A wheel speed signal anti-interference processing method, system, device and medium based on automobile broadband road noise

Abstract

The present application relates to a kind of wheel speed signal anti-interference processing method, system, equipment and medium based on automobile broadband road noise, belong to the wheel speed signal control field in automobile engineering.The method includes obtaining original wheel speed signal data set, interference received signal, according to the interference received signal, interference signal training set and interference signal test set are constructed, the interference signal training set is overlapped and framed and short-time Fourier transform is obtained Space-frequency domain features, according to the spectral characteristics of the original wheel speed signal data set, the interference signal training set is added training label, according to the space-frequency domain features and the training label, interference suppression model is trained by complex signal convolutional neural network, the performance evaluation test of the interference suppression model is carried out by the interference signal test set, and anti-interference wheel speed signal is output by the interference suppression model meeting performance evaluation test.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of automobile road safety control, and particularly relates to a wheel speed signal anti-interference processing method, system, device and medium based on automobile broadband road noise. BACKGROUND

[0002] When the automobile is running at high speed, real-time and accurate acquisition and processing of the wheel speed signal is the basis for active safety control systems such as indirect tire pressure monitoring systems, anti-lock braking systems and drive slip control systems. The frequent operation of various relays, transformers, electromagnets and other inductive circuit devices in the circuit on the vehicle causes noise in the circuit. In addition, various electromagnetic waves in the atmospheric environment also cause the sensor to sense noise with a very wide frequency band and very small amplitude. The broadband noise caused by the internal interference and external interference of the vehicle circuit makes it difficult to accurately obtain the instantaneous wheel speed signal. At the same time, inaccurate acquisition of the wheel speed signal also causes large errors and large divergence of the motion parameters such as wheel slip rate / slip rate and yaw angular velocity calculated from the wheel speed signal, thereby reducing the signal-to-noise ratio of the wheel speed sensor signal and the divergence of the wheel speed signal calculation data, which is not conducive to safe and effective control of the automobile.

[0003] Traditional Kalman filtering, wavelet denoising and digital Wiener filtering can appropriately reduce the noise in the wheel speed signal and reduce the divergence of the automobile motion parameters calculated based on the wheel speed signal, but cannot completely eliminate the noise. At present, Kalman filter and genetic iteration algorithm are usually used to suppress sensor signal errors, a fault-tolerant method based on analytical redundancy is used to solve the sensor signal oscillation distortion problem, and a frequency domain minimum mean square error-based adaptive enhancer is used as a prediction of the wheel speed sensor signal to improve the signal-to-noise ratio. However, the above methods are not suitable for the case where the automobile is running at high speed, and have limitations in processing broadband road noise with a flat spectrum. SUMMARY

[0004] The wheel speed signal anti-interference processing method based on automobile broadband road noise simultaneously extracts features from the interference received signal formed by the wheel speed target signal and the broadband noise interference source, fully learns the mutual internal relationship between the interference received signal and the original wheel speed signal through a complex signal convolutional neural network, suppresses the interference components to obtain an interference suppression model, improves the signal-to-noise ratio of the wheel speed sensor signal and the accuracy of the wheel speed signal calculation data, and ensures the safety of automobile road travel. The limitations of the traditional noise active method in processing broadband road noise with a flat spectrum when the automobile is running at high speed are solved, and the signal-to-noise ratio of the wheel speed sensor signal and the accuracy of the wheel speed signal calculation data are improved. The specific technical solution is as follows:

[0005] A wheel speed signal anti-interference processing method based on automobile broadband road noise, comprising:

[0006] S1: obtaining an original wheel speed signal data set, collecting interference receiving signals through an internal microphone array of an automobile, the interference receiving signals including a wheel speed target signal and a broadband noise interference source, and constructing an interference signal training set and an interference signal test set according to the interference receiving signals;

[0007] S2: performing overlapping frame processing on the interference signal training set to obtain discrete signal frames, performing short-time Fourier transform on the discrete signal frames to obtain amplitude spectrum and phase spectrum, splicing to obtain space-time frequency domain features according to the amplitude spectrum and the phase spectrum, and adding training labels to the interference signal training set according to the spectral features of the original wheel speed signal data set;

[0008] S3: training an interference suppression model according to the space-time frequency domain features and the training labels through a complex signal convolutional neural network;

[0009] S4: performing performance evaluation test on the interference suppression model through the interference signal test set to obtain performance evaluation results.

