Automobile road noise active control method and device based on Elman neural network

Abstract

The invention discloses an automobile road noise active control method and device based on an Elman neural network. The method comprises the following steps: acquiring vibration acceleration data and in-vehicle noise data when a vehicle runs; the vibration acceleration data and the in-vehicle noise data are used for training an Elman neural network; taking a predicted noise signal output by the Elman neural network as a reference signal, and inputting the reference signal into a filter to obtain a filtered signal; playing the filtering signal in the target noise reduction area, and synchronously collecting residual noise as an error signal; adjusting the weight parameter of the filter through the error signal, and processing the reference signal through the filter with the updated weight parameter to obtain an updated filtering signal; the steps of generating the error signal and updating the weight parameter are repeatedly executed until the average noise reduction amount in the vehicle is lower than a preset condition, and a final filtering signal is obtained; and playing the final filtering signal to counteract the noise in the vehicle. According to the invention, the acoustic feedback problem and the noise transmission time delay problem caused by directly using the noise signal as the reference signal can be avoided.

Description

Technical Field

[0001] The present application relates to the technical field of automobile noise processing, and in particular to an Elman neural network-based automobile road noise active control method, device, storage medium and electronic equipment. Background Art

[0002] In the automotive industry, in-vehicle noise has a crucial impact on vehicle comfort and quality. Vehicle NVH (Noise, Vibration, Harshness) performance has attracted increasing attention and has become a key factor in car purchase decisions. Pure electric vehicles use a combination of batteries and electric motors to replace internal combustion engines. Since there is no engine noise, road noise becomes the main source of noise during driving. There are generally two propagation paths for road noise generated by mutual excitation between the tire surface and the road: one is through air transmission, and the other is through the vibration transmission of vehicle sheet metal parts. The frequency range of this structural road noise transmitted by sheet metal vibration is approximately 20~450Hz, and the noise has strong randomness and time-varying characteristics, which can cause anxiety and irritability in the car occupants. It is also the main cause of motion sickness in passengers, which is not conducive to the riding experience and physical and mental health of the occupants.

[0003] At present, there are two main methods for noise control: passive control and active control. Passive noise reduction refers to the use of physical means such as sound insulation, vibration reduction, and sound absorption to control the noise inside the car. This method has been adopted by many automobile manufacturers due to its relatively simple design concept and uncomplicated manufacturing process. However, there are some disadvantages in the passive noise reduction method, such as expensive sound insulation materials and strict requirements on material performance. Active noise control uses the principle of sound wave superposition to eliminate noise. When the generated secondary noise signal has the same amplitude and opposite phase as the primary noise signal, the two signals will be superimposed on each other, and the residual noise signal intensity after superposition will be greatly reduced. The active noise control system can flexibly change the noise reduction strategy according to different driving conditions of the vehicle, and has a good effect on suppressing low-frequency noise, which has obvious advantages in adjusting the NVH performance of pure electric vehicles.

[0004] Active noise control systems usually use the interior noise of the vehicle as a reference signal. This method has some limitations: on the one hand, it introduces the problem of acoustic feedback, which affects the noise reduction effect of the system; on the other hand, the transmission delay of the noise signal affects the response speed of the system. Summary of the invention

[0005] The embodiments of the present application provide a method, device, storage medium and electronic device for active control of automobile road noise based on Elman neural network, which can avoid the acoustic feedback problem and noise transmission delay problem of directly using noise signal as reference signal.

[0006] The embodiment of the present application provides a method for actively controlling automobile road noise based on an Elman neural network, comprising: Obtain vibration acceleration data and in-car noise data when the car is driving; Using the vibration acceleration data and the in-vehicle noise data to train an Elman neural network to obtain a trained Elman neural network; wherein the input of the Elman neural network is the vibration acceleration data, and the output of the Elman neural network is a predicted noise signal; Using the predicted noise signal as a reference signal, inputting the reference signal into a filter to obtain a filtered signal; Playing the filtered signal in the target noise reduction area, and synchronously collecting the remaining noise, and using the remaining noise as an error signal; Adjusting the weight parameter of the filter by the error signal, and processing the reference signal by the filter after updating the weight parameter to obtain an updated filtered signal; Repeat the steps of generating an error signal and updating weight parameters until the average noise reduction amount in the vehicle is lower than a preset condition, thereby obtaining a final filtered signal; The final filtered signal is played in the target noise reduction area to offset the noise inside the vehicle.

[0007] Furthermore, the above-mentioned active control method for automobile road noise based on Elman neural network, wherein the predicted noise signal is used as a reference signal, and the reference signal is input into a filter to obtain a filtered signal, comprises: Compute the filtered signal at the filter output:

[0008] in, W(n) is the weight coefficient matrix of the filter, X(n) is the reference signal matrix.

