Noise modeling model training method and device, electronic equipment and storage medium

Abstract

The invention discloses a noise modeling model training method and device, electronic equipment and a storage medium. The method comprises the following steps: structuring an original data sample to obtain a sample data feature corresponding to the original data sample; decoding the sample data features based on each first parameter vector and the initial model to obtain a predicted noise feature corresponding to each first parameter vector; determining a competition degree corresponding to each first parameter vector; determining a plurality of second parameter vectors based on the competition degree corresponding to each first parameter vector; performing transformation processing on the plurality of second parameter vectors to obtain a plurality of third parameter vectors; decoding the sample data features based on each third parameter vector and the initial model to obtain a predicted noise feature corresponding to each third parameter vector; determining a first target parameter vector from the plurality of third parameter vectors; and training the initial model based on the first target parameter vector to obtain a noise modeling model.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a noise modeling model training method, device, electronic device and storage medium. Background Art

[0002] Low-voltage power line carrier communication technology refers to a technology that uses low-voltage distribution networks as transmission media to achieve data and information exchange. This technology isolates power frequency current and high-frequency signals to ensure that the signal and power frequency current occupy different frequency bands during the transmission process to ensure that both parties can transmit in an orderly and efficient manner. However, the existing transmission lines were not originally designed for data transmission, but only for energy transmission. At the same time, they are connected to a variety of complex electrical equipment, making the noise generated when they are used as communication channels very complex, including various narrowband, broadband, pulse and other noises. Noise will increase the signal error rate and reduce the signal-to-noise ratio, thereby affecting the communication quality. At present, in the prior art, the power line channel noise is modeled by the Back Propagation Neural Network (BPNN), but due to the performance limitations of BPNN itself, it is easy to fall into the local optimal solution and lack of global search capability, thereby reducing the accuracy of noise modeling, resulting in the inability to filter the noise in the power line channel according to the noise characteristics obtained by noise modeling, affecting the communication quality. Summary of the invention

[0003] The embodiments of the present application provide a noise modeling model training method, device, electronic device and storage medium, which improve the accuracy of noise modeling and filter the noise in the power line channel based on the noise characteristics obtained by noise modeling, thereby improving the communication quality.

[0004] In a first aspect, an embodiment of the present application provides a noise modeling model training method, the method comprising:

[0005] Collecting data on the power line channel to construct training samples, wherein the training samples include original data samples and noise characteristic samples;

[0006] Structuring the original data sample to obtain sample data features corresponding to the original data sample;

[0007] Decoding the sample data feature based on each first parameter vector and the initial model to obtain a predicted noise characteristic corresponding to each first parameter vector, wherein each first parameter vector is randomly generated;

[0008] Determine the degree of competition corresponding to each first parameter vector based on the predicted noise characteristic corresponding to each first parameter vector and the noise characteristic sample;

[0009] Based on the competitiveness corresponding to each first parameter vector, determine a plurality of second parameter vectors from the plurality of first parameter vectors, wherein the plurality of second parameter vectors are first parameter vectors whose competitiveness is greater than a first preset threshold value among the plurality of first parameter vectors;

[0010] Performing transformation processing on the plurality of second parameter vectors to obtain a plurality of third parameter vectors;

[0011] Decoding the sample data features based on each third parameter vector and the initial model to obtain predicted noise characteristics corresponding to each third parameter vector;

[0012] Determine a first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample;

[0013] The initial model is trained based on the first target parameter vector to obtain a noise modeling model.

[0014] In a second aspect, an embodiment of the present application provides a noise modeling method, the method comprising:

[0015] Collect data on the power line channel to obtain initial data;

[0016] Structuring the initial data to obtain initial data features;

[0017] The initial features are input into a noise modeling model to obtain target noise characteristics corresponding to the initial data, wherein the noise modeling model is trained by the method of the first aspect.

[0018] In a third aspect, an embodiment of the present application provides a noise modeling model training device, the device comprising: a transceiver unit and a processing unit;

[0019] The transceiver unit is used to collect data on the power line channel and construct training samples, wherein the training samples include original data samples and noise characteristic samples;

[0020] The processing unit is used to structure the original data sample to obtain sample data features corresponding to the original data sample;

[0021] The processing unit is used to decode the sample data feature based on each first parameter vector and the initial model to obtain the predicted noise characteristic corresponding to each first parameter vector, wherein each first parameter vector is randomly generated;

[0022] The processing unit is used to determine the competitiveness corresponding to each first parameter vector based on the predicted noise characteristic corresponding to each first parameter vector and the noise characteristic sample;

[0023] The processing unit is configured to determine a plurality of second parameter vectors from the plurality of first parameter vectors based on the competitiveness corresponding to each first parameter vector, wherein the plurality of second parameter vectors are first parameter vectors whose competitiveness is greater than a first preset threshold value among the plurality of first parameter vectors;

[0024] The processing unit is used to perform transformation processing on the multiple second parameter vectors to obtain multiple third parameter vectors;

[0025] The processing unit is used to decode the sample data feature based on each third parameter vector and the initial model to obtain the predicted noise characteristic corresponding to each third parameter vector;

[0026] The processing unit is configured to determine a first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample;

[0027] The processing unit is used to train the initial model based on the first target parameter vector to obtain a noise modeling model.

[0028] In a fourth aspect, an embodiment of the present application provides a noise modeling device, the device comprising: a transceiver unit and a processing unit:

[0029] The transceiver unit is used to collect data on the power line channel to obtain initial data;

[0030] The processing unit is used to structure the initial data to obtain initial data features;

[0031] The processing unit is used to input the initial features into a noise modeling model to obtain target noise characteristics corresponding to the initial data, wherein the noise modeling model is trained by the method of the first aspect.

[0032] In a fifth aspect, an embodiment of the present application provides an electronic device, comprising: a processor and a memory, the processor being connected to the memory, the memory being used to store a computer program, and the processor being used to execute the computer program stored in the memory, so that the electronic device executes the methods of the first and second aspects.

[0033] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program enables a computer to execute the methods of the first aspect and the second aspect.

[0034] In a seventh aspect, an embodiment of the present application provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and is computer-operable to enable the computer to execute the methods of the first and second aspects.

[0035] Implementing the embodiments of the present application has the following beneficial effects:

[0036] It can be seen that in an embodiment of the present application, a training sample is constructed by collecting data on a power line channel, wherein the training sample includes an original data sample and a noise characteristic sample; the original data sample is structured to obtain sample data characteristics corresponding to the original data sample; the sample data characteristics are decoded based on each first parameter vector and the initial model to obtain a predicted noise characteristic corresponding to each first parameter vector, wherein each first parameter vector is randomly generated; based on the predicted noise characteristic and the noise characteristic sample corresponding to each first parameter vector, the competitiveness corresponding to each first parameter vector is determined; based on the competitiveness corresponding to each first parameter vector, a plurality of second parameter vectors are determined from a plurality of first parameter vectors, wherein the plurality of second parameter vectors are first parameter vectors in the plurality of first parameter vectors whose competitiveness is greater than a first preset threshold; the plurality of second parameter vectors are transformed to obtain a plurality of third parameter vectors; based on each third parameter vector and the initial model, the sample data characteristics are decoded to obtain a predicted noise characteristic corresponding to each third parameter vector; based on the predicted noise characteristic and the noise characteristic sample corresponding to each third parameter vector, a first target parameter vector is determined from a plurality of third parameter vectors; the initial model is trained based on the first target parameter vector to obtain a noise modeling model. First, the sample data features are decoded by randomly generating a first parameter vector and an initial model to obtain the predicted noise characteristics corresponding to each first parameter vector, and then based on the competitiveness corresponding to each first parameter vector, multiple second parameter vectors with competitiveness greater than a first preset threshold are determined from multiple first parameter vectors. In the present application, the multiple first parameter vectors with the largest competitiveness are not directly used as the first target parameter vector to prevent the obtained first target parameter vector from being a local optimal solution. Instead, the multiple second parameter vectors are continuously transformed to expand the search range to obtain multiple third parameter vectors, and then the first target parameter vector is determined from the multiple third parameter vectors so that the obtained first target parameter vector is a global optimal solution. The initial model is then trained based on the global optimal solution to obtain a noise modeling model, which can not only improve the accuracy of noise modeling, but also filter the noise in the power line channel based on the noise characteristics obtained by noise modeling, thereby improving communication quality, and can also improve the convergence speed of the noise modeling model. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 A flowchart of a noise modeling model training method provided in an embodiment of the present application;