[0010] Preferably, the microphone array is a uniform linear array composed of microphone sensors, and the interference receiving signals are collected through the uniform linear array.

[0011] Preferably, the step S1 specifically comprises:

[0012] The uniform linear array is mounted, the uniform linear array is composed of isotropic elements, the interference receiving signals are obtained through the uniform linear array, the interference receiving signals are incident on the uniform linear array from different directions, the interference receiving signals include a wheel speed target signal and a broadband noise interference source, isotropic element variables are obtained from the interference signals through the isotropic elements, the isotropic element variables include isotropic element amplitude variation and isotropic element time delay, and the calculation formula of the interference receiving signals is:

[0013]

[0014] wherein x z is an interference receiving signal, z represents a zth isotropic element, z is a positive integer, ω l is an incident frequency of the interference receiving signal, k represents a kth group of isotropic element variables, k is a positive integer, x z (ω l , k) is the kth group of isotropic element variables, the incident frequency ω lD is the total number of the wheel speed target signals, d represents the dth wheel speed target signal, s d is the frequency domain of the wheel speed target signals, s d (ω l is the frequency domain of the dth wheel speed signal, e is the base of the logarithmic function, j is the complex unit, Φ d is the incident angle of the dth wheel speed target signal, τ z (Φ d ) is the relative time delay of the wheel speed target signals, Δ is the inter-element distance of the omnidirectional elements, c represents the sound speed, and the value is 340 m / s, Q is the total number of the broadband noise interference sources, q represents the qth broadband noise interference source, ρ is the amplitude of the broadband noise interference source, ρ z (q) is the amplitude of the qth broadband noise interference source to the zth omnidirectional element, A is a constant, σ z (q) is the distance from the qth broadband noise interference source to the zth omnidirectional element, i q (ω l is the frequency signal of the kth group of omnidirectional element variables, the qth broadband noise interference source, μ z (q) represents the propagation time delay of the qth broadband noise interference source to the zth omnidirectional element.

[0015] Preferably, the step S3 specifically comprises:

[0016] S301: the complex signal convolutional neural network is a convolutional neural network for processing complex signals;

[0017] S302: the space-time frequency domain features are obtained by complex convolution to obtain a space-time frequency domain feature map;

[0018] S303: the space-time frequency domain feature map is obtained by maximum value pooling to obtain a space-time frequency domain pooling feature map;

[0019] S304: the space-time frequency domain pooling feature map is obtained by a fully connected layer to obtain a feature transfer function, the feature transfer function is a transfer function between the interference receiving signal and the omnidirectional element, and a wheel speed test signal is obtained according to an optimal solution of the feature transfer function, the wheel speed test signal is a wheel speed target signal of the omnidirectional element;

[0020] S305: a wheel speed signal loss value is obtained according to the wheel speed test signal and the training label by a loss function, and a calculation formula of the loss function is:

[0021]

[0022] wherein, COST represents the wheel speed signal loss value, N represents the number of frequency spectrum frames inputted in each batch, w represents a complex convolution kernel, x deNoise,r(w) represents the real part of the wheel speed signal output by the complex signal convolution neural network, x deNoise,i(w) represents the imaginary part of the wheel speed signal output by the complex signal convolution neural network, x noNoise,r(w) represents the real part of the training label, x noNoise,i(w) represents the imaginary part of the training label.

[0023] S306: obtaining the gradient of each layer weight and the gradient of each layer bias of the complex signal convolution neural network by back propagation according to the wheel speed signal loss value;

[0024] S307: updating the parameters of the complex signal convolution neural network to obtain the interference suppression model by the gradient of each layer weight and the gradient of each layer bias of the complex signal convolution neural network.

[0025] Preferably, the expression of the complex convolution is:

[0026]

[0027] w = w r +jw i ,

[0028] x = x r +jx i ,

[0029] wherein, x represents a complex matrix, x r represents the real part of the interference receiving signal, x i represents the imaginary part of the interference receiving signal, w represents a complex convolution kernel, w r represents the real part parameter of the complex convolution kernel, w i represents the imaginary part parameter of the complex convolution kernel, j represents a complex unit, and * represents a convolution operation.

