CN117334177A

CN117334177A - Method and system for reducing wind noise of vehicle

Info

Publication number: CN117334177A
Application number: CN202210704715.XA
Authority: CN
Inventors: 黄智超; 何梦楠; 龚彩霞
Original assignee: Beijing Didi Infinity Technology and Development Co Ltd
Current assignee: Beijing Didi Infinity Technology and Development Co Ltd
Priority date: 2022-06-21
Filing date: 2022-06-21
Publication date: 2024-01-02

Abstract

Embodiments of the present disclosure provide a method for reducing wind noise of a vehicle, the method including: acquiring a first wind noise signal acquired by a first microphone for representing wind noise near a vehicle window, wherein the first microphone is positioned near the vehicle window; determining a second wind noise signal of an in-vehicle region according to a neural network model and the first wind noise signal, wherein the neural network model constructs nonlinear changes of wind noise transmitted from the vicinity of the vehicle window to the in-vehicle region; determining a noise reduction signal according to the second wind noise signal so as to drive a loudspeaker to generate noise reduction sound waves, wherein the loudspeaker is positioned in the vehicle; acquiring an error signal acquired by a second microphone for representing residual noise in the vehicle interior region, wherein the residual noise in the vehicle interior region is derived from superposition of wind noise and noise reduction waves, and the second microphone is positioned in the vehicle interior region; and updating the noise reduction signal according to the error signal.

Description

Method and system for reducing wind noise of vehicle

Technical Field

The present disclosure relates to the field of vehicle noise reduction, and in particular, to a method and system for reducing wind noise of a vehicle.

Background

In the course of running the vehicle, noise, such as engine noise, road noise, wind noise, etc., is inevitably generated in the vehicle. Active noise reduction of a vehicle means that a plurality of speakers are installed in the vehicle, and sound waves opposite to noise are played by the speakers, so that the opposite sound waves cancel out the noise in a noise reduction area (e.g., an in-vehicle area), thereby reducing or eliminating the noise in the noise reduction area. The active noise reduction of the vehicle can be divided into multi-region noise reduction and single-region noise reduction according to the noise reduction region division. The single-zone noise reduction is usually the noise reduction of the pointer to the cab, and the multi-zone noise reduction also takes into account the noise reduction of the passenger zone. On the one hand, the current vehicle noise reduction is mainly aimed at engine noise and road noise, but rarely involves wind noise. On the other hand, conventional vehicle noise reduction is mostly noise reduction for the cab region, and is rarely involved for other regions, which may cause noise of the cab to be suppressed while increasing noise of other regions.

Accordingly, it is desirable to provide a method of reducing wind noise in a vehicle that is capable of reducing wind noise in one or more areas of the vehicle.

Disclosure of Invention

One of the embodiments of the present specification provides a method of reducing wind noise of a vehicle, the method comprising: acquiring a first wind noise signal acquired by a first microphone for representing wind noise near a vehicle window, wherein the first microphone is positioned near the vehicle window; determining a second wind noise signal of an in-vehicle region according to a neural network model and the first wind noise signal, wherein the neural network model constructs nonlinear changes of wind noise transmitted from the vicinity of the vehicle window to the in-vehicle region; determining a noise reduction signal according to the second wind noise signal so as to drive a loudspeaker to generate noise reduction sound waves, wherein the loudspeaker is positioned in the vehicle; acquiring an error signal acquired by a second microphone for representing residual noise in the vehicle interior region, wherein the residual noise in the vehicle interior region is derived from superposition of wind noise and noise reduction waves, and the second microphone is positioned in the vehicle interior region; and updating the noise reduction signal according to the error signal.

In some embodiments, the first microphone includes a plurality of first microphones located near different windows, the second microphone includes a plurality of second microphones located in different in-vehicle regions, each of the plurality of second microphones corresponds to one of the plurality of first microphones, and the neural network model constructs a nonlinear change from any one of the first microphones to its corresponding second microphone.

In some embodiments, the neural network model includes a deep complex convolution loop network model trained from first wind noise sample data near the vehicle window and second wind noise sample data of the in-vehicle region during the vehicle driving.

In some embodiments, the method further comprises: acquiring a transfer function of the noise reduction sound wave transferred from the loudspeaker position to the in-vehicle area, wherein determining the noise reduction signal comprises: the noise reduction signal is determined based on the transfer function and the second wind noise signal.

In some embodiments, the transfer function is determined based on a swept frequency signal of the speaker and a sound signal collected by the second microphone.

In some embodiments, updating the noise reduction signal based on the error signal comprises: and updating the noise reduction signal according to the error signal by using an adaptive filter.

In some embodiments, the adaptive filter updates filter coefficients according to an LMS algorithm, the filter coefficients being related to weight values for different regions within the vehicle, the weight values for the different regions reflecting differences in noise reduction capabilities of the different regions.

In some embodiments, the iteration step of the filter coefficients is related to the vehicle speed.

Some embodiments of the present specification also provide a system for reducing wind noise in a vehicle, the system comprising at least one first microphone located near a vehicle window, at least one second microphone located in an in-vehicle region, a speaker, and a processor configured to: acquiring a first wind noise signal acquired by the first microphone and used for representing wind noise near the vehicle window; determining a second wind noise signal of the in-vehicle region according to a neural network model and the first wind noise signal, wherein the neural network model constructs nonlinear changes of wind noise transmitted from the vicinity of the vehicle window to the in-vehicle region; determining a noise reduction signal according to the second wind noise signal so as to drive the loudspeaker to generate noise reduction sound waves; acquiring an error signal acquired by the second microphone and used for representing residual noise in the vehicle interior region, wherein the residual noise in the vehicle interior region is derived from superposition of wind noise and the noise reduction wave; and updating the noise reduction signal according to the error signal.

Some embodiments of the present disclosure also provide a computer-readable storage medium storing computer instructions that, when read by a computer, perform a method as in any one of the embodiments above.

The embodiments of the present specification also provide a vehicle including: at least one first microphone configured to acquire a first wind noise signal indicative of wind noise near a vehicle window, wherein the first microphone is located near the vehicle window; a speaker configured to generate a noise reduction acoustic wave under driving of a noise reduction signal, wherein the noise reduction signal is determined based on a second wind noise signal of an in-vehicle region, the second wind noise signal being determined based on a neural network model and the first wind noise signal; at least one second microphone configured to acquire an error signal representing residual noise of the in-vehicle region, the residual noise of the in-vehicle region resulting from a superposition of wind noise and noise-reducing waves, wherein the second microphone is located in the in-vehicle region; and the neural network model constructs nonlinear changes of wind noise transmitted from the vicinity of the vehicle window to the in-vehicle area.