[0039] Figure 2 A schematic diagram of a noise collection system provided in an embodiment of the present application;

[0040] Figure 3 An internal circuit diagram of a noise collection system provided in an embodiment of the present application;

[0041] Figure 4 A schematic diagram of the convergence speed of a BPNN in the prior art provided in an embodiment of the present application;

[0042] Figure 5 A schematic diagram of an optimized BPNN convergence speed provided in an embodiment of the present application;

[0043] Figure 6 A schematic diagram of a BPNN error in the prior art provided in an embodiment of the present application;

[0044] Figure 7 A schematic diagram of an optimized BPNN error provided in an embodiment of the present application;

[0045] Figure 8 A flowchart of another noise modeling model training method provided in an embodiment of the present application;

[0046] Fig. 9 A flowchart of a noise modeling method provided in an embodiment of the present application;

[0047] Fig.10 A schematic diagram of an application of a noise modeling method provided in an embodiment of the present application;

[0048] Fig.11 A block diagram of the functional units of a noise modeling model training device provided in an embodiment of the present application;

[0049] Fig.12 A block diagram of the functional units of a noise modeling device provided in an embodiment of the present application;

[0050] Fig.13 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0051] 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 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 ordinary technicians in this field without creative work are within the scope of protection of this application.

[0052] The terms "first", "second", "third" and "fourth" etc. in the specification, claims and drawings of the present application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0053] Reference to "embodiments" herein means that a particular feature, result, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0054] In order to facilitate understanding of the technical solution of the present application, the relevant technical terms involved in the present application are first explained.

[0055] Power Line Communication (PLC) channel is a technology that uses existing AC or DC power lines as a transmission medium for high-speed data transmission.

[0056] The characteristics of noise in the power line channel can be summarized as follows:

[0057] (1) Randomness. The noise sources on power lines are very diverse, many of which are unstable. These include noise from rain and thunder, and the fact that various electrical equipment can be powered on or off at any time, which can generate a large amount of random noise.

[0058] (2) Periodicity and continuity. In addition to random noise, there are also some slowly changing periodic noises. The amplitude of this periodic noise can reach 10V, which usually comes from low-power electrical appliances on the wires. At the same time, due to the wide variety of equipment, a large number of different harmonics are superimposed together, turning into continuous noise on the communication channel;

[0059] (3) Variability. Depending on the equipment connected at different times in different regions and the structure of the power distribution network, the amplitude and frequency of the noise will vary.

[0060] Secondly, there are five types of power line noise: colored background noise, narrowband noise, periodic pulse noise synchronized with the power frequency, periodic pulse noise asynchronous with the power frequency, and asynchronous non-periodic pulse noise.

[0061] Due to the different interference capabilities and continuous changing characteristics of the above five types of noise, they can be divided into two categories. The first three categories can be unified as background noise according to their own interference capabilities; since the other two types not only have high noise intensity but are mostly continuous burst-like pulse noise, researchers classify them as pulse noise.

[0062] (1) Colored background noise is a type of combined noise with a wide spectrum that can occupy the entire channel bandwidth. Its power spectrum density is relatively small compared to other noises and it slows down over time.

[0063] (2) Narrowband noise is noise composed of sinusoidal amplitude modulated signals, and its power spectrum is relatively high within a limited frequency range;

[0064] (3) The cause of pulse noise asynchronous to the power frequency is mainly the periodic opening and closing of high-output power supply equipment, and its cycle frequency is generally 50-200kHz;

[0065] (4) The characteristics of pulse noise synchronized with the power frequency are high power, long working duration, working frequency of 50 Hz or its integer multiples, and periodicity of 10ms-20ms;

[0066] (5) Random noise is different from other types of noise. It refers to a group of pulse interference groups with relatively concentrated energy, short duration and wide spectrum range.

[0067] See also Figure 1 , Figure 1 A flowchart of a noise modeling model training method provided in an embodiment of the present application, the method includes but is not limited to steps 101-109:

[0068] 101: Collect data on the power line channel and construct training samples.

[0069] Exemplarily, the training samples include original data samples and noise characteristic samples. In the process of model training, multiple data are collected to form a training set and a test set. The training sample in this application is just any sample in the training set. This application uses the training sample to train the initial model as an example to illustrate a noise modeling model training method. When there are multiple training samples, refer to the training method with only one training sample.

[0070] This application takes the collection of training samples as an example for explanation. The original data samples are the collected measured data. Specifically, Figure 2 The noise acquisition system shown in the figure is used for measurement. A coupling circuit and a high-speed acquisition card are used to connect to the low-voltage power line on the distribution box outside the laboratory for noise acquisition, and then the collected raw data samples are sent to the computer. The coupling circuit is used to couple the noise into the high-speed acquisition card. After the high-speed acquisition card obtains the raw data samples, it is transmitted to the computer for the next step of storage and analysis. It should be noted that the raw data samples include both the transmission data originally transmitted by the transmission line and the noise data generated by the transmission line during the data transmission process. The noise characteristic sample represents the characteristic behavior of the noise in the raw data sample, that is, the performance and change law of the noise in the power line channel under different times, frequencies and conditions, including: amplitude distribution, power spectrum density, periodicity, etc.

[0071] Furthermore, the internal circuit of the noise collection system is as follows Figure 3 As shown, it includes: low-voltage power line, power frequency transformer, coupler, high-speed acquisition card and computer terminal. Among them, one end of each coupler is connected to the A channel and B channel of the power line respectively, and the other end is connected to the neutral line. One end of each power frequency transformer is connected to the A channel and B channel of the power line respectively, and the other end is connected to the live line. The high-frequency (1MHz-30MHz) interference and noise in the power line network are adjusted to -5VDC-+5VDC and the noise frequency range is adjusted to 1MHz-2.5MHz through the coupler and the power frequency transformer to isolate the high-frequency noise signal in the power line network, so as to provide a pure power supply to the acquisition card and reduce the impact of high-frequency noise in the power line on the acquisition card. In the coupling circuit, an inductive coupling coil based on the electromagnetic induction principle of the transformer is used, that is, an integrated circuit is used to transmit power or transmit signals to another integrated circuit. In the acquisition of raw data samples, the main functions of the coupling circuit are: to prevent permanent damage to the acquisition card; to increase strong electrical signals to suppress the interference of power frequency signals on noise data testing and ensure the safety of the acquisition equipment; to use inductive couplers to couple with high-frequency signals to obtain accurate power line noise data; to reduce the attenuation of high-frequency signals to ensure that noise signals within the frequency range required for the experiment are coupled into the acquisition system to obtain the raw data samples of the PLC channel.

