Wireless anti-jamming signal detection method, device, equipment and storage medium

Abstract

The present application relates to the technical field of wireless signal detection, and discloses a wireless anti-interference signal detection method, device, equipment and storage medium, wherein the method comprises: collecting a standardized interference characteristic vector through a multi-scene wireless signal receiving device; performing interference characteristic enhancement processing on the standardized interference characteristic vector to obtain an interference characteristic sample set; constructing a dual-array cooperative detection system comprising a passive detection array and an active detection array and obtaining dual-path interference signal monitoring data through real-time monitoring; performing spatial domain and time-frequency domain double feature extraction and fusion processing on the dual-path interference signal monitoring data to obtain comprehensive interference characteristics; and dynamically adjusting the receiver filtering parameter matrix and detection threshold vector of the dual-array cooperative detection system according to the comprehensive interference characteristics, which can adaptively cope with changes in different interference environments, realize comprehensive coverage of communication spectrum, and greatly expand the monitoring range and spectrum sensing capability of wireless anti-interference.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless signal detection, and in particular to a wireless anti-interference signal detection method, device, equipment and storage medium. BACKGROUND

[0002] With the rapid development of wireless communication technology and the diversification of application scenarios, modern wireless communication environment is increasingly complex. The widespread deployment of various wireless devices leads to an increasingly crowded electromagnetic environment, and various types of interference sources exist in the network simultaneously, including narrowband interference, wideband interference, impulse interference and sweep interference, etc. In the key fields of 5G / 6G communication, smart city, industrial Internet of Things, etc., these diversified interference sources and complex interference modes seriously affect the reliability, connection stability and spectrum utilization efficiency of the communication system, bringing serious challenges to the normal operation of the wireless communication network.

[0003] Traditional wireless anti-interference signal detection methods mainly rely on fixed threshold detection, energy detection or matched filtering technology, etc. These methods are still acceptable when solving the problem of single interference source or static electromagnetic environment, but they show obvious shortcomings when facing complex and variable interference environment. These traditional methods usually only focus on single features in time domain or frequency domain, and cannot effectively capture the complex time-frequency characteristics of interference signals; at the same time, they lack accurate classification ability of interference signals, and cannot take corresponding suppression strategies for different types of interference; in addition, due to the lack of actual interference samples, it is difficult to establish an accurate interference feature model, resulting in a big discount in detection performance in actual application. SUMMARY

[0004] The present application provides a wireless anti-interference signal detection method, device, equipment and storage medium, which can adaptively cope with the changes of different interference environments, realize the comprehensive coverage of communication spectrum, and greatly expand the monitoring range and spectrum sensing ability of wireless anti-interference.

[0005] In a first aspect, the present application provides a wireless anti-interference signal detection method, which comprises:

[0006] Collecting multi-band interference signal data through a multi-scene wireless signal receiving device, performing time-frequency domain feature extraction and preprocessing on the multi-band interference signal data, and obtaining a standardized interference feature vector;

[0007] Performing interference feature enhancement processing on the standardized interference feature vector to obtain an interference feature sample set;

[0008] Constructing a dual-array cooperative detection system comprising a passive detection array and an active detection array according to the interference feature sample set, and performing real-time monitoring on the wireless environment through the dual-array cooperative detection system to obtain dual-path interference signal monitoring data;

[0009] performing spatial domain and time-frequency domain dual feature extraction and fusion processing on the dual-path interference signal monitoring data to obtain comprehensive interference features;

[0010] adjusting a receiver filtering parameter matrix and a detection threshold vector of the dual-array collaborative detection system according to the comprehensive interference features.

[0011] In a second aspect, the present application provides a wireless anti-interference signal detection device, which comprises:

[0012] a collection module, configured to collect multi-band interference signal data through a multi-scene wireless signal receiving device, and perform time-frequency domain feature extraction and preprocessing on the multi-band interference signal data to obtain a standardized interference feature vector;

[0013] an enhancement module, configured to perform interference feature enhancement processing on the standardized interference feature vector to obtain an interference feature sample set;

[0014] a construction module, configured to construct a dual-array collaborative detection system comprising a passive detection array and an active detection array according to the interference feature sample set, and perform real-time monitoring on a wireless environment through the dual-array collaborative detection system to obtain dual-path interference signal monitoring data;

[0015] a fusion module, configured to perform spatial domain and time-frequency domain dual feature extraction and fusion processing on the dual-path interference signal monitoring data to obtain comprehensive interference features;

[0016] an adjustment module, configured to adjust a receiver filtering parameter matrix and a detection threshold vector of the dual-array collaborative detection system according to the comprehensive interference features.

[0017] In a third aspect, the present application provides a wireless anti-interference signal detection device, which comprises a memory and at least one processor, wherein the memory stores instructions; and the at least one processor invokes the instructions in the memory to enable the wireless anti-interference signal detection device to perform the wireless anti-interference signal detection method described above.

