Active noise reduction methods, devices, systems, and vehicles

Abstract

This application discloses a method, apparatus, system, and vehicle for active noise reduction. The method includes: downsampling a reference noise signal and decomposing it into first sub-band signals under multiple sub-bands; upsampling the first sub-band signals after processing them through an adaptive filter to obtain a noise-reduced signal; the adaptive filter is determined based on the first sub-band signals and their corresponding second sub-band signals; the second sub-band signals are obtained by second signal decomposition through downsampling of an error signal. This application effectively reduces the number of data points processed per unit time while meeting the requirement of effectively processing noise within the noise reduction frequency band by performing filtering of each sub-band signal at a low sampling rate, thereby improving the noise reduction effect.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, specifically to a vehicle active noise reduction method, device, system, and vehicle. Background Technology

[0002] Active noise cancellation in vehicles is a system that uses the principle of sound wave interference to cancel out in-vehicle noise in real time. It collects noise signals through sensors, analyzes the noise frequency and phase, and then uses a filter to generate an inverse sound wave signal. This signal is emitted through a speaker and cancels out the original noise, thus significantly reducing in-vehicle noise. Furthermore, sound pickup devices are typically placed within the vehicle's audible area to collect error signals after noise cancellation, which are then used to update the filter and further improve the noise cancellation effect.

[0003] However, the above-mentioned active noise reduction process relies on filter updates. In practical applications, it has been found that some theoretical filter update methods inevitably lead to problems such as vehicle delay and insufficient computational redundancy, which affect the active noise reduction effect of the vehicle. In order to avoid introducing additional hardware costs, it is urgent to optimize the existing noise reduction processing framework. Summary of the Invention

[0004] This application provides a method, apparatus, system, and vehicle for active noise reduction, aiming to provide an active noise reduction method with low latency and low computational load, so as to at least partially solve the above-mentioned problems.

[0005] In a first aspect, this application provides a method for active noise reduction in a vehicle, the vehicle having a reference sensor for acquiring a reference noise signal, a speaker for outputting a noise reduction signal, and an error sensor located within the audible area of ​​the vehicle for acquiring an error signal, the method comprising: The acquired reference noise signal is downsampled to obtain a first signal at a low sampling rate, and the first signal is decomposed into first sub-band signals under multiple sub-bands; Each of the first sub-band signals is processed by its corresponding adaptive filter, and the processed signal is upsampled to obtain the noise reduction signal for the speaker output. The adaptive filter corresponding to each of the first sub-band signals is determined based on the first sub-band signal and its corresponding second sub-band signal. The second sub-band signal is obtained by decomposing the second signal and by downsampling the error signal.

[0006] In one embodiment of this application, the step of decomposing the first signal into first sub-band signals under multiple sub-bands includes: The first signal is decomposed into first sub-band signals under multiple sub-bands by analyzing the filter bank; The analysis filter bank includes multiple analysis filters obtained by processing the prototype filter through cosine modulation.

[0007] In one embodiment of this application, the downsampling process of the acquired reference noise signal to obtain a first signal at a low sampling rate includes: The obtained reference noise signal is downsampled based on polyphase decomposition to obtain the first signal at a low sampling rate; The step of upsampling the processed signal to obtain the noise-reduced signal for the speaker output includes: The processed signal is upsampled based on polyphase decomposition to obtain a noise-reduced signal for the speaker output.

[0008] In one embodiment of this application, processing each first sub-band signal through its corresponding adaptive filter includes: For each of the first sub-band signals, the first sub-band signal is processed by its corresponding multiple branch adaptive filters based on polyphase decomposition.

[0009] In one embodiment of this application, the adaptive filter corresponding to each of the first sub-band signals is determined by the following steps: For each of the first sub-band signals, the first sub-band signal is processed through the corresponding sub-band secondary path to obtain the third sub-band signal; The adaptive step size for each sub-band is determined based on the third sub-band signal and the second sub-band signal corresponding to the first sub-band signal. The filtering parameters of the adaptive filter corresponding to the first sub-band signal are updated according to the adaptive step size.

[0010] In one embodiment of this application, the sub-band secondary path corresponding to the first sub-band signal is determined through the following steps: The sweep frequency signal at a low sampling rate is upsampled and then played through the speaker. The signal after playback is acquired through the error sensor; The played signal is downsampled and then decomposed together with the frequency sweep signal. The sub-band signals obtained from the decomposition are used as excitation signals and response signals, respectively, to determine the secondary paths of the sub-bands.

[0011] In one embodiment of this application, the method further includes: setting different initialization step sizes for each sub-band, and / or determining an adaptive step size for each sub-band based on the normalization result of the signal energy of the first sub-band signal.

[0012] In one embodiment of this application, the reference noise signal includes a vibration signal acquired by an accelerometer and a sound signal acquired by a pickup device.

[0013] Secondly, this application also provides a vehicle active noise reduction device, the vehicle active noise reduction device including a processor, the processor being used to achieve active noise reduction of the vehicle through the vehicle active noise reduction method as described in any of the above claims.

[0014] In one embodiment of this application, the processor includes a digital signal processor and a field-programmable gate array; The field-programmable gate array (FPGA) is used to acquire a reference noise signal collected by a reference sensor and an error signal collected by an error sensor through an interface, and to process the reference noise signal and the error signal to obtain a first sub-band signal and a second sub-band signal; to provide the first sub-band signal and the second sub-band signal to the digital signal processor (DSP), and to receive the filtering parameters of the adaptive filter returned by the DSP; to process the reference noise signal based on the filtering parameters of the adaptive filter to obtain a noise-reduced signal and to provide it to the loudspeaker; The digital signal processor is used to determine the filtering parameters of the adaptive filter based on the first sub-band signal and the second sub-band signal sent by the field-programmable gate array.