Drawings

The present specification will be further elucidated by way of example embodiments, which will be described in detail by means of the accompanying drawings. The embodiments are not limiting, in which like numerals represent like structures, wherein:

FIG. 1 is a diagram of an application scenario of a vehicle wind noise reduction system according to some embodiments of the present description;

FIG. 2 is an exemplary frame diagram of a vehicle wind noise reduction system according to some embodiments of the present disclosure;

FIG. 3 is an exemplary flow chart of a method of reducing wind noise in a vehicle according to some embodiments of the present disclosure;

FIG. 4 is an exemplary flow chart of an adaptive filter updating noise reduction signal according to some embodiments of the present description;

FIG. 5 is an exemplary flow chart of a neural network model training method, according to some embodiments of the present description.

Detailed Description

In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the drawings that are required to be used in the description of the embodiments will be briefly described below. It is apparent that the drawings in the following description are only some examples or embodiments of the present specification, and it is possible for those of ordinary skill in the art to apply the present specification to other similar situations according to the drawings without inventive effort. Unless otherwise apparent from the context of the language or otherwise specified, like reference numerals in the figures refer to like structures or operations.

It will be appreciated that "system," "apparatus," "unit" and/or "module" as used herein is one method for distinguishing between different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

As used in this specification and the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.

A flowchart is used in this specification to describe the operations performed by the system according to embodiments of the present specification. It should be appreciated that the preceding or following operations are not necessarily performed in order precisely. Rather, the steps may be processed in reverse order or simultaneously. Also, other operations may be added to or removed from these processes.

FIG. 1 is a diagram of an application scenario of a vehicle wind noise reduction system according to some embodiments of the present description. As shown in fig. 1, an application scenario 100 of a vehicle wind noise reduction system may include a processing device 110, a network 120, a storage device 130, and a vehicle 140.

In some embodiments, one or more components of the application scenario 100 may be connected and/or in communication with each other via a network 120 (e.g., a wireless connection, a wired connection, or a combination thereof). As shown in fig. 1, processing device 110 may be connected to storage device 130 through network 120. For another example, the processing device 110 may be connected to the vehicle 140 via the network 120 so that the processing device 110 can analyze and process information (e.g., a first wind noise signal, an error signal, a noise reduction signal, etc.) acquired by the vehicle 140.

The processing device 110 may be configured to process information and/or data related to the application scenario 100, such as a first wind noise signal, an error signal, a noise reduction signal, etc. In some embodiments, processing device 110 may process data, information, and/or processing results obtained from other devices or system components and execute program instructions based on such data, information, and/or processing results to perform one or more functions described herein.

The network 120 may connect components of the application scenario 100 and/or connect the application scenario 100 with external resources. The network 120 enables communication between the various components and other components outside the application scenario 100 to facilitate the exchange of data and/or information. In some embodiments, network 120 may be a local area network, a wide area network, the internet, etc., or a combination of network structures.

Storage device 130 may be used to store data and/or instructions. In some embodiments, the storage device 130 may store data and/or instructions that the processing device 110 uses to perform or use to accomplish the exemplary methods described in this specification. In some embodiments, the storage device 130 may be connected to the network 120 to communicate with one or more components of the application scenario 100 (e.g., the processing device 110, the vehicle 140).

The vehicle 140 may be used to acquire sound signals from different areas within the vehicle. In some embodiments, the vehicle 140 may include a first microphone 141 and a second microphone 142. The first microphone 141 may be located near the window to collect a sound signal (also referred to as a first wind noise signal) that characterizes wind noise near the window. The second microphone 142 may be located in an interior area to collect sound signals representative of the interior area. For example, the second microphone 142 may also collect a sound signal (also referred to as an error signal) that characterizes residual noise after noise reduction in the interior region of the vehicle.

In some embodiments, the sound signals collected by the first microphone 141 and the second microphone 142 may be transferred to the processing device 110 through the network 120 to enable the processing device 110 to process the sound signals. For example, the processing device 110 may process the sound signals collected by the first microphone 141 and/or the second microphone 142 to obtain parameter information (e.g., amplitude, phase, etc.) of the corresponding sound signals. For another example, the processing device 110 may determine a signal representative of wind noise of the in-vehicle region (i.e., a second wind noise signal) based on the sound signal collected by the first microphone 141 (i.e., the first wind noise signal), and generate a noise reduction signal therefrom. For example, the phase of the noise reduction sound wave generated by the noise reduction signal transmitted to the in-vehicle region is opposite or approximately opposite to the phase of the second wind noise signal; the amplitude of the noise-reduced sound wave generated by the noise-reduced signal transmitted to the interior region of the vehicle is equal to or approximately equal to the amplitude of the second wind noise signal. In some embodiments, vehicle 140 may also include speakers 143. The speaker 143 may be used to generate noise-reducing sound waves driven by the noise-reducing signal, where the noise-reducing sound waves and wind noise near the window may interfere with each other in the interior region (e.g., the cab position, the passenger position) when the wind noise is transmitted to the interior region, so as to cancel each other out or partially cancel each other out, thereby reducing the wind noise of the vehicle.

It should be noted that the application scenario is provided for illustrative purposes only and is not intended to limit the scope of the present description. Many modifications and variations will be apparent to those of ordinary skill in the art in light of the present description. For example, the application scenario may also include a control component to control the operational state of the microphone and speaker. As another example, application scenarios may be implemented on other devices to implement similar or different functionality. However, variations and modifications do not depart from the scope of the present description.

FIG. 2 is an exemplary frame diagram of a vehicle wind noise reduction system according to some embodiments of the present disclosure. As shown in fig. 2, the vehicle wind noise reduction system 200 may include a first microphone 210, a second microphone 220, and a speaker 230.

In some embodiments, the vehicle wind noise reduction system 200 may be applied to vehicle active noise reduction. By utilizing the principle of cancellation of acoustic wave interference, the speaker 230 may generate a sound signal that is homologous to wind noise and has a certain phase difference (for example, opposite phase or approximately opposite phase), so that when the sound signal output by the speaker 230 and wind noise are respectively transferred to the noise reduction area, the sound signal and the wind noise can be mutually offset or partially offset in the noise reduction area, thereby reducing the wind noise of the vehicle. In some embodiments, the noise reduction region may include an in-vehicle region. The in-vehicle region includes a cab region and a passenger region. The passenger area refers to an area where no driver is present, for example, a passenger area, a vehicle rear-row location area, or the like.

The first microphone 210 may be used to collect a wind noise signal. In some embodiments, the first microphone 210 may be located near the window (e.g., at a location outside the door near the window) for acquiring a first wind noise signal indicative of wind noise near the window. In some embodiments, the number of first microphones 210 may be one or more. The plurality of first microphones 210 may be located near different windows, respectively. For example, the plurality of first microphones 210 may be located near a cab window and a passenger window, respectively. The passenger windows may include passenger compartment windows and windows on both sides of the rear passenger compartment. A first microphone 210 located near a different window may collect a first wind noise signal indicative of wind noise near the corresponding window.

The second microphone 220 may be used to collect sound signals of an area within the vehicle. In some embodiments, the second microphone 220 may be located in an interior region (e.g., a position on the seat proximate to the headrest) for capturing noise remaining from the interior region after noise reduction, and generating an error signal indicative of the remaining noise in the interior region.