[0030] Preferably, the specific steps of the back propagation are:

[0031] S306-1: obtaining the gradient of each layer loss value by a sigmoid activation function according to the wheel speed signal loss value;

[0032] S306-2: obtaining the gradient of each layer weight and the gradient of each layer bias of the complex signal convolution neural network by back propagation layer by layer according to the gradient of each layer loss value by applying the chain rule.

[0033] Preferably, the step S4 specifically comprises:

[0034] The performance evaluation result is calculated by a signal-to-noise ratio according to the interference receiving signal and the original wheel speed signal, and the specific calculation formula is:

[0035] SNR improvement = SNR after -SNR before ,

[0036]

[0037] x(t) nosiy = x(t) total -x(t)ideal,

[0038]

[0039] x(t)knosiy = x(t)ktotal - x(t)ideal,

[0040] wherein, SNR improvement represents the performance evaluation result, SNR before represents the signal-to-noise ratio of the interference signal training set which is not processed by the interference suppression model, x(t) ideal represents the original wheel speed signal, x(t) nosiy represents a noise signal, x(t) total represents the interference receiving signal, SNR after represents the signal-to-noise ratio of the interference signal training set which is processed by the interference suppression model, x(t) knosiy represents a noise signal in the anti-interference wheel speed signal, x(t) ktotal represents the anti-interference wheel speed signal.

[0041] It is judged whether the performance evaluation result is a positive number, if the performance evaluation result is a positive number, the interference suppression model passes the performance evaluation test;

[0042] If the performance evaluation result is a negative number, the interference suppression model does not pass the performance evaluation test;

[0043] It is judged whether the interference suppression model passes the performance evaluation test, if the interference suppression model passes the performance evaluation test, an anti-interference wheel speed signal is obtained by the interference suppression model according to a real-time wheel speed signal, the real-time wheel speed signal is a wheel speed signal collected by a wheel speed sensor of a vehicle in real time, and the anti-interference wheel speed signal is a real-time wheel speed signal from which noise is removed by the interference suppression model.

[0044] A wheel speed signal anti-interference processing system based on automobile broadband road noise comprises a signal acquisition module, a signal preprocessing module, a model training module and a performance evaluation module, and comprises:

[0045] The signal acquisition module is used for acquiring an original wheel speed signal data set, collecting an interference receiving signal through an internal microphone array of the automobile, the interference receiving signal including a wheel speed target signal and a broadband noise interference source, and constructing an interference signal training set and an interference signal test set according to the interference receiving signal.

[0046] The signal preprocessing module is used for performing overlapping frame processing on the interference signal training set to obtain a discrete signal frame, performing short-time Fourier transform on the discrete signal frame to obtain an amplitude spectrum and a phase spectrum, splicing the amplitude spectrum and the phase spectrum to obtain a space-time frequency domain feature, and adding a training label to the interference signal training set according to a spectrum feature of the original wheel speed signal data set.

[0047] The model training module is used for training an interference suppression model through a complex signal convolutional neural network according to the space-time frequency domain feature and the training label.

[0048] The performance evaluation module is used for performing performance evaluation test on the interference suppression model through the interference signal test set to obtain a performance evaluation result.

[0049] An electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor implements the above-mentioned wheel speed signal anti-interference processing method based on automobile broadband road noise when executing the program.

[0050] A storage medium containing computer executable instructions for executing the above-mentioned wheel speed signal anti-interference processing method based on automobile broadband road noise when executed by a computer processor.

