Active noise control method based on multi-channel adaptive filtering

Abstract

The invention discloses an active noise control method based on multichannel adaptive filtering, and relates to the technical field of automobile noise control and signal processing. The invention discloses an active noise control method based on multi-channel adaptive filtering. The method comprises the following steps: acquiring original noise signals at different positions in a vehicle through a reference microphone; performing multi-band band-pass filtering on the acquired noise signal to extract noise components of a target frequency band; inputting the filtered reference signal into a multi-channel adaptive filtering controller to generate an anti-noise signal; estimating a secondary path to generate an actual anti-noise signal based on the transmission path of the loudspeaker and the error microphone; through a feedback signal collected by an error microphone, closed-loop iteration is carried out to update and control a filter coefficient, and real-time offset and dynamic convergence of a noise signal are realized; through vector superposition of an original noise signal and an actual anti-noise signal, efficient and real-time suppression of multi-channel and broadband noise is realized. The method has more excellent multi-target noise reduction performance and robustness.

Description

Technical Field

[0001] This invention relates to the field of automotive noise control and signal processing technology, and in particular to an active noise control method based on multi-channel adaptive filtering. Background Technology

[0002] With the development of the automotive industry, in-vehicle noise has become increasingly significant in its impact on passenger comfort and driving experience, making it a key issue in automotive acoustic design. Under the combined influence of multiple noise sources such as engine noise, wind noise, and road vibration, in-vehicle noise exhibits multi-channel, wide-bandwidth characteristics, significantly affecting passenger auditory comfort. Therefore, effectively suppressing in-vehicle noise and improving ride comfort has become an important direction in modern automotive noise control technology research. Active noise control technology, by generating inverse signals to cancel noise signals, provides an efficient suppression method for complex noise environments and has become an important approach to improving the in-vehicle acoustic environment.

[0003] While existing multi-channel active noise control methods can suppress in-vehicle noise to some extent, they still face many challenges. Because the frequency characteristics of noise vary at different locations within the vehicle, a single bandpass filter struggles to accurately extract the target frequency band for each channel, thus affecting the accuracy of the generated noise-resistant signal. Furthermore, traditional methods often lack explicit design for the coordinated operation of delay lines in the secondary path estimation and control filter iteration processes, leading to slow convergence or instability in adaptive noise control methods under multi-channel, wide-bandgap conditions. Summary of the Invention

[0004] The purpose of this invention is to address the above-mentioned shortcomings by proposing an active noise control method based on multi-channel adaptive filtering. This method employs a multi-band bandpass filter module to extract the reference signal by frequency division, thereby achieving multi-target noise reduction for different noise sources. Furthermore, it utilizes a closed-loop adaptive filtering structure based on delay line collaboration to improve the convergence speed and stability of the control filter.

[0005] The present invention specifically adopts the following technical solution:

[0006] An active noise control method based on multi-channel adaptive filtering includes the following steps:

[0007] 1) Collect raw noise signals from different locations inside the vehicle using a reference microphone;

[0008] 2) Perform multi-band bandpass filtering on the acquired noise signal to extract the noise components of the target frequency band;

[0009] 3) Input the filtered reference signal into the multi-channel adaptive filter controller to generate an anti-noise signal;

[0010] 4) Estimate the secondary path based on the transmission path between the loudspeaker and the error microphone to generate the actual noise-resistant signal;

[0011] 5) The feedback signal collected by the error microphone is used to iteratively update the control filter coefficients in a closed loop, thereby achieving real-time cancellation and dynamic convergence of noise signals;

[0012] 6) By vector superposition of the original noise signal and the actual anti-noise signal, efficient and real-time suppression of multi-channel, wideband noise can be achieved.

