CN114598403A - Data link broadband noise electromagnetic signal interference prediction method and system - Google Patents
Data link broadband noise electromagnetic signal interference prediction method and system Download PDFInfo
- Publication number
- CN114598403A CN114598403A CN202210336234.8A CN202210336234A CN114598403A CN 114598403 A CN114598403 A CN 114598403A CN 202210336234 A CN202210336234 A CN 202210336234A CN 114598403 A CN114598403 A CN 114598403A
- Authority
- CN
- China
- Prior art keywords
- prediction
- dual
- data link
- model
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000012549 training Methods 0.000 claims abstract description 49
- 238000001228 spectrum Methods 0.000 claims abstract description 31
- 238000012360 testing method Methods 0.000 claims abstract description 29
- 238000000605 extraction Methods 0.000 claims abstract description 23
- 230000004927 fusion Effects 0.000 claims abstract description 15
- 238000004891 communication Methods 0.000 claims description 22
- 238000012545 processing Methods 0.000 claims description 16
- 238000010606 normalization Methods 0.000 claims description 15
- 238000011156 evaluation Methods 0.000 claims description 14
- 238000006243 chemical reaction Methods 0.000 claims description 12
- 238000005457 optimization Methods 0.000 claims description 10
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 11
- 238000013527 convolutional neural network Methods 0.000 description 44
- 230000006870 function Effects 0.000 description 14
- 238000002347 injection Methods 0.000 description 3
- 239000007924 injection Substances 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 230000003930 cognitive ability Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 239000000243 solution Substances 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/309—Measuring or estimating channel quality parameters
- H04B17/336—Signal-to-interference ratio [SIR] or carrier-to-interference ratio [CIR]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/391—Modelling the propagation channel
- H04B17/3913—Predictive models, e.g. based on neural network models
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Electromagnetism (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Monitoring And Testing Of Transmission In General (AREA)
Abstract
The invention relates to a method and a system for predicting data link broadband noise electromagnetic signal interference, which relate to the field of electromagnetic interference effect, and the method comprises the following steps: acquiring a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram; constructing a sample set according to the broadband electromagnetic interference time spectrum image and the data link performance parameter histogram; dividing the sample set into a training set and a testing set; training the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model; the dual-channel CNN model comprises a feature extraction layer, a feature fusion layer and a prediction regression layer which are sequentially connected; and predicting the dual-channel CNN prediction model by using the test set to obtain a data chain performance prediction result. The invention realizes the prediction of the interference degree of the data chain by using a dual-channel CNN model.
Description
Technical Field
The invention relates to the field of electromagnetic interference effect, in particular to a method and a system for predicting data link broadband noise electromagnetic signal interference.
Background
The unmanned aerial vehicle as emerging air combat force develops a new modern war mode by the advantages of multiple loads, high efficiency and low cost. The data link is a communication junction for information interaction between the unmanned aerial vehicle and the ground, and is also a neuron which is most sensitive to electromagnetic interference. Because various electromagnetic signals are densely staggered in a complex battlefield electromagnetic environment, the data chain of the unmanned aerial vehicle is threatened by multiple parties, and the communication performance of the data chain is seriously influenced, so that the overall combat efficiency of the unmanned aerial vehicle is reduced.
The broadband noise signal is a typical interference signal pattern in a complex electromagnetic environment of a battlefield, and in the battlefield environment, when a large number of electronic devices work simultaneously and space electromagnetic signals are combined in a random staggered mode, the broadband noise signal can be regarded as low-power unintentional noise interference to an unmanned aerial vehicle data chain; in addition, enemy electromagnetic jammers can also emit high-power noise electromagnetic interference to the communication frequency band of my party data chain in a targeted manner. The intentional or unintentional noise interference can cause the signal-to-noise ratio of the data chain to be reduced and the bit error rate to be increased to different degrees, and the transmission of the control instruction of the unmanned aerial vehicle by the party is influenced.
At present, the machine learning method also draws a lot of attention and applications in the fields of electromagnetic compatibility and communication. The convolutional neural network can directly input an original image, avoids complex preprocessing on the image, and has certain deep learning capability, so that the convolutional neural network is widely applied. There is no method of applying machine learning to electromagnetic interference prediction.