[0009] Furthermore, in the above-mentioned active control method for automobile road noise based on Elman neural network, the step of adjusting the weight parameter of the filter by using the error signal comprises:

[0010] in, W(n+ 1 ) It is n+ The weight coefficient at 1 iteration, W(n) For the n The weight coefficient at the iteration, is the convergence step size, which controls the update speed of the weight coefficient. and They are the error signal and reference signal processed by the improved FxLMS algorithm.

[0011] Furthermore, in the above-mentioned active control method for automobile road noise based on Elman neural network, the weight parameter of the filter is adjusted by the error signal, and the reference signal is processed by the filter after the weight parameter is updated to obtain an updated filtered signal, including: calculating an adaptive constraint factor based on the error signal and the reference signal; Adjusting the weight parameters of the filter by the adaptive constraint factor; The reference signal is processed by the filter with updated weight parameters to obtain an updated filtered signal.

[0012] Furthermore, in the above-mentioned active control method for automobile road noise based on Elman neural network, the step of calculating the adaptive constraint factor based on the error signal and the reference signal comprises:

[0013] in, and Respectively represent the error signal e(n) and reference signal X(n) The Euclidean norm of , , , and yes and The proportion adjustment coefficient between is between [0, 1]; The step of adjusting the weight parameter of the filter by using the adaptive constraint factor comprises:

[0014] in, W(n+ 1 ) It is n+ The weight coefficient at 1 iteration, is the Euclidean norm, b is the set weight constraint factor.

[0015] Furthermore, the above-mentioned vehicle road noise active control method based on Elman neural network, wherein the method also includes: When the amplitudes of the reference signal and the error signal exceed a preset threshold, adjusting the reference signal and the error signal:

[0016]

[0017] in, and The reference signals are The lower and upper limits of and The error signal lower and upper limits.

[0018] Furthermore, the above-mentioned Elman neural network-based active control method for automobile road noise includes, after the step of obtaining vibration acceleration data and interior noise data of the automobile when the automobile is driving: In each fixed-length sliding window, a cross-correlation function of the vibration acceleration data and the in-vehicle noise data is calculated to obtain a similarity relationship, and a cross-power spectrum of the vibration acceleration data and the in-vehicle noise data is calculated to obtain a coherence relationship; The vibration acceleration data and the in-vehicle noise data are screened according to the similarity relationship and the coherence relationship.

[0019] The embodiment of the present application also provides an active control device for automobile road noise based on Elman neural network, comprising: A data acquisition module is used to obtain vibration acceleration data and in-vehicle noise data when the car is driving; An identification module, used for training an Elman neural network using the vibration acceleration data and the in-vehicle noise data to obtain a trained Elman neural network; wherein the input of the Elman neural network is the vibration acceleration data, and the output of the Elman neural network is a predicted noise signal; A noise reduction module is used to use the predicted noise signal as a reference signal, input the reference signal into a filter, and obtain a filtered signal; play the filtered signal in a target noise reduction area, and synchronously collect residual noise, and use the residual noise as an error signal; adjust the weight parameter of the filter according to the error signal, and process the reference signal through the filter after updating the weight parameter to obtain an updated filtered signal; repeatedly perform the steps of generating an error signal and updating the weight parameter until the average noise reduction amount in the vehicle is lower than a preset condition, and obtain a final filtered signal; and play the final filtered signal in the target noise reduction area to offset the noise in the vehicle.

[0020] An embodiment of the present application also provides a computer-readable storage medium, in which a plurality of instructions are stored. The instructions are suitable for being loaded by a processor to execute any of the above-mentioned methods for active control of automobile road noise based on Elman neural network.

[0021] An embodiment of the present application also provides an electronic device, including a processor and a memory, wherein the processor is electrically connected to the memory, the memory is used to store instructions and data, and the processor is used for the steps in any of the above-mentioned methods for active control of automobile road noise based on Elman neural network.

[0022] The present application provides an active control method, device, storage medium and electronic device for automobile road noise based on the Elman neural network. The present application predicts the noise signal through the Elman neural network, uses the predicted noise signal as a reference signal to obtain a filtered signal through filter processing, and uses the filtered signal to offset the noise inside the car. The present application establishes a transfer function relationship from the vibration signal to the noise signal through the Elman neural network, avoiding the acoustic feedback problem and the noise transmission delay problem of directly using the noise signal as the reference signal. And the weight parameters of the filter are adjusted through the error signal, and the reference signal is processed by the filter after the weight parameters are updated to obtain an updated filtered signal. The present application continuously adjusts the weight parameters of the filter, so that the final output filtered signal can more accurately offset the noise inside the car, thereby improving the noise reduction effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The technical solution and other beneficial effects of the present application will be made apparent by describing in detail the specific implementation methods of the present application in conjunction with the accompanying drawings.

[0024] Figure 1 A flowchart of a method for active control of automobile road noise based on Elman neural network provided in an embodiment of the present application.

[0025] Figure 2 A schematic diagram of the structure of a data acquisition system provided in an embodiment of the present application.

[0026] Figure 3 A flowchart for generating a filtering signal provided in an embodiment of the present application.