[0013] Preferably, in step 2), the multi-band bandpass filtering is based on the frequency band characteristics of noise collected by the reference microphone at different locations inside the vehicle. A corresponding bandpass filter is configured for each channel to extract the target frequency band noise of each channel and suppress interference from irrelevant frequency bands, thus achieving targeted noise suppression. The bandpass filtering process for each channel is expressed as follows:

[0014] First, let the acquired reference microphone signal be... The target frequency band is If the Hz value is such that a corresponding finite impulse response (FIR) bandpass filter is designed, then... The ideal impulse response is:

[0015] ;

[0016] in, For an ideal impulse response, Sampling frequency, For filter length, For the filter center index, For filter coefficient index;

[0017] Then, based on the ideal impulse response With Hamming window function Calculate the filter coefficients :

[0018] ;

[0019] in, For the first Window function coefficients for each sampling point This is the filter length;

[0020] Finally, a multi-band bandpass filter is used to filter the reference signal to obtain the target frequency band component. The target frequency band component of the reference signal can be expressed as:

[0021] ;

[0022] in, The signal after filtering;

[0023] By selecting different channels and The noise features at each location are extracted to provide an effective reference signal for subsequent multi-channel adaptive filtering.

[0024] Preferably, the specific estimation method for the secondary path in the secondary path estimation method in 4) is as follows:

[0025] First, let the filtered reference signal be... The corresponding error microphone signal is The construction length is Adaptive filter The filter output is:

[0026] ;

[0027] Then, by calculating the error signal Update the filter coefficients so that the filter output... Approximation error microphone signal To achieve secondary path estimation, the filter update process is as follows:

[0028] ;

[0029] in, This is the step size parameter;

[0030] Finally, the converged filter coefficients As an estimated secondary path transfer function, it is used by a multi-channel adaptive filter controller to generate accurate noise-resistant signals.

[0031] Preferably, the coefficient iteration method for the closed-loop iterative update of the control filter coefficients in step 5) is as follows:

[0032] First, the acquired signals of each reference channel Input reference signal delay line, forming a length of Sliding vector:

[0033] ;

[0034] in, To control the filter length, Used to generate the output signal of the control filter and ensure that the most recent consecutive sample can be accessed with each update;

[0035] Then, the output of the reference signal delay line is passed through a secondary path filter. Forming a filter reference signal delay line:

[0036] ;

[0037] in, This represents the convolution operation. It is used to simulate the actual effect of the reference signal after passing through the speaker-error microphone path, so as to accurately calculate the error signal in the closed loop.

[0038] Secondly, the output delay line of the control filter stores the historical values ​​of the control filter output signal:

[0039] ;

[0040] in, To control the filter coefficient vector, With error signal Alignment is used for adaptive algorithm updates;

[0041] Finally, the feedback signal was collected through the error microphone. With control filter output error Compute closed-loop update:

[0042] ;

[0043] Through the coordinated operation of the reference signal delay line, the filtered reference signal delay line, and the control filter output delay line, the entire closed-loop adaptive algorithm can stably update the control filter coefficients, ensuring real-time cancellation and dynamic convergence of noise signals in a multi-channel, wide-bandgap environment.

[0044] The present invention has the following beneficial effects:

[0045] This invention achieves frequency isolation and targeted suppression of multi-channel noise within the vehicle through a multi-band bandpass filtering mechanism; it accurately models the acoustic transmission characteristics from the speaker to the error microphone through an adaptive filtering secondary path estimation mechanism, eliminating the impact of modeling errors on the control algorithm; and it achieves real-time cancellation and dynamic convergence of multi-channel, wideband noise through closed-loop iterative updates of the control filter coefficients. Compared with existing methods, this invention effectively improves the multi-target noise reduction capability, convergence speed, and control accuracy of the active noise control system, significantly enhancing system stability and noise suppression effects, and providing reliable guarantees for in-vehicle comfort and driving experience. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the signal processing flow for an active noise control method based on multi-channel adaptive filtering.