[0018] In a fourth aspect, the present application provides a computer readable storage medium, which stores instructions; and when the instructions are run on a computer, the computer is enabled to perform the wireless anti-interference signal detection method described above.

[0019] In the technical scheme provided by the application, based on multi-scene multi-band interference signal collection and time-frequency domain feature extraction, the method comprehensively captures the feature performance of various interference sources in different environments, interference feature enhancement processing is performed by using a Wasserstein generative adversarial network, the problem of insufficient actual interference signal samples is effectively solved, the scale of the training data set is expanded to 10 times that of the original data, and the learning ability of the model for various interference features is significantly enhanced. A dual-array cooperative detection system combining a passive detection array and an active detection array is adopted, the interference detection capability of the system in a complex electromagnetic environment is greatly improved through spatial diversity technology, and the system is particularly suitable for complex scenes with multiple interference sources coexisting. A spatial domain and time-frequency domain dual feature extraction and fusion processing mechanism is combined, the feature differences of interference signals in different domains are fully utilized, and the recognition ability of the system for weak interference and complex interference modes is improved. Based on a hierarchical interference signal detection model, interference signal detection, interference type identification and interference parameter estimation are sequentially completed, and accurate identification and parameter acquisition of different types of interference signals are realized. The receiver filtering parameters and detection threshold are dynamically adjusted through iterative calculation, so that the system can adaptively cope with the changes of different interference environments. The interference signal generation-disappearance process tracking model effectively captures the random appearance and disappearance characteristics of interference signals, provides the system with interference behavior prediction ability, and enhances the perception and response ability to burst interference. Through two cooperative working modes of synchronous scanning and complementary scanning, comprehensive coverage of the communication spectrum is realized, and the monitoring range and spectrum sensing ability of the system are greatly expanded.

[0020] Other features and advantages of the present application will be set forth in the following specification, and in part will become apparent to those skilled in the art upon examination of the following specification or can be learned by practice of the application. The objects and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

[0021] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the following preferred embodiments are specifically described below, and the accompanying drawings are described as follows. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 An embodiment schematic diagram of a wireless anti-interference signal detection method in the embodiment of the present application;

[0023] Figure 2 An embodiment schematic diagram of a wireless anti-interference signal detection device in the embodiment of the present application;

[0024] Figure 3 An embodiment schematic diagram of a wireless anti-interference signal detection device in the embodiment of the present application; DETAILED DESCRIPTION

[0025] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0026] The terms "comprising" and "having" and any variations thereof mentioned in the embodiments of the present application are intended to cover the inclusions not exclusively. For example, the processes, methods, systems, products or devices comprising a series of steps or units are not limited to the listed steps or units, but optionally further comprise other steps or units not listed, or optionally further comprise other steps or units inherent to the processes, methods, products or devices.

[0027] In order to facilitate the understanding of the embodiments, first, a wireless anti-interference signal detection method disclosed by the embodiments of the present application is described in detail. As shown in Figure 1 The method comprises the following steps:

[0028] 101. Collecting multi-band interference signal data by a multi-scene wireless signal receiving device, performing time-frequency domain feature extraction and preprocessing on the multi-band interference signal data, and obtaining a standardized interference feature vector;

[0029] It can be understood that the execution subject of the present application can be a wireless anti-interference signal detection device, and can also be a terminal or a server, and the specific execution subject is not limited here. The embodiments of the present application take the server as the execution subject for example.

[0030] Specifically, in a plurality of typical electromagnetic environments, a wireless signal receiving device with full-band coverage capability is deployed, which includes a high-sensitivity receiving antenna, a low-noise amplification module, a high-speed analog-to-digital converter, and a digital signal processing unit with parallel processing capability, and monitoring nodes are laid out in urban dense areas, industrial areas, suburban areas, and traffic interference zones, etc. to perform full-spectrum continuous scanning sampling on multiple communication frequency bands such as 2.4 GHz and 5 GHz, and to obtain original data of multi-band interference signals containing various types of wireless interference sources in real time. The multi-band interference signal data is subjected to hierarchical filtering preprocessing, and through methods such as sliding window averaging and pulse removal in the time domain, burst wideband interference and transient anomalies in the sampling process are removed, and then in the frequency domain, a bandpass filter set is used to suppress non-target band interference, so as to retain the main components of the interference signals in the core frequency band and obtain the preliminary purified interference signals. The above signals are subjected to amplitude normalization processing to eliminate the intensity imbalance caused by hardware gain difference or environmental propagation attenuation, and at the same time, time synchronization calibration is performed by using synchronous time stamp and phase alignment strategy, to ensure that the multi-channel signals are highly consistent in time scale, forming standardized time domain interference data with unified time base and unified amplitude. The standardized time domain interference data is subjected to short-time Fourier transform, which is mapped to the joint time-frequency domain space to reveal the frequency structure change of the signal in different time periods, and the short-time spectrum is calculated by using Gaussian window or Hamming window function sliding frame, and a set of two-dimensional time-frequency domain interference feature distribution graphs are formed, which show the frequency energy concentration degree of the signal, and can also reveal the dynamic behavior characteristics such as periodic disturbance and frequency sweeping characteristics. Finally, representative statistical quantity indexes are selected in the time-frequency feature distribution to construct a high-dimensional feature set, including the spectral envelope for describing the energy structure, the center frequency offset reflecting the frequency shift characteristics, the spectral bandwidth reflecting the signal diffusion range, and the time-frequency distribution mode quantifying the structural complexity, and the above multiple dimensions are combined to construct the multi-dimensional feature set of the interference signal, and the standardized interference feature vector is organized based on the unified standard feature arrangement order, normalization standard and structure coding rule.