Disclosure of Invention
The invention aims to provide a method and a system for predicting data link broadband noise electromagnetic signal interference, which realize prediction of the interference degree of a data link by utilizing a dual-channel CNN model.
In order to achieve the purpose, the invention provides the following scheme:
a data link broadband noise electromagnetic signal interference prediction method comprises the following steps:
acquiring a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram;
constructing a sample set according to the broadband electromagnetic interference time spectrum image and the data link performance parameter histogram;
dividing the sample set into a training set and a testing set;
training the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model; the dual-channel CNN model comprises a feature extraction layer, a feature fusion layer and a prediction regression layer which are sequentially connected;
and predicting the dual-channel CNN prediction model by using the test set to obtain a data chain performance prediction result.
Optionally, the acquiring a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram specifically includes:
acquiring IQ information and data link performance parameters of an electromagnetic signal on a communication link;
performing STFT conversion on the IQ signal to obtain a broadband electromagnetic interference time spectrum image;
and carrying out normalization processing on the data chain performance parameters and drawing a data chain performance parameter histogram according to the data chain performance parameters after the normalization processing.
Optionally, the training of the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model specifically includes:
inputting the broadband electromagnetic interference time-frequency spectrum image of the training set into a first feature extraction module of the feature extraction layer to obtain a first local feature;
inputting the data chain performance parameter histogram of the training set into a second feature extraction module of the feature extraction layer to obtain a second local feature;
inputting the first local feature and the second local feature into the feature fusion layer to obtain a feature vector;
inputting the feature vector into the prediction regression layer to obtain a data link performance prediction value;
determining a loss function according to the data link performance prediction value and a target value in the training set;
optimizing the dual-channel CNN model by using SGDM, RMSProp and Adam optimization methods according to the loss function respectively to obtain an evaluation index and a plurality of optimization results; the evaluation index comprises a root mean square error and an accuracy;
and determining a dual-channel CNN prediction model according to the evaluation index and the plurality of optimization results.
Optionally, the expression of the loss function is:
where loss is the loss function, tiIs the target value, y, under the i-th class of prediction resultsiK is the batch number for the model prediction value.
Optionally, the root mean square error is expressed as:
where RMSE is the root mean square error, yiAs model predicted value, tiFor target values under class i prediction results, NtrainIs the total number of samples of the channel.
Optionally, the expression of the accuracy is:
where Acc (y, t) is accuracy, yiAs model predicted value, tiFor target values under class i prediction results, NtrainIs the total number of samples of the channel, sign (y)i,ti) Is a symbolic function.
A data link broadband noise electromagnetic signal interference prediction system, comprising:
the acquisition module is used for acquiring a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram;
the construction module is used for constructing a sample set according to the broadband electromagnetic interference time spectrum image and the data link performance parameter histogram;
the splitting module is used for dividing the sample set into a training set and a testing set;
the training module is used for training the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model; the dual-channel CNN model comprises a feature extraction layer, a feature fusion layer and a prediction regression layer which are sequentially connected;
and the prediction module is used for predicting the dual-channel CNN prediction model by using the test set to obtain a data chain performance prediction result.
Optionally, the obtaining module specifically includes:
the acquisition unit is used for acquiring IQ information and data link performance parameters of the electromagnetic signals on the communication link;
the STFT conversion unit is used for carrying out STFT conversion on the IQ signals to obtain a broadband electromagnetic interference time spectrum image;
and the normalization processing and drawing unit is used for performing normalization processing on the data chain performance parameters and drawing a data chain performance parameter histogram according to the data chain performance parameters after the normalization processing.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the method comprises the steps of obtaining a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram; constructing a sample set according to the broadband electromagnetic interference time spectrum image and the data link performance parameter histogram; dividing the sample set into a training set and a testing set; training the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model; the dual-channel CNN model comprises a feature extraction layer, a feature fusion layer and a prediction regression layer which are sequentially connected; and predicting the dual-channel CNN prediction model by using the test set to obtain a data chain performance prediction result. The method and the device can accurately predict the interference effect level of the data link under the broadband noise interference, so that the data link can evaluate the threat of the broadband noise interference.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a flow chart of a data link wideband noise electromagnetic signal interference prediction method provided by the present invention;
fig. 2 is a graph of interference power versus performance parameter variation provided by the present invention;
FIG. 3 is a schematic diagram of an injection data chain electromagnetic interference acquisition system;
FIG. 4 is a normalized histogram of data chain performance parameters;
fig. 5 is a structural diagram of a two-channel CNN prediction model.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a method and a system for predicting data link broadband noise electromagnetic signal interference, which realize prediction of the interference degree of a data link by utilizing a dual-channel CNN model.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
According to the method, a dual-channel CNN network is established, a broadband electromagnetic interference time-frequency spectrogram and a data link state parameter histogram are used as model input, a set data link communication performance boundary is used as output, the communication performance of the unmanned aerial vehicle data link under different broadband noise interferences is predicted, and the problem of interference effect evaluation of the data link under the broadband noise interferences is solved.