[0015] In one embodiment of this application, the field-programmable gate array processes the reference noise signal and the error signal based on polyphase decomposition to obtain a first sub-band signal and a second sub-band signal; and / or, The field-programmable gate array (FPGA) processes the reference noise signal based on polyphase decomposition and using the filtering parameters of the adaptive filter to obtain a noise-reduced signal, which is then provided to the loudspeaker.

[0016] In one embodiment of this application, the digital signal processor is used to provide a sweep signal to the speaker.

[0017] Thirdly, this application also provides a vehicle active noise cancellation system, including a reference sensor for acquiring a reference noise signal, a speaker for outputting a noise cancellation signal, an error sensor located within the audible area of ​​the vehicle for acquiring an error signal, and a vehicle active noise cancellation device as described in any of the above.

[0018] In one embodiment of this application, the reference sensor includes an accelerometer and a pickup device.

[0019] Fourthly, this application also provides a vehicle, the vehicle including a vehicle active noise cancellation device as described in any of the preceding claims, or including a vehicle active noise cancellation system as described in any of the preceding claims, or achieving active noise cancellation of the vehicle by a vehicle active noise cancellation method as described in any of the preceding claims.

[0020] The vehicle active noise reduction method provided in this application, after collecting a reference noise signal, obtains a signal at a low sampling rate through downsampling processing, and then processes the low sampling rate signal into multiple sub-band signals, which are then processed by an adaptive filter corresponding to each sub-band signal. This allows the filtering of each sub-band signal to be completed at a low sampling rate, effectively reducing the number of data points processed per unit time while meeting the requirements of effective noise processing within the noise reduction frequency band, thereby improving the noise reduction effect. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is an architecture diagram of an active noise reduction method provided by related technologies; Figure 2 This is a schematic flowchart of a vehicle active noise reduction method provided in an embodiment of this application; Figure 3 This is a flowchart illustrating the steps for updating an adaptive filter under each sub-band, as provided in an embodiment of this application. Figure 4 This is a flowchart illustrating the steps of an offline method for establishing a sub-band secondary path according to an embodiment of this application. Figure 5 This is a flowchart of a complete active noise reduction method for vehicles provided in an embodiment of this application; Figure 6a This is a schematic diagram illustrating the noise reduction effect of the vehicle active noise reduction method provided in this application embodiment under the wideband real vehicle noise frequency domain; Figure 6b This is a schematic diagram illustrating the noise reduction effect of the active vehicle noise reduction method provided by related technologies in the wideband real vehicle noise frequency domain; Figure 6c This is a schematic diagram illustrating the noise reduction effect of the vehicle active noise reduction method provided in this application embodiment under wideband real vehicle noise time domain (10 seconds); Figure 6d This is a schematic diagram illustrating the noise reduction effect of the vehicle active noise reduction method provided by the related technology in the wideband real vehicle noise time domain (10 seconds); Figure 7 This is a schematic diagram of the architecture of a processor provided in an embodiment of this application; Figure 8 This is a schematic diagram of the architecture of a vehicle active noise reduction system provided in an embodiment of this application; Figure 9 This is a schematic diagram of a headrest equipped with an error sensor provided in an embodiment of this application; Figure 10 This is a schematic diagram of the structure of a vehicle provided in an embodiment of this application. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] In this application, the term "exemplary" is used to mean "used as an example, illustration, or description." Any embodiment described as "exemplary" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes are not described in detail to avoid obscuring the description of this application with unnecessary detail. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0025] To clearly understand the vehicle active noise cancellation method, device, system, and vehicle provided in this application, the relevant implementation scheme of the active noise cancellation method will be described below. Specifically, vehicle active noise cancellation is a system that cancels in-vehicle noise in real time through the principle of sound wave interference. It typically includes four parts: a sound pickup system, a control module, a speaker system, and an adaptive algorithm. The sound pickup system is used to collect sound signals inside the vehicle, such as a collected reference noise signal and an error signal located at the human ear (used to indicate some residual noise that may remain after the reference noise signal is canceled by the noise cancellation signal played by the speaker system). The signal collected by the sound pickup system is processed by a filter in the control module to provide a noise-canceling signal for playback through the speaker system. In order to further improve the noise cancellation effect of the vehicle, the filter is usually continuously updated and iterated through an adaptive algorithm to reduce the collected error signal as much as possible, that is, to make the noise-canceling signal cancel out the reference signal as completely as possible. For details, please refer to [link to relevant documentation]. Figure 1 , Figure 1 An architecture diagram of an active noise reduction method provided for related technologies.

[0026] Among them, various noises during vehicle operation are collected by a sound pickup device as reference noise signals for the system input. via secondary path The signal obtained after processing The error signal collected by the sound pickup system Together with an improved filtered-x Least Mean Square (FxLMS) algorithm, an optimized filter W(z) is iteratively derived for the reference noise signal. After processing, the reverse noise reduction signal y(n) is broadcast through the loudspeaker system to a specific location to achieve active noise control.

[0027] In the aforementioned active noise reduction process, related technologies have proposed some improvements to the filter optimization algorithms, such as the time-frequency domain filtering x normalized minimum mean square error (TF-FxNLMS) algorithm or other improved algorithms, to improve the filter update effect and thus enhance the active noise reduction performance of vehicles. However, in practical applications, it has been found that these methods usually introduce additional computational overhead, inevitably leading to problems such as latency and insufficient computational redundancy in the vehicle, typically requiring additional hardware cost.

[0028] To address the aforementioned technical problems, this application proposes a vehicle active noise reduction method applicable to existing noise reduction frameworks, without introducing additional hardware costs and while maintaining the effectiveness of active noise reduction. Specifically, based on the Nyquist sampling theorem, it employs a low sampling rate that effectively handles noise within a specific noise reduction frequency band. Adaptive filtering of the signal, particularly sub-band signals, is performed at this low sampling rate, reducing the number of data points processed per unit of computation time, improving computational efficiency, reducing latency, and thus enhancing the noise reduction effect. To facilitate understanding, the following will provide a detailed description of the vehicle active noise reduction method provided in this application, particularly the related hardware and vehicle system architecture implemented based on this method, using specific embodiments.