In some embodiments, the number of second microphones 220 may be one or more. The plurality of second microphones 220 may be located in different in-vehicle regions, respectively. For example, the plurality of second microphones 220 may be located in a cabin area and a passenger area, respectively (e.g., a position on the driver's seat near the headrest and a position on the passenger seat near the headrest). The passenger area may include a passenger cabin area and a plurality of areas of the rear passenger compartment. In some embodiments, the second microphones 220 located in different in-vehicle regions may each collect an error signal indicative of residual noise in the corresponding in-vehicle region.

In some embodiments, each of the plurality of second microphones 220 corresponds to one of the plurality of first microphones 210, 210. For example, the second microphone 220 located in the cab area corresponds to the first microphone 210 located near the cab window. As another example, a second microphone 220 located in the passenger area corresponds to the first microphone 210 near the passenger window.

In some embodiments, the second microphone 220 corresponding to the first microphone 210 may be the one of the plurality of second microphones 220 that is closest to the first microphone 210. Taking four vehicles as an example, the in-vehicle region may include a cab region, a co-driver region, a rear left side region, and a rear right side region; the windows may include a cab window, a co-pilot window, a rear left side window, and a rear right side window. At this time, the second microphone 220 of the cab area corresponds to the first microphone 210 near the cab window; the second microphone 220 of the co-pilot zone corresponds to the first microphone 210 near the co-pilot window; the second microphone 220 of the rear left region corresponds to the first microphone 210 near the rear left window; the second microphone 220 in the rear right region corresponds to the first microphone 210 near the rear right window.

Speaker 230 may be used to generate noise reducing sound waves. In some embodiments, speaker 230 may generate noise reducing sound waves driven by the noise reducing signal, where the noise reducing sound waves interfere with wind noise in the interior region, thereby canceling or partially canceling each other to reduce or eliminate wind noise in the interior region. In some embodiments, the noise reduction signal may be derived based on the first wind noise signal acquired by the first microphone 210. For example, the processing device may determine a signal representative of wind noise of the in-vehicle region (i.e., a second wind noise signal) based on the first wind noise signal and generate a noise reduction signal therefrom. The amplitude of the noise reducing sound wave generated by the noise reducing signal transmitted to the in-vehicle region may be equal or approximately equal to the amplitude of the second wind noise signal, and the phase of the noise reducing sound wave generated by the noise reducing signal transmitted to the in-vehicle region may be opposite or approximately opposite to the phase of the second wind noise signal.

In some embodiments, the number of speakers 230 may be one or more, and one or more speakers 230 may be located in different in-vehicle regions, respectively. For example, multiple speakers 230 may be located in the cab area and the passenger area, respectively, to collectively achieve active noise reduction for different in-vehicle areas.

In some embodiments, when the noise reducing sound wave of the in-vehicle region interferes with wind noise, the wind noise and the noise reducing sound wave may not completely cancel each other, thereby generating residual noise in the in-vehicle region. At this time, the second microphone 220 may be used to collect the residual noise signal (i.e., an error signal representing residual noise in the vehicle interior region), and the processing device 110 further updates the noise reduction signal according to the error signal collected by the second microphone 220, so that the noise reduction sound wave output by the speaker 230 can better cancel the noise in the vehicle interior region, thereby improving the noise reduction capability of the vehicle wind noise system 200. For more on updating the noise reduction signal see the relevant description of fig. 3-5.

It should be noted that the above description of the system and its components is for descriptive convenience only and is not intended to limit the present disclosure to the scope of the illustrated embodiments. It will be understood by those skilled in the art that, given the principles of the system, it is possible to combine the individual components arbitrarily or to connect the constituent subsystems with other components without departing from such principles. For example, the individual components may share a single memory device, or the individual components may each have a separate memory device. Such variations are within the scope of the present description.

FIG. 3 is an exemplary flow chart of a method of reducing wind noise in a vehicle according to some embodiments of the present description. In some embodiments, the process 300 may be performed by the processing device 110.

At step 310, a first wind noise signal representative of wind noise near a vehicle window is acquired by a first microphone, wherein the first microphone is located near the vehicle window.

In some embodiments, a first microphone (e.g., first microphone 210) may be located near the window to collect a first wind noise signal indicative of wind noise near the window. In some embodiments, the number of first microphones may be one or more, the one or more first microphones being located in proximity to different windows, respectively, to collect first wind noise signals indicative of wind noise in the vicinity of the corresponding windows. A further description of the first microphone and the first wind noise signal may be found in fig. 2.

And 320, determining a second wind noise signal of the in-vehicle region according to the neural network model and the first wind noise signal, wherein the neural network model constructs nonlinear changes of wind noise transmitted to the in-vehicle region from the vicinity of the vehicle window.

In some embodiments, during the transmission of the first wind noise signal from the vicinity of the vehicle window to the interior region, the parameter information, e.g., amplitude, phase, of the first wind noise signal may change such that the parameter information of the wind noise signal reaching the interior region is different from the parameter information of the first wind noise signal, such that the first wind noise signal cannot be used to characterize the wind noise of the interior region. It is also understood that the first wind noise signal cannot be used as a wind noise to be eliminated or reduced in the in-vehicle region. Therefore, in order to realize active noise reduction of the vehicle, it is necessary to acquire parameter information of the wind noise signal of the in-vehicle region after the first wind noise signal is transmitted from the vicinity of the vehicle window to the in-vehicle region, so as to obtain the wind noise signal for representing the wind noise of the in-vehicle region, thereby performing noise reduction processing on the wind noise signal of the in-vehicle region. For convenience of description, the wind noise signal of the in-vehicle region may be referred to as a second wind noise signal. The first wind noise signal is transmitted from the vicinity of the window to the in-vehicle region, and the window and the in-vehicle region may have a correspondence relationship. Similar to the correspondence between the first microphone and the second microphone, the window may correspond to an in-vehicle region closest to the window. For example, the cab window corresponds to a cab area, and the passenger window corresponds to a passenger area.

In some embodiments, the second wind noise signal for a different in-vehicle region may be determined based on the wind noise signal near its corresponding window. For example, the processing device 110 may determine a second wind noise signal for the cab region based on a first wind noise signal acquired by a first microphone near a cab window. For another example, the processing device 110 may determine a second wind noise signal for the passenger area based on a first wind noise signal acquired by a first microphone near the passenger window.

In some embodiments, the change in the parameter of the wind noise signal is non-linear during the transmission of the wind noise signal from near the vehicle window to the corresponding in-vehicle region. The neural network model may construct a nonlinear variation in the wind noise transmitted from near the vehicle window to the in-vehicle region. In some embodiments, the processing device 110 may determine a second wind noise signal for the in-vehicle region from the neural network model and the first wind noise signal. For example, the processing device 110 may take parameter information of the first wind noise signal as input to a neural network model that outputs the magnitude and phase of the second wind noise signal for the in-vehicle region.

