CN117033922A - Electromagnetic environment influence analysis method and system based on artificial intelligence - Google Patents
Electromagnetic environment influence analysis method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN117033922A CN117033922A CN202310471414.1A CN202310471414A CN117033922A CN 117033922 A CN117033922 A CN 117033922A CN 202310471414 A CN202310471414 A CN 202310471414A CN 117033922 A CN117033922 A CN 117033922A
- Authority
- CN
- China
- Prior art keywords
- frequency
- layer
- electromagnetic
- frequency band
- signals
- 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.)
- Pending
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 79
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 18
- 230000036541 health Effects 0.000 claims abstract description 33
- 238000000034 method Methods 0.000 claims abstract description 30
- 238000004891 communication Methods 0.000 claims abstract description 23
- 210000002569 neuron Anatomy 0.000 claims description 52
- 230000008569 process Effects 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 17
- 230000004913 activation Effects 0.000 claims description 14
- 238000012806 monitoring device Methods 0.000 claims description 13
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 11
- 238000003062 neural network model Methods 0.000 claims description 9
- 230000006978 adaptation Effects 0.000 claims description 4
- 230000007613 environmental effect Effects 0.000 claims description 4
- 238000012544 monitoring process Methods 0.000 description 8
- 238000001228 spectrum Methods 0.000 description 4
- 238000013480 data collection Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000000977 initiatory effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005553 drilling Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/001—Measuring interference from external sources to, or emission from, the device under test, e.g. EMC, EMI, EMP or ESD testing
-
- 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
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Electromagnetism (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an electromagnetic environment influence analysis method and system based on artificial intelligence, wherein the system comprises an equipment management module, an intelligent analysis module and a result display module; the method comprises the following steps: 1) Acquiring an electromagnetic signal; 2) Identifying a current signal of the frequency-consuming device; 3) Judging whether the signals are dense signals or not, if not, recording the frequency band of the group of effective signals; 4) Acquiring an available frequency band; 5) Combining the continuous available frequency bands into a new frequency band; 6) Calculating health values of all frequency bands; 7) And finally, taking the frequency band with the maximum health value as the recommended frequency band. The invention can rapidly analyze the signal condition in the electromagnetic environment, and intelligently propose the frequency using suggestion by combining the data such as the frequency using plan, the frequency using threat, the frequency using rule and the like, thereby rapidly solving the problem and recovering the communication.
Description
Technical Field
The invention relates to the technical field of electromagnetic environment analysis, in particular to an electromagnetic environment influence analysis method and system based on artificial intelligence.
Background
The spectrum plays a very critical role in various industries, especially in some special applications, where the fields of drilling, navigation, railways, transportation, etc. require the use of special communication devices for communication. Without the proper frequency band, the call and data download are never talking. In the actual production and use process, the communication environment is very complex in practice, because in a certain regional space, the communication environment is formed by overlapping a plurality of electromagnetic signals which are densely distributed in space domain, time domain, frequency domain and energy, are numerous, have complex patterns and are dynamically alternated, so that the normal operation of an information system and electronic equipment is seriously hindered, and the field use and the performance exertion of the professional frequency electronic communication equipment at the far end side are obviously influenced.
In general, the frequency-using device is affected by multiple factors such as frequency spectrum parameter drift, frequency misuse, electromagnetic interference and the like in a complex electromagnetic environment, so that frequency-using problems such as unavailable frequency-using plan, poor standby frequency, insufficient frequency spectrum resources and the like are caused. The personnel using the communication equipment to communicate on site are not professionals in the communication field, do not have technical analysis of the communication frequency interference problem, but are professionals, and can spend a great deal of time on finding the problem reason, so that great challenges and low working efficiency are brought to on-site work.