[0029] Specifically, please refer to Figure 2 , Figure 2 This is a flowchart illustrating the steps of an active noise reduction method for a vehicle provided in this application. This method is mainly applied in vehicles with active noise reduction functionality. The vehicle typically includes a reference sensor for acquiring a reference noise signal, a speaker for outputting a noise reduction signal, and an error sensor located within the audible area of ​​the vehicle for acquiring an error signal. Descriptions of the reference noise signal, noise reduction signal, and error signal can be found in the foregoing descriptions; these will not be repeated in this application. Furthermore, descriptions of the reference sensor, speaker, and error sensor will be provided in detail in subsequent embodiments. Specifically, the active noise reduction method for a vehicle provided in this application includes steps S210-S220: S210, the acquired reference noise signal is downsampled to obtain a first signal at a low sampling rate, and the first signal is decomposed into first sub-band signals under multiple sub-bands.

[0030] In one embodiment of this application, the reference noise signal is typically acquired by relevant sensors mounted on the vehicle to reflect environmental noise, including but not limited to wind noise, road noise, and engine noise. For example, in one possible embodiment, the reference noise signal provided by this application includes vibration signals acquired by an acceleration sensor and sound signals acquired by a pickup device, which can more comprehensively reflect environmental noise and improve the noise reduction effect of the vehicle.

[0031] Building upon the foregoing, in the embodiments of this application, considering that this application is primarily for noise reduction in a specific noise reduction frequency band, the Nyquist sampling theorem is utilized. The signal can be processed into a first signal at a low sampling rate through downsampling for subsequent filtering, thus reducing the amount of data processed per unit computation time while matching the noise reduction frequency band. Specifically, taking the target noise reduction frequency band of 0-1.5kHz as an example, in one experimental scheme of this application, the acquired reference noise signal can be downsampled to a low sampling rate of 3kHz. That is, a low-pass filter with a cutoff frequency of 1.5kHz is used in the downsampling design. The low-pass filter used for downsampling can be designed and obtained through resampling functions in relevant software, such as the `resample` function in Matlab. Of course, based on actual needs, it is also feasible to perform other degrees of downsampling on the acquired reference noise signal to match the noise processing within the actual noise reduction frequency band. This application embodiment does not limit this. For ease of description, the embodiments in this application will be described using a low sampling rate of 3kHz.

[0032] Of course, in order to update the filter parameters at a low sampling rate, the error signal used to update the filter usually needs to be processed in the same way. That is, in the embodiments of this application, the error signal collected by the error sensor will also be processed into a second signal at a low sampling rate in the same way. This will not be repeated in the embodiments of this application.

[0033] Building upon the aforementioned foundation, while reducing the amount of data processed per unit computation time to improve the update rate of filter parameters, some embodiments of this application can appropriately improve the filter update algorithm to enhance the filter update effect, thereby improving the active noise reduction effect of the vehicle. For example, in one embodiment of this application, delay-free subband filtering can be used, that is, the signal at a low sampling rate is processed into multiple subbands and updated according to its corresponding filter parameters in each subband, and finally the updated signals of each subband are summed and output, and finally upsampled to restore the noise-reduced signal that can be played by the speaker. It should be noted that if delay-free subband filtering is not used at a low sampling rate, that is, the reference error signal is directly decomposed into multiple subband signals and updated through their corresponding filters, it has been found in practical applications that there is often an unavoidable delay problem, which is intolerable for the actual application of vehicles and seriously affects the user experience.

[0034] Specifically, there are various possible processing methods for decomposing a first signal at a low sampling rate into multiple sub-band signals. For example, different analysis filter banks can be used to process the first signal at the low sampling rate. These analysis filter banks can be implemented based on Discrete Fourier Transform (DFT) or Discrete Cosine Transform (DFT). However, considering that in practical applications, DFT-based filter banks output complex sub-band signals, requiring simultaneous calculation and storage of both real and imaginary parts, involving double multiplication and addition operations and hardware storage, resulting in high hardware resource consumption and low real-time performance, one embodiment of this application uses cosine modulation to generate the analysis filter bank. This outputs only the sub-band signal containing the real signal, reducing unnecessary complex number operations and storage, saving hardware storage space, improving signal transmission efficiency, and enhancing the real-time performance of the noise reduction system. In other words, in one embodiment of this application, decomposing the first signal into multiple sub-band signals includes: The first signal is decomposed into first sub-band signals under multiple sub-bands by analyzing the filter bank; The analysis filter bank includes multiple analysis filters obtained by processing the prototype filter through cosine modulation.

[0035] Specifically, the analysis filter bank obtained through cosine modulation mainly includes multiple analysis filters obtained by processing the prototype filter, i.e., a low-pass FIR filter (Finite Impulse Response), with cosine modulation. The low-pass FIR filter here can be implemented by designing a Kaiser filter or an equiripple filter; this application does not limit this implementation.

[0036] The subband decomposition of the signal is achieved using the analysis filter bank obtained by cosine modulation provided in this application embodiment. The filter bank can be perfectly reconstructed, satisfying the aliasing cancellation condition of adjacent channels without considering complex number operations. In particular, the analysis filter bank provided in this application can also be perfectly mapped to a high-efficiency polyphase structure for subsequent hardware, such as data processing of a Field Programmable Gate Array (FPGA), which will be described in detail in subsequent related embodiments.

[0037] Of course, similar to downsampling the error signal to obtain a second signal at a low sampling rate, in one embodiment of this application, in order to update the filters within each sub-band, it is also necessary to process the second signal at the low sampling rate into multiple sub-band signals by analyzing the filter bank to update the sub-band filters within each sub-band respectively. That is, in one embodiment of this application, the method further includes: The second signal is decomposed into multiple sub-band signals by analyzing the filter bank.