In some embodiments, where the first microphone comprises a plurality of first microphones located near different vehicle windows and the second microphone comprises a plurality of second microphones located in different in-vehicle regions, the neural network model may construct a nonlinear change from any one of the first microphones to its corresponding second microphone. In some embodiments, the neural network model may construct a nonlinear variation from a first microphone to the one closest thereto. The nonlinear change relation between the first microphone and the corresponding second microphone constructed by the neural network model can reflect the change condition of parameters (such as amplitude and phase) of wind noise signals in the process of transmitting the wind noise signals from the vicinity of the vehicle window to the corresponding in-vehicle area. The parameter variation can be represented by a transfer function (also called a first transfer function), which can be denoted as H (n, m). In some embodiments, the first transfer function of the wind noise signal from near the different windows to the corresponding in-vehicle region may be the same or different.

In some embodiments, the neural network model may include a deep-roll convolution loop network model (Deep Complex Convolution Recurrent Network, DCCRN) that may be trained from first wind noise sample data near a window of a vehicle and second wind noise sample data of an in-vehicle region during vehicle travel.

The first wind noise sample data may be a microphone signal (also referred to as a reference signal) near a window of a vehicle recorded during running of the vehicle. In some embodiments, the first wind noise sample data may be a data set of reference signals near different windows during the vehicle driving.

The second wind noise sample data may be a microphone signal (also called an in-vehicle signal) of an in-vehicle region recorded during running of the vehicle. In some embodiments, the second wind noise sample data may be a data set formed by in-vehicle signals of different in-vehicle regions during the running of the vehicle.

When the deep complex convolution network model is trained by using the first wind noise sample data and the second wind noise sample data, the first wind noise sample data and the second wind noise sample data are in one-to-one correspondence, that is, a reference signal near a vehicle window recorded in the running process of the vehicle corresponds to an in-vehicle signal of an in-vehicle area nearest to the vehicle window. For example, the reference signal near the cab window corresponds to the in-vehicle signal of the cab region. For more details on deep-roll-up cyclic network models and their training process, see the relevant description of fig. 5 of the present specification.

And 330, determining a noise reduction signal according to the second wind noise signal to drive a loudspeaker to generate noise reduction sound waves, wherein the loudspeaker is positioned in the vehicle.

In some embodiments, the second wind noise signal may be used to characterize wind noise of the in-vehicle region. The processing device 110 may determine the amplitude and phase of the noise reduction signal according to the parameter information of the second wind noise signal. For example, a speaker may generate noise-reduced sound waves driven by a noise reduction signal. The processing device may determine the amplitude and phase of the noise reduction signal such that when the noise reduction sound wave generated by the speaker is transferred to the in-vehicle region, the amplitude of the noise reduction sound wave is equal or approximately equal to the amplitude of the wind noise of the in-vehicle region, and the phase of the noise reduction sound wave is opposite or approximately opposite to the phase of the wind noise of the in-vehicle region. Thus, noise reduction sound waves and wind noise are interfered in the vehicle interior area, so that the noise reduction sound waves and the wind noise are mutually offset or partially offset, the wind noise of the vehicle interior area is weakened or eliminated, and the active noise reduction of the vehicle is realized.

In some embodiments, the second wind noise signals characterizing the wind noise of the different in-vehicle regions are different, and the processing device 110 may determine the noise reduction signals from the wind noise of the different in-vehicle regions and drive speakers (e.g., speakers located in the different in-vehicle regions) based on the noise reduction signals to generate corresponding noise reduction sound waves such that the wind noise of the different in-vehicle regions is reduced to the same extent or to different extents.

In some embodiments, the magnitude and phase of the wind noise will change as the wind noise is transferred from the vicinity of the vehicle window to the corresponding in-vehicle region, and similarly, the magnitude and phase of the noise reducing sound wave will also change as the noise reducing sound wave is transferred from the speaker location to the in-vehicle region. In some embodiments, the change in parameters (such as amplitude and phase) of the noise reducing sound wave during the transmission from the speaker position to the interior region of the vehicle may be represented by a transfer function (also referred to as a second transfer function), which may be denoted as H (s, m). In some embodiments, the processing device 110 may obtain a second transfer function of the noise reduction sound wave transferred from the speaker location to the in-vehicle region and determine the noise reduction signal based on the second transfer function and the second wind noise signal. For example, the processing device 110 may determine the wind noise signal for the speaker location based on the conjugate function of the second transfer function and the second wind noise signal, and the processing device 110 may further determine the noise reduction signal from the wind noise signal for the speaker location.

In some embodiments, the second transfer function may be determined from a swept frequency signal output by the speaker and a sound signal collected by an in-vehicle microphone. The sweep frequency signal may refer to a constant amplitude signal whose frequency varies periodically over a range. The in-vehicle microphone may be disposed in an in-vehicle region to collect sound signals of the in-vehicle region. During the driving of the vehicle, the sweep signal may be played by the speaker, and the sound signal of the in-vehicle region may be collected by the in-vehicle microphone located in the in-vehicle region, and the processing device 110 may construct a change in the transmission of the noise reduction sound wave from the speaker position to the in-vehicle region according to the sweep signal and the sound signal collected by the in-vehicle microphone, so as to determine a second transfer function of the noise reduction sound wave transmitted from the speaker position to the in-vehicle region. For example, the second transfer function can be obtained by performing convolution operation on the sound signal collected by the microphone in the vehicle and the inverse sweep frequency signal.

In some embodiments, to take into account the effect of in-car white noise, the second transfer function may be calculated by convolving the sound signal collected by the in-car microphone with the maximum white noise sequence and then convolving with the inverse swept frequency signal, thereby simulating the transfer function from the speaker location to the in-car region in a real scene (i.e., where the user would hear other unpleasant sounds besides wind noise).

In some embodiments, the greater the number of speakers, the better the effect of determining the second transfer function based on the swept frequency signal and the sound signal collected by the in-vehicle microphone, the more accurately the second transfer function reflects the change in parameters of the noise reduction sound wave transferred from the speaker location to the in-vehicle region. During the running process of the vehicle, the plurality of speakers can be controlled to sequentially play the sweep frequency signals, and sound signals collected by the plurality of in-vehicle microphones are recorded, so that the transfer function from each speaker to each in-vehicle microphone is calculated. By way of example only, the number of speakers and the number of in-vehicle microphones may both be 4. The second transfer function of the speaker to the in-car microphone may be expressed as h (s _i ,m _j ) Where i denotes the number of the speaker, and values from 1 to 4,j denote the number of the in-vehicle microphone, and values from 1 to 4.