As shown in fig. 1, a method for predicting data link broadband noise electromagnetic signal interference provided by the present invention includes:
as shown in fig. 2, when the interference power is within 1dB from the out-of-lock power, the signal-to-noise ratio of the data link is sharply reduced, the bit error rate is exponentially increased, and the communication performance may deteriorate to the communication interruption at any time; when the distance between the interference power and the unlocking power is 1-3 dB, the data link still can work normally, but the signal-to-noise ratio is reduced compared with the normal working state, the error code rises quickly along with the increase of the interference power, and the communication quality is reduced; when the interference power distance is unlocked to 3-6 dB, the data chain can work normally, but a lower code error number exists; when the interference power distance is out-of-lock greater than 6dB, the data chain works normally, and the error code number is 0.
A drone communication performance boundary is specified. And dividing the electromagnetic interference level of the data link according to the power value of the distance loss lock by taking the loss lock as a sign of data link communication interruption according to the data link communication mechanism and the electromagnetic compatibility sensitivity test experience. The interference levels are respectively: interrupting: losing the lock; disturbance: the interference power distance is unlocked by 1 dB; interference: the interference power distance is unlocked by 3 dB; early warning: the interference power distance is 6dB out of lock.
As shown in fig. 3, an injection type unmanned aerial vehicle data link electromagnetic interference data acquisition system is adopted to obtain a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram sample set, the sample set is preprocessed, and the preprocessed sample set is processed according to the following steps of: the ratio of 3 is divided into a training set and a test set. The airborne data chain and the ground data chain are connected through a radio frequency cable, and two 40dB fixed attenuators and an adjustable attenuator are connected in series in the middle: the fixed attenuator is used for attenuating a larger working signal output by the radio frequency port, simulating a long-distance transmission condition and protecting the radio frequency front end from being broken down by a high-power signal; the adjustable attenuator is used for adjusting the working signal and simulating the working signal of the UAV at different flight distances. The interference signal is generated by the signal generator and injected into the communication link through the combiner, and electromagnetic interference to the unmanned aerial vehicle data link in the space is simulated. The ground end data chain is connected with a computer through a network cable, and state information of the data chain is read in real time through state monitoring software, wherein the state information comprises qualitative information and three quantitative data: the qualitative information is whether the data chain is unlocked or not; the quantitative information is: automatic gain control voltage (AGC) indication, Bit Error Rate (BER), signal to noise ratio (SNR). The electromagnetic space signal is measured by a monitoring receiver, and the receiver is connected to a receiving end of the airborne data chain antenna and records IQ data of the analog space electromagnetic signal.
Step 101: and acquiring a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram. Step 101, specifically comprising:
IQ information and data link performance parameters of an electromagnetic signal on a communication link are obtained.
And carrying out STFT conversion on the IQ signal to obtain a broadband electromagnetic interference time spectrum image.
And carrying out normalization processing on the data chain performance parameters and drawing a data chain performance parameter histogram according to the data chain performance parameters after normalization processing.
IQ data and data link performance parameters of the electromagnetic signals on the communication link are obtained through the acquisition system. Due to the fact that IQ data and data chain performance parameter structures are different, fusion modeling is not convenient to conduct directly, and therefore visualization processing is conducted on the two kinds of data. STFT conversion is carried out on time domain signals formed by IQ data, then the square of the modulus value is taken, and then combination is carried out according to the time sequence to obtain a time-frequency spectrogram of the electromagnetic signals; the data chain performance parameters AGC, SNR and BER are normalized and a histogram is plotted as shown in fig. 4. The time-frequency spectrogram and the data link performance parameter histogram change in real time along with the interference.