[0051] The beneficial effects of the present application are:

[0052] (1) The interference receiving signal is segmented on the time axis by overlapping frame to obtain the discrete signal frame, and there is partial overlap between frames, which guarantees the information continuity between frames, can more accurately analyze the spectrum characteristics of the signal, and reduces the computational complexity;

[0053] (2) The discrete signal frame is converted from the time domain to the frequency domain by short-time Fourier transform, the spectrum information of the signal in each time period is analyzed through a sliding window, the amplitude spectrum and the phase spectrum are obtained, and it is beneficial to subsequent differentiation of the wheel speed signal and the broadband noise;

[0054] (3) The amplitude spectrum, phase spectrum, real part and imaginary part of the interference received signal are processed by complex convolution, which well guarantees the integrity of the signal, avoids the problem of losing the original inherent characteristics of the signal caused by simply separating the real part and imaginary part or processing the amplitude spectrum and phase spectrum respectively, and improves the ability to suppress the interference components in the interference received signal;

[0055] (4) The deep complex convolution neural network considers the amplitude and phase information of the signal at the same time, fully learns the mutual internal relationship between the interference received signal and the original wheel speed signal, suppresses the interference components in the interference received signal, outputs an anti-interference wheel speed signal, and improves the signal-to-noise ratio of the wheel speed sensor signal and the accuracy of the wheel speed signal calculation data. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to facilitate the understanding of those skilled in the art, the present application will be further described below with reference to the accompanying drawings.

[0057] Figure 1 A flow chart of a wheel speed signal anti-interference processing method based on automobile broadband road noise according to the present application. DETAILED DESCRIPTION

[0058] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined application purpose, the specific embodiments, structures, features and effects according to the present application are described in detail as follows in combination with the drawings and preferred embodiments.

[0059] Please refer to Figure 1 A wheel speed signal anti-interference processing method based on automobile broadband road noise, comprising:

[0060] S1: obtaining an original wheel speed signal data set, collecting an interference received signal through an internal microphone array of an automobile, the interference received signal including a wheel speed target signal and a broadband noise interference source, constructing an interference signal training set and an interference signal test set according to the interference received signal;

[0061] S2: performing overlapping frame processing on the interference signal training set to obtain a discrete signal frame, performing short-time Fourier transform on the discrete signal frame to obtain an amplitude spectrum and a phase spectrum, splicing the amplitude spectrum and the phase spectrum to obtain a space-time frequency domain feature, and adding a training label to the interference signal training set according to the frequency spectrum feature of the original wheel speed signal data set;

[0062] S3: training an interference suppression model according to the space-time frequency domain feature and the training label through a complex signal convolution neural network;

[0063] S4: performing performance evaluation test on the interference suppression model through the interference signal test set to obtain a performance evaluation result.

[0064] In step S1, the microphone array is a uniform linear array composed of microphone sensors, and the interference receiving signal is collected through the uniform linear array.

[0065] In this embodiment, a uniform linear array composed of isotropic elements is mounted to obtain the interference receiving signal, the interference receiving signal is incident on the uniform linear array from different directions, the interference receiving signal includes a wheel speed target signal and a broadband noise interference source, isotropic element variables are obtained from the isotropic elements according to the interference signal, the isotropic element variables include isotropic element amplitude variation and isotropic element time delay, and the interference receiving signal calculation formula is:

[0066]

[0067] wherein x z is the interference receiving signal, z represents the zth isotropic element, z is a positive integer, ω l is the incident frequency of the interference receiving signal, k represents the kth group of isotropic element variables, k is a positive integer, x z (ω l ,k) is the kth group of isotropic element variables, and the incident frequency ω l corresponding to the received signal in the zth isotropic element receiving data, D is the total number of wheel speed target signals, d represents the dth wheel speed target signal, s d is the frequency domain of the wheel speed target signal, s d (ω l ,k) represents the kth group of isotropic element variables, and the frequency domain of the dth wheel speed signal, e is the base of the logarithmic function, j is the complex unit, Φ d is the incident angle of the dth wheel speed target signal, τ z (Φ d ) is the relative time delay of the wheel speed target signal, Δ is the isotropic element spacing, c represents the sound speed, and the value is 340 m / s, Q is the total number of broadband noise interference sources, q represents the qth broadband noise interference source, ρ is the amplitude of the broadband noise interference source, ρ z (q) is the amplitude of the qth broadband noise interference source to the zth isotropic element, A is a constant, σ z (q) is the distance from the qth broadband noise interference source to the zth isotropic element, i q (ω l ,k) represents the kth group of isotropic element variables, and the frequency signal of the qth broadband noise interference source, μ z(q) represents the propagation delay of the qth wideband noise interference source to the zth isotropic element.