[0027] Figure 4 Another flow chart for generating a filtering signal provided in an embodiment of the present application.

[0028] Figure 5 A schematic diagram of the structure of an active vehicle road noise control device based on an Elman neural network provided in an embodiment of the present application.

[0029] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0030] Figure 7 Another schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0032] The embodiment of the present application provides a method, device, storage medium and electronic device for active control of automobile road noise based on Elman neural network. The embodiment of the present application provides an active control device for automobile road noise based on Elman neural network, which can be integrated in an electronic device, which can be a terminal, a server and other devices, wherein the terminal can include a tablet computer, a laptop computer, a personal computer (PC, Personal Computer), a micro processing box, or other devices.

[0033] See also Figure 1 , Figure 1 The flowchart of the active control method of automobile road noise based on Elman neural network provided in the embodiment of the present application is applied in electronic equipment. The active control method of automobile road noise based on Elman neural network includes the following steps: S1, obtaining vibration acceleration data and in-car noise data when the car is driving.

[0034] Specifically, the vehicle data acquisition system is used to obtain the vibration acceleration data and the in-vehicle noise data when the vehicle is driving. Figure 2 A schematic diagram of the structure of the data acquisition system provided in the embodiment of the present application is shown in FIG. Figure 2 As shown in the figure, the vibration sensor is responsible for measuring the vibration acceleration of the vehicle and converting it into a corresponding vibration electrical signal. The sensor drive interface supplies the necessary power supply and input interface for the vibration sensor by providing working current and high-impedance input interface. The signal conditioning circuit processes the vibration signal output by the vibration sensor and generates a voltage signal suitable for subsequent steps. The A / D converter converts the input electrical signal into a digital signal. The converted digital signal is transmitted to the ARM processor via the SPI bus. The ARM processor receives the vibration signal collected by the A / D converter and encapsulates it into a UDP datagram. The encapsulated UDP data is transmitted to the data logger via the Ethernet controller and the RJ45 100M network cable.

[0035] In addition, it is necessary to collect the vibration acceleration data and in-car noise data of the car under different working conditions. The specific car driving conditions are divided into three categories: low-speed 30km / h constant speed driving condition, medium-speed 60km / h constant speed driving condition, high-speed 80km / h constant speed driving condition. Among them, the audio sensor is installed on the driver's seat headrest to measure the noise signal in the front row of the car. The vibration acceleration sensor is installed at the connection between the body sheet metal and the suspension system to collect vibration signals at the four suspensions.

[0036] As an example, five vibration acceleration sensors (numbered 1, 2, 3, 4, and 5 respectively) may be installed on the left front suspension, the right front suspension, the left rear suspension, the right rear suspension, and the driver's seat headrest, respectively.

[0037] In one embodiment, after step S1, a multiple coherence analysis method is used to perform validity analysis on the in-vehicle noise data and the vibration acceleration data, which specifically includes the following steps: SA1: In each fixed-length sliding window, the cross-correlation function of the vibration acceleration data and the in-vehicle noise data is calculated to obtain the similarity relationship, and the cross-power spectrum of the vibration acceleration data and the in-vehicle noise data is calculated to obtain the coherence relationship.

[0038] Specifically, during the period when the noise level is higher than a threshold (e.g., 50 dB), a fixed-length sliding window (e.g., 1 second) is applied to the collected data. Within each window, the cross-correlation function of the vibration acceleration data and the interior noise data is calculated to determine the similarity between the two, and the cross-power spectrum of the vibration acceleration data and the interior noise data is calculated to determine the coherence of a specific frequency.

[0039] SA2, screening the vibration acceleration data and the in-vehicle noise data according to the similarity relationship and the coherence relationship.

[0040] If the similarity between the vibration acceleration data and the in-vehicle noise data is obvious, and the coherence within the target noise reduction frequency band reaches a significant level, it indicates that under the corresponding working condition, the acceleration signal can effectively construct a reference signal and is considered to meet the system requirements. The significance threshold can be set based on the experience of technicians in this field.

[0041] S2, using the vibration acceleration data and the in-vehicle noise data to train the Elman neural network to obtain a trained Elman neural network; wherein the input of the Elman neural network is the vibration acceleration data, and the output of the Elman neural network is the predicted noise signal.

[0042] In one embodiment, step S2 comprises the following steps: S21, intercept the vibration acceleration data and in-car noise data collected when the car is in a stable driving state to create a data set, and divide it into a training set and test set .

[0043] Specifically, step S21 includes: S211, respectively intercept the relatively stable parts of the vibration acceleration data and the in-vehicle noise data, filter out the data that is greatly interfered by the environmental noise, and finally obtain the processed sample data, reduce unnecessary interference in the test, and ensure the reliability and accuracy of the data.

[0044] S212, randomly extract 4000 data segments from the 4 sets of vibration acceleration data collected, each data segment has a length of 8000, that is, randomly extract 4000 data segments with a duration of about 7.5 seconds.