[0072] 102: Structuring the original data samples to obtain sample data features corresponding to the original data samples.

[0073] Exemplarily, the original data samples are structured to obtain sample data features corresponding to the original data samples, including: data cleaning of the original data samples, removing or correcting outliers, missing values ​​and duplicate values ​​in the data set to ensure the quality of the data. For missing values, interpolation or filling methods can be used; for outliers, they can be identified and eliminated by setting thresholds. Then, the cleaned original data samples are normalized to scale the data to a specific range, usually [0,1] or [-1,1]. This helps to speed up the learning speed of the neural network because data in a small value range is easier to process. Finally, the dimension of the data is adjusted according to the structural requirements of the BPNN to obtain the sample data features.

[0074] 103: Decode the sample data feature based on each first parameter vector and the initial model to obtain the predicted noise characteristic corresponding to each first parameter vector.

[0075] Exemplarily, the sample data features are decoded based on each first parameter vector and the initial model to obtain the predicted noise characteristics corresponding to each first parameter vector, wherein each first parameter vector is randomly generated.

[0076] It should be noted that the initial model refers to BPNN, and the number of BPNN layers can be set to 1 for the input layer and the output layer. The number of hidden layers has a great influence on the neural network. Fewer hidden layers will slow down the convergence speed of the network and weaken the generalization ability; more hidden layers will increase the training time. Furthermore, the number of neurons in the input layer and the output layer can be determined by the dimension of the data. In this article, the input data can be 7-dimensional, and the output data can be 1-dimensional, that is, the number of neurons in the input layer can be 7, and the number of neurons in the output layer can be 1. Furthermore, the training function of the initial model is the Levenberg-Marquardt algorithm in the Trainlm function, which has a faster convergence speed.

[0077] Specifically, the first parameter vector is a set of BPNN weights and thresholds. Before model training, multiple first parameter vectors are randomly generated, and the randomly generated first parameter vectors are applied to the initial model. The sample data features are decoded based on each first parameter vector, and the predicted noise characteristics corresponding to each first parameter vector are output through the initial model. Among them, the dimensions of the predicted noise characteristics and the noise characteristic samples are the same. The predicted noise characteristics represent the characteristics and behaviors of the noise predicted by the initial model.

[0078] 104: Determine the degree of competition corresponding to each first parameter vector based on the predicted noise characteristic and the noise characteristic sample corresponding to each first parameter vector.

[0079] Exemplarily, based on the predicted noise characteristics and noise characteristic samples corresponding to each first parameter vector, determining the competitiveness corresponding to each first parameter vector includes: obtaining the matching degree between the predicted noise characteristics corresponding to each first parameter vector and the sample data characteristics and; obtaining the second matching degree between the predicted noise characteristics corresponding to each first parameter vector; based on the second matching degree and the second preset threshold, obtaining the first similarity between the predicted noise characteristics corresponding to each first parameter vector and the predicted noise characteristics corresponding to other first parameter vectors; based on the first similarity, determining the frequency of the predicted noise characteristics corresponding to each first parameter vector among the predicted noise characteristics corresponding to multiple first parameter vectors; based on the matching degree and frequency between the predicted noise characteristics corresponding to each first parameter vector and the sample data characteristics and, determining the competitiveness corresponding to each first parameter vector.

[0080] Optionally, obtain the matching degree between the predicted noise characteristic corresponding to each first parameter vector and the sample data characteristic and. The matching degree can be represented by the Euclidean distance between the predicted noise characteristic corresponding to each first parameter vector and the sample data characteristic and. The matching degree between the predicted noise characteristic corresponding to each first parameter vector and the sample data characteristic and can be determined by formula (1):

[0081]

[0082] Among them, M(a i ,c m ) means a i and c m The matching degree, a i represents the predicted noise characteristics corresponding to the i-th first parameter vector among the predicted noise characteristics corresponding to the multiple first parameter vectors, a i,k represents the kth dimension of the predicted noise characteristic corresponding to the i-th first parameter vector, c m represents the sample data characteristics, c m,k represents the kth dimension of the sample data features, and L is the total dimension of the predicted noise characteristics and sample data features.

[0083] The first matching degree between the predicted noise characteristic corresponding to each first parameter vector and the sample data characteristic is obtained, which represents the first matching degree between the predicted noise characteristic obtained by applying each first parameter vector and the sample characteristic. The larger the first matching degree is, the smaller the error of the predicted noise characteristic obtained by decoding the first parameter vector in the initial model is. Therefore, the first matching degree can be used to measure the error between the predicted noise characteristic obtained by the initial model applying each first parameter vector and the sample data characteristic.

[0084] Optionally, a second matching degree between the predicted noise characteristics corresponding to each first parameter vector is obtained. Specifically, the second matching degree between the predicted noise characteristics corresponding to each first parameter vector can be determined by formula (2).

[0085]

[0086] Among them, M(a i ,b j ) means a i and b j The matching degree, a i represents the predicted noise characteristics corresponding to the i-th first parameter vector among the predicted noise characteristics corresponding to the multiple first parameter vectors, a i,k represents the kth dimension of the predicted noise characteristic corresponding to the i-th first parameter vector, b j represents the predicted noise characteristic corresponding to the jth first parameter vector among the predicted noise characteristics corresponding to the multiple first parameter vectors, where i and j are not equal, and b j,k represents the kth dimension of the predicted noise characteristic corresponding to the jth first parameter vector, and L is the total dimension of the predicted noise characteristic.

[0087] Optionally, based on the second matching degree and the second preset threshold, a first similarity between the predicted noise characteristics corresponding to each first parameter vector and the predicted noise characteristics corresponding to other first parameter vectors is obtained. The first similarity between the predicted noise characteristics corresponding to each first parameter vector and the predicted noise characteristics corresponding to other first parameter vectors can be determined by formula (3):

[0088]

[0089] Among them, S(a i ,b j ) means a i and b j The similarity of i represents the predicted noise characteristics corresponding to the i-th first parameter vector among the predicted noise characteristics corresponding to the multiple first parameter vectors, b j represents the predicted noise characteristic corresponding to the j-th first parameter vector among the predicted noise characteristics corresponding to the multiple first parameter vectors, wherein i and j are not equal, and x represents the second preset threshold.

[0090] Optionally, based on the first similarity, the frequency of the predicted noise characteristic corresponding to each first parameter vector in the predicted noise characteristics corresponding to multiple first parameter vectors is determined. Specifically, it can be determined by formula (4):

[0091]

[0092] Among them, ai represents the predicted noise characteristic corresponding to the i-th first parameter vector among the predicted noise characteristics corresponding to the multiple first parameter vectors, P(a i ) represents the frequency of the predicted noise characteristic corresponding to the i-th first parameter vector in the predicted noise characteristics corresponding to the multiple first parameter vectors, N is the number of predicted noise characteristics corresponding to the multiple first parameter vectors, S(a i ,b j ) means a i and b j The similarity of j Represents the predicted noise characteristic corresponding to the jth first parameter vector among the predicted noise characteristics corresponding to multiple first parameter vectors.