[0068] In step S3, an interference suppression model is trained by a complex signal convolutional neural network according to the spatio-temporal frequency domain features and the training labels. Specifically, the following steps can be implemented:

[0069] S301: The complex signal convolutional neural network is a convolutional neural network for processing complex signals.

[0070] S302: The spatio-temporal frequency domain features are obtained by complex convolution to obtain a spatio-temporal frequency domain feature map.

[0071] The expression of the complex convolution is:

[0072]

[0073] w = w r +jw i ,

[0074] x = x r +jx i ,

[0075] wherein x is a complex matrix, x r is the real part of the interference received signal, x i is the imaginary part of the interference received signal, w is a complex convolution kernel, w r is the real part parameter of the complex convolution kernel, w i is the imaginary part parameter of the complex convolution kernel, j is a complex unit, and * represents convolution operation.

[0076] S303: The spatio-temporal frequency domain feature map is obtained by maximum value pooling to obtain a spatio-temporal frequency domain pooling feature map.

[0077] S304: The spatio-temporal frequency domain pooling feature map is obtained by a fully connected layer to obtain a feature transfer function. The feature transfer function is a transfer function between the interference received signal and the isotropic element. A wheel speed test signal is obtained according to the optimal solution of the feature transfer function. The wheel speed test signal is a wheel speed target signal of the isotropic element.

[0078] S305: A wheel speed signal loss value is obtained according to the wheel speed test signal and the training label by a loss function. The calculation formula of the loss function is:

[0079]

[0080] wherein COST is a wheel speed signal loss value, N represents the number of frequency spectrum frames input in each batch, w is a complex convolution kernel, x deNoise,r(w)represents a real part of the wheel speed signal output by the complex signal convolutional neural network, x deNoise,i(w) represents an imaginary part of the wheel speed signal output by the complex signal convolutional neural network, x noNoise,r(w) represents a real part of the training label, x noNoise,i(w) represents an imaginary part of the training label;

[0081] S306: obtaining, according to the wheel speed signal loss value, a gradient of each layer weight of the complex signal convolutional neural network and a gradient of each layer bias of the complex signal convolutional neural network through back propagation;

[0082] S307: updating parameters of the complex signal convolutional neural network through the gradient of each layer weight of the complex signal convolutional neural network and the gradient of each layer bias of the complex signal convolutional neural network to obtain the interference suppression model.

[0083] It should be noted that in step S301, the complex signal convolutional neural network comprises 3 convolutional layers, 3 pooling layers and 1 fully connected layer; in step S302, the complex signal convolutional neural network uses amplitude and phase, and real and imaginary parts of frequency spectrum to perform complex convolution operation to obtain the spatio-temporal frequency domain feature map when processing the spatio-temporal frequency domain feature.

[0084] In this embodiment, according to the wheel speed signal loss value, a gradient of each layer weight of the complex signal convolutional neural network and a gradient of each layer bias of the complex signal convolutional neural network are obtained through back propagation, which is specifically implemented by the following steps:

[0085] S306-1: obtaining a loss value gradient of each layer through a sigmoid activation function according to the wheel speed signal loss value;

[0086] S306-2: obtaining the gradient of each layer weight of the complex signal convolutional neural network and the gradient of each layer bias of the complex signal convolutional neural network through back propagation layer by layer according to the loss value gradient of each layer by applying the chain rule.

[0087] The step S4 specifically comprises:

[0088] The performance evaluation result is obtained through SNR calculation according to the interference receiving signal and the original wheel speed signal, and the specific calculation formula is:

[0089] SNR improvement = SNR after -SNR before ,

[0090]

[0091] x(t) nosiy = x(t) total -x(t)ideal,

[0092]

[0093] x(t)knosiy=x(t)ktotal-x(t)ideal,

[0094] where SNR improvement represents the performance evaluation result, SNR before represents the signal-to-noise ratio of the interference signal training set processed by the interference suppression model, x(t) ideal represents the original wheel speed signal, x(t) nosiy represents the noise signal, x(t) total represents the interference received signal, SNR after represents the signal-to-noise ratio of the interference signal training set processed by the interference suppression model, x(t) knosiy represents the noise signal in the anti-interference wheel speed signal, x(t) ktotal represents the anti-interference wheel speed signal;