Step 102: and constructing a sample set according to the broadband electromagnetic interference time spectrum image and the data link performance parameter histogram.
Acquiring a broadband electromagnetic interference time-frequency spectrum image and a data chain performance parameter histogram sample set, carrying out visualization and pretreatment on the sample set, and dividing the pretreated sample set into a training set and a testing set. The electromagnetic interference time spectrum image is obtained by performing STFT conversion on IQ data and then overlapping the IQ data in a time domain, and the data chain performance histogram is drawn by normalized data chain performance parameters in real time.
Step 103: the sample set is divided into a training set and a test set.
And obtaining a model sample through visual conversion of the data, and then preprocessing a sample set. 440 data chain performance parameter samples and 440 electromagnetic interference spectrogram samples are obtained through testing, images of the samples are divided by 255, and data are normalized to be between (0, 1), so that interference caused by difference of value ranges of all dimensional data is reduced. The samples are marked according to the specified communication performance boundary and interference level of the unmanned aerial vehicle, and 440 samples are divided into four groups of {0, 1, 3 and 6}, wherein each group comprises 110 samples. Samples for each channel were run as 7: the ratio of 3 is divided into a training set and a test set.
Step 104: training the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model; the dual-channel CNN model comprises a feature extraction layer, a feature fusion layer and a prediction regression layer which are sequentially connected.
As shown in fig. 5, a two-channel CNN model is constructed for predicting the interference degree of the data chain. The double-channel CNN model consists of a feature extraction layer, a feature fusion layer and a prediction regression layer. The characteristic extraction layer is composed of a convolution layer, a pooling layer and a full-connection layer, and is two parallel branches which are respectively used for extracting a data chain performance parameter sample and local characteristics of an electromagnetic interference spectrogram; the feature fusion layer consists of an addition layer and a full connection layer and is used for fusing the image features to obtain a feature vector corresponding to each image; and the prediction regression layer calculates the semimean square error loss of the regression task and obtains a prediction result.
The characteristic extraction layer is composed of a convolution layer, a pooling layer and a full-connection layer, and is composed of two parallel CNN network branches which are respectively used for extracting data chain performance parameter samples and local characteristics of an electromagnetic interference spectrogram. The specific process of extracting the features is as follows: the feature mapping of the input image is convolved with the convolution kernel, the local features of the input image sample set are extracted, and then the pooling layer is connected and used for reducing the dimension of the image features, so that the overfitting of the model is reduced to a certain extent. The convergence speed of the network is accelerated by a modified linear unit (RELU) activation function, and the image features are further extracted by the combination of a convolution layer and a full connection layer.
The feature fusion layer is composed of an addition layer and a full connection layer and used for fusing the image features to obtain a feature vector corresponding to each image, and for the regression problem, the feature number of the full connection layer is 1.
The prediction regression layer calculates the half mean square error loss of the regression task. Through the back propagation of the neural network, each neuron is continuously trained to update the network weight and the deviant, so that the error gradient is reduced, and the model is continuously optimized.
inputting the broadband electromagnetic interference time-frequency spectrum image of the training set into a first feature extraction module of the feature extraction layer to obtain a first local feature;
inputting the data chain performance parameter histogram of the training set into a second feature extraction module of the feature extraction layer to obtain a second local feature;
inputting the first local feature and the second local feature into the feature fusion layer to obtain a feature vector;
inputting the feature vector into the prediction regression layer to obtain a data link performance prediction value;
determining a loss function according to the data link performance prediction value and a target value in the training set; wherein the expression of the loss function is:
where loss is the loss function, tiIs the target value, y, under the i-th class of prediction resultsiK is the batch number for the model prediction value.
The loss function loss is used for measuring the degree of inconsistency between the predicted value and the target value, small batches are set for accelerating the iteration speed, the batch number is assumed to be k, and the half-mean square error in each small batch observation is taken as the loss function.