In some embodiments, the processing device 110 may determine the noise reduction signal based on the second transfer function and the second wind noise signal. In some embodiments, the processing device 110 may obtain a conjugate function of the second transfer function based on the second transfer function. The conjugate function of the second transfer function may reflect changes in the phase and amplitude of the sound wave as it passes from the in-vehicle region to the speaker location. For example, a wind noise signal (e.g., a second wind noise signal) that characterizes wind noise in an area within the vehicle may be processed by a conjugate function of the second transfer function as a wind noise signal that characterizes wind noise at a location of the speaker. In some embodiments, the processing device 110 may determine the noise reduction signal from the wind noise signal of the speaker location. For example, the amplitude of the noise reduction signal may be equal or approximately equal to the amplitude of the wind noise signal at the speaker location, and the phase of the noise reduction signal may be opposite or approximately opposite to the phase of the wind noise signal at the speaker location.

Step 340, obtaining an error signal collected by a second microphone for characterizing residual noise in the vehicle interior region, wherein the residual noise in the vehicle interior region is derived from superposition of wind noise and noise reduction waves, and the second microphone is located in the vehicle interior region.

In some embodiments, the error signal collected by the second microphone may be a noise signal that is transmitted to an area within the vehicle and a noise signal that remains after the wind noise is cancelled. When noise reduction sound waves and wind noise interfere and cancel in an in-vehicle area, the noise reduction sound waves and the wind noise can not be completely counteracted, so that partial wind noise can be remained in the in-vehicle area, and residual noise is formed. That is, the residual noise is derived from the superposition of wind noise and noise reduction waves in the in-vehicle region. For example, in the process of transmitting the noise-reducing sound wave to the in-vehicle region, the amplitude and phase of the noise-reducing sound wave may change, so that the noise-reducing sound wave and wind noise cannot be completely counteracted in the in-vehicle region.

Step 350, updating the noise reduction signal according to the error signal.

In order to better eliminate wind noise in the in-vehicle region, the processing device 110 may update the noise reduction signal according to the error signal collected by the second microphone (for characterizing residual noise in the in-vehicle region), so that the noise reduction sound wave generated by the speaker can better eliminate the noise in the in-vehicle region, reduce the noise residual in the in-vehicle region, and thereby improve the active noise reduction effect of the vehicle.

In some embodiments, the processing device 110 may adjust the phase and amplitude of the noise reduction signal according to the error signal to enable updating of the noise reduction signal. In some embodiments, after the processing device 110 updates the noise reduction signal according to the error signal, the speaker generates an updated noise reduction sound wave under the drive of the updated noise reduction signal, and the updated noise reduction sound wave can be better offset with the wind noise of the in-vehicle area, so as to improve the wind noise reduction capability of the vehicle.

In some embodiments, when updating the noise reduction signal according to the error signal, the influence of the noise reduction sound wave generated by each speaker in the plurality of speakers on the sound signal of the in-vehicle area (for example, the noise reduction sound wave generated by each speaker affects the residual noise of the in-vehicle area) can be comprehensively considered so as to better eliminate the noise of the in-vehicle area. Taking the cab area as an example, the noise reduction sound waves generated by the speakers at different positions are transferred from the corresponding positions to the cab area, each noise reduction sound wave affects the sound signal of the cab area, at this time, the processing device 110 may obtain an error signal of the in-vehicle area based on the effect of the noise reduction sound waves generated by each speaker on the residual noise of the in-vehicle area, and update the noise reduction signal based on the error signal.

In some embodiments, the processing device 110 may update the noise reduction signal based on the error signal using an adaptive filter. Fig. 4 is an exemplary flow chart of updating a noise reduction signal for an adaptive filter according to some embodiments of the present description.

In some embodiments, the adaptive filter may be a filter that is capable of changing filter parameters using an adaptive algorithm. In some embodiments, the adaptive filter may iteratively adjust filter parameters using an adaptive least mean square algorithm (Least Mean Square, LMS). The adaptive filter may update the filter coefficients via an LMS algorithm to update parameters (e.g., amplitude and phase) of the noise reduction signal such that an error signal indicative of residual noise in the interior region is gradually reduced, eventually to zero or approaching zero, thereby improving the ability of the vehicle to actively reduce noise.

Referring to fig. 4, a wind noise signal X (n) near a window may be transmitted along a wind noise path (as indicated by the "wind noise path" arrow in fig. 4) to an in-vehicle region (e.g., where a second microphone of the in-vehicle region is located). After the wind noise signal X (n) near the window is transmitted to the in-vehicle region, the amplitude and phase of the wind noise signal X (n) change, and at this time, the wind noise signal of the in-vehicle region may be represented as d (n). In some embodiments, in conjunction with the foregoing description, the transmission process of wind noise may also be understood as that after the first wind noise signal X (n) near the vehicle window passes through the first transfer function H (n, m), the wind noise signal reaching the in-vehicle area is the second wind noise signal d (n).

The processing device 110 may determine the noise reduction signal based on the second wind noise signal d (n). As shown in fig. 4, it may also be understood that the second wind noise signal d (n) may be input to the filter W, and the filter determines the noise reduction signal based on the second wind noise signal d (n) and causes the speaker to output the noise reduction sound wave under the driving of the noise reduction signal. Further, the noise reduction sound wave may be transmitted to the in-vehicle region along a noise reduction sound wave path (indicated by an arrow of "noise reduction sound wave path" in fig. 4). In some embodiments, the noise reduction signal determined by the filter W is a noise reduction signal y (n) after being processed by the second transfer function H (s, m) to reach the in-vehicle region. And the wind noise signal d (n) of the in-vehicle region is overlapped with the noise reduction signal y (n) to obtain an error signal e (n) of the in-vehicle region. In some embodiments, the adaptive filter may update the filter coefficients of the filter W based on the error signal e (n) using an LMS algorithm, thereby updating the noise reduction signal of the speaker.

In some embodiments, the speaker outputs an updated noise reduction sound wave under the drive of the updated noise reduction signal, the updated noise reduction sound wave reaches the in-vehicle region again through the second transfer function H (s, m), at this time, the noise reduction signal of the in-vehicle region is updated y (n), and y (n) is overlapped with the wind noise signal d (n) again, so as to obtain an error signal e (n) of the in-vehicle region; the adaptive filter repeatedly updates the filter coefficients based on the updated error signal e (n) using the LMS algorithm, thereby repeatedly updating the noise reduction signal of the speaker. The adaptive filter automatically updates the filter coefficient through a plurality of iterations in the manner, and the error signal e (n) of the in-vehicle region can be made small enough or close to zero in the process of repeatedly updating the filter coefficient.