[0038] For details regarding the analysis of filter banks and subband filtering, please refer to the foregoing embodiments; these will not be repeated in the embodiments of this application.

[0039] Furthermore, in the process of downsampling the reference noise signal or error signal, in order to conform to the design of existing chip architectures, in one embodiment of this application, polyphase decomposition technology can be further incorporated, that is, the long filter of the signal is split into parallel short branch filters for processing in a field-programmable gate array. In other words, in one embodiment of this application, the downsampling of the acquired reference noise signal to obtain a first signal at a low sampling rate includes: The obtained reference noise signal is downsampled based on polyphase decomposition to obtain the first signal at a low sampling rate.

[0040] In the embodiments of this application, compared with the filtering process of long filters, the mode of short branch multi-group processing is more suitable for efficient parallel computing in real-time embedded systems (such as systems containing field-programmable gate arrays).

[0041] S220: After processing each of the first sub-band signals through its corresponding adaptive filter, the processed signals are upsampled to obtain a noise-reduced signal for the speaker output.

[0042] In one embodiment of this application, by passing each first sub-band signal through its corresponding adaptive filter, such as the filter obtained by updating the filter parameters through the aforementioned FxLMS algorithm, the output signals in each sub-band are summed to obtain the total output signal at a low sampling rate, such as the output signal at a 3kHz sampling rate. Based on this, the total output signal at the low sampling rate is upsampled to obtain the final signal adapted to the speaker output mode and provided to the speaker as a noise reduction signal for output. For example, if the speaker needs to output a 48kHz signal, in this embodiment of the application, the processed signal can be upsampled to 48kHz and provided to the speaker for output.

[0043] Furthermore, similar to the parallel processing using polyphase decomposition technology during downsampling, in one embodiment of this application, the upsampling process can also be further combined with polyphase decomposition technology for parallel processing to further improve data processing efficiency, thereby improving the active noise reduction effect of the vehicle. That is, the upsampling of the processed signal to obtain the noise-reduced signal for the speaker output includes: The processed signal is upsampled based on polyphase decomposition to obtain a noise-reduced signal for the speaker output.

[0044] Specifically, for the target noise reduction frequency band, such as 0-1.5kHz, a low-pass filter with a cutoff frequency of 1.5kHz can also be introduced in the upsampling process of this application. In this case, after the downsampling and upsampling of the polyphase decomposition, the accuracy of the output signal is 0-1.5kHz, which still meets the requirements for noise reduction. The relevant description of the low-pass filter can be referred to the above downsampling description, and will not be repeated here.

[0045] Furthermore, in the aforementioned process of processing each first sub-band signal using an adaptive filter corresponding to each sub-band signal, the iterative update effect of the adaptive filter will affect the noise reduction signal, thereby affecting the final noise reduction effect. Based on this, in one embodiment of this application, the adaptive filter corresponding to each first sub-band signal is determined based on the first sub-band signal and its corresponding second sub-band signal. The second sub-band signal is obtained by decomposing the second signal, and the second signal is obtained by downsampling the error signal. Specifically, please refer to... Figure 3 , Figure 3 A flowchart illustrating the steps for updating an adaptive filter in each sub-band, as provided in this application embodiment, specifically includes steps S310 to S330: S310, for each of the first sub-band signals, the first sub-band signal is processed through the corresponding sub-band secondary path to obtain the third sub-band signal.

[0046] Specifically, in one embodiment of this application, the obtained sub-band signal is processed through its corresponding sub-band secondary path to obtain a third sub-band signal. The secondary path under each sub-band typically needs to be established offline beforehand. In particular, considering that the filter update provided in this application is completed at a low sampling rate, this application also provides an offline establishment method for determining the sub-band secondary path corresponding to each sub-band. For details, please refer to [link to relevant documentation]. Figure 4 , Figure 4 A flowchart illustrating the steps of an offline method for establishing a sub-band secondary path provided in this application embodiment specifically includes steps S410 to S440: S410 upsamples the sweep frequency signal at a low sampling rate and plays it through the speaker.

[0047] In one embodiment of this application, in order to estimate the secondary path of the system interference term at a low sampling rate, it is necessary to first generate an excitation signal at the low sampling rate. Specifically, this excitation signal can be a frequency sweep signal provided by the frequency sweep function of relevant software, such as the sweeptone function in Matlab. For example, in one possible implementation, the frequency sweep signal can be a signal with a frequency of 0-1.5kHz at a sampling rate of 3kHz.

[0048] Based on this, the sweep frequency signal at a low sampling rate can be further upsampled and played through a corresponding speaker. For example, in one possible implementation, the sweep frequency signal can be upsampled to 48kHz. This application will not repeat the description of this.

[0049] In addition, to further improve the offline establishment effect of secondary paths, the obtained sweep frequency signal can be silenced for a period of time and then upsampled to the speaker for playback. This application embodiment does not impose any restrictions on this.

[0050] S420, the signal after playback is acquired through the error sensor.

[0051] After playing the upsampled sweep signal through the speaker, the signal after playback can be obtained through the error sensor. The signal after playback and the upsampled sweep signal can be understood, which roughly estimates the secondary path, that is, the entire system response of sound propagation from the speaker to the user's ear.

[0052] S430: After downsampling the played signal, it is decomposed into the frequency sweep signal and the frequency sweep signal, respectively.

[0053] In addition to the above, in order to estimate the secondary path of the downsampled system, the acquired signal after playback needs to be downsampled again to the sampling rate corresponding to the sweep signal. For example, if the sweep signal is a 0-1.5kHz frequency signal with a 3kHz sampling rate, the signal after playback will be downsampled again to 3kHz.

[0054] In addition to the above, in order to determine the secondary path within each sub-band, this application will further perform sub-band decomposition synchronously with the downsampled signal and the aforementioned sweep signal, for example, by performing sub-band decomposition through the aforementioned analysis filter bank, thereby obtaining the next set of signals corresponding to each sub-band, namely the playback signal and sweep signal after downsampling of each sub-band.