Optimizing the dual-channel CNN model by using SGDM, RMSProp and Adam optimization methods according to the loss function respectively to obtain an evaluation index and a plurality of optimization results; the evaluation index comprises a root mean square error and an accuracy; wherein the root mean square error is expressed as:
where RMSE is the root mean square error, yiAs model predicted value, tiFor target values under class i prediction results, NtrainIs the total number of samples of the channel.
And optimizing the value of a loss function in the training set by using an SGDM (generalized mean square deviation), RMSProp and Adam optimization method, and then evaluating the prediction accuracy through index parameters. The evaluation indexes are root mean square error RMSE and accuracy.
Accuracy is defined as the ratio of the number of correctly predicted samples to the total number of samples. The samples in the model are evenly distributed, so the prediction accuracy index can be used for evaluation. In the present model, it is defined that the prediction result is correct when the absolute value of the difference between the prediction result and the true value is less than 1. Suppose yiAs model predicted value, tiIs a target value, NtrainIs the total number of samples of the channel.
The accuracy expression is:
where Acc (y, t) is accuracy, yiAs model predicted value, tiFor target values under class i prediction results, NtrainSign (y) is the total number of samples of a channeli,ti) Is a function of the sign.
RMSE was used to evaluate the predicted effect of the final model. The smaller the RMSE, the better the prediction of the model.
The predicted structure of the two-channel CNN model optimized using SGDM, RMSProp and Adam is shown in table 1. It can be shown that the prediction accuracy of the dual-channel convolution model using the SGDM solver is the highest, the RMSE is the lowest, and then the RMSProp optimized model, the Adam optimized model result is relatively poor. The accuracy of the two-channel CNN model optimized by the three solvers is over 90 percent.
TABLE 1 prediction of results of a test set by a network
| Solver | Rate of accuracy | RMSE |
| SGDM | 0.9697 | 0.5189 |
| Adam | 0.9091 | 0.7542 |
| RMSProp | 0.9621 | 0.5214 |
And determining a dual-channel CNN prediction model according to the evaluation index and the plurality of optimization results.
Step 105: and predicting the dual-channel CNN prediction model by using the test set to obtain a data chain performance prediction result.
The method specifies the electromagnetic interference effect grade of the data link, and obtains a visual sample set by obtaining a broadband noise electromagnetic interference time-frequency spectrogram and a data link performance parameter normalized histogram; preprocessing the sample set, dividing the preprocessed sample set into a training set and a test set, then constructing a dual-channel CNN prediction model, inputting the training set and the test set into the model for training, and obtaining a trained CNN prediction model; and finally, inputting the test set into the dual-channel CNN prediction model, and predicting the input image of the test set to obtain a prediction result of the performance boundary grade of the data chain. The method can accurately predict the interference effect grade of the data link under the broadband noise interference, so that the data link can evaluate the threat of the broadband noise interference, certain anti-interference measures can be taken in time, and the intelligent interference cognitive ability of the data link is improved.
The invention also provides a data link broadband noise electromagnetic signal interference prediction system, which comprises:
and the acquisition module is used for acquiring the broadband electromagnetic interference time-frequency spectrum image and the data link performance parameter histogram. The obtaining module specifically includes: the acquisition unit is used for acquiring IQ information and data link performance parameters of the electromagnetic signals on the communication link; the STFT conversion unit is used for carrying out STFT conversion on the IQ signals to obtain a broadband electromagnetic interference time spectrum image; and the normalization processing and drawing unit is used for performing normalization processing on the data chain performance parameters and drawing a data chain performance parameter histogram according to the data chain performance parameters after the normalization processing.
And the construction module is used for constructing a sample set according to the broadband electromagnetic interference time spectrum image and the data link performance parameter histogram.
And the splitting module is used for dividing the sample set into a training set and a testing set.
The training module is used for training the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model; the dual-channel CNN model comprises a feature extraction layer, a feature fusion layer and a prediction regression layer which are sequentially connected.
And the prediction module is used for predicting the dual-channel CNN prediction model by using the test set to obtain a data chain performance prediction result.