With continued reference to fig. 4, the wind noise signal X (n) near the window may be a wind noise signal representing wind noise in the vehicle interior region after passing through the first transfer function H (n, m), and the wind noise signal in the vehicle interior region may further pass through a conjugate function of the second transfer function H (s, m)The wind noise signal that characterizes the wind noise at the loudspeaker position may then be referred to as: />

In some embodiments, the updated equation for the filter coefficients of the adaptive filter may be expressed as equation (1):

W(n+1)＝W(n)-2μe(n)*Xf(n)， (1)

Where n represents the number of iterations and μ is the iteration step. In some embodiments, 0 < 1/2 μ xxf (n) jjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj can be used to ensure convergence of the adaptive filter iteratively solving the filter coefficients ² ＜1。

In some embodiments, the wind noise signal Xf (n) of the speaker position is proportional to the wind speed, where the wind speed is proportional to the vehicle speed. Specifically, the method can be expressed as follows: the faster the vehicle speed, the greater the wind speed, and the greater the wind noise signal at the speaker position; the slower the vehicle speed, the lower the wind speed and the lower the wind noise signal at the speaker location. In some embodiments, different iteration steps μmay be matched for different Xf (n) in order to speed up convergence in the iterative updating of the filter coefficients. That is, at different vehicle speeds, the adaptive filter may employ different ranges of iteration steps. For example, the wind speeds (energies) corresponding to the different vehicle speeds, and the ranges of the iteration steps μ at the different vehicle speeds may be stored in advance in a storage device (e.g., the storage device 130).

In some embodiments, when the adaptive filter updates the filter coefficients according to the LMS algorithm, the filter coefficients may be related to weight values of different regions within the vehicle, which may reflect differences in noise reduction capabilities of the different regions. In some embodiments, the weight values of different in-vehicle regions may be set to be the same or different. For example, to ensure that the noise reduction effect of the cab region is best, the weight value of the cab region may be set to be larger than that of other regions (e.g., passenger regions).

Specifically, considering the noise reduction effect of different in-vehicle regions, the iterative update formula of the filter coefficient of the adaptive filter for each speaker can be expressed as formula (2):

wherein n represents the iteration number, i represents the number of in-vehicle regions, j represents the number of speakers, and P _i Weight value mu representing different areas in vehicle _i Iteration step length e for representing different in-vehicle regions _i Error signal, xf, representing different in-vehicle zones _i (n) wind noise signals representing different speaker positions. In some embodiments, the values of i and j may be set according to the actual situation (e.g., the number of vehicle seats). For example, i and j may each take a value of 4. Equation (2) shows an iterative update of the filter coefficients of the j-th speaker by the adaptive filter.

In some embodiments, P may be assigned according to noise reduction capability of different in-vehicle regions _i Different weight values. For example, when the noise reduction capacities of the 4 in-vehicle regions are the same or substantially the same, the weight values of the 4 in-vehicle regions may be set to be equal, e.g., P _i May all have a value of 0.25, i.e. P ₁ ＝P ₂ ＝P ₃ ＝P ₄ =0.25. For another example, if it is desired to make the noise reduction capability of the cab region stronger than that of other regions (e.g., passenger regions), the weight value of the cab region may be set to be larger than that of the other regions, such as P _i The value may be taken as [0.4,0.2,0.2，0.2]i.e. P ₁ ＝0.4，P ₂ ＝0.2，P ₃ ＝0.2，P ₄ ＝0.2。

In some embodiments, according to the formula (2), the filter coefficient is iteratively updated by summing the i representing the number of the areas in the vehicle, and the influence of noise reduction of each area in the vehicle on the filter coefficient can be comprehensively considered in the process of updating the filter coefficient, so that the effect of reducing wind noise of the whole vehicle is improved. On the other hand, by setting that different in-vehicle regions have the same or different weight values, the different importance of the in-vehicle regions can be realized to flexibly adjust the noise reduction of each in-vehicle region.

In some embodiments, the adaptive filter iteratively updates the filter coefficients in the manner shown in fig. 4, so that the error signal e (n) of the in-vehicle area is sufficiently small or approaches zero, and the noise reduction sound wave generated by the speaker under the drive of the noise reduction signal after the update can be completely or approximately completely cancelled out by the in-vehicle area and the wind noise signal, so that no or almost no residual signal exists in the in-vehicle area, thereby improving the wind noise reduction capability of the vehicle.

It should be noted that the above description of the process 300 is for purposes of example and illustration only and is not intended to limit the scope of applicability of the present disclosure. Various modifications and changes to flow 300 will be apparent to those skilled in the art in light of the present description. For example, step 310 and step 320 may be combined into the same step, and for example, step 330 may be split into two steps, however, such modifications and variations are still within the scope of the present description.

FIG. 5 is an exemplary flow chart of a neural network model training method, according to some embodiments of the present description. In some embodiments, the process 500 may be performed by the processing device 110.

Step 510, obtaining first wind noise sample data and second wind noise sample data, calculating spectral features on a complex domain of the first wind noise sample data, calculating mask features CRM as labels of the first wind noise sample data through time domain sequence data of the second wind noise sample data, and using the spectral feature data on the complex domain of the first wind noise sample data with labels as a training set required by network training.

In some embodiments, the first wind noise sample data may include a microphone signal (also referred to as a reference signal) near a window of a vehicle during travel of the vehicle. The second wind noise sample data may include microphone signals (also called in-vehicle signals) of an in-vehicle region recorded during the running of the vehicle. In some embodiments, the first wind noise sample data and the second wind noise sample data are in one-to-one correspondence, i.e. the reference signal recorded during the running of the vehicle in the vicinity of the window corresponds to the in-vehicle signal of the in-vehicle area nearest to the window. In some embodiments, the first wind noise sample data and the second wind noise sample data may be stored in a storage device (e.g., storage device 130), from which processing device 110 may retrieve the first wind noise sample data and the second wind noise sample data.

In some embodiments, the processing device 110 may calculate spectral features over the first wind noise sample data complex domain. In some embodiments, the processing device 110 may perform a short-time fourier transform on the first wind noise sample data to obtain a frequency spectrum over the complex domain of the first wind noise sample data. In some embodiments, the processing device 110 may perform frame sampling processing on the first wind noise sample data to obtain a first waveform diagram corresponding to the first wind noise sample data; and performing short-time Fourier transform based on the first waveform diagram to obtain a frequency spectrum on the first wind noise sample data complex domain. In some embodiments, the first wind noise sample data spectrum may be represented by a real part and an imaginary part.

In some embodiments, the mask feature CRM may be computed from the second wind noise sample data time domain sequence data as a tag of the first wind noise sample data. In some embodiments, mask feature CRM may be calculated with equation (3):

wherein Y is _r And Y _i Representing the real part and the imaginary part of the data spectrum of the first wind noise sample respectively, S _r And S is _i Respectively represent the firstReal and imaginary parts of the two wind noise sample data spectrum. In some embodiments, the training set required for network training may include spectral feature data over a plurality of tagged first wind noise sample data complex fields.

And step 520, training the deep convolution cyclic network model by taking the frequency spectrum characteristics on the first wind noise sample data complex domain in the training set required by the network training as the input data of the deep convolution cyclic network model to obtain prediction data.

In some embodiments, the prediction data output by the deep-roll convolution loop network model may be spectral features over the prediction data complex domain. In some embodiments, the deep-convolution cyclic network model may also perform a short-time fourier transform on the spectral features on the complex domain of the predicted data to obtain the predicted data time-domain sequence data.