[0055] S440, the decomposed subband signals are used as excitation signals and response signals respectively to determine the subband secondary path.

[0056] Building upon the aforementioned foundation, after decomposing the played signal and the sweep signal into multiple sub-band signals, for each sub-band, a set of sub-band signals (including the sub-band decomposition signals of the downsampled played signal and the sweep signal) are used as the excitation signal (sub-band decomposition signal of the sweep signal) and the response signal (sub-band decomposition signal of the downsampled played signal), respectively. This allows the determination of the secondary path describing the relationship between the two signals, i.e., the sub-band secondary path under the corresponding sub-band. For example, in one possible implementation, this can be achieved through relevant software, such as the impzest function in Matlab. This application will not repeat the description of this in the embodiments.

[0057] It should be noted that, compared with directly playing and acquiring the sweep frequency signal at a high sampling rate, and using it as the excitation signal and response signal respectively after subband decomposition to estimate the subband secondary path, the offline establishment method of the secondary path provided in this application can more accurately describe the overall response error of the system at a low sampling rate, making the analysis of filter parameters more accurate.

[0058] S320, determine the adaptive step size for each sub-band based on the third sub-band signal and the second sub-band signal corresponding to the first sub-band signal.

[0059] After the secondary path obtained by the aforementioned offline estimation method processes the subband signal under each subband to obtain the third subband signal, the corresponding adaptive step size under each subband can be calculated according to the third subband signal and its corresponding second subband signal, that is, the subband signal within the corresponding subband obtained after downsampling and subband decomposition of the error signal, according to the preset formula. The adaptive step size is usually used to describe the adjustment step size when updating the filtering parameters of the adaptive filter.

[0060] Specifically, in one embodiment of this application, the adaptive step size under each sub-band can also be determined by the signal energy of the sub-band decomposition signal of the first sub-band signal, that is, the reference noise signal under low sampling rate. In particular, it can be determined based on the normalization result of its signal energy. The specific calculation formula will not be repeated in this embodiment of the application.

[0061] Furthermore, in one embodiment of this application, different initialization step sizes can be set for each sub-band based on actual needs. Specifically, the initialization step size can be based on user experience or experimentally measured under specific experimental conditions. Of course, in some embodiments, the initialization step size of each sub-band can be set based on the signal energy of the first sub-band signal within each sub-band. This application does not limit this.

[0062] S330, update the filtering parameters of the adaptive filter corresponding to the first sub-band signal according to the adaptive step size.

[0063] Based on the aforementioned foundation, and using the adaptive step size obtained in each sub-band, the filtering parameters of the adaptive filter in each sub-band are updated accordingly. This yields the updated adaptive filtering parameters for processing the first sub-band signal, thereby outputting the corresponding signal.

[0064] It should be noted that the adaptive filter update process provided in this application is an iterative update process. That is, after the vehicle's active noise cancellation function is activated, reference noise signals and error signals are continuously collected, and the adaptive filter is continuously updated according to the aforementioned streaming computation process. The updated filter is used for noise reduction processing in the next time step, and error signals continue to be collected to update the adaptive filter. The specific update process can be referred to in the foregoing. Figure 1 The relevant architecture diagrams are not repeated in this application embodiment.

[0065] To clearly understand the complete process of the vehicle active noise reduction method provided in this application's embodiments, the following will describe the vehicle active noise reduction method of this application in conjunction with specific embodiments and specific application scenarios. Please refer to... Figure 5 , Figure 5 The flowchart of a complete vehicle active noise reduction method provided in this application embodiment specifically includes the following steps.

[0066] Specifically, a complete active noise reduction method for vehicles mainly includes two processes: "filtering" and "iterative optimal filter".

[0067] The first step in filtering is to collect broadband non-steady-state noise as the input signal using a reference microphone and accelerometer. Specifically, this application, by placing the sensor inside the vehicle, can collect noise signals from vehicles traveling at 90 km / h on a highway as a reference signal.

[0068] The second step in filtering is to perform polyphase decomposition on the input signal and downsample it to 3kHz. A lower sampling rate results in fewer data points processed per unit time, leading to higher computational efficiency. Furthermore, according to the Nyquist sampling theorem, a 3kHz sampling rate is also suitable for effectively handling noise in the target noise reduction frequency band of 0-1.5kHz. Here, the low-pass filter designed for downsampling has a cutoff frequency of 1.5kHz. The downsampled low-pass filter can be designed and obtained using the `resample` function in Matlab.

[0069] The third step in filtering is to calculate the analysis filter bank for the 3kHz subband using cosine modulation. The effective frequency range of this filter bank is also 0-1.5kHz. Cosine modulation is used here because it can perfectly reconstruct the filter bank, satisfy the aliasing cancellation condition between adjacent channels, eliminates the need for complex number operations, and can perfectly map to an efficient polyphase structure. The prototype filter for the analysis filter bank (a low-pass FIR filter) can be implemented by designing a Kaiser filter or an equiripple filter.

[0070] The fourth step of filtering is to input the 3kHz input signal into the analysis filter bank to obtain the corresponding subband signal.

[0071] The fifth step of filtering is to pass each subband signal through the corresponding adaptive filter to obtain the subband output signal at 3kHz.

[0072] The sixth step of filtering is to sum the signals to obtain the total output signal at 3kHz.

[0073] The seventh step in filtering is to perform polyphase decomposition on the total output signal and upsample it to 48kHz, while also introducing a low-pass filter with a cutoff frequency of 1.5kHz. The output signal after polyphase decomposition, downsampling, and upsampling has an accuracy of 0-1.5kHz, corresponding to the target noise reduction frequency band. Upsampling to 48kHz is necessary because the speaker's output signal needs to be 48kHz.