[0047] Figure 2This is an embodiment of the present invention. In the figure, (a) is the original dual-channel noise signal, (b) is the signal processed by the active noise control method, and (c) is the residual noise signal. Detailed Implementation

[0048] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings and specific examples:

[0049] An active noise control method based on multi-channel adaptive filtering, combined with body Figure 1 This includes the following steps:

[0050] 1) Acquire raw noise signals from different locations inside the vehicle using a reference microphone. Noise signal acquisition: Acquire noise signals from different locations inside the vehicle using a reference microphone, such as... Figure 2 The original noise signal matrix shown in (a) The aforementioned signal acquisition matrix Represented as:

[0051] ;

[0052] in, and These represent the noise signals at the two sampling points, , This represents the number of sampling points.

[0053] 2) Multi-band bandpass filtering is applied to the acquired noise signal to extract the noise components of the target frequency band. Multi-band bandpass filtering is based on the frequency band characteristics of noise collected by the reference microphone at different locations inside the vehicle. A corresponding bandpass filter is configured for each channel to extract the target frequency band noise of each channel and suppress interference from irrelevant frequency bands, achieving targeted noise suppression. The bandpass filtering process for each channel is represented as follows:

[0054] First, let the acquired reference microphone signal be... The target frequency band is If the Hz value is such that a corresponding finite impulse response (FIR) bandpass filter is designed, then... The ideal impulse response is:

[0055] ;

[0056] in, For an ideal impulse response, Sampling frequency, For filter length, For the filter center index, For filter coefficient index;

[0057] Then, based on the ideal impulse response With Hamming window function Calculate the filter coefficients :

[0058] ;

[0059] in, For the first Window function coefficients for each sampling point This is the filter length;

[0060] Finally, a multi-band bandpass filter is used to filter the reference signal to obtain the target frequency band component. The target frequency band component of the reference signal can be expressed as:

[0061] ;

[0062] in, The signal after filtering;

[0063] By selecting different channels and The noise features at each location are extracted to provide an effective reference signal for subsequent multi-channel adaptive filtering.

[0064] 3) Input the filtered reference signal into the multi-channel adaptive filter controller to generate an anti-noise signal;

[0065] Let the control filter coefficients be: The corresponding output signal is:

[0066] ;

[0067] At this time, the control signal It is generated by superimposing the outputs of all channels. Control signal. It can be represented as:

[0068]

[0069] in This represents the number of channels.

[0070] 4) Estimate the secondary path based on the transmission path between the loudspeaker and the error microphone to generate the actual noise-resistant signal;

[0071] The specific estimation method for secondary paths is as follows:

[0072] First, let the filtered reference signal be... The corresponding error microphone signal is The construction length is Adaptive filter The filter output is:

[0073] ;

[0074] Then, by calculating the error signal Update the filter coefficients so that the filter output... Approximation error microphone signal To achieve secondary path estimation, the filter update process is as follows:

[0075] ;

[0076] in, This is the step size parameter;

[0077] Finally, the converged filter coefficients As an estimated secondary path transfer function, it is used by a multi-channel adaptive filter controller to generate accurate noise-resistant signals.

[0078] 5) The feedback signal acquired by the error microphone is used to iteratively update the control filter coefficients in a closed loop, achieving real-time noise cancellation and dynamic convergence. The iteration method for updating the control filter coefficients in a closed loop is as follows:

[0079] First, the acquired signals of each reference channel Input reference signal delay line, forming a length of Sliding vector:

[0080] ;

[0081] in, To control the filter length, Used to generate the output signal of the control filter and ensure that the most recent consecutive sample can be accessed with each update;

[0082] Then, the output of the reference signal delay line is passed through a secondary path filter. Forming a filter reference signal delay line:

[0083] ;

[0084] in, This represents the convolution operation. It is used to simulate the actual effect of the reference signal after passing through the speaker-error microphone path, so as to accurately calculate the error signal in the closed loop.