[0093] It should be noted that determining the frequency of the predicted noise characteristic corresponding to each first parameter vector in the predicted noise characteristics corresponding to multiple first parameter vectors indicates the uniqueness of the predicted noise characteristic corresponding to each first parameter vector. The lower the frequency, the more unique the predicted noise characteristic corresponding to the first parameter vector is. The predicted noise characteristic corresponding to the first parameter vector can increase the diversity of the predicted noise characteristics corresponding to multiple first parameter vectors. The higher the frequency, the more frequently the predicted noise characteristic corresponding to the first parameter vector appears. The first parameter vector may not be able to increase the diversity of the predicted noise characteristics corresponding to multiple first parameter vectors, which will make the predicted noise characteristics corresponding to multiple first parameter vectors gradually tend to be single, which may cause the model to fall into a local optimal solution, thereby affecting the global search capability of the model.

[0094] Optionally, based on the first matching degree and the frequency, the competitiveness corresponding to each first parameter vector is determined. Specifically, it can be determined by formula (5):

[0095] V(a i )=y*M(a i ,c m )-z*P(a i ) Formula (5)

[0096] Among them, a i represents the predicted noise characteristics corresponding to the i-th first parameter vector among the predicted noise characteristics corresponding to the multiple first parameter vectors, V(a i ) represents the competitiveness corresponding to the i-th first parameter vector, y is the third preset threshold, z is the fourth preset threshold, P(a i ) represents the frequency of the predicted noise characteristic corresponding to the i-th first parameter vector in the predicted noise characteristics corresponding to multiple first parameter vectors, M(a i ,c m ) means a i and c m The matching degree, c mRepresents the characteristics of sample data.

[0097] The competitiveness corresponding to each first parameter vector represents the degree of optimization of the performance of the initial model by each first parameter vector. The higher the competitiveness corresponding to the first parameter vector, the stronger the performance of the initial model using the first parameter vector.

[0098] It can be seen that in the present application, based on the first matching degree and frequency, the competitiveness corresponding to each first parameter vector is determined. The higher the corresponding first matching degree and the lower the frequency, the higher the competitiveness of the first parameter vector, and the stronger the performance of the initial model using the first parameter vector. The higher the first matching degree, the smaller the error of the predicted noise characteristics obtained by decoding the first parameter vector in the initial model, and the lower the frequency, the more unique the predicted noise characteristics corresponding to the first parameter vector. The predicted noise characteristics corresponding to the first parameter vector can increase the diversity of the predicted noise characteristics corresponding to multiple first parameter vectors, and improve the global search capability of the model. Therefore, based on the first matching degree and frequency, determining the competitiveness corresponding to each first parameter vector can determine the corresponding first parameter vector with high first matching degree and low frequency, so as to facilitate the subsequent training of the initial model based on the first parameter vector. On the basis of improving the accuracy of the model, the global search capability of the model can also be improved, so that the model can find the global optimal solution and ensure the global convergence capability of the model.

[0099] 105: Determine a plurality of second parameter vectors from the plurality of first parameter vectors based on the competitiveness corresponding to each first parameter vector.

[0100] Exemplarily, based on the competitiveness corresponding to each first parameter vector, multiple second parameter vectors are determined from multiple first parameter vectors, wherein the multiple second parameter vectors are first parameter vectors whose competitiveness is greater than a first preset threshold value among the multiple first parameter vectors. Specifically, the competitiveness corresponding to each first parameter vector is sorted in order from large to small, the median value is used as the first preset threshold value, and the first parameter vectors whose corresponding competitiveness is greater than the first preset threshold value among the multiple first parameter vectors are used as multiple second parameter vectors. It should be noted that the first parameter vectors corresponding to the first N / 2 competitiveness after sorting can also be directly used as multiple second parameter vectors.

[0101] It can be seen that in the embodiment of the present application, based on the degree of competitiveness corresponding to each first parameter vector, multiple second parameter vectors are determined from multiple first parameter vectors, wherein the multiple second parameter vectors are first parameter vectors whose degree of competitiveness is greater than a first preset threshold value among the multiple first parameter vectors, and the corresponding first parameter vector with a large first matching degree and a small frequency is screened out through the size of the degree of competitiveness corresponding to each first parameter vector, so that the corresponding first parameter vector with a high frequency is suppressed, and the corresponding first parameter vector with a high matching degree and a small frequency is retained, thereby ensuring the diversity of the predicted noise characteristics corresponding to the first parameter vectors. On the basis of improving the accuracy of the model, the global search capability of the model can also be improved, so that the model can find the global optimal solution and ensure the global convergence ability of the model.

[0102] 106: Perform transformation processing on the multiple second parameter vectors to obtain multiple third parameter vectors.

[0103] Exemplarily, multiple second parameter vectors are transformed to obtain multiple third parameter vectors, including: copying multiple second parameter vectors to obtain multiple fourth parameter vectors; cross-merging multiple fourth parameter vectors to obtain multiple fifth parameter vectors; mutating multiple fifth parameter vectors to obtain multiple third parameter vectors. First, multiple second parameter vectors are copied to obtain multiple fourth parameter vectors, and the selected better solutions (i.e., multiple second parameter vectors) are copied to generate multiple identical copies, and the number of copies can be fixed. By copying the better solution, the obtained multiple fourth parameter vectors can increase the propagation speed of high-quality parameter vectors and accelerate convergence while maintaining the diversity of parameter vectors.

[0104] Optionally, multiple fourth parameter vectors are cross-fused to obtain multiple fifth parameter vectors, and one or more parameter vectors are randomly selected from the multiple fourth parameter vectors for pairing. According to a certain crossover probability and crossover strategy (such as single-point crossover, multi-point crossover, uniform crossover, etc.), part of the data of the paired multiple parameter vectors is exchanged to generate new parameter vectors, that is, multiple fifth parameter vectors. The crossover operation can combine the advantages of two parameter vectors to produce a new and possibly better solution. In addition, new areas in the solution space can be explored through crossover to increase the chance of finding a global optimal solution.

[0105] Optionally, multiple fifth parameter vectors are mutated to obtain multiple third parameter vectors. The fifth parameter vector is selected for mutation according to a certain mutation probability. The values ​​at one or more vector positions of the selected fifth parameter vector are randomly changed. The mutation method can be random inversion, random addition or subtraction of a constant, etc. The mutation operation can introduce new values, increase the diversity of the fifth parameter vector, prevent the model from falling into a local optimal solution, and help the model to perform a fine search in a local range. It can also jump out of the local optimum. When the model falls into a local optimum, an appropriate mutation operation can help it jump out of the local area and continue to search for a better solution.

[0106] It can be seen that in the embodiment of the present application, multiple second parameter vectors are copied to obtain multiple fourth parameter vectors; multiple fourth parameter vectors are cross-fused to obtain multiple fifth parameter vectors; multiple fifth parameter vectors are mutated to obtain multiple third parameter vectors. Multiple second parameter vectors are copied, cross-fused and mutated to obtain multiple third parameter vectors. First, by copying the better solution (i.e., multiple second parameter vectors), the obtained multiple fourth parameter vectors can increase the propagation speed of high-quality parameter vectors while maintaining the diversity of parameter vectors. Secondly, by crossing, new areas in the solution space are explored to increase the chance of finding the global optimal solution; again, new values ​​are introduced through mutation operations to increase the diversity of the fifth parameter vector, prevent the model from falling into the local optimal solution, and help the model to perform fine search in the local range. It is also possible to jump out of the local optimum. When the model falls into the local optimum, appropriate mutation operations can help it jump out of the local area and continue to search for a better solution until the optimal solution or the approximate optimal solution is found.