[0074] The eighth step in filtering is to acquire the error signal at the error microphone location. This error signal is a full-frequency 48kHz signal.

[0075] It is important to note that the adaptive filter used in the fifth step of filtering needs to be implemented through an "iterative optimal filter" process, as follows: The "iterative optimal filter" process operates at a sampling rate of 3kHz. The reference signal portion directly uses the sub-band signal obtained in the fourth step of the "filtering" process. The error signal portion undergoes polyphase decomposition of the error signal acquired in the final step of the "filtering" process, downsampling it to 3kHz, and then the sub-band error signal is obtained through filter bank analysis. In other words, the first step of the iterative optimal filter is to perform polyphase decomposition of the high-sampling-rate error signal, downsample it, and obtain the sub-band error signal through filter bank analysis.

[0076] Based on this, the second step of the iterative optimal filter is to use the subband input signal of the low sampling rate overanalysis filter bank, and the reference signal part directly adopts the subband signal obtained in the fourth step of the "filtering" process.

[0077] Furthermore, the secondary path part needs to introduce the secondary path corresponding to each sub-band at 3kHz. That is, the third step of the iterative optimal filter is to introduce the secondary path corresponding to each sub-band at a low sampling rate.

[0078] Specifically, the step size part designs a different initial step size for each sub-band and introduces normalization processing so that the step size is adjusted according to the signal energy during the iteration process. That is, the fourth step of the iterative optimal filter is to have an independent initial step size for each sub-band, and the iteration part performs normalization processing.

[0079] Finally, the fifth step of the iterative optimal filter is to solve for the sub-band adaptive filter and output it to the filtering process. In other words, after determining the above, the adaptive filter for each sub-band can be iteratively solved at 3kHz and output to the "filtering" process, which is the fifth step of the filtering process.

[0080] To clearly understand the difference in performance between the vehicle active noise reduction method provided in this application and the active noise reduction methods provided in related technologies, please refer to one embodiment of this application. Figures 6a-6d , Figures 6a-6d The diagram illustrates the effects of active noise reduction implemented using different algorithms.

[0081] Specifically, Figure 6a This diagram illustrates the noise reduction effect of the vehicle active noise reduction method provided in this application embodiment under a wideband real vehicle noise frequency domain. Figure 6b This is a schematic diagram illustrating the noise reduction effect of the active vehicle noise reduction method provided by the related technology in the wideband real vehicle noise frequency domain. Figure 6c This is a schematic diagram illustrating the noise reduction effect of the vehicle active noise reduction method provided in this application embodiment under wideband real vehicle noise time domain (10 seconds). Figure 6dThe diagram illustrates the noise reduction effect of the vehicle active noise reduction method provided by the related technologies in the time domain (10 seconds) of broadband real vehicle noise. It can be seen that, for broadband non-steady-state noise, compared with the active noise reduction methods provided by the related technologies, the vehicle active noise reduction method provided in this application can achieve faster convergence of noise suppression, and the noise control frequency band is significantly widened, and the noise control depth is also significantly increased.

[0082] The vehicle active noise reduction method provided in this application obtains a signal at a low sampling rate by downsampling after collecting a reference noise signal, and then processes the low sampling rate signal into multiple sub-band signals, which are then processed by an adaptive filter corresponding to each sub-band signal. This allows the filtering of each sub-band signal to be completed at a low sampling rate, effectively reducing the number of data points processed per unit time while meeting the requirements of effective noise processing within the noise reduction frequency band, thereby improving the noise reduction effect.

[0083] Building upon the aforementioned solutions, and to better implement the aforementioned active vehicle noise reduction method, one embodiment of this application further provides a vehicle active noise reduction device for implementing the aforementioned active vehicle noise reduction method. Specifically, the vehicle active noise reduction device may exist in the form of a processor, used to process the steps of the vehicle active noise reduction method provided in any of the aforementioned embodiments, thereby achieving active noise reduction for the vehicle. In particular, the vehicle active noise reduction device, i.e., the processor architecture, provided in this application is particularly suitable for processing the vehicle active noise reduction method of this application. For ease of description, specific embodiments will be described below.

[0084] Please see Figure 7 , Figure 7 This is a schematic diagram of the architecture of a processor provided in an embodiment of this application. Specifically, the processor includes a digital signal processor and a field-programmable gate array (FPGA). The digital signal processor, also known as a DSP (Digital Signal Processing), is mainly used for processing real-time signals, while the field-programmable gate array (FPGA) is a programmable logic processor that can perform general-purpose functions.

[0085] Based on this, the field-programmable gate array (FPGA) is used to acquire a reference noise signal collected by a reference sensor and an error signal collected by an error sensor through an interface, and to process the reference noise signal and the error signal to obtain a first sub-band signal and a second sub-band signal; to provide the first sub-band signal and the second sub-band signal to the digital signal processor (DSP), and to receive the filtering parameters of the adaptive filter returned by the DSP; to process the reference noise signal based on the filtering parameters of the adaptive filter to obtain a noise-reduced signal and to provide it to the loudspeaker; The digital signal processor is used to determine the filtering parameters of the adaptive filter based on the first sub-band signal and the second sub-band signal sent by the field-programmable gate array.

[0086] Furthermore, in one embodiment of this application, the field-programmable gate array (FPGA) processes the reference noise signal and the error signal based on polyphase decomposition to obtain a first sub-band signal and a second sub-band signal; and / or, the FPGA processes the reference noise signal based on polyphase decomposition and using the filtering parameters of the adaptive filter to obtain a noise-reduced signal and provides it to the loudspeaker. The FPGA provided by this application embodiment can achieve polyphase decomposition of the signal, enabling high-speed signal processing using a parallel filter structure (short branches), greatly improving data processing efficiency, reducing system latency, and enhancing the effect of active noise reduction.

[0087] Furthermore, in one embodiment of this application, the digital signal processor is used to provide a sweep signal to the speaker, that is, for offline estimation of the secondary path.