[0085] Secondly, the output delay line of the control filter stores the historical values ​​of the control filter output signal:

[0086] ;

[0087] in, To control the filter coefficient vector, With error signal Alignment is used for adaptive algorithm updates;

[0088] Finally, the feedback signal was collected through the error microphone. With control filter output error Compute closed-loop update:

[0089] ;

[0090] Through the coordinated operation of the reference signal delay line, the filtered reference signal delay line, and the control filter output delay line, the entire closed-loop adaptive algorithm can stably update the control filter coefficients, ensuring real-time cancellation and dynamic convergence of noise signals in a multi-channel, wide-bandgap environment.

[0091] 6) By vector superposition of the original noise signal and the actual anti-noise signal, efficient and real-time suppression of multi-channel, broadband noise is achieved. For example... Figure 2 The noise immunity signal output by the system shown in (b) is superimposed on the original noise signal. This superposition process can be represented as:

[0092] ;

[0093] Then, the residual noise is minimized. The process of minimizing the residual noise is as follows:

[0094] ;

[0095] Finally, the residual noise of the system is as follows Figure 2 As shown in (c), an active noise control method based on multi-channel adaptive filtering is used to achieve real-time suppression and stable convergence of multi-channel, wideband noise.

[0096] This solution also provides a target frequency band extraction mechanism based on multi-band bandpass filtering. This mechanism configures a corresponding bandpass filter for each channel based on the frequency band characteristics of noise signals collected by the reference microphone at different locations inside the vehicle, thereby achieving frequency division extraction of noise from different locations and effective suppression of interference from irrelevant frequency bands. By rationally dividing the passband range of each channel, the main noise characteristics of each location can be selectively preserved, enhancing the effectiveness of the reference signal. In specific implementation, the bandpass filter adopts a finite impulse response structure and is combined with a Hamming window function design to balance frequency selectivity and filtering smoothness. By performing multi-band bandpass filtering on the collected signal, noise components of different target frequency bands can be obtained, forming multiple reference signal inputs. This multi-band bandpass filtering mechanism can achieve frequency division isolation and targeted extraction of multi-channel noise, effectively avoiding the limitation of a single filter being unable to simultaneously process multiple target noises under wide bandwidth conditions, and providing a more representative input signal basis for subsequent multi-channel adaptive filtering processes.

[0097] This solution also provides a secondary path estimation mechanism based on adaptive filtering. This mechanism is used to accurately model the acoustic transfer characteristics between the loudspeaker and the error microphone in an active noise control system, providing a basis for the control filter to generate accurate noise-resistant signals. In this mechanism, an adaptive filter model is first constructed using the filtered reference signal and the feedback signal collected by the error microphone. By continuously adjusting the filter coefficients, the model output gradually approximates the actual error signal, thereby achieving dynamic estimation of the secondary path transfer function. This dynamic estimation process continuously optimizes and converges the filter parameters by minimizing the difference between the reference signal and the error signal. Once the filter parameters converge, their final coefficients characterize the acoustic transfer characteristics of the path from the loudspeaker to the error microphone, which can be used to simulate the response of the real secondary path in an active noise control closed-loop system. This adaptive estimation mechanism effectively eliminates the influence of secondary path modeling errors on the control algorithm, improving the accuracy of the control filter update process and the overall stability of the system.

[0098] This solution also provides a closed-loop iterative adaptive update mechanism for the control filter. This mechanism is used to adjust the control filter coefficients in real time in a multi-channel active noise control system, enabling the system to dynamically cancel noise based on feedback signals collected by the error microphone. In this mechanism, a reference signal delay line is first used to retain continuous sampling information for each channel, ensuring the control filter accesses the latest input features in each iteration. Then, a filtered reference signal delay line simulates the actual impact of the reference signal after passing through the speaker-error microphone path, making the calculation of the error signal closer to the actual acoustic response. Simultaneously, the control filter output delay line records the historical values ​​of the filter output, providing a basis for iterative updates to the adaptive algorithm. Through the coordinated operation of the reference signal delay line, the filtered reference signal delay line, and the control filter output delay line, the update mechanism can stably and in real-time update the control filter coefficients, achieving dynamic noise cancellation in multi-channel, wideband environments. This closed-loop iterative update mechanism effectively improves the system's convergence speed and noise reduction accuracy, providing reliable closed-loop adjustment capabilities for multi-channel active noise control.