Therefore, how to help the frequency personnel to make the judgment of the frequency situation in time and react quickly, ensure the orderly and effective expansion of various frequency activities, scientifically decide the service frequency department and dynamically adjust the frequency by the frequency spectrum management and control department, promote the efficient exertion of the combat efficiency of the frequency weapon equipment, and become the technical problem to be solved by the personnel in the field.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide an electromagnetic environment influence analysis method and system based on artificial intelligence, which can rapidly perform AI analysis on signal conditions in an electromagnetic environment when the frequency is influenced by the environment, and intelligently propose frequency using suggestions by combining data such as a frequency using plan, frequency using threat, frequency using rules and the like, so that the problem can be rapidly solved, and communication is recovered.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows: an electromagnetic environment influence analysis method based on artificial intelligence is characterized in that: the method specifically comprises the following steps:
1) Acquiring a current signal of the frequency-taking device and an electromagnetic signal existing in the environment;
2) Identifying the current signal of the frequency-using equipment in the acquired electromagnetic environment;
3) Dividing the electromagnetic signals after the current signals are identified into a plurality of groups according to the frequency order, sequentially inputting the frequency of each group of electromagnetic signals into a trained concentration analysis model, judging whether the group of signals are dense signals, and if so, repeating the step; if not, recording the frequency band of the group of effective signals, and then repeating the step; until the judgment of a plurality of groups of electromagnetic signals is completed;
4) Comparing the frequency band recorded in the step 3) with a preset frequency plan, a frequency threat and a frequency rule in sequence to obtain a frequency band which meets the frequency plan and the frequency rule and is beyond the frequency threat, and recording the frequency band as an available frequency band; repeating the step until all the frequency bands recorded in the step 3) are judged;
5) Arranging the available frequency bands obtained in the step 4) according to the frequency order, and combining the continuous available frequency bands into a new frequency band;
6) Counting the number of electromagnetic signals of each frequency band processed in the step 5), calculating a signal intensity average value, calculating a health value of each frequency band, and arranging each frequency band in descending order according to the health value;
7) And finally, taking the frequency band with the maximum health value as the recommended frequency band.
Further, in step 6), the health value calculation model of each frequency band is: frequency band health value=1/(signal intensity average value noise threshold value electromagnetic signal number).
Preferably, in step 3), the concentration analysis model is a neural network model, which includes a three-layer neural network, wherein: the first layer is an input layer with k neurons, the second layer is an intermediate layer with p neurons, and the third layer is an input layer or an output layer with 1 neuron;
the activation function activation between network layers adopts a linear rectification function relu, and finally, the cross entropy of an output tensor and a target tensor is calculated to be used as the loss of a model, and rmsprop is used as an optimizer of the model;
in the training process, sample data comprising noise is input to neurons of a first layer, and corresponding concentration values 1 or 0 are input to neurons of a third layer, wherein the concentration values are concentration values of effective signals after noise is identified, the concentration values 1 are dense, and the concentration values 0 are non-dense;
in the working process, after a group of electromagnetic signals are input to neurons of the first layer, the second layer is processed according to the trained operation model, noise and effective signals can be automatically identified, and the third layer outputs a concentration value of the effective signals.
Preferably, in step 3), a noise threshold is calculated according to the intensity of the current signal of the frequency-consuming device and the noise proportionality coefficient, and then the intensity of the electromagnetic signal in the environment is compared with the noise threshold to obtain an electromagnetic signal with the intensity greater than the noise threshold, and the electromagnetic signal is recorded as an effective signal; and then dividing the effective signals into a plurality of groups according to the frequency order.
Further, in step 3), the concentration analysis model is a neural network model, which includes a three-layer neural network, wherein: the first layer is an input layer with k neurons, the second layer is an intermediate layer with p neurons, and the third layer is an input layer or an output layer with 1 neuron;
the activation function activation between network layers adopts a linear rectification function relu, and finally, the cross entropy of an output tensor and a target tensor is calculated to be used as the loss of a model, and rmsprop is used as an optimizer of the model;
in the training process, after a group of effective signal sample data is input to neurons of a first layer, a corresponding density value of 1 or 0 is input to neurons of a third layer at the same time, so as to train neurons of a second layer, wherein the effective signals are electromagnetic signals with the intensity larger than a noise threshold, the density value of 1 is dense, and the density value of 0 is non-dense;
in the working process, after a group of effective signals are input to neurons of the first layer, the second layer is processed according to the trained operation model, and the third layer outputs a concentration result value of 1 or 0.
Further, each set of signals includes 10 electromagnetic signals or effective signals.