[0045] S213, 4000 data segments are also extracted from a set of in-vehicle noise data, and the length of each data segment is 8000.

[0046] S214, using 4 sets of vibration acceleration data as sample input and 1 set of in-vehicle noise data as sample output, thus obtaining 4000 sample pairs, each sample pair having 8000 data points, and the data duration is about 7.5 seconds.

[0047] S215, split the obtained 4000 sample pairs into training sets and test sets, and randomly select 3500 sample pairs as training sets and 500 sample pairs as test sets.

[0048] S216, repeating steps S211-S215 to obtain a vibration acceleration data set and a vehicle interior noise data set when the vehicle is driving under different working conditions, and separating the vibration acceleration data set and the vehicle interior noise data set according to a certain ratio to form a training set and test set Among them, the training set and test set It also includes two types of data: vibration acceleration data and in-vehicle noise data.

[0049] S22, the training set and test set Import into the MATLAB workspace.

[0050] S23 In order to make the sample data magnitude in the same range and to make the network process data more quickly, the training set in the MATLAB workspace and test set The data in is normalized, and the Mapminmax() function can be used for processing.

[0051] S24, initialize the weights of each layer of the network.

[0052] Specifically, the weight parameters of the Elman neural network are initialized, the number of input layer nodes is 4, the output layer outputs the front row noise signal, the number of output layer nodes is 1, the transfer function of the hidden layer uses the Sigmoid function, the transfer functions of the output layer and the receiving layer use the hyperbolic tangent (tanh) function, and the weights of the network are determined by initializing random values.

[0053] S25, use the newelm() function of the MATLAB toolbox to establish a neural network, specify the number of nodes in the input layer, hidden layer, receiving layer and output layer, and select the activation function. The specific function is:

[0054] Among them, inputs is the number of nodes in the input layer, outputs is the number of nodes in the output layer, hiddenSizes is the number of nodes in the hidden layer, tanh is the hyperbolic tangent transfer function, sigmoid is the sigmoid transfer function, and trainlm is the training algorithm used.

[0055] S26, training is performed through the train() function, which takes the input data of the training set and the output data of the training set as sample input. The specific function is as follows

[0056] Among them, inputs is the input data of the training set, that is, the vibration acceleration data, and targets is the target data of the training set, that is, the interior noise data.

[0057] S27, network performance test is performed through the sim() function.

[0058] The specific approach is to use the vibration acceleration data in the test set as the input of the Elman neural network, perform a denormalization operation on the output of the Elman neural network, and finally obtain the output predicted noise signal.

[0059] S28, compare the predicted noise signal obtained in the previous step with the in-vehicle noise data in the test set data to check whether it meets the preset network model accuracy requirements. If the accuracy requirements are not met, you can try to adjust the network structure by increasing or decreasing the number of hidden layer nodes of the Elman neural network, increasing the number of layers of the Elman neural network, or adjusting the activation function, and repeat the above steps until the error requirements are met.

[0060] S29, through the save() function, save the Elman neural network model whose training accuracy meets the preset requirements.

[0061] S3, taking the predicted noise signal as a reference signal, inputting the reference signal into a filter to obtain a filtered signal.

[0062] Figure 3 A flowchart of generating a filtering signal provided in an embodiment of the present application, Figure 4 Another flow chart of generating a filtering signal provided in an embodiment of the present application. The present application actively controls road noise through an improved FxLMS algorithm. Steps S3-S6 can refer to Figure 3 and Figure 4 Step S3 specifically includes: Compute the filtered signal at the filter output:

[0063] in, W(n) is the weight coefficient matrix of the filter, X(n) is the reference signal matrix.

[0064] S4, playing the filtered signal in the target noise reduction area, and synchronously collecting the remaining noise, and using the remaining noise as the error signal.

[0065] Specifically, the filtered signal After passing through the secondary sound channel S(z), it becomes a filtered signal The purpose of the embodiment of the present application is to filter the signal It is superimposed with the error signal to achieve noise reduction. Among them, the secondary sound channel S(z) represents the filtered signal From the perspective of signal transmission, the changes that occur during the actual playback process include DAC playback, air medium transmission, and ADC acquisition. Among them, the ADC acquisition link has a small impact and is often ignored.

[0066] In an embodiment of the present application, a speaker is disposed in a target noise reduction area (i.e., a seat headrest), a filtered signal is played through the speaker, and at the same time, the remaining noise after cancellation is collected through the speaker.

[0067] S5, adjusting the weight parameters of the filter according to the error signal, and processing the reference signal through the filter with the updated weight parameters to obtain an updated filtered signal.

[0068] In one embodiment, step S5 comprises the following steps: S51, adjust the weight coefficients of the adaptive FIR filter based on the improved FxLMS (Filtered-X Least Mean Square) algorithm W(n) :

[0069] in, W(n+ 1 )It is n+ The weight coefficient at 1 iteration, W(n) For the n The weight coefficient at the iteration, is the convergence step size, which controls the update speed of the weight coefficient. and They are the error signal and reference signal processed based on the improved FxLMS algorithm respectively.