And 530, calculating the frequency spectrum characteristics on the second wind noise sample data complex domain, and comparing the frequency spectrum characteristics on the predicted data complex domain with the frequency spectrum characteristics on the second wind noise sample data complex domain to obtain a comparison result.

In some embodiments, the processing device 110 performs a short-time fourier transform on the second wind noise sample data to obtain a spectrum over the complex domain of the second wind noise sample data. In some embodiments, the processing device 110 may perform frame sampling processing on the second wind noise sample data to obtain a second waveform diagram corresponding to the second wind noise sample data; and performing short-time Fourier transform based on the second waveform diagram to obtain a frequency spectrum on a data complex domain of the second wind noise sample. In some embodiments, the second wind noise sample data spectrum may be represented by a real part and an imaginary part.

In some embodiments, processing device 110 may compare the spectrum on the complex domain of the predicted data with the spectrum on the complex domain of the second wind noise sample data to obtain a difference in the complex domain of the predicted data and the second wind noise sample data. The difference in complex domain between the predicted data and the second wind noise sample data may be used to characterize the difference in amplitude and phase of the predicted signal output by the deep-convolution cyclic network model and the real signal (i.e., the in-vehicle signal). In some embodiments, the amplitude and phase of the predicted signal may be adjusted according to the real and imaginary part correspondence in the complex domain, such that the predicted signal more closely fits the real signal.

And step 540, updating parameters of the deep-rolling circulation network model according to the comparison result to obtain the trained deep-rolling circulation network model.

In some embodiments, the processing device 110 may update parameters of the deep-roll convolution loop network model based on differences in the complex domain between the prediction data and the second wind noise sample data. In the process of continuously updating parameters, the difference between the predicted data and the second wind noise sample data can be gradually reduced until the trained deep complex convolution cyclic network model meets the preset condition, so that the trained deep complex convolution cyclic network model is obtained. In some embodiments, the preset condition may be that the loss function is less than a threshold, converges, or that the training period reaches a threshold. In some embodiments, the loss function may use a normalized signal-to-noise ratio (SI-SNR). The normalized signal-to-noise ratio may be determined according to a second waveform map corresponding to the second wind noise sample data and a predicted waveform map corresponding to the predicted signal. In some embodiments, the normalized signal-to-noise ratio (SI-SNR) may be expressed as equation (4):

Where s may represent a second waveform,a predicted waveform map may be represented. In some embodiments, the predicted waveform map may be obtained by performing a frame sampling process on the predicted signal.

Some embodiments of the present disclosure further provide a computer readable storage medium, where the storage medium stores computer instructions, and when the computer reads the computer instructions in the storage medium, the computer performs the method described in any embodiment of the present disclosure, and detailed descriptions thereof with reference to fig. 1 to 5 are omitted herein.

Some embodiments of the present specification also provide a vehicle including: at least one first microphone configured to acquire a first wind noise signal indicative of wind noise near a vehicle window, wherein the first microphone is located near the vehicle window; a speaker configured to generate a noise reduction sound wave under the driving of a noise reduction signal, wherein the noise reduction signal is determined based on a second wind noise signal of the in-vehicle region, the second wind noise signal is determined based on the neural network model and the first wind noise signal; at least one second microphone configured to acquire an error signal representing residual noise of the in-vehicle region, the residual noise of the in-vehicle region resulting from a superposition of wind noise and noise-reducing waves, wherein the second microphone is located in the in-vehicle region; the neural network model builds nonlinear changes of wind noise transmitted from the vicinity of the vehicle window to the region in the vehicle.

It should be noted that the above description of the process 500 is for purposes of illustration and description only, and is not intended to limit the scope of applicability of the present disclosure. Various modifications and changes to flow 500 will be apparent to those skilled in the art in light of the present description. For example, steps 510, 530 may be split into multiple steps. However, such modifications and variations are still within the scope of the present description.

While the basic concepts have been described above, it will be apparent to those skilled in the art that the foregoing detailed disclosure is by way of example only and is not intended to be limiting. Although not explicitly described herein, various modifications, improvements, and adaptations to the present disclosure may occur to one skilled in the art. Such modifications, improvements, and modifications are intended to be suggested within this specification, and therefore, such modifications, improvements, and modifications are intended to be included within the spirit and scope of the exemplary embodiments of the present invention.

Meanwhile, the specification uses specific words to describe the embodiments of the specification. Reference to "one embodiment," "an embodiment," and/or "some embodiments" means that a particular feature, structure, or characteristic is associated with at least one embodiment of the present description. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the present description may be combined as suitable.

Furthermore, those skilled in the art will appreciate that the various aspects of the specification can be illustrated and described in terms of several patentable categories or circumstances, including any novel and useful procedures, machines, products, or materials, or any novel and useful modifications thereof. Accordingly, aspects of the present description may be performed entirely by hardware, entirely by software (including firmware, resident software, micro-code, etc.), or by a combination of hardware and software. The above hardware or software may be referred to as a "data block," module, "" engine, "" unit, "" component, "or" system. Furthermore, aspects of the specification may take the form of a computer product, comprising computer-readable program code, embodied in one or more computer-readable media.

The computer storage medium may contain a propagated data signal with the computer program code embodied therein, for example, on a baseband or as part of a carrier wave. The propagated signal may take on a variety of forms, including electro-magnetic, optical, etc., or any suitable combination thereof. A computer storage medium may be any computer readable medium that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code located on a computer storage medium may be propagated through any suitable medium, including radio, cable, fiber optic cable, RF, or the like, or a combination of any of the foregoing.

The computer program code necessary for operation of portions of the present description may be written in any one or more programming languages, including an object oriented programming language such as Java, scala, smalltalk, eiffel, JADE, emerald, C ++, c#, vb net, python and the like, a conventional programming language such as C language, visual Basic, fortran 2003, perl, COBOL 2002, PHP, ABAP, a dynamic programming language such as Python, ruby and Groovy, or other programming languages and the like. The program code may execute entirely on the user's computer or as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or processing device. In the latter scenario, the remote computer may be connected to the user's computer through any form of network, such as a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet), or the use of services such as software as a service (SaaS) in a cloud computing environment.

Furthermore, the order in which the elements and sequences are processed, the use of numerical letters, or other designations in the description are not intended to limit the order in which the processes and methods of the description are performed unless explicitly recited in the claims. While certain presently useful inventive embodiments have been discussed in the foregoing disclosure, by way of various examples, it is to be understood that such details are merely illustrative and that the appended claims are not limited to the disclosed embodiments, but, on the contrary, are intended to cover all modifications and equivalent arrangements included within the spirit and scope of the embodiments of the present disclosure. For example, while the system components described above may be implemented by hardware devices, they may also be implemented solely by software solutions, such as installing the described system on an existing processing device or mobile device.

Likewise, it should be noted that in order to simplify the presentation disclosed in this specification and thereby aid in understanding one or more inventive embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof. This method of disclosure, however, is not intended to imply that more features than are presented in the claims are required for the present description. Indeed, less than all of the features of a single embodiment disclosed above.