An electromagnetic environment influence analysis system based on the analysis method is characterized in that: the intelligent analysis system comprises a device management module, an intelligent analysis module and a result display module;
the device management module is used for being connected with the frequency-using device and the electromagnetic environment monitoring device to acquire electromagnetic signals in the frequency-using device and the environment, carrying out data acquisition and control on the connected device, and then transmitting the acquired data to the intelligent analysis module;
the intelligent analysis module comprises a signal analysis module, a frequency consumption management module, an analysis judging module and a frequency consumption suggestion module; the signal analysis module acquires a current signal of the frequency-using equipment and an electromagnetic signal in the environment according to the data acquired by the equipment management module, and then analyzes the concentration of the electromagnetic signal in the environment to acquire a non-dense signal frequency band; the frequency use management module is used for presetting or inputting the existing frequency use plan, frequency use threat and frequency use rule in the frequency use area by a user; the analysis and judgment module is used for combining the frequency band where the non-dense signal is located with a frequency utilization plan, a frequency utilization threat and a frequency utilization rule, and analyzing to obtain the available frequency band of the frequency utilization equipment; then calculating the health value of each available frequency band, and taking the available frequency band with the maximum health value as a suggested frequency band to be used;
the result display module is used for displaying data acquired by the intelligent analysis module through the equipment management module, frequency usage rules input by a user, frequency usage plans and frequency usage threats, and suggested frequency usage bands of the frequency usage equipment obtained through analysis and processing of the intelligent analysis module.
Further, the device management module is in communication connection with the frequency-using device and the electromagnetic environment monitoring device through the Internet of things, and comprises a device information management module, a device communication protocol adaptation module, a data acquisition module, a data processing module and an information interface module.
Further, the electromagnetic environment monitoring device comprises a frequency monitoring device.
Compared with the prior art, the invention has the following advantages: the method comprises the steps of intelligently collecting data of electromagnetic environment monitoring equipment through the technology of the Internet of things, obtaining current signals of the frequency equipment and electromagnetic signals in the environment, analyzing and processing the data through a density analysis model based on Artificial Intelligence (AI), realizing density distribution condition analysis of effective signals in the environment, and then combining a frequency using range of the frequency equipment, a frequency using plan, a frequency using threat, a frequency using rule and the like input by a user to give out frequency bands which can be used by the frequency using equipment, so that on-site frequency using and frequency tube personnel can be helped to timely make frequency using condition judgment and quickly react, various frequency using activities can be orderly and effectively unfolded, and on-site communication work can be promoted to be efficiently exerted; the system can reduce the communication technical requirements on field workers, and improve the field work efficiency, so that the stability and the rationality of normal work of the frequency-using equipment are ensured.
Drawings
FIG. 1 is a flow chart of the present invention.
FIG. 2 is an analysis flow chart of a concentration analysis model.
Fig. 3 is a functional block diagram of a concentration analysis model.
Detailed Description
The invention will be further described with reference to the drawings and examples.
Example 1: an electromagnetic environment influence analysis method based on artificial intelligence specifically comprises the following steps:
1) Acquiring a current signal of the frequency-taking device and an electromagnetic signal existing in the environment;
2) Identifying the current signal of the frequency-using equipment in the acquired electromagnetic environment; because the current signal of the frequency-using device is also in the environment, the current signal of the frequency-using device in the detected environment signal needs to be identified, so that the influence on the subsequent processing is avoided.
3) The electromagnetic signals after the current signals are identified are divided into a plurality of groups according to the sequence of the frequency, and each group of signals comprises k electromagnetic signals. Then sequentially inputting the frequency of each group of signals into a trained density analysis model based on Artificial Intelligence (AI), and judging whether the group of signals are dense signals or not; if yes, repeating the step; if not, recording the frequency band of the group of effective signals, and then repeating the step; until the judgment of the electromagnetic signals is completed.
The density analysis model is a neural network model and is built based on an artificial neural network library keras, and comprises three layers of neural networks: the first layer is an input layer with k neurons, the second layer is an intermediate layer with p neurons, and the third layer is an input layer or an output layer with 1 neuron. In particular, as an example, the first layer has 10 neurons and the second layer has 64 neurons; i.e. take k=10, p=64.
The activation function activation between network layers adopts a linear rectification function relu, and finally, the cross entropy of the output tensor and the target tensor is calculated to be used as the loss of a model, and rmsprop is used as an optimizer of the model.
In the training process, sample data comprising noise is input to neurons of a first layer, and corresponding concentration values 1 or 0 are input to neurons of a third layer, wherein the concentration values are concentration values of effective signals after noise is identified, the concentration values 1 are dense, and the concentration values 0 are non-dense. In the working process of the neural network model, the second layer firstly acquires a noise threshold calculation model according to the sample data input by the first layer and the corresponding concentration value input by the third layer, and continuously corrects the noise threshold calculation model in the continuous training process; and then calculating according to the noise threshold calculation model to obtain a noise threshold, acquiring noise and effective signals according to the noise threshold, and finally analyzing the effective signals in the group of sample data to obtain a density calculation model of the effective signals, and continuously correcting the density calculation model. The neurons in the second layer neural network store the noise threshold calculation model and the density calculation model after training is completed, so that input data can be processed in actual work. In training, sample data are data considered to be calculated and designed according to theoretical values, experimentally obtained data or other empirical data; the more sample data, the better, preferably more than 5000 sample data.
In the working process, after a group of electromagnetic signals (including noise and effective signals) are input to neurons of a first layer, a second layer is processed according to a trained operation model, the noise and the effective signals can be automatically identified according to a noise threshold calculation model stored after training, and finally, the concentration value of the effective signals is obtained according to a density calculation model stored after training, and the concentration value of the effective signals is output through a third layer.
4) Comparing the frequency band recorded in the step 3) with a preset frequency plan, a frequency threat and a frequency rule in sequence to obtain a frequency band which meets the frequency plan and the frequency rule and is beyond the frequency threat, and recording the frequency band as an available frequency band; and repeating the step until all the frequency bands recorded in the step 3) are judged.
5) And 4) arranging the available frequency bands obtained in the step 4) according to the frequency order, and combining the continuous available frequency bands into a new frequency band.
6) Counting the number of electromagnetic signals of each frequency band processed in the step 5), calculating a signal intensity average value, calculating a health value of each frequency band, and arranging each frequency band in descending order according to the health value; the health value calculation model of each frequency band is as follows: frequency band health value=1/(signal intensity average value noise threshold value electromagnetic signal number).
7) And finally, taking the frequency band with the maximum health value as the recommended frequency band.
Example 2: referring to fig. 1 to 3, an electromagnetic environment influence analysis method based on artificial intelligence specifically includes the following steps:
1) The current signal of the frequency-taking device and the electromagnetic signal present in the environment are obtained.
2) Identifying a current signal of a frequency-consuming device in the acquired environmental electromagnetic signal; because the current signal of the frequency-using device is also in the environment, the current signal of the frequency-using device in the detected environment signal needs to be identified, so that the influence on the subsequent processing is avoided.
3) Firstly, calculating to obtain a noise threshold according to the intensity of a current signal of the frequency-using equipment and a noise proportion coefficient, wherein the noise proportion coefficient is a critical signal-to-noise ratio which does not influence the frequency-using state and is usually 0.4-0.5; then comparing the intensity of the electromagnetic signals in the environment with a noise threshold value to obtain electromagnetic signals with the intensity larger than the noise threshold value, and marking the electromagnetic signals as effective signals; the effective signals are then divided into groups according to the order of frequency, and each group of signals comprises k effective signals when the method is implemented. Then sequentially inputting the frequency of each group of signals into a trained density analysis model based on Artificial Intelligence (AI), judging whether the group of effective signals are dense signals, and if so, repeating the step; if not, recording the frequency band of the group of effective signals, and then repeating the step; and obtaining all non-dense signal frequency bands until a plurality of groups of effective signals are judged.
The density analysis model is a neural network model and is built based on an artificial neural network library keras, and comprises three layers of neural networks, wherein: the first layer is an input layer with k neurons, the second layer is an intermediate layer with p neurons, and the third layer is an input layer or an output layer with 1 neuron. In particular, as an example, the first layer has 10 neurons and the second layer has 64 neurons; i.e. take k=10, p=64.
The activation function activation between network layers adopts a linear rectification function relu, and finally, the cross entropy of the output tensor and the target tensor is calculated to be used as the loss of a model, and rmsprop is used as an optimizer of the model.
In the training process, after a group of (with known concentration) sample data (only including effective signals) is input to the neurons of the first layer, a corresponding concentration value of 1 or 0 is input to the neurons of the third layer at the same time, so as to train the neurons of the second layer, namely, the p neurons of the second layer in the trained neural network are known quantities; wherein the density value 1 is dense and 0 is non-dense. In training, the more sample data, the better, preferably more than 5000 sample data. In the working process of the neural network model, the second layer acquires a noise density calculation model according to the sample data input by the first layer and the corresponding density value input by the third layer, and continuously corrects the density calculation model in the continuous training process; and then storing the density calculation model after training is finished so as to process the input data in actual work. In training, sample data are data considered to be calculated and designed according to theoretical values, experimentally obtained data or other empirical data; the more sample data, the better, preferably more than 5000 sample data.
In the working process, after a group of effective signals are input to neurons of a first layer, a second layer is processed according to a trained operation model (a density calculation model) to obtain a density value of the effective signals, and a third layer is used for outputting a density result value of 1 or 0.
4) Comparing the frequency band recorded in the step 3) with a preset frequency plan, a frequency threat and a frequency rule in sequence to obtain a frequency band which meets the frequency plan and the frequency rule and is beyond the frequency threat, and recording the frequency band as an available frequency band; and repeating the step until all the frequency bands recorded in the step 3) are judged.
5) And 4) arranging the available frequency bands obtained in the step 4) according to the frequency order, and combining the continuous available frequency bands into a new frequency band.
6) Counting the number of electromagnetic signals of each frequency band processed in the step 5), calculating a signal intensity average value, calculating a health value of each frequency band, and arranging each frequency band in descending order according to the health value; the health value calculation model of each frequency band is as follows: frequency band health value=1/(signal intensity average value noise threshold value electromagnetic signal number).
7) And finally, taking the frequency band with the maximum health value as the recommended frequency band.
The invention also discloses an electromagnetic environment influence analysis system based on the analysis method, which comprises an equipment management module, an intelligent analysis module and a result display module.
The device management module is used for being connected with the frequency-using device and the electromagnetic environment monitoring device to collect electromagnetic signals in the frequency-using device and the environment, and carrying out data collection and control on the connected device, and then transmitting the collected data to the intelligent analysis module. In practice, the electromagnetic environment monitoring device comprises a frequency monitoring device. As optimization, the device management module also collects data information of the frequency-using device and the detection device, specifically, the device management module is in communication connection with the frequency-using device and the electromagnetic environment monitoring device through the internet of things, and the device management module comprises a device information management module, a device communication protocol adaptation module, a data collection module, a data processing module and an information interface module. After the device management module establishes connection with the devices (the frequency-using device and the monitoring device), the device information management module configures basic information of the devices, wherein the basic information comprises the device type, the device name, the device acquisition parameters, the device IP, the device port and the device protocol. After the information of the monitoring equipment (fixed monitoring station or mobile monitoring machine, etc.) is configured in the equipment information management module, the monitoring equipment is accessed into the same network as the system through the Ethernet, so that the system can automatically identify the monitoring equipment. The device communication protocol adaptation module is driven by the loading device, so that the device communication protocol can be adapted, and the communication format, the communication content and the like of the device can be identified from the communication packet sent by the device to the system. The data acquisition module is used for initiating a data acquisition instruction to the monitoring equipment after the system identifies the equipment and can normally communicate with the equipment, and feeding back the acquired data to the data acquisition module according to the instruction requirement after the monitoring equipment receives the instruction. The data processing module internally processes the acquired original data into data meeting the analysis requirement of the system application, namely, the data is converted into data which can be identified and processed by the intelligent analysis module. The information interface module is responsible for receiving instructions needing to be subjected to data acquisition from other modules, initiating a request to the data acquisition module according to the instructions, and carrying out data acquisition by the data acquisition module; and then, the data processed by the data processing module is transmitted to the intelligent analysis module through an API interface of the system internal standard.
The intelligent analysis module comprises a signal analysis module, a frequency consumption management module, an analysis judging module and a frequency consumption suggestion module. The signal analysis module acquires the current signal of the frequency-using equipment and the electromagnetic signal in the environment according to the data acquired by the equipment management module, then acquires the noise threshold according to the intensity of the current signal of the frequency-using equipment, and analyzes the concentration of the electromagnetic signal above the noise threshold in the environment to acquire a non-dense signal frequency band.
As an embodiment, the specific analysis process is implemented as follows:
the noise threshold is calculated first, then the noise is removed, and the process can be realized through a filter or an oscilloscope, and the effective signal is reserved.
The obtained effective signals are then divided into a plurality of groups according to the frequency order, and each group of signals comprises 10 effective signals when the method is implemented. Then sequentially inputting the frequency of each group of signals into a trained concentration analysis model, judging whether the group of effective signals are dense signals, and recording the frequency range of the group of effective signals when the group of signals are non-dense signals; the steps are repeated until the judgment of the signals of the groups is completed.
The concentration analysis model is a neural network module and comprises three layers of neural network models, wherein: the first layer is an input layer with 10 neurons, the second layer is an intermediate layer with 64 neurons, and the third layer is an input layer or an output layer with 1 neuron.
The activation function activation between network layers adopts a linear rectification function relu, and finally, the cross entropy of the output tensor and the target tensor is calculated to be used as the loss of a model, and rmsprop is used as an optimizer of the model.
In the training process, after a group of sample data is input to the neurons of the first layer, a corresponding concentration value of 1 or 0 is input to the neurons of the third layer at the same time, so as to train the neurons of the second layer, wherein the concentration value of 1 is dense, and the concentration value of 0 is non-dense.
In the working process, after a group of training effective signals are input to neurons of the first layer, the second layer is processed according to the trained operation model, and the third layer outputs a concentration result value of 1 or 0.
The frequency use management module is used for presetting or inputting the existing frequency use plan, the frequency use threat and the frequency use rule in the frequency use area by a user. The analysis and judgment module is used for combining the frequency band where the non-dense signals are located with a frequency utilization plan, a frequency utilization threat and a frequency utilization rule, and analyzing and obtaining the available frequency band of the frequency utilization equipment. And then calculating the health value of each available frequency band, and taking the available frequency band with the maximum health value as the recommended frequency band.
In the implementation process, the continuous frequency bands are combined, then the electromagnetic signal quantity statistics is carried out on each frequency band, the signal intensity average value is calculated, then the health value of each frequency band is calculated, and the frequency bands are arranged in descending order according to the health value; the health value calculation model of each frequency band is as follows: frequency band health value=1/(signal intensity average value noise threshold value electromagnetic signal number). And taking the available frequency band with the maximum health value as the recommended frequency band.
The result display module is used for displaying data acquired by the intelligent analysis module through the equipment management module, a frequency using plan, a frequency using threat and a frequency using rule input by a user, and a suggested frequency using band of the frequency using equipment obtained through analysis and processing by the intelligent analysis module.
Finally, it should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the technical solution, and those skilled in the art should understand that modifications and equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the present invention, and all such modifications and equivalents are included in the scope of the claims.
Claims (9)
1. An electromagnetic environment influence analysis method based on artificial intelligence is characterized in that: the method specifically comprises the following steps:
1) Acquiring a current signal of the frequency-taking device and an electromagnetic signal existing in the environment;
2) Identifying the current signal of the frequency-using equipment in the acquired electromagnetic environment;
3) Dividing the electromagnetic signals after the current signals are identified into a plurality of groups according to the frequency order, sequentially inputting the frequency of each group of electromagnetic signals into a trained concentration analysis model, judging whether the group of signals are dense signals, and if so, repeating the step; if not, recording the frequency band of the group of effective signals, and then repeating the step; until the judgment of a plurality of groups of electromagnetic signals is completed;
4) Comparing the frequency band recorded in the step 3) with a preset frequency plan, a frequency threat and a frequency rule in sequence to obtain a frequency band which meets the frequency plan and the frequency rule and is beyond the frequency threat, and recording the frequency band as an available frequency band; repeating the step until all the frequency bands recorded in the step 3) are judged;
5) Arranging the available frequency bands obtained in the step 4) according to the frequency order, and combining the continuous available frequency bands into a new frequency band;
6) Counting the number of electromagnetic signals of each frequency band processed in the step 5), calculating a signal intensity average value, calculating a health value of each frequency band, and arranging each frequency band in descending order according to the health value;
7) And finally, taking the frequency band with the maximum health value as the recommended frequency band.
2. The electromagnetic environment impact analysis method based on artificial intelligence according to claim 1, wherein: in the step 6), the health value calculation model of each frequency band is as follows: frequency band health value=1/(signal intensity average value noise threshold value electromagnetic signal number).
3. The electromagnetic environment impact analysis method based on artificial intelligence according to claim 1, wherein: in step 3), the concentration analysis model is a neural network model, which includes three layers of neural networks, wherein: the first layer is an input layer with k neurons, the second layer is an intermediate layer with p neurons, and the third layer is an input layer or an output layer with 1 neuron;
the activation function activation between network layers adopts a linear rectification function relu, and finally, the cross entropy of an output tensor and a target tensor is calculated to be used as the loss of a model, and rmsprop is used as an optimizer of the model;
in the training process, sample data comprising noise is input to neurons of a first layer, and corresponding concentration values 1 or 0 are input to neurons of a third layer, wherein the concentration values are concentration values of effective signals after noise is identified, the concentration values 1 are dense, and the concentration values 0 are non-dense;
in the working process, after a group of electromagnetic signals are input to neurons of the first layer, the second layer is processed according to the trained operation model, noise and effective signals can be automatically identified, and the third layer outputs a concentration value of the effective signals.
4. The electromagnetic environment impact analysis method based on artificial intelligence according to claim 1, wherein: in the step 3), a noise threshold value is calculated according to the intensity of the current signal of the frequency-using equipment and the noise proportion coefficient, then the intensity of an electromagnetic signal in the environment is compared with the noise threshold value, and an electromagnetic signal with the intensity larger than the noise threshold value is obtained and is recorded as an effective signal; and then dividing the effective signals into a plurality of groups according to the frequency order.
5. The electromagnetic environment impact analysis method based on artificial intelligence according to claim 4, wherein: in step 3), the concentration analysis model is a neural network model, which includes three layers of neural networks, wherein: the first layer is an input layer with k neurons, the second layer is an intermediate layer with p neurons, and the third layer is an input layer or an output layer with 1 neuron;
the activation function activation between network layers adopts a linear rectification function relu, and finally, the cross entropy of an output tensor and a target tensor is calculated to be used as the loss of a model, and rmsprop is used as an optimizer of the model;
in the training process, after a group of effective signal sample data is input to neurons of a first layer, a corresponding density value of 1 or 0 is input to neurons of a third layer at the same time, so as to train neurons of a second layer, wherein the effective signals are electromagnetic signals with the intensity larger than a noise threshold, the density value of 1 is dense, and the density value of 0 is non-dense;
in the working process, after a group of effective signals are input to neurons of the first layer, the second layer is processed according to the trained operation model, and the third layer outputs a concentration result value of 1 or 0.
6. The electromagnetic environment impact analysis method based on artificial intelligence according to claim 3 or 4, wherein: each set of signals includes 10 electromagnetic signals or effective signals.
7. An electromagnetic environmental impact analysis system based on the analysis method according to any one of claims 1 to 6, characterized in that: the intelligent analysis system comprises a device management module, an intelligent analysis module and a result display module;
the device management module is used for being connected with the frequency-using device and the electromagnetic environment monitoring device to acquire electromagnetic signals in the frequency-using device and the environment, carrying out data acquisition and control on the connected device, and then transmitting the acquired data to the intelligent analysis module;
the intelligent analysis module comprises a signal analysis module, a frequency consumption management module, an analysis judging module and a frequency consumption suggestion module; the signal analysis module acquires a current signal of the frequency-using equipment and an electromagnetic signal in the environment according to the data acquired by the equipment management module, and then analyzes the concentration of the electromagnetic signal in the environment to acquire a non-dense signal frequency band; the frequency use management module is used for presetting or inputting the existing frequency use plan, frequency use threat and frequency use rule in the frequency use area by a user; the analysis and judgment module is used for combining the frequency band where the non-dense signal is located with a frequency utilization plan, a frequency utilization threat and a frequency utilization rule, and analyzing to obtain the available frequency band of the frequency utilization equipment; then calculating the health value of each available frequency band, and taking the available frequency band with the maximum health value as a suggested frequency band to be used;
the result display module is used for displaying data acquired by the intelligent analysis module through the equipment management module, a frequency using plan, a frequency using threat and a frequency using rule input by a user, and a suggested frequency using band of the frequency using equipment obtained through analysis and processing by the intelligent analysis module.
8. The electromagnetic environmental impact analysis system of claim 7, wherein: the device management module is in communication connection with the frequency-using device and the electromagnetic environment monitoring device through the Internet of things, and comprises a device information management module, a device communication protocol adaptation module, a data acquisition module, a data processing module and an information interface module.
9. The electromagnetic environmental impact analysis system of claim 7, wherein: the electromagnetic environment monitoring device includes a frequency monitoring device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310471414.1A CN117033922A (en) | 2023-04-27 | 2023-04-27 | Electromagnetic environment influence analysis method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310471414.1A CN117033922A (en) | 2023-04-27 | 2023-04-27 | Electromagnetic environment influence analysis method and system based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117033922A true CN117033922A (en) | 2023-11-10 |
Family
ID=88601010
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310471414.1A Pending CN117033922A (en) | 2023-04-27 | 2023-04-27 | Electromagnetic environment influence analysis method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117033922A (en) |
-
2023
- 2023-04-27 CN CN202310471414.1A patent/CN117033922A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104396188B (en) | System and method for carrying out basic reason analysis to mobile network property problem | |
| Lichodzijewski et al. | Dynamic intrusion detection using self-organizing maps | |
| US11570060B2 (en) | Artificial intelligence real-time microseism monitoring node | |
| CN116488750A (en) | Electromagnetic interference influence analysis method and system based on neural network | |
| US11513245B2 (en) | Artificial intelligence calculation method and apparatus for monitoring earthquake in real time based on edge cloud cooperation, and storage medium | |
| CN107979810A (en) | Indoor orientation method, server and system | |
| CN110138468B (en) | Lora signal monitoring system and method of multi-node software radio platform | |
| CN105592487A (en) | Evaluation method and apparatus for LTE network service flow | |
| CN102083116A (en) | Method for constructing intelligent communication alarm fault positioning expert system based on meta-knowledge | |
| CN111124852A (en) | Fault prediction method and system based on BMC health management module | |
| CN103490849A (en) | Method and device for analyzing signaling traffic | |
| CN106325140A (en) | Multi-channel short wave/ultrashort wave signal processing system and processing method based on FPGA | |
| CN105897488A (en) | Visualization method of radio signal data | |
| CN110855478A (en) | Single-sensor topology sensing method and device for unreliable information | |
| CN112732802A (en) | Enterprise data mining system and method based on big data | |
| CN105553574A (en) | Support-vector-machine-based MAC protocol identification method in cognitive radio | |
| CN117033922A (en) | Electromagnetic environment influence analysis method and system based on artificial intelligence | |
| CN104573864A (en) | Data analysis alarm method based on autoregressive prediction | |
| US20150058466A1 (en) | Device for server grouping | |
| CN102118200B (en) | Spectrum detecting method based on covariance absolute value method | |
| CN103746781B (en) | Method and apparatus for processing response information in reception sensitivity test | |
| CN101494873A (en) | Method for listening wireless real time multi-channel data packet in domain network | |
| CN106034092A (en) | Serial port baud rate automation identification method, equipment and system thereof | |
| CN106937254A (en) | Zigbee positions networking equipment timer-controlled self-inspection system and method | |
| CN113677027A (en) | Method and system for predicting future transmission of node based on machine 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 |