[0070] S52, processing the reference signal through the filter with updated weight parameters to obtain an updated filtered signal.

[0071] In another embodiment, the reference signal in the noise reduction process is considered X(n) and error signal e(n) Refer to the Sigmoid function form to see the changes in , and add it to the update formula of the weight constraint factor b.

[0072] Step S5 may further include the following steps: S53, calculating an adaptive constraint factor based on the error signal and the reference signal.

[0073]

[0074] in, and Respectively represent the error signal e(n) and reference signal X(n) The Euclidean norm of , , , and yes and The weight adjustment coefficient between is between [0, 1].

[0075] S54, adjusting the weight parameters of the filter through the adaptive constraint factor.

[0076]

[0077] in, W(n+ 1 ) It is n+ The weight coefficient at 1 iteration, is the Euclidean norm, b is the set weight constraint factor.

[0078] S55, processing the reference signal through the filter with updated weight parameters to obtain an updated filtered signal.

[0079] In one embodiment, and It can be calculated by the improved FxLMS algorithm: When the amplitude of the reference signal and the error signal exceeds the preset threshold, the reference signal and the error signal are adjusted:

[0080]

[0081] in, and The reference signals are The lower and upper limits of and The error signal lower and upper limits.

[0082] Since the noise signal may diverge when a large outlier appears, the embodiment of the present application proposes the above method for adjusting the reference signal and the error signal to avoid the influence of the outlier in the noise signal on the stability of noise control.

[0083] S6, repeatedly executing the steps of generating an error signal and updating weight parameters until the average noise reduction amount in the vehicle is lower than a preset condition, thereby obtaining a final filtering signal.

[0084] Specifically, it is determined whether the noise reduction effect meets the preset requirement (for example, the average noise reduction ANR is lower than -20dB). If the requirement is not met, steps S3-S5 are repeated until the preset requirement is met.

[0085] S7 plays the final filtered signal in the target noise reduction area to offset the noise inside the car.

[0086] Furthermore, with the standard symmetric stable distribution ( ) The simulated impulse noise is used as the input signal of the Elman neural network to compare and verify the noise reduction performance of the road noise active control method proposed in this application and the mainstream noise reduction algorithm. The algorithm noise reduction performance evaluation index is the average noise reduction amount ANR, which is specifically defined as follows:

[0087]

[0088]

[0089] in, For the error signal e(n) The absolute value estimate of For the expected signal d(n) The absolute value of the estimate is set as and , forgetting factor Set to 0.999 Specifically, strong impulse noise ( =1.3) is the input signal, the evaluation index is the number of iterations when the algorithm converges and the average noise reduction amount ANR, the simulation results are shown in Table 1; Table 1 Comparison of noise reduction performance of various algorithms under strong impulse noise environment

[0090] It can be seen from Table 1 that the road noise active control method proposed in this application is slightly better than the latest active noise control algorithm in the industry in terms of algorithm convergence speed and noise reduction effect.

[0091] Furthermore, the noise signals collected under different working conditions are used as reference signals, and the noise reduction performance is simulated and analyzed in combination with the Elman neural network. The average noise reduction amount is used as the evaluation index, and the final results are shown in Table 2: Table 2 Comparison of noise reduction effects of active road noise control methods based on Elman neural network

[0092] Table 2 uses the vibration acceleration data and in-vehicle noise data collected under different working conditions as the reference signals of the system to simulate and verify the noise reduction effect of the road noise active control method based on the Elman neural network. It can be seen from the table that regardless of the speed, the designed road noise active control method has a good noise reduction effect. Compared with the noise reduction method using the noise signal as the reference signal, the road noise active control method based on Elman neural network identification has a better noise reduction effect and a larger noise reduction amount.

[0093] First, the present application uses the Elman neural network to establish the transfer function relationship from the vibration signal to the noise signal, avoiding the acoustic feedback problem and the noise transmission delay problem of directly using the noise signal as the reference signal. Secondly, the present application actively controls the road noise through the improved FxLMS algorithm. This method has two advantages: (1) The adaptive constraint factor b is directly determined by the formula, avoiding the process of manual continuous testing and debugging of the b value, solving the problem of inaccurate b value caused by errors in the experimental process, and avoiding tedious experimental work; (2) The value of the weight constraint factor b determined by this method is closely related to the reference signal and the error signal. The value of b is adaptively adjusted during the noise reduction process, enhancing the real-time adjustment capability.

[0094] According to the method described in the above embodiment, this embodiment will be further described from the perspective of an active control device for automobile road noise based on an Elman neural network. The active control device for automobile road noise based on an Elman neural network can be implemented as an independent entity or integrated in an electronic device. The electronic device can be a terminal, a server, and the like. The terminal can include a tablet computer, a laptop computer, a personal computer (PC), a micro processing box, or other devices.