[0099] Of course, the above description is not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. An active noise control method based on multi-channel adaptive filtering, characterized in that, Includes the following steps: 1) Collect raw noise signals from different locations inside the vehicle using a reference microphone; 2) Perform multi-band bandpass filtering on the acquired noise signal to extract the noise components of the target frequency band; 3) Input the filtered reference signal into the multi-channel adaptive filter controller to generate an anti-noise signal; 4) Estimate the secondary path based on the transmission path between the loudspeaker and the error microphone to generate the actual noise-resistant signal; 5) The feedback signal collected by the error microphone is used to iteratively update the control filter coefficients in a closed loop, thereby achieving real-time cancellation and dynamic convergence of noise signals; 6) By vector superposition of the original noise signal and the actual anti-noise signal, efficient and real-time suppression of multi-channel, wideband noise can be achieved.

2. The active noise control method based on multi-channel adaptive filtering as described in claim 1, characterized in that, In section 2), the multi-band bandpass filtering is based on the frequency band characteristics of noise collected by the reference microphone at different locations inside the vehicle. A corresponding bandpass filter is configured for each channel to extract the target frequency band noise of each channel and suppress interference from irrelevant frequency bands, achieving targeted noise suppression. The bandpass filtering process for each channel is represented as follows: First, let the acquired reference microphone signal be... The target frequency band is If the Hz value is such that a corresponding finite impulse response (FIR) bandpass filter is designed, then... The ideal impulse response is: ； in, For an ideal impulse response, Sampling frequency, For filter length, For the filter center index, For filter coefficient index; Then, based on the ideal impulse response With Hamming window function Calculate the filter coefficients : ； in, For the first Window function coefficients for each sampling point This is the filter length; Finally, a multi-band bandpass filter is used to filter the reference signal to obtain the target frequency band component. The target frequency band component of the reference signal can be expressed as: ； in, The signal after filtering; By selecting different channels and The noise features at each location are extracted to provide an effective reference signal for subsequent multi-channel adaptive filtering.

3. The active noise control method based on multi-channel adaptive filtering as described in claim 1, characterized in that, The specific estimation method for the secondary path in 4) is as follows: First, let the filtered reference signal be... The corresponding error microphone signal is The construction length is Adaptive filter The filter output is: ； Then, by calculating the error signal Update the filter coefficients so that the filter output... Approximation error microphone signal To achieve secondary path estimation, the filter update process is as follows: ； in, This is the step size parameter; Finally, the converged filter coefficients As an estimated secondary path transfer function, it is used by a multi-channel adaptive filter controller to generate accurate noise-resistant signals.

4. The active noise control method based on multi-channel adaptive filtering as described in claim 1, characterized in that, 5) The coefficient iteration method for updating the control filter coefficients in the closed-loop iteration is as follows: First, the acquired signals of each reference channel Input reference signal delay line, forming a length of Sliding vector: ； in, To control the filter length, Used to generate the output signal of the control filter and ensure that the most recent consecutive sample can be accessed with each update; Then, the output of the reference signal delay line is passed through a secondary path filter. Forming a filter reference signal delay line: ； in, This represents the convolution operation. It is used to simulate the actual effect of the reference signal after passing through the speaker-error microphone path, so as to accurately calculate the error signal in the closed loop.

5. Secondly, the output delay line of the control filter stores the historical values ​​of the control filter output signal: ； in, To control the filter coefficient vector, With error signal Alignment is used for adaptive algorithm updates; Finally, the feedback signal was collected through the error microphone. With control filter output error Compute closed-loop update: 。

Images