[0107] 107: Decode the sample data feature based on each third parameter vector and the initial model to obtain the predicted noise characteristic corresponding to each third parameter vector.

[0108] It should be noted that the sample data features are decoded based on each third parameter vector and the initial model to obtain the predicted noise characteristics corresponding to each third parameter vector. Each third parameter vector is applied to the initial model, and the sample data features are decoded based on each third parameter vector, and the predicted noise characteristics corresponding to each third parameter vector output by the initial model are obtained. Among them, the dimensions of the predicted noise characteristics and the noise characteristic samples are the same. The predicted noise characteristics represent the characteristics and behaviors of the noise predicted by the initial model.

[0109] 108: Determine a first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic and the noise characteristic sample corresponding to each third parameter vector.

[0110] Exemplarily, based on the predicted noise characteristics and noise characteristic samples corresponding to each third parameter vector, determining the first target parameter vector from multiple third parameter vectors includes: determining the competitiveness corresponding to each third parameter vector based on the predicted noise characteristics and noise characteristic samples corresponding to each third parameter vector; and taking the third parameter vector with the largest corresponding competitiveness among the multiple third parameter vectors as the first target parameter vector. It should be noted that the method for determining the competitiveness corresponding to each third parameter vector based on the predicted noise characteristics and noise characteristic samples corresponding to each third parameter vector is similar to the method for determining the competitiveness corresponding to each first parameter vector based on the predicted noise characteristics and noise characteristic samples corresponding to each first parameter vector recorded in step 104, and the present application will not repeat them here.

[0111] It should be noted that before using the third parameter vector with the greatest competitiveness among multiple third parameter vectors as the first target parameter vector, the method also includes: updating the multiple third parameter vectors, sorting them in descending order based on the competitiveness corresponding to each third parameter vector, retaining the first N / 2 third parameter vectors, removing the last N / 2 third parameter vectors, and randomly generating N / 2 sixth parameter vectors, and reusing the first N / 2 third parameter vectors and N / 2 sixth parameter vectors as multiple first parameter vectors. This process can be understood as an update of the parameter vector, which is updated every time the model is iterated.

[0112] Further, after updating multiple third parameter vectors, it is also determined whether the iterative process meets the preset conditions. Specifically, the preset condition can be whether the number of iterations is greater than the fifth preset threshold, or whether the error is less than the sixth preset threshold. In the present application, the preset condition is whether the error is less than the sixth preset threshold as an example for explanation. The third parameter vector with the highest degree of competition among multiple third parameter vectors is obtained, and the third matching degree between the predicted noise characteristic corresponding to the third parameter vector with the highest degree of competition and the sample data characteristic and is obtained. If the third matching degree is greater than the sixth preset threshold, the corresponding third parameter vector with the highest degree of competition is used as the first target parameter vector.

[0113] Furthermore, if the third matching degree is less than a sixth preset threshold, the sample data feature is decoded again based on the new plurality of first parameter vectors, and the process of steps 103 - 108 is repeated until the first target parameter vector is determined.

[0114] It can be seen that in an embodiment of the present application, based on each first parameter vector and the initial model, the sample data features are decoded to obtain the predicted noise characteristics corresponding to each first parameter vector, wherein each first parameter vector is randomly generated; based on the predicted noise characteristics and noise characteristic samples corresponding to each first parameter vector, the competitiveness corresponding to each first parameter vector is determined; based on the competitiveness corresponding to each first parameter vector, multiple second parameter vectors are determined from multiple first parameter vectors, wherein the multiple second parameter vectors are first parameter vectors in the multiple first parameter vectors whose competitiveness is greater than a first preset threshold; the multiple second parameter vectors are transformed to obtain multiple third parameter vectors; based on each third parameter vector and the initial model, the sample data features are decoded to obtain the predicted noise characteristics corresponding to each third parameter vector; based on the predicted noise characteristics and noise characteristic samples corresponding to each third parameter vector, the first target parameter vector is determined from the multiple third parameter vectors. It can be seen that the present application starts from a plurality of randomly generated first parameter vectors (initial solutions) to search for the first target parameter vector (optimal solution). This determination method is not limited to a specific problem, nor does it emphasize the quality of the algorithm parameter settings and the initial solution. By utilizing its heuristic intelligent search mechanism, even if it starts with a poor quality initial solution, it can eventually search for the global optimal solution to the problem. It has low dependence on the problem and the initial solution, and has strong adaptability and robustness.

[0115] 109: Train the initial model based on the first target parameter vector to obtain a noise modeling model.

[0116] Exemplarily, the initial model is trained based on the first target parameter vector to obtain a noise modeling model, including: decoding sample data features based on the first target parameter vector to obtain predicted noise characteristics corresponding to the first target parameter vector; obtaining the loss between the predicted noise characteristics corresponding to the first target parameter vector and the noise characteristic sample; and training the model based on the loss to obtain a noise modeling model.

[0117] It should be noted that after obtaining the first target parameter vector through steps 101-108, the BPNN needs to be trained through the first target parameter vector. Specifically, the sample data features are decoded based on the first target parameter vector to obtain the predicted noise characteristics corresponding to the first target parameter vector. The first target parameter vector is applied to the initial model, the sample data is input into the initial model, and the initial model decodes the sample data features to obtain the predicted noise characteristics corresponding to the first target parameter vector. The loss between the predicted noise characteristics corresponding to the first target parameter vector and the noise characteristic sample is obtained, wherein the loss can be the mean square error, cross entropy loss, etc. between the predicted noise characteristics corresponding to the first target parameter vector and the noise characteristic sample, and this application does not limit this. If the loss meets the preset end condition, the model converges to obtain a trained noise modeling model. If the loss does not meet the preset end condition, the first target parameter is adjusted by the loss to obtain a new first target parameter until the model converges to obtain a trained noise modeling model. In practical applications, the losses corresponding to multiple training samples can be weighted to obtain a weighted loss, and the weighted loss can be used to adjust the first target parameter of the initial model; or the first target parameter of the initial model can be adjusted one by one using the loss of each training sample, which is not limited in this application.

[0118] It can be seen that in an embodiment of the present application, a training sample is constructed by collecting data on a power line channel, wherein the training sample includes an original data sample and a noise characteristic sample; the original data sample is structured to obtain sample data characteristics corresponding to the original data sample; the sample data characteristics are decoded based on each first parameter vector and the initial model to obtain a predicted noise characteristic corresponding to each first parameter vector, wherein each first parameter vector is randomly generated; based on the predicted noise characteristic and the noise characteristic sample corresponding to each first parameter vector, the competitiveness corresponding to each first parameter vector is determined; based on the competitiveness corresponding to each first parameter vector, a plurality of second parameter vectors are determined from a plurality of first parameter vectors, wherein the plurality of second parameter vectors are first parameter vectors in the plurality of first parameter vectors whose competitiveness is greater than a first preset threshold; the plurality of second parameter vectors are transformed to obtain a plurality of third parameter vectors; based on each third parameter vector and the initial model, the sample data characteristics are decoded to obtain a predicted noise characteristic corresponding to each third parameter vector; based on the predicted noise characteristic and the noise characteristic sample corresponding to each third parameter vector, a first target parameter vector is determined from a plurality of third parameter vectors; the initial model is trained based on the first target parameter vector to obtain a noise modeling model. First, the sample data features are decoded by randomly generating a first parameter vector and an initial model to obtain the predicted noise characteristics corresponding to each first parameter vector, and then based on the competitiveness corresponding to each first parameter vector, multiple second parameter vectors with competitiveness greater than a first preset threshold are determined from multiple first parameter vectors. In the present application, the multiple first parameter vectors with the largest competitiveness are not directly used as the first target parameter vector to prevent the obtained first target parameter vector from being a local optimal solution. Instead, the multiple second parameter vectors are continuously transformed to expand the search range to obtain multiple third parameter vectors, and then the first target parameter vector is determined from the multiple third parameter vectors so that the obtained first target parameter vector is a global optimal solution. The initial model is then trained based on the global optimal solution to obtain a noise modeling model, which can not only improve the accuracy of noise modeling, but also filter the noise in the power line channel based on the noise characteristics obtained by noise modeling, thereby improving communication quality, and can also improve the convergence speed of the noise modeling model.