[0095] See also Figure 5 , Figure 5 The active control device for automobile road noise based on Elman neural network provided by the embodiment of the present application is specifically described, which is applied to electronic equipment. The active control device for automobile road noise based on Elman neural network may include: A data acquisition module is used to obtain vibration acceleration data and in-vehicle noise data when the car is driving; The identification module is used to train the Elman neural network using the vibration acceleration data and the in-vehicle noise data to obtain a trained Elman neural network; wherein the input of the Elman neural network is the vibration acceleration data, and the output of the Elman neural network is a predicted noise signal; The noise reduction module is used to use the predicted noise signal as a reference signal, input the reference signal into the filter, and obtain a filtered signal; play the filtered signal in the target noise reduction area, and synchronously collect the remaining noise, and use the remaining noise as an error signal; adjust the weight parameters of the filter through the error signal, and process the reference signal through the filter after updating the weight parameters to obtain an updated filtered signal; repeat the steps of generating the error signal and updating the weight parameters until the average noise reduction amount in the car is lower than the preset condition, and obtain the final filtered signal; and play the final filtered signal in the target noise reduction area to offset the noise in the car.

[0096] During specific implementation, the above modules and / or units can be implemented as independent entities, or can be arbitrarily combined to be implemented as the same or several entities. The specific implementation of the above modules and / or units can refer to the previous method embodiments. The specific beneficial effects that can be achieved can also refer to the beneficial effects in the previous method embodiments, which will not be repeated here.

[0097] In addition, the embodiment of the present application also provides an electronic device, which may be a computer, a tablet computer, or other device. Figure 6 As shown, the electronic device 400 includes a processor 401 and a memory 402. The processor 401 is electrically connected to the memory 402.

[0098] The processor 401 is the control center of the electronic device 400. It uses various interfaces and lines to connect various parts of the entire electronic device. It executes various functions of the electronic device and processes data by running or loading applications stored in the memory 402 and calling data stored in the memory 402, thereby monitoring the electronic device as a whole.

[0099] In this embodiment, the processor 401 in the electronic device 400 loads instructions corresponding to the processes of one or more application programs into the memory 402 according to the following steps, and the processor 401 runs the application programs stored in the memory 402 to implement various functions: Obtain vibration acceleration data and in-car noise data when the car is driving; Using the vibration acceleration data and the in-vehicle noise data to train an Elman neural network to obtain a trained Elman neural network; wherein the input of the Elman neural network is the vibration acceleration data, and the output of the Elman neural network is a predicted noise signal; Using the predicted noise signal as a reference signal, inputting the reference signal into a filter to obtain a filtered signal; Playing the filtered signal in the target noise reduction area, and synchronously collecting the remaining noise, and using the remaining noise as an error signal; Adjusting the weight parameter of the filter by the error signal, and processing the reference signal by the filter after updating the weight parameter to obtain an updated filtered signal; Repeat the steps of generating an error signal and updating weight parameters until the average noise reduction amount in the vehicle is lower than a preset condition, thereby obtaining a final filtered signal; The final filtered signal is played in the target noise reduction area to offset the noise inside the vehicle.

[0100] The electronic device can implement the steps in any embodiment of the active control method for automobile road noise based on Elman neural network provided in the embodiments of the present application. Therefore, it can achieve the beneficial effects that can be achieved by any active control method for automobile road noise based on Elman neural network provided in the embodiments of the present invention. For details, please refer to the previous embodiments and will not be repeated here.

[0101] Figure 7 The specific structural block diagram of the electronic device provided by the embodiment of the present invention is shown, and the electronic device can be used to implement the active control method of automobile road noise based on Elman neural network provided in the above embodiment. The electronic device 500 can be a terminal, a server and other devices, wherein the terminal can include a tablet computer, a notebook computer, a personal computer (PC, Personal Computer), a micro processing box, or other devices.

[0102] The RF circuit 510 is used to receive and send electromagnetic waves, realize the mutual conversion between electromagnetic waves and electrical signals, and thus communicate with a communication network or other devices. The RF circuit 510 may include various existing circuit elements for performing these functions, such as antennas, radio frequency transceivers, digital signal processors, encryption / decryption chips, user identity module (SIM) cards, memories, etc. The RF circuit 510 can communicate with various networks such as the Internet, corporate intranets, wireless networks, or communicate with other devices through wireless networks. The above-mentioned wireless networks may include cellular telephone networks, wireless local area networks, or metropolitan area networks. The above-mentioned wireless networks may use various communication standards, protocols and technologies, including but not limited to Global System for Mobile Communication (GSM), Enhanced Data GSM Environment (EDGE), Wideband Code Division Multiple Access (WCDMA), Code Division Access (CDMA), Time Division Multiple Access (TDMA), Wireless Fidelity (Wi-Fi) (such as Institute of Electrical and Electronics Engineers standards IEEE 802.11a, IEEE 802.11b, IEEE802.11g and / or IEEE802.11n), Voice over Internet Protocol (VoIP), Worldwide Interoperability for Microwave Access (Wi-Max), other protocols for email, instant messaging and short messages, and any other suitable communication protocols, even those that have not yet been developed.