[0095] determining whether the performance evaluation result is positive, if the performance evaluation result is positive, the interference suppression model passes the performance evaluation test;

[0096] if the performance evaluation result is negative, the interference suppression model fails the performance evaluation test;

[0097] determining whether the interference suppression model passes the performance evaluation test, if the interference suppression model passes the performance evaluation test, an anti-interference wheel speed signal is obtained from the interference suppression model according to a real-time wheel speed signal, the real-time wheel speed signal is a wheel speed signal collected by a vehicle wheel speed sensor in real time, and the anti-interference wheel speed signal is a real-time wheel speed signal from which noise is removed by the interference suppression model.

[0098] A wheel speed signal anti-interference processing system based on vehicle broadband road noise includes a signal acquisition module, a signal preprocessing module, a model training module, and a performance evaluation module.

[0099] The signal acquisition module is configured to obtain an original wheel speed signal dataset, collect an interference received signal through an array of microphones inside a vehicle, the interference received signal including a wheel speed target signal and a broadband noise interference source, and construct an interference signal training set and an interference signal test set according to the interference received signal.

[0100] The signal preprocessing module is used for performing overlapping frame processing on the interference signal training set to obtain discrete signal frames, performing short-time Fourier transform on the discrete signal frames to obtain amplitude spectrum and phase spectrum, splicing the amplitude spectrum and the phase spectrum to obtain a space-time frequency domain feature, and adding a training label to the interference signal training set according to the spectrum feature of the original wheel speed signal data set.

[0101] The model training module is used for training an interference suppression model through a complex signal convolutional neural network according to the space-time frequency domain feature and the training label.

[0102] The performance evaluation module is used for performing performance evaluation test on the interference suppression model through the interference signal test set to obtain a performance evaluation result.

[0103] The computer storage medium of the embodiment of the application can adopt any combination of one or more computer readable media. The computer readable medium can be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium may, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination thereof. More specific examples (non-exhaustive list) of the computer readable storage medium include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, the computer readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or component.

[0104] The computer readable signal medium can include a data signal propagating in a baseband or as part of a carrier wave propagating through a transmission medium, in which the computer readable program code is carried. Such a propagating data signal can take on many forms, including but not limited to electromagnetic signals, optical signals or any suitable combination thereof. The computer readable signal medium can also be any computer readable medium that is not a computer readable storage medium and that can transmit, propagate or transport a program for use by or in connection with an instruction execution system, device or component.

[0105] The computer readable media on which the program code can be carried by any suitable medium, including but not limited to wireless, wired, optical fiber cable, RF, and the like, or any suitable combination of these. Computer program code for carrying out operations of the present application can be written in one or more programming languages, or combinations of languages, including object oriented, such as Java, Smalltalk, C++, and conventional procedural, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, 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 server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0106] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application in any form. Although the present application has been disclosed as above with the preferred embodiment, it is not intended to limit the present application, and any person skilled in the art can make some changes or modifications to the equivalent embodiments within the scope of the technical solution of the present application, without departing from the technical solution of the present application. Any modification, equivalent change and modification of the above embodiments, which does not depart from the technical solution of the present application, is still within the scope of the technical solution of the present application.

Claims

1. A method for anti-interference processing of wheel speed signals based on vehicle broadband road noise, characterized in that, include: S1: Obtain the original wheel speed signal dataset, collect interference received signals through the car's internal microphone array, the interference received signals include wheel speed target signals and broadband noise interference sources, and construct interference signal training set and interference signal test set based on the interference received signals; The microphone array is a uniform linear array composed of microphone sensors, and the interference received signal is collected through the uniform linear array; The specific method for constructing the uniform linear array is as follows: Equipped with a uniform linear array, the uniform linear array is composed of isotropic array elements. The interference received signal is obtained through the uniform linear array. The interference received signal is incident on the uniform linear array from different directions. The interference received signal includes a wheel speed target signal and a broadband noise interference source. Isotropic array element variables are obtained from the interference received signal through the isotropic array elements. The isotropic array element variables include isotropic array element amplitude variation and isotropic array element time delay. S2: The interference signal training set is subjected to overlapping frame processing to obtain discrete signal frames. The discrete signal frames are subjected to short-time Fourier transform to obtain amplitude spectrum and phase spectrum. The amplitude spectrum and phase spectrum are spliced ​​to obtain spatiotemporal frequency domain features. Training labels are added to the interference signal training set according to the spectral features of the original wheel speed signal dataset. S3: Based on the spatiotemporal frequency domain features and the training labels, an interference suppression model is obtained by training a complex signal convolutional neural network; S4: The performance evaluation results are obtained by performing a performance evaluation test on the interference suppression model using the interference signal test set.

2. The method for anti-interference processing of wheel speed signals based on vehicle broadband road noise according to claim 1, characterized in that, Step S1 specifically includes: The formula for calculating the interference received signal is as follows: Where, x z To interfere with the received signal, z represents the z-th isotropic array element, where z takes the value of a positive integer, ω. l Let x be the incident frequency of the interference received signal, k represent the k-th group of isotropic array element variables, and k takes the value of a positive integer. z (ω l (k) represents the variable of the k-th isotropic array element and the incident frequency ω in the data received by the z-th isotropic array element. l For the corresponding received signals, D is the total number of the wheel speed target signals, d represents the d-th wheel speed target signal, and s d For the frequency domain of the wheel speed target signal, s d (ω l (k) represents the frequency domain of the isotropic array element variables of the k-th group and the wheel speed signal of the d-th group, e is the base of the logarithmic function, j is the complex unit, and Φ d Let τ be the incident angle of the d-th wheel speed target signal. z (Φ d ) represents the relative time delay of the target signal of the wheel speed, Δ represents the spacing between the isotropic array elements, c represents the speed of sound, which is 340 m / s, Q represents the total number of broadband noise interference sources, q represents the q-th broadband noise interference source, and ρ represents the amplitude of the broadband noise interference source. z (q) represents the amplitude of the q-th broadband noise interference source when it reaches the z-th isotropic array element, where A is a constant and σ is a constant. z (q) represents the distance from the q-th broadband noise interference source to the z-th isotropic array element, i q (ω l (k) represents the frequency signal of the isotropic array element variable of the k-th group and the broadband noise interference source of the q-th group, μ z (q) represents the propagation delay of the q-th broadband noise interference source to the z-th isotropic array element.

3. The method for anti-interference processing of wheel speed signals based on vehicle broadband road noise according to claim 1, characterized in that, Step S3 specifically includes: S301: The complex signal convolutional neural network is a convolutional neural network for processing complex signals; S302: Obtain a spatiotemporal frequency domain feature map by performing complex convolution on the spatiotemporal frequency domain features; S303: Obtain a spatiotemporal frequency domain pooled feature map by max pooling the spatiotemporal frequency domain feature map; S304: The spatiotemporal frequency domain pooling feature map is passed through a fully connected layer to obtain a feature transfer function. The feature transfer function is the transfer function between the interference received signal and the isotropic array element. The wheel speed test signal is obtained by optimally resolving the feature transfer function. The wheel speed test signal is the wheel speed target signal of the isotropic array element. S305: Obtain the wheel speed signal loss value based on the wheel speed test signal and the training label using a loss function. The formula for calculating the loss function is as follows: Where COST is the wheel speed signal loss value, N represents the number of spectral frames in each batch of input, w is the complex convolution kernel, and x deNoise,r(w) x represents the real part of the wheel speed signal output by the complex signal convolutional neural network. deNoise,i(w) x represents the imaginary part of the wheel speed signal output by the complex signal convolutional neural network. noNoise,r(w) x represents the real part of the training label. noNoise,i(w) Represents the imaginary part of the training labels; S306: Based on the wheel speed signal loss value, the weights of each layer of the complex signal convolutional neural network and the bias gradients of each layer of the complex signal convolutional neural network are obtained through backpropagation; S307: The interference suppression model is obtained by updating the parameters of the complex signal convolutional neural network through the weights of each layer and the bias gradient of each layer of the complex signal convolutional neural network.

4. The method for anti-interference processing of wheel speed signals based on vehicle broadband road noise according to claim 3, characterized in that, The expression for the complex convolution is: w=w r +jw i , x=x r +jx i , Where x is a complex matrix, x r Let x be the real part of the interference received signal. i Let w be the imaginary part of the interference received signal, and w be the complex convolution kernel. r Let w be the real part parameter of the complex convolution kernel. i is the imaginary part parameter of the complex convolution kernel, j is the complex unit, and * represents the convolution operation.

5. The method for anti-interference processing of wheel speed signals based on vehicle broadband road noise according to claim 3, characterized in that, The specific steps of the back propagation are as follows: S306-1: Obtain the gradient of the loss value of each layer by using the sigmoid activation function based on the wheel speed signal loss value; S306-2: Based on the loss value gradient of each layer, the weights of each layer of the complex signal convolutional neural network and the bias gradient of each layer of the complex signal convolutional neural network are obtained by backpropagation layer by layer using the chain rule.

6. The method for anti-interference processing of wheel speed signals based on vehicle broadband road noise according to claim 1, characterized in that, Step S4 specifically includes: The performance evaluation result is obtained by calculating the signal-to-noise ratio based on the interference received signal and the original wheel speed signal. The specific calculation formula is as follows: SNR improvement =SNR after -SNR before , x(t) nosiy =x(t) total -x(t)ideal, x(t)knosiy=x(t)ktotal-x(t)ideal, Among them, SNR improvement The SNR represents the performance evaluation result. before x(t) represents the signal-to-noise ratio of the training set of interference signals that were not processed by the interference suppression model. ideal Let x(t) represent the original wheel speed signal. nosiy Let x(t) represent the noise signal. total The SNR indicates the interference received signal. after Let x(t) represent the signal-to-noise ratio of the training set of the interference signals processed by the interference suppression model. knosiy Let x(t) represent the noise signal in the anti-interference wheel speed signal. ktotal Indicates anti-interference wheel speed signal; Determine whether the performance evaluation result is positive. If the performance evaluation result is positive, then the interference suppression model passes the performance evaluation test. If the performance evaluation result is negative, then the interference suppression model has failed the performance evaluation test. Determine whether the interference suppression model passes the performance evaluation test. If the interference suppression model passes the performance evaluation test, then obtain an anti-interference wheel speed signal based on the real-time wheel speed signal through the interference suppression model. The real-time wheel speed signal is the wheel speed signal collected in real time by the vehicle wheel speed sensor, and the anti-interference wheel speed signal is the real-time wheel speed signal with noise removed by the interference suppression model.

7. A wheel speed signal anti-interference processing system based on vehicle broadband road noise, comprising a signal acquisition module, a signal preprocessing module, a model training module, and a performance evaluation module, characterized in that, include: The signal acquisition module is used to acquire the original wheel speed signal dataset, and to acquire interference received signals through the microphone array inside the car. The interference received signals include wheel speed target signals and broadband noise interference sources. An interference signal training set and an interference signal test set are constructed based on the interference received signals. The microphone array is a uniform linear array composed of microphone sensors, and the interference received signal is collected through the uniform linear array; The specific method for constructing the uniform linear array is as follows: Equipped with a uniform linear array, the uniform linear array is composed of isotropic array elements. The interference received signal is obtained through the uniform linear array. The interference received signal is incident on the uniform linear array from different directions. The interference received signal includes a wheel speed target signal and a broadband noise interference source. Isotropic array element variables are obtained from the interference received signal through the isotropic array elements. The isotropic array element variables include isotropic array element amplitude variation and isotropic array element time delay. The signal preprocessing module is used to perform overlapping frame processing on the interference signal training set to obtain discrete signal frames, perform short-time Fourier transform on the discrete signal frames to obtain amplitude spectrum and phase spectrum, splice the amplitude spectrum and phase spectrum to obtain spatiotemporal frequency domain features, and add training labels to the interference signal training set according to the spectral features of the original wheel speed signal dataset. The model training module is used to train an interference suppression model using a complex signal convolutional neural network based on the spatiotemporal frequency domain features and the training labels. The performance evaluation module is used to perform performance evaluation tests on the interference suppression model using the interference signal test set to obtain performance evaluation results.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the anti-interference processing method for wheel speed signals based on vehicle broadband road noise as described in any one of claims 1-6.

9. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the wheel speed signal anti-interference processing method based on vehicle broadband road noise as described in any one of claims 1-6.

Images