The invention is suitable for the technical field of electromagnetic interference effect evaluation, and provides an unmanned aerial vehicle data link broadband noise electromagnetic interference effect evaluation method and device. And (3) a specified unmanned aerial vehicle communication performance boundary is adopted, an injection type unmanned aerial vehicle data chain electromagnetic interference data acquisition system is adopted, a broadband electromagnetic interference time-frequency spectrum image and a data chain performance parameter histogram sample set are obtained, and the preprocessed sample set is divided into a training set and a testing set. And then constructing a dual-channel CNN regression prediction model, inputting the training set into the dual-channel CNN prediction model to obtain a trained dual-channel CNN prediction model, finally inputting the test set into the trained dual-channel CNN prediction model, and performing regression prediction on an input image of the test set to obtain a prediction result of the performance boundary grade of the data chain. The method and the device can accurately predict the interference effect grade of the data chain under the broadband noise interference, so that the data chain can evaluate the threat of the broadband noise interference, certain anti-interference measures can be taken in time, and the intelligent interference cognitive ability of the data chain is improved.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the description of the method part.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (8)
1. A method for predicting data link broadband noise electromagnetic signal interference is characterized by comprising the following steps:
acquiring a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram;
constructing a sample set according to the broadband electromagnetic interference time spectrum image and the data link performance parameter histogram;
dividing the sample set into a training set and a test set;
training the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model; the dual-channel CNN model comprises a feature extraction layer, a feature fusion layer and a prediction regression layer which are sequentially connected;
and predicting the dual-channel CNN prediction model by using the test set to obtain a data chain performance prediction result.
2. The method for predicting interference of a broadband noise electromagnetic signal of a data link according to claim 1, wherein the obtaining of a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram specifically includes:
acquiring IQ information and data link performance parameters of an electromagnetic signal on a communication link;
performing STFT conversion on the IQ signal to obtain a broadband electromagnetic interference time spectrum image;
and carrying out normalization processing on the data chain performance parameters and drawing a data chain performance parameter histogram according to the data chain performance parameters after the normalization processing.
3. The method for predicting data chain broadband noise electromagnetic signal interference according to claim 1, wherein the training of the two-channel CNN model by using the training set to obtain the two-channel CNN prediction model specifically comprises:
inputting the broadband electromagnetic interference time-frequency spectrum image of the training set into a first feature extraction module of the feature extraction layer to obtain a first local feature;
inputting the data chain performance parameter histogram of the training set into a second feature extraction module of the feature extraction layer to obtain a second local feature;
inputting the first local feature and the second local feature into the feature fusion layer to obtain a feature vector;
inputting the feature vector into the prediction regression layer to obtain a data link performance prediction value;
determining a loss function according to the data link performance prediction value and a target value in the training set;
optimizing the dual-channel CNN model by using SGDM, RMSProp and Adam optimization methods according to the loss function respectively to obtain an evaluation index and a plurality of optimization results; the evaluation index comprises a root mean square error and an accuracy;
and determining a dual-channel CNN prediction model according to the evaluation index and the plurality of optimization results.
4. The method of claim 3, wherein the loss function is expressed as:
where loss is the loss function, tiIs the target value, y, under the i-th class of prediction resultsiK is the batch number for the model prediction value.
5. The method of claim 3, wherein the root mean square error is expressed as:
where RMSE is the root mean square error, yiAs model predicted value, tiFor target values under class i prediction results, NtrainIs a summary of the channelsThis number.
6. The method of claim 3, wherein the accuracy is expressed as:
where Acc (y, t) is accuracy, yiAs model predicted value, tiFor target values under class i prediction results, NtrainSign (y) is the total number of samples of a channeli,ti) Is a symbolic function.
7. A data link wideband noise electromagnetic signal interference prediction system, comprising:
the acquisition module is used for acquiring a broadband electromagnetic interference time-frequency spectrum image and a data link performance parameter histogram;
the construction module is used for constructing a sample set according to the broadband electromagnetic interference time spectrum image and the data link performance parameter histogram;
the splitting module is used for dividing the sample set into a training set and a testing set;
the training module is used for training the dual-channel CNN model by using the training set to obtain a dual-channel CNN prediction model; the dual-channel CNN model comprises a feature extraction layer, a feature fusion layer and a prediction regression layer which are sequentially connected;
and the prediction module is used for predicting the dual-channel CNN prediction model by using the test set to obtain a data chain performance prediction result.
