reduction capability and overall performance in dynamic wind noise environments.

[0054] Figure 8 This is a frequency domain schematic diagram illustrating the noise reduction performance of a multi-channel adaptive active wind noise reduction car seat and its noise reduction method according to an embodiment of the present invention. Figure 8 As shown in the diagram, this schematic diagram visually demonstrates the wind noise suppression effect before and after applying the system and method described in this invention through 1 / 3 octave band analysis.

[0055] The horizontal axis usually represents frequency, measured in Hertz (Hz), covering the main energy distribution range of wind noise (e.g., 100Hz to 1500Hz).

[0056] The vertical axis typically represents sound pressure level, measured in decibels (dB), used to quantify the intensity of noise. The baseline noise is represented by red bars. This represents the original wind noise sound pressure level collected at the occupant's ear error microphone under preset test conditions when the active wind noise reduction system of this invention is not enabled. The residual noise curve is represented by green bars, representing the residual noise spectrum measured at the same error microphone under the same test conditions after enabling the multi-channel adaptive active wind noise reduction system described in this invention.

[0057] A comparison of the two sets of bars clearly shows that, within the main frequency band of wind noise, the sound pressure level of the residual noise is significantly lower than that of the reference noise, especially in the 100Hz to 200Hz range where the noise reduction effect is most obvious, fully demonstrating the effective wind noise suppression capability of the system and method of this invention in a wide frequency range.

[0058] Therefore, the system proposed in this invention can effectively reduce modeling errors and crosstalk interference, overcome the limitations of headrest speaker layout, and significantly improve the active noise reduction performance of the in-vehicle wind noise environment.

[0059] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0060] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the invention. The advantages of the present invention have been fully and effectively realized. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments; any variations or modifications can be made to the implementation of the present invention without departing from these principles.

Claims

1. A multi-channel adaptive active wind noise reduction car seat, characterized in that, The system includes a multi-channel adaptive active wind noise reduction system, comprising a reference microphone, a speaker group, an error microphone group, and a multi-channel controller. The reference microphone is located near the noise source and in a windy environment to collect wind noise signals as a reference signal. The speaker group is located in or near the headrest of the car seat. The error microphone group is located in or near the headrest of the car seat to collect residual noise signals in the noise reduction area as error signals. The input terminals of the multi-channel controller are connected to the reference microphone and the error microphone group, respectively, and the output terminal of the multi-channel controller is connected to the speaker group. The multi-channel controller suppresses wind noise through adaptive filters and decoupling filters. The filtering method of the decoupling filter includes: The secondary channel decoupling IIR filtering is specifically as follows: Step 1: Generate inverse model estimation signal. In the initial stage of system operation, the identification signal is input to the speaker group through the initialization module. And use the error microphone group to collect the response. ; Secondary channel transfer function matrix between the speaker group and the error microphone group Each element is subscript Indicates the secondary speaker serial number. Indicates the error microphone serial number; Step 2, inverse model estimation and iterative update, based on the identified... Calculate its approximate inverse model This makes its components ;Will The model is constructed as an autoregressive model, and the coefficients are updated in real time using recursive least squares. The inverse model estimation and iterative update include: at each sampling time... The error of the system is calculated, and the output of the inverse filter is calculated as follows: ,in For order is L The modeling coefficient vector and the inverse model estimation error signal are: ; The filtering method of the adaptive filter includes: Step 1: In each control cycle, residual noise signals are collected using the error microphone group. ,in For the current moment, ` represents the channel number; initializes the adaptive filters for each channel. J order coefficient vector And set the initial step size. Step size variation parameters and minimum step size and maximum step size ; Step 2, refer to the signal Through coefficient The adaptive filter generates the secondary control signal. The secondary control signal After being filtered by the decoupling filter, the signal is fed back to the power amplifier to drive the speaker group to generate a secondary sound field to cancel out wind noise. Step 3: Calculate the instantaneous power of the error signal. ,in It is a smoothing factor; Dynamically adjust the step size based on error power. ,in Forgetting factor of step size This is the gain coefficient; Limit the step size ; Step 4, using variable step size Update adaptive filter coefficients ,in for J 1st order vector It is a vector composed of the current and past reference signals after pre-filtering; Step 5, update the filter coefficients For the filtering calculation at the next time step, repeat steps 2' to 4' to reduce the residual noise signal. As the mean square value gradually decreases, the system tends to converge and stabilize.

2. The multi-channel adaptive active wind noise reduction car seat as described in claim 1, characterized in that, The multi-channel controller includes: An initialization module is used to initialize the internal parameters and system state variables of the multi-channel controller, and to measure the secondary channel response from each speaker in the speaker group to the corresponding error microphone. A decoupling filter, based on an IIR filter, performs inverse modeling of the secondary channel response from each of the loudspeakers to the corresponding error microphone, to generate a predicted inverted wind noise signal; An adaptive filter, based on a variable step-size adaptive method for FIR filters, dynamically adjusts the coefficients of the FIR filter according to the reference signal and the error signal; The output module is used to input the output signal of the adaptive filter into the decoupling filter to obtain a control signal output to the corresponding speaker, so as to drive the speaker to generate a secondary sound field to cancel wind noise.

3. The multi-channel adaptive active wind noise reduction car seat as described in claim 2, characterized in that, The coefficients of the multi-channel controller specifically include: The IIR filter has pole coefficients. , is the result of calculation for the decoupling filter. Each channel coefficient K The number of channels, and the coefficient for each channel includes... L Modeling coefficients; The FIR filter has the value obtained from the adaptive filter calculation. K Each channel coefficient includes [number] channel coefficients, and each channel coefficient includes [number] channel coefficients. J Zero-point coefficients.

4. The multi-channel adaptive active wind noise reduction car seat as described in claim 3, characterized in that, The reference microphone is located on the outer window frame of the car window, or at the junction of the window glass and the car body, or on the inner surface of the car body adjacent to the window glass, and is used to directly or indirectly collect wind noise signals transmitted into the car body through the window structure as the reference signal.

5. The multi-channel adaptive active wind noise reduction car seat as described in claim 4, characterized in that, The error microphone group and the speaker group are arranged symmetrically on the left and right sides of the headrest; The speaker assembly is built into the forward-curving structure on both sides of the headrest, and its sound radiation direction is towards the left and right ear areas of the occupant. The error microphone is positioned in front of the corresponding speaker and closer to the occupant's ear than the corresponding speaker, and is used to collect residual noise signals near the occupant's corresponding ear.

6. A multi-channel adaptive active wind noise reduction method, wherein wind noise is reduced using a multi-channel adaptive active wind noise reduction car seat as described in any one of claims 3 to 5, characterized in that, The multi-channel adaptive active wind noise reduction method includes the following steps: S10, the multi-channel controller issues a command to acquire wind noise, the reference microphone picks up the wind noise and feeds it back to the multi-channel controller, and the multi-channel controller generates a predicted inverse wind noise signal through the adaptive filter; S20. The multi-channel controller sends the predicted anti-phase wind noise signal to the decoupling filter, then sends the filtered anti-phase wind noise signal to the speaker group, and feeds back the real-time wind noise error signal collected by the error microphone to the adaptive filter. The anti-phase wind noise emitted by the speaker group reaches the error microphone group through the secondary channel to suppress the wind noise at the error microphone.

7. The multi-channel adaptive active wind noise reduction method as described in claim 6, characterized in that, The filtering method of the decoupling filter includes: The secondary channel decoupling IIR filtering is specifically as follows: Step 1: Generate inverse model estimation signal. In the initial stage of system operation, the initialization module inputs identification signal to the speaker group. And use the error microphone group to collect the response. ; Secondary channel transfer function matrix from the speaker group to the error microphone group Each element is subscript Indicates the secondary speaker serial number. Indicates the error microphone serial number; Step 2, inverse model estimation and iterative update, based on the identified... Calculate its approximate inverse model This makes its components ;Will The model is an autoregressive model, and the coefficients are updated in real time using the recursive least squares method. The update formula is as follows: ； ； ； in, For the driving signal vector, for The equivalent inverse matrix of the autocorrelation matrix of the time-driven signal vector is obtained through iterative updating of the formula. It is the identity matrix of the corresponding dimension.

8. The multi-channel adaptive active wind noise reduction method as described in claim 7, characterized in that, The inverse model estimation and iterative update in step 2 specifically include: At each sampling time : Step 2.1, calculate the system error, the inverse filter output is calculated as follows: ,in For order is L The modeling coefficient vector and the inverse model estimation error signal are: ； Step 2.2, for the coefficients The update is performed using the recursive least squares method, and the update formula is as follows: ； ； ； in, For the driving signal vector, for The equivalent inverse matrix of the autocorrelation matrix of the time-driven signal vector is obtained through iterative updating of the formula. It is the identity matrix of the corresponding dimension; Step 2.3 involves applying the modeling coefficients obtained in steps 2.1 and 2.

2. As coefficients of the secondary channel inverse model, the output signal of the adaptive filter serves as the input of the decoupling filter and is processed by the secondary channel inverse model. Filtering is performed to obtain decoupled, independent signals for each channel, which are then fed back to the corresponding loudspeakers.

Images