In some embodiments, numbers describing the components, number of attributes are used, it being understood that such numbers being used in the description of embodiments are modified in some examples by the modifier "about," approximately, "or" substantially. Unless otherwise indicated, "about," "approximately," or "substantially" indicate that the number allows for a 20% variation. Accordingly, in some embodiments, numerical parameters set forth in the specification and claims are approximations that may vary depending upon the desired properties sought to be obtained by the individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and employ a method for preserving the general number of digits. Although the numerical ranges and parameters set forth herein are approximations that may be employed in some embodiments to confirm the breadth of the range, in particular embodiments, the setting of such numerical values is as precise as possible.

Each patent, patent application publication, and other material, such as articles, books, specifications, publications, documents, etc., referred to in this specification is incorporated herein by reference in its entirety. Except for application history documents that are inconsistent or conflicting with the content of this specification, documents that are currently or later attached to this specification in which the broadest scope of the claims to this specification is limited are also. It is noted that, if the description, definition, and/or use of a term in an attached material in this specification does not conform to or conflict with what is described in this specification, the description, definition, and/or use of the term in this specification controls.

Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments of this specification. Other variations are possible within the scope of this description. Thus, by way of example, and not limitation, alternative configurations of embodiments of the present specification may be considered as consistent with the teachings of the present specification. Accordingly, the embodiments of the present specification are not limited to only the embodiments explicitly described and depicted in the present specification.

Claims

1. A method of reducing wind noise in a vehicle, the method comprising:

Acquiring a first wind noise signal acquired by a first microphone for representing wind noise near a vehicle window, wherein the first microphone is positioned near the vehicle window;

determining a second wind noise signal of an in-vehicle region according to a neural network model and the first wind noise signal, wherein the neural network model constructs nonlinear changes of wind noise transmitted from the vicinity of the vehicle window to the in-vehicle region;

determining a noise reduction signal according to the second wind noise signal so as to drive a loudspeaker to generate noise reduction sound waves, wherein the loudspeaker is positioned in the vehicle;

acquiring an error signal acquired by a second microphone for representing residual noise in the vehicle interior region, wherein the residual noise in the vehicle interior region is derived from superposition of wind noise and noise reduction waves, and the second microphone is positioned in the vehicle interior region; and

and updating the noise reduction signal according to the error signal.

2. The method of claim 1, wherein the first microphones comprise a plurality of first microphones positioned adjacent different windows of the vehicle, the second microphones comprise a plurality of second microphones positioned in different in-vehicle regions, each of the plurality of second microphones corresponds to one of the plurality of first microphones, and the neural network model constructs a nonlinear change from any one of the first microphones to its corresponding second microphone.

3. The method of claim 2, wherein the neural network model comprises a deep complex convolution loop network model trained from first wind noise sample data near the vehicle window and second wind noise sample data of the in-vehicle region during the vehicle driving.

4. The method according to claim 1, wherein the method further comprises: acquiring a transfer function of the noise reduction sound wave transferred from the loudspeaker position to the in-vehicle area, wherein determining the noise reduction signal comprises:

the noise reduction signal is determined based on the transfer function and the second wind noise signal.

5. The method of claim 4, wherein the transfer function is determined based on a swept frequency signal of the speaker and a sound signal collected by the second microphone.

6. The method of claim 1, wherein updating the noise reduction signal based on the error signal comprises:

and updating the noise reduction signal according to the error signal by using an adaptive filter.

7. The method of claim 6, wherein the adaptive filter updates the filter coefficients according to an LMS algorithm, the filter coefficients being related to weight values for different regions within the vehicle, the weight values for the different regions reflecting differences in noise reduction capabilities of the different regions.

8. The method of claim 6, wherein the iteration step of the filter coefficients is related to vehicle speed.

9. A system for reducing wind noise in a vehicle, the system comprising at least one first microphone located near a vehicle window, at least one second microphone located in an interior region of the vehicle, a speaker, and a processor,

the processor is configured to perform the following operations:

acquiring a first wind noise signal acquired by the first microphone and used for representing wind noise near the vehicle window;

determining a second wind noise signal of the in-vehicle region according to a neural network model and the first wind noise signal, wherein the neural network model constructs nonlinear changes of wind noise transmitted from the vicinity of the vehicle window to the in-vehicle region;

determining a noise reduction signal according to the second wind noise signal so as to drive the loudspeaker to generate noise reduction sound waves;

acquiring an error signal acquired by the second microphone and used for representing residual noise in the vehicle interior region, wherein the residual noise in the vehicle interior region is derived from superposition of wind noise and the noise reduction wave; and

and updating the noise reduction signal according to the error signal.

10. The system of claim 9, wherein the first microphones comprise a plurality of first microphones positioned adjacent different windows of the vehicle, the second microphones comprise a plurality of second microphones positioned in different in-vehicle regions, each of the plurality of second microphones corresponds to one of the plurality of first microphones, and the neural network model constructs a nonlinear change from any one of the first microphones to its corresponding second microphone.

11. The system of claim 10, wherein the neural network model comprises a deep complex convolution loop network model trained from first wind noise sample data near the vehicle window and second wind noise sample data of the in-vehicle region during travel of the vehicle.

12. The system of claim 9, wherein the processor is further configured to: acquiring a transfer function of the noise reduction sound wave transferred from the loudspeaker position to the in-vehicle area, wherein determining the noise reduction signal comprises:

13. The system of claim 12, wherein the transfer function is determined based on a swept frequency signal of the speaker and a sound signal collected by the second microphone.

14. The system of claim 9, wherein updating the noise reduction signal based on the error signal comprises:

15. The system of claim 14, wherein the adaptive filter updates the filter coefficients according to an LMS algorithm, the filter coefficients being related to weight values for different regions within the vehicle, the weight values for the different regions reflecting differences in noise reduction capabilities of the different regions.

16. The system of claim 14, wherein the iteration step of the filter coefficients is related to vehicle speed.

17. A computer readable storage medium storing computer instructions which, when read by a computer, perform the method of any one of claims 1-8.

18. A vehicle, the vehicle comprising:

at least one first microphone configured to acquire a first wind noise signal indicative of wind noise near a vehicle window, wherein the first microphone is located near the vehicle window;

a speaker configured to generate a noise reduction acoustic wave under driving of a noise reduction signal, wherein the noise reduction signal is determined based on a second wind noise signal of an in-vehicle region, the second wind noise signal being determined based on a neural network model and the first wind noise signal;

at least one second microphone configured to acquire an error signal representing residual noise of the in-vehicle region, the residual noise of the in-vehicle region resulting from a superposition of wind noise and noise-reducing waves, wherein the second microphone is located in the in-vehicle region;

the neural network model is used for constructing nonlinear changes of wind noise transmitted from the vicinity of the vehicle window to the in-vehicle area.