8. The system according to claim 7, wherein the obtaining module specifically includes:
the acquisition unit is used for acquiring IQ information and data link performance parameters of the electromagnetic signals on the communication link;
the STFT conversion unit is used for carrying out STFT conversion on the IQ signals to obtain a broadband electromagnetic interference time spectrum image;
and the normalization processing and drawing unit is used for performing normalization processing on the data chain performance parameters and drawing a data chain performance parameter histogram according to the data chain performance parameters after the normalization processing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210336234.8A CN114598403B (en) | 2022-03-31 | 2022-03-31 | Data link broadband noise electromagnetic signal interference prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210336234.8A CN114598403B (en) | 2022-03-31 | 2022-03-31 | Data link broadband noise electromagnetic signal interference prediction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114598403A true CN114598403A (en) | 2022-06-07 |
| CN114598403B CN114598403B (en) | 2024-03-12 |
Family
ID=81811545
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210336234.8A Active CN114598403B (en) | 2022-03-31 | 2022-03-31 | Data link broadband noise electromagnetic signal interference prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114598403B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2019122008A (en) * | 2018-01-11 | 2019-07-22 | 株式会社Ｎｔｔドコモ | Device, method and program for estimating radio wave propagation |
| JP2019122039A (en) * | 2018-01-10 | 2019-07-22 | 富士通株式会社 | Interference source prediction model generation method, interference source recognition method, device, and terminal device |
| CN110996343A (en) * | 2019-12-18 | 2020-04-10 | 中国人民解放军陆军工程大学 | Interference recognition model based on deep convolutional neural network and intelligent recognition algorithm |
| CN111404849A (en) * | 2020-03-20 | 2020-07-10 | 北京航空航天大学 | OFDM channel estimation and signal detection method based on deep learning |
| CN113271162A (en) * | 2021-05-12 | 2021-08-17 | 中国人民解放军陆军工程大学 | Method and system for predicting boundary response of electromagnetic interference performance in data link of unmanned aerial vehicle |
| CN113542180A (en) * | 2021-06-30 | 2021-10-22 | 北京频谱视觉科技有限公司 | Frequency domain identification method of radio signal |
| CN113569742A (en) * | 2021-07-29 | 2021-10-29 | 西南交通大学 | Broadband electromagnetic interference source identification method based on convolutional neural network |
| WO2021230586A1 (en) * | 2020-05-11 | 2021-11-18 | Samsung Electronics Co., Ltd. | Method and system for beamforming for at least one transmitting antenna in a radio-communication networking environment |
| US20210374553A1 (en) * | 2020-06-02 | 2021-12-02 | Salesforce.Com, Inc. | Systems and methods for noise-robust contrastive learning |
| CN114154545A (en) * | 2021-12-07 | 2022-03-08 | 中国人民解放军32802部队 | Intelligent unmanned aerial vehicle measurement and control signal identification method under strong mutual interference condition |
-
2022
- 2022-03-31 CN CN202210336234.8A patent/CN114598403B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2019122039A (en) * | 2018-01-10 | 2019-07-22 | 富士通株式会社 | Interference source prediction model generation method, interference source recognition method, device, and terminal device |
| JP2019122008A (en) * | 2018-01-11 | 2019-07-22 | 株式会社Ｎｔｔドコモ | Device, method and program for estimating radio wave propagation |
| CN110996343A (en) * | 2019-12-18 | 2020-04-10 | 中国人民解放军陆军工程大学 | Interference recognition model based on deep convolutional neural network and intelligent recognition algorithm |
| CN111404849A (en) * | 2020-03-20 | 2020-07-10 | 北京航空航天大学 | OFDM channel estimation and signal detection method based on deep learning |
| WO2021230586A1 (en) * | 2020-05-11 | 2021-11-18 | Samsung Electronics Co., Ltd. | Method and system for beamforming for at least one transmitting antenna in a radio-communication networking environment |
| US20210374553A1 (en) * | 2020-06-02 | 2021-12-02 | Salesforce.Com, Inc. | Systems and methods for noise-robust contrastive learning |