[0119] In one embodiment of the present application, in order to verify the effectiveness of the training method in the present application, a simulation operation is also performed, specifically including: forming a training set, constructing a simulation experimental platform using MATLAB language, and using the built-in BPNN toolbox, using the algorithm process recorded in steps 101-107 to continuously optimize the initial weight threshold (i.e., parameter vector) of BPNN, obtain the first target parameter vector, and use the first target parameter vector to input into the initial model BPNN for subsequent power line channel noise modeling. BPNN relies on different excitation functions for nonlinear fitting. When the network has no excitation function, its learning form is a linear relationship, and it can only process some simple two-dimensional functions, etc. Therefore, the excitation function is selected to set so that the fitting effect of BPNN on nonlinear problems is determined to be the tan-sigmoid function after multiple simulations.

[0120] Then, the initial model (ie, BPNN) is trained using the method described in step 108 to obtain a trained noise modeling model.

[0121] like Figure 4 and Figure 5 As shown, compared with the noise modeling by BPNN in the prior art, the noise modeling by the optimized BPNN in this application can significantly improve the iterative convergence speed of the model. Figure 4 for Figure 5 The horizontal axis is the number of iterations, and the vertical axis is the error. Figure 4 To model noise using the existing BPNN, the iterative convergence speed and error variation of the model; Figure 5 For noise modeling through the optimized BPNN in this application (i.e., optimized by the method described in steps 101-109), the iterative convergence speed and error change of the model can be clearly seen from the figure. The optimized BPNN for noise modeling can significantly improve the iterative convergence speed of the model, reduce the error, and improve the accuracy.

[0122] In addition, if Figure 6 and Figure 7 As shown, the horizontal axis represents the test sample number and the vertical axis represents the specific value. Figure 4 The expected value and error value are obtained by noise modeling through BPNN of the prior art; Figure 5 The expected value and error value obtained by noise modeling through the optimized BPNN in the present application (i.e., optimized by the method described in steps 101-109) can be clearly seen from the figure that the gap between the expected value and the error value obtained by noise modeling through the optimized BPNN is significantly smaller than the expected value and the error value obtained by noise modeling through the BPNN of the prior art.

[0123] Further, as shown in Table 1, two different noise modeling methods are compared, namely the BPNN of the prior art and the BPNN optimized in this application (represented by the optimized BPNN in Table 1), the number of test set samples is 500. It can be seen from Table 1 that the average error of the BPNN of the prior art and the optimized BPNN prediction data is 15% and 3% respectively; the mean square error of the BPNN of the prior art and the optimized BPNN prediction data is 7.60E-04 and 9.10E-05 respectively; the correlation coefficients between the BPNN of the prior art and the optimized BPNN and the actual data are 73% and 97% respectively. Therefore, it can be seen that the optimized BPNN has the characteristics of small error and high accuracy.

[0124] Table 1

[0125]

[0126] In one embodiment of the present application, another noise modeling model training method is also provided. Figure 8 , Figure 8 A flowchart of another noise modeling model training method provided in an embodiment of the present application, the method includes but is not limited to steps 801-813:

[0127] 801: Input raw data sample.

[0128] 802: Randomly generate multiple first parameter vectors.

[0129] 803: Determine a first matching degree and frequency corresponding to each first parameter vector.

[0130] 804: Determine the degree of competition corresponding to each primary parameter vector based on the first matching degree and frequency corresponding to each first parameter vector.

[0131] 805: Obtain multiple second parameter vectors based on the degree of competition corresponding to each primary parameter vector.

[0132] Specifically, the first parameter vectors corresponding to the first N / 2 competitive degrees are stimulated, and the first parameter vectors corresponding to the last N / 2 competitive degrees are suppressed, so as to obtain a plurality of second parameter vectors.

[0133] 806: Perform transformation processing on the multiple second parameter vectors to obtain multiple third parameter vectors.

[0134] 807: Update multiple third parameter vectors.

[0135] 808: Whether the iteration process meets the preset conditions.

[0136] 809: If so, use the corresponding third parameter vector with the greatest competitiveness as the first target parameter vector.

[0137] If not, the process of steps 803 - 809 is repeated until the first target parameter vector is determined.

[0138] 810: Initial model.

[0139] Specifically, the initial model decodes the sample data features based on the first target parameter vector to obtain the predicted noise characteristics corresponding to the first target parameter vector.

[0140] 811: Obtain the loss between the predicted noise characteristic and the noise characteristic sample corresponding to the first target parameter vector.

[0141] 812: Determine whether the loss meets the preset conditions.

[0142] 813: If yes, the model converges and a trained noise modeling model is obtained.

[0143] If not, adjust the initial model based on the losses.

[0144] The method of steps 801-813 has been specifically described in steps 101-109, and this application will not repeat them here.

[0145] It should be noted that each iteration of the model will go through an immune memory process.

[0146] See also Fig. 9 , Fig. 9 A flowchart of a noise modeling method provided in an embodiment of the present application, the method includes but is not limited to steps 901-903:

[0147] 901: Collect data on the power line channel to obtain initial data.

[0148] like Fig.10 As shown, the data is collected by the method described in step 101, which will not be described in detail in this application.

[0149] 902: Structuring the initial data to obtain initial data features.

[0150] like Fig.10 , data structuring is performed, which is similar to the structuring method recorded in step 102, and this application will not repeat it here.

[0151] 903: Input the initial features into the noise modeling model to obtain the target noise characteristics corresponding to the initial data.

[0152] The noise modeling model is trained by the method described in steps 101-109.

[0153] After 903, the method further includes: performing spectrum analysis on the target noise characteristics to obtain frequency components of noise data in the initial data; and filtering the initial data based on the frequency components to obtain actual transmission data in the initial data.

[0154] Specifically, Fig.10 As shown, the target noise characteristics are obtained through the noise modeling model, and the frequency components of the noise data in the initial data are obtained by performing spectrum analysis on the target noise characteristics through the noise model. Specifically, the frequency components of the main noise are found by performing spectrum analysis on the target noise characteristics using fast Fourier transform or other spectrum analysis methods.

[0155] Furthermore, based on the frequency components, the initial data is filtered to obtain the actual transmission data in the initial data. Specifically, according to the results of the spectrum analysis (i.e., the frequency components), a band-stop filter is designed, whose notch frequency is aligned with the main frequency of the noise. Fig.10 As shown, the designed filter is applied to the target noise characteristics, the noise component of a certain frequency is filtered out, the initial data is filtered, and the actual transmission data in the initial data is obtained.