[0103] The memory 520 can be used to store software programs and modules, such as the corresponding program instructions / modules in the above embodiments. The processor 580 executes various functional applications and data processing by running the software programs and modules stored in the memory 520, that is, realizing the functions of taking pictures with the front camera, processing the captured images, and switching the display color of the display content on the display screen. The memory 520 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 520 may further include a memory remotely arranged relative to the processor 580, and these remote memories may be connected to the electronic device 500 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0104] The input unit 530 may be used to receive input digital or character information, and generate a keyboard and a mouse related to user settings and function control. The display unit 540 may be used to display information input by the user or information provided to the user and various graphical user interfaces, which may be composed of graphics, text, icons, videos, and any combination thereof. The display unit 540 may include a display panel 541, which may be configured in the form of an LCD (Liquid Crystal Display), an OLED (Organic Light-Emitting Diode), or the like.

[0105] The audio circuit 560, the speaker 561, and the microphone 562 can provide an audio interface between the user and the electronic device 500. The audio circuit 560 can transmit the electrical signal converted from the received audio data to the speaker 561, which is converted into a sound signal for output; on the other hand, the microphone 562 converts the collected sound signal into an electrical signal, which is received by the audio circuit 560 and converted into audio data, and then the audio data is output to the processor 580 for processing, and then sent to another terminal through the RF circuit 510, or the audio data is output to the memory 520 for further processing. The audio circuit 560 may also include an earplug jack to provide communication between an external headset and the electronic device 500.

[0106] The electronic device 500 can help the user receive requests, send information, etc. through the transmission module 570 (such as a Wi-Fi module), which provides the user with wireless broadband Internet access. Although the transmission module 570 is shown in the figure, it can be understood that it is not a necessary component of the electronic device 500 and can be omitted as needed without changing the essence of the invention.

[0107] The processor 580 is the control center of the electronic device 500. It uses various interfaces and lines to connect various parts of the entire mobile phone. By running or executing software programs and / or modules stored in the memory 520, and calling data stored in the memory 520, it executes various functions of the electronic device 500 and processes data, thereby monitoring the electronic device as a whole. Optionally, the processor 580 may include one or more processing cores; in some embodiments, the processor 580 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, etc., and the modem processor mainly processes wireless communications. It can be understood that the above-mentioned modem processor may not be integrated into the processor 580.

[0108] The electronic device 500 also includes a power supply 590 (such as a battery) for supplying power to various components. In some embodiments, the power supply can be logically connected to the processor 580 through a power management system, so that the power management system can manage charging, discharging, and power consumption management. The power supply 590 can also include any components such as one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, and power status indicators.

[0109] Although not shown, the electronic device 500 also includes a camera (such as a front camera, a rear camera), a Bluetooth module, etc., which will not be described in detail here. Specifically in this embodiment, the display unit of the electronic device is a touch screen display, and the mobile terminal also includes a memory, and one or more programs, wherein one or more programs are stored in the memory, and are configured to be executed by one or more processors. One or more programs include instructions for performing the following operations: Obtain vibration acceleration data and in-car noise data when the car is driving; The vibration acceleration data and the in-vehicle noise data are used to train the Elman neural network to obtain a trained Elman neural network; wherein the input of the Elman neural network is the vibration acceleration data, and the output of the Elman neural network is the predicted noise signal; The predicted noise signal is used as a reference signal, and the reference signal is input into a filter to obtain a filtered signal; Play the filtered signal in the target noise reduction area and simultaneously collect the remaining noise, using the remaining noise as the error signal; The weight parameters of the filter are adjusted by the error signal, and the reference signal is processed by the filter after the weight parameters are updated to obtain an updated filtered signal; Repeat the steps of generating an error signal and updating weight parameters until the average noise reduction amount in the vehicle is lower than a preset condition, thereby obtaining a final filtered signal; The final filtered signal is played in the target noise reduction area to cancel out the noise inside the car.

[0110] In specific implementation, the above modules can be implemented as independent entities, or can be arbitrarily combined and implemented as the same or several entities. The specific implementation of the above modules can be found in the previous method embodiments, which will not be repeated here.

[0111] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by controlling related hardware through instructions, and the instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. To this end, an embodiment of the present invention provides a storage medium, which stores a plurality of instructions, and the instructions can be loaded by a processor to execute the steps of any embodiment of the active control method for automobile road noise based on Elman neural network provided in the embodiment of the present invention.

[0112] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0113] Since the instructions stored in the storage medium can execute the steps in any embodiment of the active control method for automobile road noise based on Elman neural network provided in the embodiments of the present invention, the beneficial effects that can be achieved by any active control method for automobile road noise based on Elman neural network provided in the embodiments of the present invention can be achieved. For details, please refer to the previous embodiments and will not be repeated here.