| CN113271162A (en) * | 2021-05-12 | 2021-08-17 | 中国人民解放军陆军工程大学 | Method and system for predicting boundary response of electromagnetic interference performance in data link of unmanned aerial vehicle |
| CN113542180A (en) * | 2021-06-30 | 2021-10-22 | 北京频谱视觉科技有限公司 | Frequency domain identification method of radio signal |
| CN113569742A (en) * | 2021-07-29 | 2021-10-29 | 西南交通大学 | Broadband electromagnetic interference source identification method based on convolutional neural network |
| CN114154545A (en) * | 2021-12-07 | 2022-03-08 | 中国人民解放军32802部队 | Intelligent unmanned aerial vehicle measurement and control signal identification method under strong mutual interference condition |
Non-Patent Citations (6)
| Title |
|---|
| DONGXIAO ZHANG: "GPR-Based EMI Prediction for UAV\'s Dynamic Datalink", 《IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY 》 * |
| DONGXIAO ZHANG: "Prediction of Electromagnetic Compatibility for Dynamic Datalink of UAV", 《IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY》 * |
| 刘瑜岚;吴坤荣;樊玉琦;温鹏飞;沈飞;: "基于深度学习的战场电磁态势研究综述", 东莞理工学院学报, no. 03 * |
| 廖晓龙;张志厚;姚禹;路润琪;范祥泰;曹云勇;冯涛;石泽玉;: "基于卷积神经网络的大地电磁反演", 中南大学学报(自然科学版), no. 09 * |
| 张薇玮;丁文锐;刘春辉;: "复杂环境中无人机数据链干扰效果预测方法", 系统工程与电子技术, no. 04 * |
| 牟效乾;陈小龙;关键;周伟;刘宁波;董云龙;: "基于INet的雷达图像杂波抑制和目标检测方法", 雷达学报, no. 04 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114598403B (en) | 2024-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110222748B (en) | OFDM radar signal identification method based on 1D-CNN multi-domain feature fusion | |
| CN108845306B (en) | Laser radar echo signal denoising method based on variational modal decomposition | |
| CN111913156B (en) | Radar radiation source individual identification method based on deep learning model and feature combination | |
| CN109471074B (en) | Radar radiation source identification method based on singular value decomposition and one-dimensional CNN network | |
| CN110213003B (en) | Wireless channel large-scale fading modeling method and device | |
| CN111126471A (en) | Microseism event detection method and system | |
| CN109766791B (en) | Communication signal modulation identification method based on self-encoder | |
| CN103941254A (en) | Soil physical property classification recognition method and device based on geological radar | |
| CN112949820B (en) | Cognitive anti-interference target detection method based on generation of countermeasure network | |
| CN112949387A (en) | Intelligent anti-interference target detection method based on transfer learning | |
| CN112364729A (en) | Modulation identification method based on characteristic parameters and BP neural network | |
| CN112949383B (en) | Waveform agility radar radiation source identification method based on HyDeep-Att network | |
| CN109061577A (en) | A kind of recognition methods of different types of covering type interference and Deceiving interference | |
| CN112036239A (en) | Radar signal working mode identification method and system based on deep learning network | |
| CN110929842A (en) | Accurate intelligent detection method for burst time region of non-cooperative radio signal | |
| CN113315593A (en) | Frequency spectrum sensing algorithm based on FLOM covariance matrix and LSTM neural network | |
| CN112560342A (en) | DNN-based atmospheric waveguide parameter estimation method | |
| CN114117912A (en) | Sea clutter modeling and inhibiting method under data model dual drive | |
| CN102867104B (en) | A kind of quantitative analysis device of electromagnetic environmental impact factor of electronic equipment and method | |
| CN112014801A (en) | Composite interference identification method based on SPWVD and improved AlexNet | |
| CN106971066A (en) | Method based on Neural Network Models To Prediction geomagnetic storm | |
| CN113466807B (en) | Radar critical intercepted distance test system | |
| CN114598403B (en) | Data link broadband noise electromagnetic signal interference prediction method and system | |
| CN112947080B (en) | Scene parameter transformation-based intelligent decision model performance evaluation system | |
| CN113962264A (en) | Aero-engine rotor system fault diagnosis algorithm based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |