An intelligent scanning detection system for full-band electromagnetic signals

Abstract

The present invention discloses an intelligent scanning detection system for full-band electromagnetic signals, comprising: a signal acquisition module: adopting a three-stage cascaded radio frequency chain design to acquire full-band electromagnetic signals; a signal integration module: optimizing and selecting the optimal unit spacing through an improved genetic algorithm, and receiving full-band electromagnetic signals by an integrated millimeter-wave antenna array based on the optimal unit spacing, and outputting optimized electromagnetic signals; an intelligent scanning module: dividing different frequency band signals according to the signals of different frequency bands in the optimized electromagnetic signals and formulating corresponding scanning strategies, and outputting scanned electromagnetic signals; a signal processing module: extracting three-dimensional comprehensive features including time-frequency-space from the scanned electromagnetic signals, determining three-dimensional comprehensive features with potential connections based on similarity matching, and performing enhancement processing through an improved CAPON beamforming algorithm to obtain the best signal representation, realizing intelligent signal scanning and analysis, and capable of quickly and accurately obtaining valuable signal representations from complex electromagnetic environments.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular, to an intelligent scanning and detecting system for full-band electromagnetic signals. Background Art

[0002] At present, electromagnetic signals are increasingly widely used in various fields. From communication, radar, navigation to industrial production, medical equipment, etc., electromagnetic signals are everywhere. However, with the increasingly complex electromagnetic environment, there are many challenges in efficiently and accurately scanning and detecting full-band electromagnetic signals;

[0003] At present, the scanning and detecting accuracy and efficiency of traditional systems are relatively low. When processing signals, traditional systems adopt a uniform scanning strategy, and the residence time of the key frequency band and the non-key frequency band is the same, which will lead to low scanning efficiency. Measured data shows that the full scan in the 24GHz frequency band takes more than 200ms, which cannot meet the real-time detection requirements. When using mean statistics (such as V2) in the non-key frequency band, the time-domain details of transient signals (such as pulse width <1μs) are lost, resulting in missed detection of low-probability threat signals. Moreover, traditional electromagnetic signal scanning and detecting systems lack intelligence. They usually require a large amount of manual parameter setting and data analysis work to obtain the final signal representation, and it is difficult to automatically adjust the detection strategy according to the changes in the actual electromagnetic environment. Facing the increasingly complex and changeable electromagnetic environment, this manual intervention method is inefficient and error-prone, and cannot give full play to the performance advantages of the system. Therefore, an intelligent scanning and detecting system for full-band electromagnetic signals is proposed here. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art and achieve the above object, the present invention proposes the following technical solutions:

[0005] An intelligent scanning and detecting system for full-band electromagnetic signals, comprising:

[0006] A signal acquisition module: adopting a three-stage cascaded radio frequency chain design to acquire full-band electromagnetic signals;

[0007] A signal integration module: optimizing and selecting the optimal unit spacing through an improved genetic algorithm, and receiving full-band electromagnetic signals through an integrated millimeter-wave antenna array based on the optimal unit spacing, and outputting optimized electromagnetic signals;

[0008] An intelligent scanning module: dividing different frequency band signals in the optimized electromagnetic signals and formulating corresponding scanning strategies, and outputting scanned electromagnetic signals;

[0009] Signal processing module: Extract three-dimensional comprehensive features including time-frequency-space from the scanned electromagnetic signals, determine three-dimensional comprehensive features with potential connections based on similarity matching, and perform enhancement processing through an improved CAPON beamforming algorithm to obtain the best signal representation;

[0010] The improved genetic algorithm comprehensively considers and optimizes the basic genetic algorithm by introducing the index gain and sidelobe suppression ratio;

[0011] The improved CAPON beamforming algorithm introduces a weighting factor on the basis of the traditional CAPON algorithm to optimize the acquisition of the weight vector of beamforming.

[0012] The full-band electromagnetic signals include low-band electromagnetic signals, mid-band electromagnetic signals, and high-band electromagnetic signals.

[0013] The low-band electromagnetic signals adopt a direct sampling architecture, use a 16-bit ADC, and the sampling frequency is f = 10 GS / s. The mid-band electromagnetic signals are collected by superheterodyne frequency conversion, and the local oscillator step accuracy of the collection is less than or equal to 1 Hz. The high-band electromagnetic signals adopt the method of second harmonic mixing combined with a cryogenic low-noise amplifier, and the noise figure of the cryogenic low-noise amplifier is less than 3 dB.

[0014] The process of obtaining the optimal element spacing is as follows:

[0015] Select the highest frequency within the full-band electromagnetic signal frequency band Calculate the corresponding wavelength to determine the preliminary value of the element spacing. Through the basic genetic algorithm, use the preliminary value of the element spacing as the individual gene in the genetic algorithm, and initialize a population to obtain the antenna array performance index gain corresponding to each individual in the population And the sidelobe suppression ratio , and use the sidelobe suppression ratio And the performance index gain For comprehensive consideration, define a comprehensive performance index function ;

[0016] Based on the comprehensive performance index function , calculate the fitness of each individual in the initial population according to the comprehensive performance index function, and perform selection according to the fitness of the individual, and finally obtain the element spacing value that optimizes the antenna array performance under the current conditions .

[0017] The process of obtaining the optimized electromagnetic signals is as follows:

[0018] Based on the optimal element spacing Design the antenna structure and beam scanning basis. The integrated millimeter-wave antenna array receives electromagnetic signals of different frequency bands from space, converts them into electrical signals and outputs them to obtain optimized electromagnetic signals. .

[0019] The process of obtaining the scanned electromagnetic signal is as follows:

[0020] Based on the frequency band set covered by the optimized electromagnetic signal Define the signal density function and the importance weight function , based on the signal density function and the importance weight function Obtain the comprehensive index of each frequency band , based on the comprehensive index Divide the frequency bands into key frequency bands and non-key frequency bands;

[0021] Adopt the fine-scanning method for each frequency point of the key frequency band to obtain the key signal characteristics ;

[0022] Adopt the frequency-hopping fast-scanning method for the non-key frequency band to obtain the basic statistical characteristics ;

[0023] Based on the key signal characteristics and the basic statistical characteristics Obtain the scanned electromagnetic signal .

[0024] The three-dimensional comprehensive characteristics include time-domain characteristic signals, frequency-domain characteristic signals and space-domain characteristic signals.

[0025] The process of obtaining the best signal representation is as follows:

[0026] Obtain the similarity between different three-dimensional comprehensive characteristics through the cosine similarity algorithm, and perform signal matching through the similarity threshold to obtain the same signal source;

[0027] Obtain the weighting factor Optimize the beamforming weight vector of the traditional CAPON algorithm , through the weight vector Perform weighted processing on the three-dimensional comprehensive characteristics of the same signal source received to obtain the enhanced best signal representation .

[0028] The process of obtaining the weighting factor is as follows:

[0029] The different features in the three-dimensional comprehensive features are linearly combined according to weights to obtain a preliminary weighting factor. A reward rule is set based on the improvement of the signal-to-noise ratio of the signal and the improvement of the target signal and the interference signal. Using the reinforcement learning algorithm, according to the reward rule, the mapping relationship between the preliminary weighting factor and each feature of the signal is adjusted to obtain the weighting factor 。

[0030] The present invention has the following beneficial effects:

[0031] In the present invention, first of all, by adopting a three-stage cascaded RF chain design and combining targeted processing methods for different frequency bands, the present invention can comprehensively collect electromagnetic signals in the full frequency band. At the same time, through the signal integration module, the integrated millimeter-wave antenna array with the optimized selection of the optimal unit spacing using the improved genetic algorithm receives the signal. This algorithm comprehensively considers the gain of the antenna array performance index and the side lobe suppression ratio, dynamically adjusts the calculation weight, and can significantly improve the gain of the antenna array and enhance the signal reception ability compared with the traditional design. At the same time, it can effectively suppress the side lobe level and reduce signal interference;

[0032] Secondly, the improved CAPON beamforming algorithm introduces an adaptive weighting factor, which is dynamically adjusted according to the three-dimensional comprehensive features of the signal and system requirements, further enhancing the target signal and suppressing the interference signal, making the finally obtained signal quality higher and more conducive to subsequent signal recognition, classification, demodulation and other processes;

[0033] Finally, the intelligent scanning module calculates the comprehensive index according to the signal density function and the importance weight function, intelligently divides the signals in different frequency bands, formulates corresponding scanning strategies, uses frequency-by-frequency fine scanning to obtain detailed feature vectors for key frequency bands, and uses frequency hopping fast scanning to obtain basic statistical features for non-key frequency bands, which not only improves the scanning efficiency, but also can effectively analyze according to the characteristics of signals in different frequency bands. The signal processing module can accurately judge the potential connection of signals based on the signal matching mechanism of time-frequency-space three-dimensional features. The whole system realizes intelligent signal scanning and analysis, and can quickly and accurately obtain valuable signal representations from complex electromagnetic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a system block diagram of an intelligent scanning and detecting system for full-frequency-band electromagnetic signals proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] 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. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0036] Embodiment: As Figure 1 shown, an intelligent full-band electromagnetic signal scanning and detection system proposed by the present invention includes:

[0037] Signal acquisition module: Adopting a three-stage cascaded RF chain design to acquire full-band electromagnetic signals;

[0038] The full-band electromagnetic signals include low-frequency band electromagnetic signals (DC - 6 GHz), medium-frequency band electromagnetic signals (6 - 18 GHz), and high-frequency band electromagnetic signals (18 - 24 GHz);

[0039] For the low-frequency band (DC - 6 GHz), a direct sampling architecture is adopted, using a 16-bit ADC with a sampling frequency f = 10 GS / s. Assume the input low-frequency signal is a continuous-time signal , according to the sampling theorem, a discrete low-frequency signal in the low-frequency band is obtained after ADC sampling , and the sampling process satisfies , where , is the acquisition period;

[0040] Specifically, the direct sampling architecture avoids the problems of frequency conversion loss and image interference brought by the traditional frequency conversion architecture, and can simply and effectively obtain low-frequency band signal information, and finally outputs a sampled discrete low-frequency signal ;

[0041] For the medium-frequency band (6 - 18 GHz), the superheterodyne frequency conversion method is adopted, and the local oscillator signal , where the acquisition local oscillator step accuracy is less than or equal to 1 Hz, and the input intermediate-frequency signal is , after mixing by a double-balanced mixer, according to the mixing principle, the output intermediate-frequency band signal is: , where is the gain coefficient of the mixer;

[0042] Specifically, by precisely controlling the local oscillator frequency , precise selection and processing of intermediate-frequency signals with different frequencies can be achieved, and the RF signal is converted into an intermediate-frequency signal convenient for subsequent processing ;

[0043] For the high-frequency band (18 - 24 GHz), the method of second-harmonic mixing combined with a cryogenic low-noise amplifier is adopted. The input high-frequency signal is , and the local oscillator signal is , and the signal after second-harmonic mixing is: where is the gain coefficient of the second-harmonic mixer;

[0044] Specifically, second harmonic mixing improves the mixing efficiency and reduces the requirement for the local oscillator frequency. Then, the signal is amplified by a cryogenic low-noise amplifier (with a noise figure less than 3 dB), and the signal-to-noise ratio of the output signal is significantly improved. The output is a high-frequency signal after mixing and amplification processing. ;

[0045] The full-band RF subsystem finally outputs a set of signals after being processed in different frequency bands. , and these signals will be used as the input for the next module.

[0046] Signal integration module: An improved genetic algorithm is used to optimize and select the optimal unit spacing. The integrated millimeter-wave antenna array based on the optimal unit spacing receives the full-band electromagnetic signal and outputs an optimized electromagnetic signal.

[0047] The process of obtaining the optimal unit spacing is as follows:

[0048] Determine the frequency range in which the system operates (DC - 24 GHz). Within this frequency band, the frequency changes continuously. Select the highest frequency within the full-band electromagnetic signal frequency band. Calculate the corresponding wavelength to determine the preliminary value of the unit spacing. The calculation formula is: , where is the signal wavelength, is the maximum scanning angle;

[0049] Specifically, select the highest frequency within the frequency band because its corresponding wavelength is the shortest, and the calculated unit spacing is the most stringent value that meets the requirements of the entire frequency band. For example, for the 18 - 24 GHz frequency band, select Calculate the wavelength;

[0050] The implementation process of the improved genetic algorithm is as follows:

[0051] Through a basic genetic algorithm, the unit spacing is used as the individual gene in the genetic algorithm, and a population containing multiple different values is initialized;

[0052] By introducing the comprehensive consideration of the index gain and the sidelobe suppression ratio to optimize the basic genetic algorithm to obtain the improved genetic algorithm;

[0053] Calculate the antenna array performance index gain corresponding to each individual (i.e., each value), denoted as , where is the index gain, is the number of antenna elements, is the acquisition frequency;

[0054] Obtain the sidelobe suppression ratio , calculate the pattern function of the antenna array through Fourier transform , based on the pattern function , calculate by calculating the ratio of the main lobe peak value to the sidelobe peak value, that is , where is the angle corresponding to the main lobe peak is the angle corresponding to the sidelobe peak;

[0055] At the same time, comprehensively consider the sidelobe suppression ratio and the performance index gain , and define a comprehensive performance index function: , is the number of antenna elements;

[0056] Among them, is the weight coefficient of the performance index gain , is the weight coefficient of the sidelobe suppression ratio ;

[0057] Based on the comprehensive performance index function , for each individual in the initial population, calculate its fitness according to the comprehensive performance index function, select according to the fitness of the individual, and individuals with higher fitness have a greater probability of being selected into the next generation, and through selection, crossover, and mutation genetic operations, iteratively update the population to make the population gradually evolve towards the optimal element spacing , and finally obtain the element spacing value that optimizes the performance of the antenna array under the current conditions;

[0058] Specifically, optimize the element spacing through an improved genetic algorithm, no longer only focusing on the gain in the antenna array performance index, but introducing the sidelobe suppression ratio and adaptive parameter adjustment, and dynamically adjust the weights for calculating the gain and sidelobe suppression according to the distribution of individuals in the current population and the convergence trend of the objective function;

[0059] For example: when it is found that the population has approached convergence in the gain index, but there is still much room for improvement in the sidelobe suppression ratio index, automatically increase the weight value, strengthen the optimization of the sidelobe suppression ratio, so that the gain of the antenna array is significantly improved. Under the same signal input and environmental conditions, compared with the antenna array designed traditionally, the received signal strength may be increased by 20%-30% or even higher, effectively enhancing the signal reception ability. At the same time, the sidelobe level can be effectively suppressed, which can be reduced by 5-10dB, reducing signal interference and improving the purity and quality of the signal;

[0060] Based on the optimal element spacing Design the antenna structure and beam scanning foundation. The integrated millimeter-wave antenna array receives electromagnetic signals of different frequency bands from space, converts them into electrical signals and outputs to obtain optimized electromagnetic signals, expressed as ;

[0061] Example: In the 24GHz frequency band, a specific optimal element spacing can maximize the induction efficiency of the antenna for signals in this frequency band. At the same time, when performing full-band electromagnetic signal scanning, the optimal element spacing combined with a specific antenna structure design can achieve three-dimensional beam scanning, and the scanning range is azimuth , pitch , during the scanning process, based on the optimal element spacing the antenna array can flexibly adjust the beam direction, accurately align the signal sources in different directions, ensure efficient signal acquisition throughout the scanning range, and there will be no signal omission or reception blind area.

[0062] Intelligent scanning module: According to the signals of different frequency bands in the optimized electromagnetic signals, divide the signals of different frequency bands and formulate corresponding scanning strategies, and output scanning electromagnetic signals;

[0063] Suppose the optimized electromagnetic signal received by the system covers the frequency band set as ( ), is the index;

[0064] Define a signal density function , used to measure the number or intensity distribution of signals within a unit frequency range, define an importance weight function , indicating the importance degree of different frequency bands set artificially according to application requirements. Through the signal density function and the importance weight function calculate the comprehensive index , calculate for each frequency band , according to the value of , divide the frequency band into key frequency bands and non-key frequency bands;

[0065] Set a threshold , when , divide the frequency band where is located into key frequency bands; when , divide it into non-key frequency bands;

[0066] For the divided key frequency bands, adopt the fine-scanning method for each frequency point, and set the dwell time as (for example , at each frequency point, perform detailed feature extraction on the collected signals, adopt time-frequency analysis methods, obtain the characteristics of the signals in the time domain and frequency domain (such as signal amplitude change, frequency component, pulse width), and form key signal characteristics ;

[0067] For non-key frequency bands, a frequency hopping and fast scanning method is adopted, and the dwell time is set to (for example ), to quickly obtain the signal overview of these frequency bands. At each frequency hopping point, directly obtain the basic statistical characteristics of the signal (mean signal strength, frequency range, probability of signal occurrence);

[0068] For example, for some low-frequency non-key frequency bands, quickly scan to obtain the approximate signal strength and frequency distribution, and judge whether there are interference signals or abnormal signals;

[0069] After frequency band division and scanning strategy processing, output the scanned electromagnetic signal ;

[0070] Specifically, the signal feature vectors obtained by fine scanning of the key frequency bands in the scanned electromagnetic signal These feature vectors contain detailed information of the key frequency band signals and can be used for subsequent processing such as signal recognition, classification, and demodulation; The basic statistical characteristics obtained by fast scanning of non-key frequency bands

[0071] are used to quickly understand the signal status of non-key frequency bands and judge whether there are signal changes or abnormal conditions that require further attention. These scanned electromagnetic signals will be used as the input of the subsequent signal processing unit for more in-depth analysis and processing. Signal processing module: Extract three-dimensional comprehensive features including time-frequency-space from the scanned electromagnetic signal, determine the three-dimensional comprehensive features with potential connections based on similarity matching, and perform enhancement processing through an improved CAPON beamforming algorithm to obtain the best signal representation;

[0072] The process of obtaining three-dimensional comprehensive features is as follows:

[0073] The three-dimensional time-frequency-space features include time-domain feature signals, frequency-domain feature signals, and space-domain feature signals;

[0074] For the signals in different frequency bands in the scanned electromagnetic signal

[0075] (the signal feature vectors of key frequency bands and the basic statistical characteristics of non-key frequency bands ), extract their time-domain features (amplitude changes of the signal) respectively. Let the time-domain feature vector of the key frequency band in the scanned electromagnetic signal be in the time domain, and the non-key frequency band feature vector be , then the time-domain feature signal of the scanned electromagnetic signal is ;

[0076] Extract frequency-domain features (such as frequency components). The feature vector in the key frequency bands in the frequency domain is , and the feature vector in the non-key frequency bands in the frequency domain is . Then the frequency-domain feature signal of the scanned electromagnetic signal is ;

[0077] Extract spatial-domain features (direction of arrival of the signal). The feature vector in the key frequency bands in the spatial domain is , and the feature vector in the non-key frequency bands in the spatial domain is . Then the spatial-domain feature signal of the scanned electromagnetic signal is ;

[0078] Then the three-dimensional comprehensive feature is represented as . Calculate the similarity between different three-dimensional comprehensive features for signal matching. The similarity calculation uses the cosine similarity algorithm. Let the similarity between two three-dimensional comprehensive features be expressed as:

[0079]

[0080] Set a similarity threshold . When the similarity between two signals is greater than the similarity threshold , it is considered that there is a potential connection between these two signals, that is, they come from the same signal source, and it is uniformly expressed as ;

[0081] Enhance the same signal source through an improved CAPON beamforming algorithm; the process of obtaining the improved CAPON beamforming algorithm is as follows:

[0082] The improved CAPON beamforming algorithm introduces a weighting factor on the basis of the traditional CAPON algorithm to optimize the weight vector of the beamforming of the traditional CAPON algorithm . Let the steering vector of the antenna array be , and the covariance matrix of the received signal be . The improved weight vector is:

[0083]

[0084] where is a weighting factor that is dynamically adjusted according to the spatial-domain features of the signal (such as the DOA of the signal) and system requirements. represents the conjugate transpose. For example, for a key signal from a specific direction, increase the value corresponding to this direction to enhance the gain of the signal in this direction;

[0085] Weighting factor The acquisition process is as follows:

[0086] Introduce an adaptive learning mechanism to determine the weighting factor , in the initial operation stage, preliminarily set the weighting factor based on the three-dimensional comprehensive characteristics of the signal according to preset rules;

[0087] Specifically, the preset rules are obtained by linearly combining different characteristics in the three-dimensional comprehensive characteristics with weights. The preliminarily set weighting factor is expressed as , where is the time-domain weight, is the frequency-domain weight, is the spatial-domain weight;

[0088] Continuously monitor the improvement of the signal-to-noise ratio of the signal and the improvement of the target signal and interference signal to set the reward rules:

[0089] Let the signal-to-noise ratio of the target signal before processing be , and after processing be , the power of the interference signal before processing be , and after processing be ;

[0090] After beamforming processing, if the signal-to-noise ratio of the target signal improves by more than a certain threshold, or the interference signal is successfully separated, a positive reward is given. Specifically:

[0091] If (is the set signal-to-noise ratio improvement threshold), and ( ) is the interference signal power reduction ratio threshold), a positive reward is given. If the above conditions are not met, a negative reward is given;

[0092] Use the reinforcement learning algorithm to adjust the preliminary weighting factor according to the reward situation and obtain the weighting factor according to the mapping relationship between the signal characteristics ;

[0093] Specifically: If it is found that for a certain type of time-domain feature signal, the signal-to-noise ratio improves significantly after increasing the weighting factor, the reinforcement learning algorithm will increase the weight of the time-domain feature in the generation of the weighting factor, so that in subsequent processing, it is more inclined to adjust the weighting factor according to this time-domain feature. Through this adaptive learning mechanism, the weighting factor can be continuously optimized, and the beamforming effect can be continuously improved in different electromagnetic signal scenarios;

[0094] Then, according to the weight vector perform weighting processing on the three-dimensional comprehensive characteristics of the same signal source received to obtain the enhanced best signal representation , the formula is: = , where represents the conjugate transpose of the weight vector ;

[0095] Specifically, the signal represented after being processed by the above improved CAPON beamforming algorithm is , which is the optimal signal representation of the improved CAPON beamforming algorithm. The result of this signal representation can be used in subsequent signal recognition, classification, demodulation and other applications. The signal obtained after time-frequency-space three-dimensional comprehensive feature extraction, similarity matching and enhanced processing by the improved CAPON beamforming algorithm is the result of optimization processing on the basis of considering the multi-dimensional features of the signal and the complex electromagnetic environment. This result can provide more accurate and clearer signal information and improve the accuracy of signal detection;

[0096] Example: In signal recognition, pattern matching is performed based on the clearer features of the optimal signal representation ;

[0097] When classifying signals, the optimal signal representation features are classified according to rules or machine learning models;

[0098] In signal demodulation, the original signal is accurately restored with a high-quality signal representation .

[0099] In the application, several formulas involved are calculated by taking their numerical values after dimensionless. The establishment of the formula is a formula obtained by collecting a large amount of data for software simulation to approximate the real situation as much as possible. Some coefficients or weights in the formula are set by those skilled in the art according to the actual situation, so no more details will be given here.

[0100] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution.

[0101] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A full-band electromagnetic signal intelligent scanning detection system, characterized in that: include: Signal acquisition module: adopts a three-stage cascaded RF chain design to collect full-band electromagnetic signals; Signal integration module: An improved genetic algorithm is used to optimize and select the optimal unit spacing. The integrated millimeter-wave antenna array based on the optimal unit spacing receives full-band electromagnetic signals and outputs optimized electromagnetic signals. The optimal cell spacing acquisition process is as follows: Select the highest frequency in the full-band electromagnetic signal band Calculate the corresponding wavelength to determine the initial value of the unit spacing. Use the genetic algorithm to use the initial value of the unit spacing as the individual gene in the genetic algorithm, initialize a population, and obtain the antenna array performance index gain corresponding to each individual in the population. SLS , and the sidelobe suppression ratio Gain with performance indicators Take a comprehensive approach and define a comprehensive performance indicator function ; Based on comprehensive performance index function , calculate the fitness of each individual in the initial population according to the comprehensive performance index function, and select according to the fitness of the individual to obtain the optimal unit spacing value for the best antenna array performance ; Intelligent scanning module: According to the different frequency band signals in the optimized electromagnetic signal, the signals of different frequency bands are divided and corresponding scanning strategies are formulated to output scanning electromagnetic signals; Signal processing module: extracts three-dimensional comprehensive features including time, frequency and space from the scanning electromagnetic signal, determines the three-dimensional comprehensive features with potential connections based on similarity matching, and performs enhancement processing through the improved CAPON beamforming algorithm to obtain the best signal representation; The improved genetic algorithm optimizes the basic genetic algorithm by introducing the index gain and the sidelobe suppression ratio for comprehensive consideration; The improved CAPON beamforming algorithm introduces a weighting factor based on the traditional CAPON algorithm to optimize the acquisition of the beamforming weight vector in the traditional CAPON algorithm; Among them, the improved CAPON beamforming algorithm acquisition process is: The improved CAPON beamforming algorithm introduces a weighting factor based on the traditional CAPON algorithm. To optimize the weight vector of the traditional CAPON algorithm beamforming , assuming the steering vector of the antenna array is , the covariance matrix of the received signal is , the improved weight vector for: ； in It is a weighting factor that is dynamically adjusted according to the spatial characteristics of the signal and system requirements. represents conjugate transpose; The weighting factor The acquisition process is: The different features in the three-dimensional comprehensive features are linearly combined according to the weights to obtain the preliminary weighting factor. The reward rules are set based on the improvement of the signal-to-noise ratio of the signal and the improvement of the target signal and the interference signal. The reinforcement learning algorithm is used to adjust the mapping relationship between the preliminary weighting factor and each signal feature according to the reward rule to obtain the weighting factor. .

2. The full-band electromagnetic signal intelligent scanning detection system according to claim 1 is characterized in that: The full-band electromagnetic signal includes a low-band electromagnetic signal, a mid-band electromagnetic signal and a high-band electromagnetic signal.

3. The full-band electromagnetic signal intelligent scanning detection system according to claim 2 is characterized in that: The low-frequency electromagnetic signal adopts a direct sampling architecture, uses a 16-bit ADC, and has a sampling frequency of f=10GS / s. The intermediate-frequency electromagnetic signal is collected by superheterodyne frequency conversion, and the collected local oscillator step accuracy is less than or equal to 1Hz. The high-frequency electromagnetic signal adopts second harmonic mixing combined with a low-temperature low-noise amplifier, and the noise coefficient of the low-temperature low-noise amplifier is less than 3dB.

4. The full-band electromagnetic signal intelligent scanning detection system according to claim 1 is characterized in that: The acquisition process of the optimized electromagnetic signal is as follows: Based on the optimal cell spacing Design antenna structure and beam scanning basis, integrated millimeter wave antenna array receives electromagnetic signals of different frequency bands from space, converts them into electrical signals and outputs them to obtain optimized electromagnetic signals .

5. The full-band electromagnetic signal intelligent scanning detection system according to claim 1 is characterized in that: The acquisition process of the scanning electromagnetic signal is as follows: Based on optimizing electromagnetic signal The set of frequency bands covered ( ), As the index, define the signal density function and importance weight function ; Based on the signal density function and importance weight function Get comprehensive metrics for each frequency band , based on comprehensive indicators Divide the frequency bands into key frequency bands and non-key frequency bands; Use frequency-by-frequency scanning to obtain key signal characteristics for key frequency bands ; Use frequency hopping fast scanning to obtain basic statistical characteristics for non-key frequency bands ; Based on key signal characteristics And basic statistical characteristics Acquiring scanning electromagnetic signals .

6. The full-band electromagnetic signal intelligent scanning detection system according to claim 1 is characterized in that: The three-dimensional comprehensive features include time domain feature signals, frequency domain feature signals and space domain feature signals.

7. The full-band electromagnetic signal intelligent scanning detection system according to claim 1, characterized in that: The process of obtaining the optimal signal representation is as follows: The cosine similarity algorithm is used to obtain the similarity between different 3D comprehensive features, and the similarity threshold is used Perform signal matching to obtain the same signal source; Get weighting factor Optimizing the beamforming weight vector of the traditional CAPON algorithm , through the weight vector The three-dimensional comprehensive features of the same signal source received are weighted to obtain the enhanced optimal signal representation .

Images