[0156] It can be seen that in the embodiment of the present application, data on the power line channel is collected to obtain initial data; the initial data is structured to obtain initial data features; the initial features are input into the noise modeling model to obtain target noise features corresponding to the initial data, wherein the noise modeling model is trained by the noise modeling model training method; the target noise features are spectrally analyzed to obtain the frequency components of the noise data in the initial data; based on the frequency components, the initial data is filtered to obtain the actual transmission data in the initial data. By extracting the target noise features in the initial data through the trained noise modeling model, the accuracy and efficiency of noise feature extraction can be improved, and then spectrally analyzing and filtering based on the target noise features can effectively remove the noise in the initial data and filter the noise in the power line channel, thereby improving the communication quality.

[0157] See also Fig.11 , Fig.11 A block diagram of the functional units of a noise modeling model training device provided in an embodiment of the present application. The noise modeling model training device 1100 includes: a transceiver unit 1101 and a processing unit 1102;

[0158] The transceiver unit 1101 is used to collect data on the power line channel and construct training samples, wherein the training samples include original data samples and noise characteristic samples;

[0159] A processing unit 1102, the processing unit is used to structure the original data sample to obtain sample data features corresponding to the original data sample;

[0160] Used to decode the sample data feature based on each first parameter vector and the initial model to obtain the predicted noise characteristic corresponding to each first parameter vector, wherein each first parameter vector is randomly generated;

[0161] for determining the degree of competition corresponding to each first parameter vector based on the predicted noise characteristic corresponding to each first parameter vector and the noise characteristic sample;

[0162] for determining a plurality of second parameter vectors from a plurality of first parameter vectors based on the competitiveness corresponding to each first parameter vector, wherein the plurality of second parameter vectors are first parameter vectors whose competitiveness is greater than a first preset threshold value among the plurality of first parameter vectors;

[0163] Used to perform transformation processing on the multiple second parameter vectors to obtain multiple third parameter vectors;

[0164] Used to decode the sample data characteristics based on each third parameter vector and the initial model to obtain the predicted noise characteristics corresponding to each third parameter vector;

[0165] for determining a first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample;

[0166] Used to train the initial model based on the first target parameter vector to obtain a noise modeling model.

[0167] In one embodiment of the present application, in determining the degree of competition corresponding to each first parameter vector based on the predicted noise characteristic corresponding to each first parameter vector and the noise characteristic sample, the processing unit 1102 is specifically configured to:

[0168] Obtaining a first matching degree between the predicted noise characteristic corresponding to each first parameter vector and the sample data characteristic;

[0169] Obtaining a second matching degree between the predicted noise characteristics corresponding to each first parameter vector;

[0170] Based on the second matching degree and the second preset threshold, obtaining a first similarity between the predicted noise characteristic corresponding to each first parameter vector and the predicted noise characteristics corresponding to other first parameter vectors;

[0171] Based on the first similarity, determining the frequency of the predicted noise characteristic corresponding to each first parameter vector among the predicted noise characteristics corresponding to the plurality of first parameter vectors;

[0172] Based on the first matching degree and the frequency, a competitiveness corresponding to each first parameter vector is determined.

[0173] In one embodiment of the present application, in the aspect of performing transformation processing on the plurality of second parameter vectors to obtain a plurality of third parameter vectors, the processing unit 1102 is specifically configured to:

[0174] Copying the multiple second parameter vectors to obtain multiple fourth parameter vectors;

[0175] Cross-merging the multiple fourth parameter vectors to obtain multiple fifth parameter vectors;

[0176] The plurality of fifth parameter vectors are mutated to obtain a plurality of third parameter vectors.

[0177] In one embodiment of the present application, in determining the first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample, the processing unit 1102 is specifically configured to:

[0178] Determine the degree of competition corresponding to each third parameter vector based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample;

[0179] The third parameter vector with the greatest competitiveness among the multiple third parameter vectors is used as the first target parameter vector.

[0180] In one embodiment of the present application, in the aspect of training the initial model based on the first target parameter vector to obtain the noise modeling model, the processing unit 1102 is specifically configured to:

[0181] Decoding the sample data feature based on the first target parameter vector to obtain a predicted noise characteristic corresponding to the first target parameter vector;

[0182] Obtaining a loss between a predicted noise characteristic corresponding to a first target parameter vector and the noise characteristic sample;

[0183] Model training is performed based on the loss to obtain the noise modeling model.

[0184] See also Fig.12 , Fig.12 A block diagram of the functional units of a noise modeling device provided in an embodiment of the present application. The noise modeling device 1200 includes: a transceiver unit 1201 and a processing unit 1202;

[0185] The transceiver unit 1201 is used to collect data on the power line channel to obtain initial data;

[0186] The processing unit 1202 is used to structure the initial data to obtain initial data features;

[0187] The processing unit 1202 is used to input the initial features into a noise modeling model to obtain target noise characteristics corresponding to the initial data, wherein the noise modeling model is trained by a noise modeling model training device.

[0188] In one embodiment of the present application, the processing unit 1202 is specifically used to: perform spectrum analysis on the target noise characteristics to obtain the frequency components of the noise data in the initial data; and filter the initial data based on the frequency components to obtain the actual transmission data in the initial data.

[0189] See also Fig.13 , Fig.13 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Fig.13 As shown, the electronic device 1300 includes a transceiver 1301, a processor 1302 and a memory 1303. They are connected via a bus 1304. The memory 1303 is used to store computer programs and data, and can transmit the data stored in the memory 1303 to the processor 1302.

[0190] Optionally, the processor 1302 is configured to read a computer program in the memory 1303 to perform the following operations:

[0191] Controlling the transceiver 1301 to collect data on the power line channel and construct training samples, wherein the training samples include original data samples and noise characteristic samples;

[0192] Structuring the original data sample to obtain sample data features corresponding to the original data sample;

[0193] Decoding the sample data feature based on each first parameter vector and the initial model to obtain a predicted noise characteristic corresponding to each first parameter vector, wherein each first parameter vector is randomly generated;

[0194] Determine the degree of competition corresponding to each first parameter vector based on the predicted noise characteristic corresponding to each first parameter vector and the noise characteristic sample;

[0195] Based on the competitiveness corresponding to each first parameter vector, determine a plurality of second parameter vectors from the plurality of first parameter vectors, wherein the plurality of second parameter vectors are first parameter vectors whose competitiveness is greater than a first preset threshold value among the plurality of first parameter vectors;

[0196] Performing transformation processing on the plurality of second parameter vectors to obtain a plurality of third parameter vectors;

[0197] Decoding the sample data features based on each third parameter vector and the initial model to obtain predicted noise characteristics corresponding to each third parameter vector;

[0198] Determine a first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample;

[0199] The initial model is trained based on the first target parameter vector to obtain a noise modeling model.

[0200] Specifically, the transceiver 1301 may be Fig.11 The noise modeling model training device 1100 of the embodiment of the transceiver unit 1101, the processor 1302 can be Fig.11 The processing unit 1102 of the noise modeling model training device 1100 of an embodiment.

[0201] Specifically, the transceiver 1301 may be Fig.11 The noise modeling model training device 1100 of the embodiment of the transceiver unit 1101, the processor 1302 can be Fig.11 The processing unit 1102 of the noise modeling model training device 1100 of the embodiment. Therefore, the specific functions of the processor 1302 can refer to the specific functions of the processing unit 1102, and the specific functions of the transceiver 1301 can refer to the specific functions of the transceiver unit 1101.

[0202] Optionally, the processor 1302 is configured to read a computer program in the memory 1303 to perform the following operations:

[0203] Control the transceiver 1301 to collect data on the power line channel to obtain initial data;

[0204] Structuring the initial data to obtain initial data features;

[0205] The initial features are input into a noise modeling model to obtain target noise characteristics corresponding to the initial data, wherein the noise modeling model is trained by a noise modeling model training device.

[0206] Specifically, the transceiver 1301 may be Fig.12 The transceiver unit 1201 of the noise modeling device 1200 of the embodiment, the processor 1302 may be Fig.12 The processing unit 1202 of the noise modeling device 1200 of an embodiment.

[0207] Specifically, the transceiver 1301 may be Fig.12 The transceiver unit 1201 of the noise modeling device 1200 of the embodiment, the processor 1302 may be Fig.12Therefore, the specific functions of the processor 1302 can refer to the specific functions of the processing unit 1202, and the specific functions of the transceiver 1301 can refer to the specific functions of the transceiver 1201.

[0208] It should be understood that the electronic devices in this application may include smart phones (such as Android phones, iOS phones, Windows Phone phones, etc.), tablet computers, PDAs, laptops, mobile Internet devices MID (Mobile Internet Devices, referred to as: MID) or wearable devices, etc. The above electronic devices are only examples, not exhaustive, and include but are not limited to the above electronic devices. In practical applications, the above electronic devices may also include: smart vehicle terminals, computer equipment, etc.

[0209] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. The computer program is executed by a processor to implement part or all of the steps of any noise modeling model training method recorded in the above method embodiments.

[0210] An embodiment of the present application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute part or all of the steps of any noise modeling model training method recorded in the above method embodiments.

[0211] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action splicing, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by this application.

[0212] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0213] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0214] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0215] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of a software program module.

[0216] If the integrated unit is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of each embodiment method of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0217] A person 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 instructing related hardware through a program, and the program can be stored in a computer-readable memory, and the memory can include: a flash drive, a read-only memory (English: Read-Only Memory, abbreviated as: ROM), a random access memory (English: Random Access Memory, abbreviated as: RAM), a magnetic disk or an optical disk, etc.

[0218] The embodiments of the present application are introduced in detail above. 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 of the present application and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A noise modeling model training method, characterized in that: The method comprises: Collecting data on the power line channel to construct training samples, wherein the training samples include original data samples and noise characteristic samples; Structuring the original data sample to obtain sample data features corresponding to the original data sample; Decoding the sample data feature based on each first parameter vector and the initial model to obtain a predicted noise characteristic corresponding to each first parameter vector, wherein each first parameter vector is randomly generated; Determine the degree of competition corresponding to each first parameter vector based on the predicted noise characteristic corresponding to each first parameter vector and the noise characteristic sample; Based on the competitiveness corresponding to each first parameter vector, determining a plurality of second parameter vectors from the plurality of first parameter vectors, wherein the plurality of second parameter vectors are first parameter vectors whose competitiveness is greater than a first preset threshold value among the plurality of first parameter vectors; Performing transformation processing on the plurality of second parameter vectors to obtain a plurality of third parameter vectors; Decoding the sample data features based on each third parameter vector and the initial model to obtain predicted noise characteristics corresponding to each third parameter vector; Determine a first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample; The initial model is trained based on the first target parameter vector to obtain a noise modeling model.

2. The method according to claim 1, characterized in that The determining, based on the predicted noise characteristic corresponding to each first parameter vector and the noise characteristic sample, the degree of competition corresponding to each first parameter vector comprises: Obtaining a first matching degree between the predicted noise characteristic corresponding to each first parameter vector and the sample data characteristic; Obtaining a second matching degree between the predicted noise characteristics corresponding to each first parameter vector; Based on the second matching degree and the second preset threshold, obtaining a first similarity between the predicted noise characteristic corresponding to each first parameter vector and the predicted noise characteristics corresponding to other first parameter vectors; Based on the first similarity, determining the frequency of the predicted noise characteristic corresponding to each first parameter vector among the predicted noise characteristics corresponding to the plurality of first parameter vectors; Based on the first matching degree and the frequency, a competitiveness corresponding to each first parameter vector is determined.

3. The method according to claim 1, characterized in that The transforming process is performed on the plurality of second parameter vectors to obtain a plurality of third parameter vectors, including: Copying the multiple second parameter vectors to obtain multiple fourth parameter vectors; Cross-merging the multiple fourth parameter vectors to obtain multiple fifth parameter vectors; The plurality of fifth parameter vectors are mutated to obtain a plurality of third parameter vectors.

4. The method according to claim 3, characterized in that The step of determining a first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample comprises: Determine the degree of competition corresponding to each third parameter vector based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample; The third parameter vector with the greatest competitiveness among the multiple third parameter vectors is used as the first target parameter vector.

5. The method according to claim 2, characterized in that: The step of training the initial model based on the first target parameter vector to obtain a noise modeling model includes: Decoding the sample data feature based on the first target parameter vector to obtain a predicted noise characteristic corresponding to the first target parameter vector; Obtaining a loss between a predicted noise characteristic corresponding to a first target parameter vector and the noise characteristic sample; Model training is performed based on the loss to obtain the noise modeling model.

6. A noise modeling method, characterized in that: The method comprises: Collect data on the power line channel to obtain initial data; Structuring the initial data to obtain initial data features; The initial features are input into a noise modeling model to obtain target noise characteristics corresponding to the initial data, wherein the noise modeling model is trained by any one of the methods of claims 1-5.

7. The method according to claim 6, characterized in that The method further comprises: Performing spectrum analysis on the target noise characteristics to obtain frequency components of the noise data in the initial data; Based on the frequency component, the initial data is filtered to obtain actual transmission data in the initial data.

8. A noise modeling model training device, characterized in that: The device includes: Transceiver unit and processing unit; The transceiver unit is used to collect data on the power line channel and construct training samples, wherein the training samples include original data samples and noise characteristic samples; The processing unit is used to structure the original data sample to obtain sample data features corresponding to the original data sample; The processing unit is used to decode the sample data feature based on each first parameter vector and the initial model to obtain the predicted noise characteristic corresponding to each first parameter vector, wherein each first parameter vector is randomly generated; The processing unit is used to determine the competitiveness corresponding to each first parameter vector based on the predicted noise characteristic corresponding to each first parameter vector and the noise characteristic sample; The processing unit is configured to determine a plurality of second parameter vectors from the plurality of first parameter vectors based on the competitiveness corresponding to each first parameter vector, wherein the plurality of second parameter vectors are first parameter vectors whose competitiveness is greater than a first preset threshold value among the plurality of first parameter vectors; The processing unit is used to perform transformation processing on the multiple second parameter vectors to obtain multiple third parameter vectors; The processing unit is used to decode the sample data feature based on each third parameter vector and the initial model to obtain the predicted noise characteristic corresponding to each third parameter vector; The processing unit is configured to determine a first target parameter vector from a plurality of third parameter vectors based on the predicted noise characteristic corresponding to each third parameter vector and the noise characteristic sample; The processing unit is used to train the initial model based on the first target parameter vector to obtain a noise modeling model.

9. An electronic device, characterized in that: include: A processor and a memory, the processor is connected to the memory, the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device executes the method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method according to any one of claims 1 to 7.

Images