[0114] The above is a detailed introduction to an active control method, device, storage medium and electronic device for automobile road noise based on Elman neural network provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for technicians in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for active control of automobile road noise based on Elman neural network, characterized in that: The method comprises: Obtain vibration acceleration data and in-car noise data when the car is driving; Using the vibration acceleration data and the in-vehicle noise data to train an Elman neural network to obtain a trained Elman neural network; wherein the input of the Elman neural network is the vibration acceleration data, and the output of the Elman neural network is a predicted noise signal; Using the predicted noise signal as a reference signal, inputting the reference signal into a filter to obtain a filtered signal; Playing the filtered signal in the target noise reduction area, and synchronously collecting the remaining noise, and using the remaining noise as an error signal; Adjusting the weight parameter of the filter by the error signal, and processing the reference signal by the filter after updating the weight parameter to obtain an updated filtered signal; Repeat the steps of generating an error signal and updating weight parameters until the average noise reduction amount in the vehicle is lower than a preset condition, thereby obtaining a final filtered signal; The final filtered signal is played in the target noise reduction area to offset the noise inside the vehicle.

2. The method for active control of automobile road noise based on Elman neural network according to claim 1, characterized in that: The step of using the predicted noise signal as a reference signal and inputting the reference signal into a filter to obtain a filtered signal comprises: Compute the filtered signal at the filter output: in, W(n) is the weight coefficient matrix of the filter, X(n) is the reference signal matrix.

3. The method for active control of automobile road noise based on Elman neural network according to claim 1, characterized in that: The step of adjusting the weight parameter of the filter by using the error signal comprises: in, W(n+ 1 ) It is n+ The weight coefficient at 1 iteration, W(n) For the n The weight coefficient at the iteration, is the convergence step size, which controls the update speed of the weight coefficient. and They are the error signal and reference signal processed by the improved FxLMS algorithm.

4. The method for active control of automobile road noise based on Elman neural network according to claim 1, characterized in that: The step of adjusting the weight parameter of the filter by the error signal and processing the reference signal by the filter after updating the weight parameter to obtain an updated filtered signal comprises: calculating an adaptive constraint factor based on the error signal and the reference signal; Adjusting the weight parameters of the filter by the adaptive constraint factor; The reference signal is processed by the filter with updated weight parameters to obtain an updated filtered signal.

5. The active control method for automobile road noise based on Elman neural network according to claim 4 is characterized in that: The calculating the adaptive constraint factor based on the error signal and the reference signal comprises: in, and Respectively represent the error signal e(n) and reference signal X(n) The Euclidean norm of , , , and yes and The proportion adjustment coefficient between is between [0, 1]; The step of adjusting the weight parameter of the filter by using the adaptive constraint factor comprises: in, W(n+ 1 ) It is n+ The weight coefficient at 1 iteration, is the Euclidean norm, b is the set weight constraint factor.

6. The active control method for automobile road noise based on Elman neural network according to claim 1, characterized in that: The method further comprises: When the amplitudes of the reference signal and the error signal exceed a preset threshold, adjusting the reference signal and the error signal: in, and The reference signals are The lower and upper limits of and The error signal lower and upper limits.

7. The method for active control of automobile road noise based on Elman neural network according to claim 1, characterized in that: After the step of obtaining the vibration acceleration data and the interior noise data of the vehicle when the vehicle is driving, it includes: In each fixed-length sliding window, a cross-correlation function of the vibration acceleration data and the in-vehicle noise data is calculated to obtain a similarity relationship, and a cross-power spectrum of the vibration acceleration data and the in-vehicle noise data is calculated to obtain a coherence relationship; The vibration acceleration data and the in-vehicle noise data are screened according to the similarity relationship and the coherence relationship.

8. An active control device for automobile road noise based on Elman neural network, characterized in that: include: A data acquisition module is used to obtain vibration acceleration data and in-vehicle noise data when the car is driving; an identification module, used for training an Elman neural network using the vibration acceleration data and the in-vehicle noise data to obtain a trained Elman neural network; wherein the input of the Elman neural network is the vibration acceleration data, and the output of the Elman neural network is a predicted noise signal; A noise reduction module is used to use the predicted noise signal as a reference signal, input the reference signal into a filter, and obtain a filtered signal; play the filtered signal in a target noise reduction area, and synchronously collect residual noise, and use the residual noise as an error signal; adjust the weight parameter of the filter according to the error signal, and process the reference signal through the filter after updating the weight parameter to obtain an updated filtered signal; repeatedly perform the steps of generating an error signal and updating the weight parameter until the average noise reduction amount in the vehicle is lower than a preset condition, and obtain a final filtered signal; and play the final filtered signal in the target noise reduction area to offset the noise in the vehicle.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the vehicle road noise active control method based on Elman neural network according to any one of claims 1 to 7.

10. An electronic device, characterized in that: It comprises a processor and a memory, the processor is electrically connected to the memory, the memory is used to store instructions and data, and the processor is used to execute the steps in the active control method of automobile road noise based on Elman neural network as described in any one of claims 1 to 7.

Images