[0088] Specifically, the processor can also be integrated on a hardware development board and embedded in the system to implement the algorithm for the vehicle active noise reduction method provided in this application. Specifically, the hardware development board can also integrate more related functional modules, including but not limited to the speaker output, network port, and various interfaces shown in the figure, such as A2B 1, A2B 2, A2B 3, and A2B 4. For example, these interfaces can be assigned to accelerometers and various sound pickup devices to enable input of various signals, while the network port uses a specific protocol to capture signals from the corresponding node. Furthermore, other circuit connections will not be described again in this embodiment.

[0089] The vehicle active noise reduction device provided in this application obtains a signal at a low sampling rate by downsampling after collecting a reference noise signal, and then processes the low sampling rate signal into multiple sub-band signals, which are then processed by an adaptive filter corresponding to each sub-band signal. This allows the filtering of each sub-band signal to be completed at a low sampling rate, effectively reducing the number of data points processed per unit time while meeting the requirements of effective noise processing within the noise reduction frequency band, thereby improving the noise reduction effect.

[0090] This application also provides an active noise reduction system for vehicles; for details, please refer to [link to relevant documentation]. Figure 8 , Figure 8 This is a schematic diagram of the architecture of a vehicle active noise reduction system provided in an embodiment of this application. Specifically, it includes a reference sensor 810, a speaker 820, an error sensor 830, and a vehicle active noise reduction device 840.

[0091] The reference sensor 810 is used to acquire noise signals. It typically includes an accelerometer (e.g., an accelerometer) and a pickup device (e.g., a reference microphone) for collecting noise signals at various locations. Specifically, taking the deployment of an active noise cancellation system on a vehicle as an example, the accelerometer can be a three-axis (x, y, z) accelerometer. Two accelerometers (one on each side) can be deployed on the front suspension towers to collect road noise and engine vibration transmitted into the passenger compartment. Additionally, two more accelerometers (one on each side) can be deployed at the rear suspension mounting points to collect road noise and vibration transmitted into the passenger compartment from the rear seats. It is important to note that the accelerometer measures vibrations transmitted from the wheels to the suspension via the vehicle body. Therefore, the accelerometer needs to be mounted on the "sprung mass" side, i.e., the body portion supported above the suspension springs and shock absorbers. Otherwise, the signal may be too "local" and "sharp," unsuitable as a reference noise signal for full-vehicle active noise cancellation. Furthermore, the reference microphone can be an automotive-grade digital microphone, and two reference microphones can be deployed near the A-pillars on the driver's and passenger's sides respectively. This can effectively capture wind noise sources on both sides and provide the system with reference signals differentiating between the left and right sides. In addition, a reference microphone can be placed near the front footwell to acquire engine and road noise signals, and a reference microphone can be placed under the rear seats to capture rear wheel road noise and exhaust noise. The speaker 820 is used to output noise reduction signals, which are usually obtained by the vehicle's active noise cancellation device through the aforementioned active noise cancellation method. Specifically, the speaker includes, but is not limited to, a speaker array, a horn, or other devices that can output sound signals. They can usually be symmetrically arranged on both sides of the vehicle compartment, and this application embodiment does not limit this.

[0092] The error sensor 830 is primarily located within the vehicle's audible area to collect error signals reflecting the actual noise reduction effect. It is typically a sound pickup device, such as an error microphone. Specifically, in one possible implementation, the error microphone could be an automotive-grade digital microphone installed at the bottom of the headrest to monitor the sound pressure level near the passenger's ear, ensuring stability and accuracy for active noise cancellation in that area. For example, please refer to... Figure 9 , Figure 9 This is a schematic diagram of a headrest equipped with an error sensor, provided as an embodiment of this application.

[0093] Of course, the above-described implementation scheme is merely a feasible embodiment of a vehicle active noise cancellation system using a "4+4+4" arrangement, namely 4 reference microphones, 4 accelerometers, and 4 error microphones, and should not be construed as a limitation on the technical solution of this application. Any conventional design made by those skilled in the art based on actual needs regarding the number or deployment location of any reference sensor, speaker, and error sensor in this active noise cancellation system should be within the scope of protection claimed in this application.

[0094] Furthermore, the architecture of the vehicle active noise reduction device 840 in the system and the active noise reduction algorithm used to implement vehicle active noise reduction can be referred to the aforementioned related embodiments, and this application embodiment will not provide any related descriptions.

[0095] The vehicle active noise reduction system provided in this application obtains a low sampling rate signal by collecting a reference noise signal and then downsampling it. The low sampling rate signal is then processed into multiple sub-band signals, which are then processed by an adaptive filter corresponding to each sub-band signal. This allows the filtering of each sub-band signal to be completed at a low sampling rate. Under the premise of effectively processing noise in the noise reduction frequency band, this system effectively reduces the number of data points processed per unit time and improves the noise reduction effect.

[0096] In one embodiment of this application, a vehicle is also provided, the vehicle including a vehicle active noise cancellation device as described in any of the preceding claims, or including a vehicle active noise cancellation system as described in any of the preceding claims, or achieving active noise cancellation of the vehicle through a vehicle active noise cancellation method as described in any of the preceding claims. Specifically, please refer to... Figure 10 , Figure 10 This is a structural diagram of a vehicle provided in an embodiment of this application, which can be used to achieve active noise reduction inside the vehicle.

[0097] The descriptions of active noise cancellation devices, active noise cancellation systems, and active noise cancellation methods can be found in the descriptions of the foregoing related embodiments, and will not be repeated in the embodiments of this application.

[0098] It should be noted that, in the data processing stage, the technical solution of this application has strictly limited the scope of data collection to the minimum necessary to achieve the technical objectives, preventing the acquisition of irrelevant information. For any user information to be collected, the data subject will be clearly informed and their consent obtained. Furthermore, technologies such as encrypted storage and access control are employed to strengthen data security and ensure the security and compliance of the entire data processing process. The technical model and decision-making mechanism are based on objective technical parameters and do not introduce unnecessary parameters such as gender or age that may lead to discrimination, resolutely eliminating algorithmic discrimination and upholding public order and good morals. In addition, the specification fully describes the technical implementation methods, application scenarios, and compliance protection details. The claims are consistent with the content of the specification, key compliance designs are clear and verifiable, and the overall technical design is guided by the protection of public interests and adherence to social ethics, without any circumstances that harm public interests or violate public order and good morals.

[0099] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0100] The above provides a detailed description of a vehicle active noise reduction method, device, system, and vehicle provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for active noise reduction in vehicles, characterized in that, The vehicle has a reference sensor for acquiring a reference noise signal, a speaker for outputting a noise-reducing signal, and an error sensor located within the audible area of ​​the vehicle for acquiring an error signal. The method includes: The acquired reference noise signal is downsampled to obtain a first signal at a low sampling rate, and the first signal is decomposed into first sub-band signals under multiple sub-bands; Each of the first sub-band signals is processed by its corresponding adaptive filter, and the processed signal is upsampled to obtain the noise reduction signal for the speaker output. The adaptive filter corresponding to each of the first sub-band signals is determined based on the first sub-band signal and its corresponding second sub-band signal. The second sub-band signal is obtained by decomposing the second signal and by downsampling the error signal.

2. The method according to claim 1, characterized in that, The step of decomposing the first signal into multiple sub-band signals includes: The first signal is decomposed into first sub-band signals under the plurality of sub-bands by analyzing the filter bank; The analysis filter bank includes multiple analysis filters obtained by processing the prototype filter through cosine modulation.

3. The method according to claim 1, characterized in that, The step of downsampling the acquired reference noise signal to obtain a first signal at a low sampling rate includes: The reference noise signal is downsampled based on polyphase decomposition to obtain a first signal at a low sampling rate. The step of upsampling the processed signal to obtain the noise-reduced signal for the speaker output includes: The processed signal is upsampled based on polyphase decomposition to obtain a noise-reduced signal for the speaker output.

4. The method according to claim 1, characterized in that, The step of processing each of the first sub-band signals through its corresponding adaptive filter includes: For each of the first sub-band signals, the first sub-band signal is processed by its corresponding multiple branch adaptive filters based on polyphase decomposition.

5. The method according to claim 1, characterized in that, The adaptive filter corresponding to each of the first sub-band signals is determined through the following steps: For each of the first sub-band signals, the first sub-band signal is processed through the corresponding sub-band secondary path to obtain the third sub-band signal; The adaptive step size for each sub-band is determined based on the third sub-band signal and the second sub-band signal corresponding to the first sub-band signal. The filtering parameters of the adaptive filter corresponding to the first sub-band signal are updated according to the adaptive step size.

6. The method according to claim 5, characterized in that, The sub-band secondary path corresponding to the first sub-band signal is determined through the following steps: The sweep frequency signal at a low sampling rate is upsampled and then played through the speaker. The signal after playback is acquired through the error sensor; After downsampling the played signal, it is then combined with the frequency sweep signal for sub-band decomposition. The sub-band signals obtained from the decomposition are used as excitation signals and response signals, respectively, to determine the secondary paths of the sub-bands.

7. The method according to claim 5, characterized in that, The method further includes: setting different initialization step sizes for each sub-band, and / or determining an adaptive step size for each sub-band based on the normalization result of the signal energy of the first sub-band signal.

8. The method according to any one of claims 1 to 7, characterized in that, The reference noise signal includes vibration signals acquired by an accelerometer and sound signals acquired by a radio receiver.

9. A vehicle active noise reduction device, characterized in that, The vehicle active noise reduction device includes a processor, which is used to achieve active noise reduction of the vehicle by means of the vehicle active noise reduction method as described in any one of claims 1 to 8.

10. The apparatus according to claim 9, characterized in that, The processor includes a digital signal processor and a field-programmable gate array; The field-programmable gate array (FPGA) is used to acquire a reference noise signal collected by a reference sensor and an error signal collected by an error sensor through an interface, and to process the reference noise signal and the error signal to obtain a first sub-band signal and a second sub-band signal; to provide the first sub-band signal and the second sub-band signal to the digital signal processor (DSP), and to receive the filtering parameters of the adaptive filter returned by the DSP; to process the reference noise signal based on the filtering parameters of the adaptive filter to obtain a noise-reduced signal and to provide it to the loudspeaker; The digital signal processor is used to determine the filtering parameters of the adaptive filter based on the first sub-band signal and the second sub-band signal sent by the field-programmable gate array.

11. The apparatus according to claim 10, characterized in that, The field-programmable gate array (FPGA) processes the reference noise signal and the error signal based on polyphase decomposition to obtain a first sub-band signal and a second sub-band signal; and / or, The field-programmable gate array (FPGA) processes the reference noise signal based on polyphase decomposition and using the filtering parameters of the adaptive filter to obtain a noise-reduced signal, which is then provided to the loudspeaker.

12. The apparatus according to claim 10, characterized in that, The digital signal processor is used to provide a sweep frequency signal to the speaker.

13. A vehicle active noise reduction system, characterized in that, It includes a reference sensor for acquiring a reference noise signal, a speaker for outputting a noise reduction signal, an error sensor located within the audible area of ​​the vehicle for acquiring an error signal, and a vehicle active noise reduction device as described in any one of claims 9 to 12.

14. The system according to claim 13, characterized in that, The reference sensor includes an accelerometer and a radio receiver.

15. A vehicle, characterized in that, The vehicle includes a vehicle active noise reduction device as described in any one of claims 9 to 12, or includes a vehicle active noise reduction system as described in any one of claims 13 to 14, or achieves active noise reduction of the vehicle by a vehicle active noise reduction method as described in any one of claims 1 to 8.

Images