[0031] 102. performing interference feature enhancement processing on the standardized interference feature vector to obtain an interference feature sample set;

[0032] Specifically, statistical feature analysis is performed on the standardized interference feature vectors, the distribution of all samples in the high-dimensional feature space is modeled, the mean, variance, skewness, kurtosis and other statistical quantities of each feature dimension are obtained, and the correlation between the dimensions is analyzed in combination with the overall covariance matrix of the samples to construct an interference feature distribution parameter model. On this basis, standard normal distribution random noise vectors are generated by inverse transformation method, which are used as latent variables input to simulate the interference feature patterns that may occur but are not observed in the disturbance space. The batch of standard normal random noise vectors are input into the constructed Wasserstein generative adversarial network (WGAN-GP), and the generator network contained therein performs nonlinear deep mapping on the input vectors. The generator network is composed of multiple layers of fully connected neural networks, each layer contains a number of nodes and is assisted by LeakyReLU activation function, and has the ability to gradually construct a mapping that meets the interference feature distribution structure from the latent space. The noise samples are projected to the interference feature space at the output end of the generator to form a batch of candidate interference feature samples with complex structure but not yet verified for authenticity. The discriminator network in WGAN is used to score the authenticity of the candidate feature samples. The discriminator is also composed of multiple layers of neural networks, and its task is to distinguish between real samples (i.e. standardized interference feature vectors) and generated samples, and to optimize the discrimination accuracy during training. The scoring process calculates the minimum transportation cost between the generated samples and the real data distribution based on the Wasserstein distance, and combines a gradient penalty term to constrain the discriminator to satisfy the 1-Lipschitz continuity, so as to stabilize the training and accurately express the distribution difference between the samples. The scoring result reflects the credibility of each generated sample. After obtaining the authenticity score of the candidate samples, the scoring results are sorted and thresholded to retain high-quality synthetic samples that are closest to the real samples in high-dimensional feature structure, have the smallest Wasserstein distance, and have a smooth gradient penalty term. The synthetic samples are combined with the original standardized interference feature vector to construct a representative interference feature sample set with wide coverage.

[0033] 103. Construct a dual-array cooperative detection system comprising a passive detection array and an active detection array according to the interference feature sample set, and perform real-time monitoring of the wireless environment through the dual-array cooperative detection system to obtain dual-path interference signal monitoring data.

[0034] Specifically, time-series modeling is performed on the interference feature sample set to understand the occurrence and disappearance patterns of interference signals in the actual environment. Specifically, statistical analysis is conducted on the occurrence probability of each type of interference signal in the sample set under different time periods, frequency bands, and environmental conditions to construct an interference state sequence. Based on this, a state transition mechanism is established using a hidden Markov model, where the state space is defined as two categories: "interference exists" and "interference does not exist". The state transition probability matrix is ​​learned using the maximum likelihood estimation method, thereby accurately depicting the generation and attenuation evolution path of interference signals in a dynamic environment. Based on this transfer matrix, the system configuration and functional division of the passive and active detection arrays are guided. The passive detection array uses M receiving antennas to form a uniform linear array, which is mainly responsible for unbiased listening to signals from all directions in the wireless environment. Its receiving parameters need to be precisely set, including the gain value of the low-noise amplifier to enhance weak interference signals, the local oscillator frequency and transformation method of the mixer to ensure spectrum compatibility, and the filter bandwidth to control the range of received signals and improve the ability to resist adjacent channel interference. The active detection array is configured as a phased array with beamforming capability, consisting of N array elements. Each array element is equipped with a controllable phase and amplitude modulation module to form a digital phased array unit, which can actively focus on interference signals in specific directions and frequency bands. The array is configured with dynamic beam scanning path and corner domain resource allocation strategy according to interference transfer characteristics, thereby improving the response sensitivity and discrimination capability of interference sources in the target area. To achieve collaborative sensing between the two arrays in the system, a high-speed fiber optic or synchronous Ethernet link is established to enable low-latency communication between sampling data and control commands. A unified clock synchronization mechanism and data aggregation rules are constructed through a central processing unit. The collaborative mechanism employs two alternating modes: synchronous scanning and complementary scanning. In synchronous scanning mode, the passive and active detection arrays simultaneously focus on the same frequency band and angle for refined detection. In complementary scanning mode, the two arrays monitor different frequency bands or spatial regions, thereby expanding the overall coverage and detection breadth. In actual operation, the dual-array collaborative monitoring mechanism, through the above configuration, performs high-precision, wide-angle real-time acquisition of continuous or sudden interference signals in the wireless environment. It records the generalized omnidirectional interference energy distribution obtained by the passive detection array and the directional strong interference beam information captured by the active detection array. Combined with the multi-dimensional features such as timing, intensity, direction, and spectrum provided by the dual signal sources, a dual-channel interference signal monitoring dataset is constructed.

[0035] 104. Perform spatial and time-frequency domain feature extraction and fusion processing on the dual-channel interference signal monitoring data to obtain comprehensive interference features;

[0036] Specifically, the dual-path interference signal monitoring data is uniformly formatted, the complex signal amplitude and phase information collected by the two arrays at the same time is mapped to a unified data matrix structure, and on this basis, the spatial covariance matrix of the full-array received signal is constructed, which reflects the correlation between the signal strength and phase of different array elements, and its mathematical form is the sample average of the complex matrix product. By performing eigenvalue decomposition on the spatial covariance matrix, a set of eigenvalue sequences arranged in descending order and the corresponding eigenvector set are obtained, the first k eigenvectors constitute the interference signal subspace, and the remaining part belongs to the noise subspace. Combined with the minimum description length criterion, the eigenvalue sequence is analyzed in information theory to estimate the number of potential interference sources in the current wireless environment and extract the spatial domain interference features; At the same time of completing the spatial domain modeling, S transform processing is performed on the original time domain signal received by each array element. S transform is a time-frequency analysis tool combining the advantages of short-time Fourier transform and wavelet analysis, which can extract the frequency spectrum evolution process of the signal under different time scales without sacrificing frequency resolution. By extracting representative statistics from the transformed time-frequency distribution matrix, including spectral centroid to measure the position of frequency energy concentration, spectral bandwidth to reflect the degree of frequency spread, spectral skewness to represent the asymmetry of spectral shape, spectral kurtosis to measure the degree of frequency mutation, and spectral entropy to depict the overall chaos, a fine-grained time-frequency domain interference feature set is constructed. The spatial domain features and time-frequency domain features are uniformly processed, and normalization conversion is performed on all feature dimensions through Z-score standardization, maximum and minimum normalization or other standardization strategies to eliminate the dimension difference of physical quantities, and a standardized multi-domain feature set is obtained. Then the feature set is combined with the front-end dual-array detection results for joint reasoning processing, the trust degree fusion value of the two detection sources is calculated through Dempster-Shafer evidence theory, the trust function of each spatial feature and time-frequency feature is defined, and the outputs of the passive detection array and the active detection array are respectively assigned with initial trust degrees m1 (interference exists), m2 (interference exists), etc. The global fusion trust degree m (interference exists) is obtained by synthesizing through D-S rule, which improves the decision robustness. An adaptive weight adjustment mechanism for interference environment is introduced, the fusion weights of spatial domain and time-frequency domain features are set according to the dynamic level of signal-to-noise ratio and the feature complexity of interference signal, and the comprehensive interference features are obtained through weighted aggregation.

[0037] 105. Dynamically adjusting the receiver filter parameter matrix and detection threshold vector of the dual-array cooperative detection system according to the comprehensive interference features.

[0038] Specifically, the generalized likelihood ratio test is calculated for the comprehensive interference feature, which analyzes the statistical difference between the current signal and the interference-free background to determine the existence of the interference signal, and compares the calculation result with the preset initial threshold to obtain a clear conclusion about whether there is interference at a certain time point. The judgment result provides a basis for the subsequent clustering recognition link. Based on the interference existence result, the density peak clustering module is activated to structure the comprehensive interference feature data set. By evaluating the density of each feature point in its local area and the distance between it and the high-density area, the clustering center in the feature space is automatically found, and different interference types are divided accordingly, so as to identify whether the system is currently facing continuous interference, frequency sweeping interference or pulse interference and other different forms. After completing the interference type identification, based on the clustering result, the corresponding parameter estimation module is activated to deeply extract the specific features of the current interference. For continuous narrowband interference, the main frequency component and power level are estimated; for frequency sweeping interference, the frequency hopping trajectory is tracked; for pulse interference, the key parameters such as repetition period, pulse width and energy characteristics are extracted. These interference source parameter information will be used as an important input variable of the system state and transmitted to the adaptive optimization module. In the optimization stage, the filter parameter configuration matrix and the decision threshold vector of the receiver are adjusted by alternating iteration to improve the accuracy and sensitivity of interference detection. In the optimization process, the system first resets the filter parameters by fixing the threshold parameters, then adjusts the decision threshold based on the updated filter, and continuously repeats the process until the performance index converges, obtaining the optimal system parameter combination under the current interference environment. Based on the optimal system parameter combination, the filter structure in the passive detection array and the active detection array is adjusted, including gain control, passband range, suppression of sideband interference parameters, and the threshold judgment logic in the interference discrimination module is updated synchronously, so as to distinguish the boundaries between normal signals and weak interference and strong interference more finely. Through the closed-loop mechanism of data-driven, model feedback and parameter coordination, the dual-array cooperative detection system can respond to multiple types of interference signals in real time in complex electromagnetic environments, and continuously optimize its receiving and decision-making ability according to environmental changes, improving the overall anti-interference performance and system intelligence level.

[0039] With the interference source parameters as the initial conditions of system configuration, a set of reasonable and representative starting values for the receiver filtering parameter matrix and detection threshold vector are provided according to the key attributes of the current interference signal, such as frequency band, bandwidth, power level, time characteristics and spatial distribution, and on this basis, the definition of the optimization objective function is determined, that is, the minimum interference detection probability is maximized within the full frequency range of the system coverage, so as to ensure that the system can maintain the minimum detection capability at any frequency point and there is no detection blind area. In order to make the optimization problem more solvable and computationally stable, auxiliary variables are introduced to reconstruct the original objective function, convert the non-convex maximum-minimum optimization problem into an equivalent form, establish an iterative solvable structure, and at the same time set the upper and lower bounds and precision requirements of the bisection search, so as to form a bounded and convergent search interval as the basis for algorithm iteration. In each iteration cycle, the system takes the midpoint value of the current bisection interval as the temporary target detection probability, and under this condition, the receiver filtering parameter matrix is optimized by fixing the current detection threshold vector. Since the optimization problem of the receiver filtering parameter is essentially nonlinear and non-convex, the first-order Taylor expansion method is used to construct its linear approximation problem near the current point, and the original complex structure is converted into a convex approximation model that can be handled in form. For this convex approximation problem, the interior point method is called to solve it, which significantly improves the convergence speed while ensuring the calculation accuracy. After the filtering parameter matrix is updated, the system optimizes the detection threshold vector under the premise that the matrix is fixed, adjusts the decision limit under each frequency band or interference condition to match the signal response mode under the new filter structure, thereby forming a complete round of "fixed one type of parameter, optimized another type of parameter" alternating optimization steps. The performance of the current filtering parameter matrix and detection threshold combination is verified, the interference detection capability of the full frequency band is calculated, and a comparison is made with the current target detection probability. If the target requirement is met, it means that the combination has strong generalization detection capability, the lower bound of the search interval is updated to the current midpoint value, otherwise the upper bound is updated, thereby gradually narrowing the search range. Iterative calculation and performance judgment are repeatedly performed, the system optimal parameter configuration interval is constantly approached, and finally the receiver filtering parameter matrix and detection threshold vector output are the system parameter combination with the best performance and the strongest detection robustness under the current interference conditions.

[0040] In the embodiment of the present application, based on multi-scene multi-band interference signal collection and time-frequency domain feature extraction, the method comprehensively captures the feature performance of various interference sources in different environments, uses the Wasserstein generative adversarial network for interference feature enhancement processing, effectively solves the problem of insufficient actual interference signal samples, expands the training data set scale to 10 times the original data, and significantly enhances the learning ability of the model for various interference features. A dual-array cooperative detection system combining passive detection array and active detection array greatly improves the interference detection capability of the system in complex electromagnetic environments through spatial diversity technology, and is particularly suitable for complex scenes with multiple interference sources coexisting. Combined with spatial domain and time-frequency domain dual feature extraction and fusion processing mechanism, the feature differences of interference signals in different domains are fully utilized to improve the recognition ability of the system to weak interference and complex interference patterns. Based on the hierarchical interference signal detection model, interference signal detection, interference type recognition and interference parameter estimation are completed in turn to realize accurate recognition and parameter acquisition of different types of interference signals. Through iterative calculation, the receiver filter parameters and detection threshold are dynamically adjusted, so that the system can adapt to the changes of different interference environments. The interference signal generation-disappearance process tracking model effectively captures the random appearance and disappearance characteristics of interference signals, provides the system with interference behavior prediction ability, and enhances the perception and response ability to burst interference. Through two kinds of cooperative working modes of synchronous scanning and complementary scanning, the system realizes comprehensive coverage of the communication spectrum, greatly expands the monitoring range and spectrum sensing ability of the system.

[0041] In a specific embodiment, the process of performing step 101 can specifically include the following steps:

[0042] Obtain multi-band interference signal data by performing full-spectrum scanning sampling in multiple different electromagnetic environments through the multi-scene wireless signal receiving device;

[0043] Perform time domain filtering and frequency domain filtering on the multi-band interference signal data to obtain preliminary purified interference signals;

[0044] Perform amplitude normalization and time synchronization calibration processing on the preliminary purified interference signals to obtain standardized time domain interference data;

[0045] Perform short-time Fourier transform on the standardized time domain interference data to obtain time-frequency domain interference feature distribution;

[0046] Extract the spectral envelope, center frequency offset, spectral bandwidth and time-frequency distribution mode from the time-frequency domain interference feature distribution to obtain a multi-dimensional feature set, and construct a standardized interference feature vector based on the multi-dimensional feature set.

[0047] Specifically, a wireless signal receiving system with high sampling accuracy, wide frequency response capability and strong environmental adaptability is constructed, and the system is deployed in multiple representative electromagnetic environments, including urban dense areas, suburban open areas, industrial plant areas, transportation hubs and high noise background areas, to ensure the comprehensiveness and diversity of the collected data. The receiving system has an adjustable frequency band front-end receiver, a high-sensitivity low-noise amplifier, a wideband antenna array, a high-speed analog-to-digital conversion module, and a supporting digital signal processing unit. The sampling rate of the analog-to-digital converter is at least twice the target signal bandwidth to meet the Nyquist sampling theorem and ensure spectral integrity and fidelity. The system also supports full-spectrum scanning, which continuously switches between multiple typical frequency bands such as 2.4 GHz, 5 GHz and 900 MHz within a given time period to complete spectral coverage. During the collection process, the time waveform data of the signal is recorded, and its power spectral density, signal-to-noise ratio, frequency change rate and other instantaneous characteristics are captured simultaneously to reflect the performance of the interference signal in the real environment. The multi-band raw data obtained is subjected to time domain filtering operation, and techniques such as moving average, median filtering and adaptive noise removal are used to remove burst-type impulse interference, sharp peaks and power jumps and other noise components. Subsequently, frequency domain filtering is used to suppress out-of-band non-target signals, and the frequency domain filter set is flexibly configured with low-pass, band-pass or notch filters according to the expected interference frequency band to eliminate interference from external or background electromagnetic noise sources, thereby preserving the main characteristics of the interference signal to the greatest extent. After joint time and frequency domain processing, the preliminary purified interference signal is obtained. To eliminate intensity shifts caused by hardware differences, environmental attenuation, gain unevenness and other factors, the preliminary purified signal is subjected to amplitude normalization processing to make the amplitude range of all signals fall within a unified interval. Time synchronization calibration is performed based on the system sampling time reference or GPS reference timestamp to eliminate sampling relative deviation and timing misalignment between multi-channel received signals, ensuring the consistency and comparability of the signals in the time scale, and obtaining standardized time domain interference data. The standardized time domain interference data is mapped to the time-frequency domain to reveal its frequency evolution characteristics and short-time energy distribution structure. Short-time Fourier transform is used to process the signal in frames and perform Fourier transform within each frame to generate a two-dimensional time-frequency spectrum, which shows the synchronization of signal changes in the time axis and frequency axis. Selecting appropriate window functions such as Gaussian window or Hamming window can effectively balance the time resolution and frequency resolution, and the overlap rate and window length of the sliding window are dynamically adjusted according to the change rate of the interference signal to ensure the detail integrity and spectral structure clarity of the transformed time-frequency distribution. After transformation, a set of matrix-form time-frequency domain interference characteristic distribution graphs is obtained, which reflect the interference duration and intermittent mode in the time axis and reveal the spectral distribution range and energy concentration area in the frequency axis.The high-dimensional statistical features are extracted from the time-frequency feature distribution. The spectral envelope is used to depict the overall profile of the frequency energy, which helps to identify whether the interference has a typical narrowband, wideband or sweep structure. The center frequency offset reflects the moving degree of the interference spectrum center relative to the standard frequency point, which is a key indicator for measuring the stability and drift trend of the spectrum. The spectral bandwidth quantifies the spread degree of the interference signal in the frequency, which is related to the signal propagation range and system intrusion degree. The time-frequency distribution pattern is the overall expression of the signal form structure, which is modeled by combining multi-dimensional features such as time duration, frequency sparsity and change speed, to reveal whether the interference signal has periodicity, modulation type, burst type or continuity characteristics. The spectral kurtosis, spectral skewness and spectral entropy and other auxiliary statistical quantities are extracted to depict the nonlinear characteristics and structural complexity of the interference. All the extracted spectral form, distribution position, energy change and time-frequency behavior features are combined into a multi-dimensional feature set with consistent structure according to the preset order, and each dimension feature is normalized to make it available for subsequent training modeling and pattern discrimination in a unified scale, thereby constructing a standardized interference feature vector.

[0048] In a specific embodiment, the process of performing step 102 can specifically include the following steps:

[0049] Performing statistical distribution analysis on the standardized interference feature vector to obtain interference feature distribution parameters, and generating a random noise vector of standard normal distribution based on the interference feature distribution parameters;

[0050] Inputting the random noise vector of standard normal distribution into the Wasserstein generative adversarial network, performing nonlinear transformation mapping by the generator network in the Wasserstein generative adversarial network, mapping the random noise vector of standard normal distribution to the interference feature space, and obtaining a candidate interference feature sample;

[0051] Calculating the Wasserstein distance and gradient penalty value of the candidate interference feature sample and the standardized interference feature vector by the discriminator network in the Wasserstein generative adversarial network, to obtain a sample authenticity score;

[0052] According to the sample authenticity score, high-quality synthesized interference feature samples are screened and reserved, and the synthesized interference feature samples and the standardized interference feature vector are merged to obtain an interference feature sample set.

[0053] Specifically, in the processing system, an interference feature distribution modeling mechanism is established, which takes the standardized interference feature vector as the basic data, calculates the distribution of each dimension feature in the sample space through statistical analysis method, including the mean, variance, skewness, kurtosis and other statistical parameters of each dimension, and calculates the overall covariance matrix combined with the joint distribution relationship between the features, forming a high-dimensional joint probability distribution model, which is used to derive the standard normal distribution generation mechanism that meets the distribution characteristics of each dimension, so that the subsequent generated random noise can be close to the potential distribution characteristics of the original interference data in the statistical structure. After completing the distribution parameter modeling, the sample space is expanded and enhanced by constructing a Wasserstein generative adversarial network, which consists of two substructures of generator and discriminator. The generator maps the random noise vector of the standard normal distribution to the synthesized sample in the interference feature space, while the discriminator evaluates the closeness of each synthesized sample and the real sample in the statistical distribution. A number of random vectors of the same dimension are sampled from the standard normal distribution, which are input into the generator network of the Wasserstein generative adversarial network as input in the latent space. The generator performs nonlinear transformation layer by layer through a multi-layer neural network structure, each layer including full connection operation, activation function mapping, Dropout operation to prevent overfitting, etc., so as to project the low-dimensional noise input to the output space of the same dimension as the real interference feature step by step. The output result is the preliminary generated candidate interference feature sample. The candidate sample is input into the discriminator network in the WGAN at the same time, which receives and discriminates the real sample and the generated sample. Unlike the traditional GAN using cross-entropy loss for binary classification, the Wasserstein generative adversarial network uses a distribution distance-based evaluation mechanism. The discriminator does not output a class label of 0 or 1, but outputs a real value score, which represents the distance of the current sample from the real sample set in the distribution distance dimension. The Wasserstein generative adversarial network measures the optimal transportation cost between sample distributions through the Wasserstein distance, that is, it judges the minimum effort required to transform a generated sample into a real sample. The smaller the distance, the more real and reliable the generated sample is. In order to ensure the stability of distance calculation and the robustness of model training, a gradient penalty term is introduced as a regular constraint. By calculating the partial derivative of the gradient of the discriminator with respect to the sample and limiting its norm, the network satisfies the 1-Lipschitz continuity, thereby avoiding the problems of gradient explosion or overfitting of the discrimination function. Through the score output by the discriminator, the authenticity evaluation of each candidate interference feature sample is obtained, and on this basis, an adaptive score threshold is set. According to the relationship between the Wasserstein score of the candidate sample and the threshold, the screening operation is performed, and only the generated samples with a score better than the threshold and highly close to the standardized real sample in distribution are retained.To avoid sample distribution collapse or pattern unification, the distribution parameters of input noise are constantly adjusted in the multi-round generation process, so as to guide the generator to explore more kinds of interference patterns, and ensure the sample quality while preserving diversity. After screening, qualified synthetic interference feature samples are collected and fused with the original standardized interference feature vector set to form a new interference feature sample set.

[0054] In a specific embodiment, the process of performing step 103 can specifically include the following steps:

[0055] Based on the interference feature sample set, interference signal generation and disappearance process tracking is performed to obtain an interference state transition probability matrix;

[0056] According to the interference state transition probability matrix, a passive detection array with M receiving antennas and an active detection array with N elements are configured to obtain a dual-array cooperative detection system;

[0057] The receiving parameters of the passive detection array are set, including low noise amplifier gain, mixer parameters and filter bandwidth; the digital phased unit and beam forming parameters of the active detection array are set;

[0058] A high-speed data transmission link between the passive detection array and the active detection array is established, and two cooperative working modes of synchronous scanning and complementary scanning are adopted to obtain a dual-array cooperative monitoring mechanism;

[0059] Through the dual-array cooperative monitoring mechanism, interference signals in the wireless environment are collected in real time, and omnidirectional interference data received by the passive detection array and directional interference data obtained by the active detection array are recorded respectively to obtain dual-channel interference signal monitoring data.

[0060] Specifically, based on the interference feature sample set, a dynamic modeling mechanism is constructed to describe the trend of interference behavior change. This mechanism needs to have the ability to track the start-stop state of the interference signal in the time dimension, and can depict its statistical variation pattern. A state modeling method based on Markov property is adopted to analyze the distribution rule of interference feature samples in consecutive time segments, identify the state transition path between frames, and define the state space as two basic states of "interference exists" and "interference does not exist". Then, by using frequency statistics and maximum likelihood estimation method, the state transition frequency is extracted from a large number of samples, and after normalization, the interference state transition probability matrix is constructed, which reflects the holding probability and jump probability of the interference signal in a given time period. Based on the state transition probability matrix, the construction and parameter initialization of the dual-array collaborative detection architecture are guided. The passive detection array is set as an omnidirectional monitoring array, and M receiving antennas are arranged in a uniform linear manner, each array element is connected to an independent RF front-end circuit, and the core receiving parameters need to be optimized and configured for different interference intensities and background noise levels, including selecting appropriate low-noise amplifier gain to improve weak interference receiving sensitivity, setting frequency offset value of the mixer to ensure that the spectrum transformation is not distorted, and setting the filter bandwidth to cover the current main interference frequency band and suppress out-of-band spurious signals, while ensuring that the system dynamic range is sufficient to match the power variation of the interference signal; The active detection array consists of N digitally controlled array elements, which is characterized by beamforming and angle control capability, and can actively point to the direction of the interference source under the support of interference positioning information, improving the directional signal capture accuracy, and its configuration parameters include beam scanning step, main lobe width control, side lobe suppression strategy, phase vector control, etc. All configurations need to focus on improving the resolution and response sensitivity of the target directional interference. In order to realize the collaborative work capability between the two arrays, a high-speed and stable data interaction link is established in the system architecture, which adopts synchronous optical fiber communication, clock phase-locked network or low-latency Ethernet bus to realize the passive detection array and the active detection array report the data frames collected by themselves to the central control processing module through the link, and receive the frequency band instructions, beam control parameters and scheduling priority instructions distributed by the system. In the dual-array collaborative operation mechanism, two working modes of synchronous scanning and complementary scanning are introduced to adapt to different application scenarios, among which the synchronous scanning mode requires the passive detection array and the active detection array to monitor the same frequency band and space area at the same time, to realize fine tracking and collaborative discrimination of specific interference sources, while the complementary scanning mode requires the two arrays to cover different frequency resources or spatial dimensions, to realize the parallelization of wide spectrum coverage and fast response. The mechanism dynamically switches through the task queue scheduler, and the specific switching strategy is dynamically adjusted based on the prediction trend of the interference state transition probability matrix, so as to enhance synchronization during interference outbreak or burst period, and adopt complementary monitoring during interference sparse or unknown stage, thereby ensuring the coverage integrity and response flexibility of the monitoring system.After the system is put into operation, the dual-array cooperative monitoring mechanism performs real-time sensing on the target area. The passive detection array continuously receives electromagnetic signals in all directions in an omnidirectional passive manner, records the distribution of general interference energy and the trend of background noise change, forms a global interference detection data stream, and the active detection array performs spatial directional reception according to the beam direction allocated by the system, captures the main components of interference signals in high-confidence directions, and provides directional interference feature information. The combination of the two constitutes a dual-path interference signal monitoring data, which contains the time domain, frequency domain and spatial domain three-dimensional structure information of the signal, and also covers the dynamic evolution, spatial propagation path and frequency spectrum variation characteristics of the signal, with high information density and discrimination ability.

[0061] In a specific embodiment, the process of performing step 104 can specifically include the following steps:

[0062] According to the dual-path interference signal monitoring data, a spatial covariance matrix is calculated, and the spatial covariance matrix is subjected to eigenvalue decomposition to obtain a sequence of eigenvalues and a corresponding set of eigenvectors;

[0063] Based on the minimum description length criterion, the sequence of eigenvalues is analyzed to determine the number of interference sources existing in the wireless environment, and a spatial domain interference feature is obtained;

[0064] Perform S transform on each array element signal in the dual-path interference signal monitoring data to generate a time-frequency distribution matrix, and extract the spectral centroid, spectral bandwidth, spectral skewness, spectral kurtosis and spectral entropy from the time-frequency distribution matrix to obtain a time-frequency domain interference feature;

[0065] The spatial domain interference feature and the time-frequency domain interference feature are subjected to dimension normalization processing to obtain a standardized multi-domain feature set;

[0066] The standardized multi-domain feature set and the dual-array detection result are fused by the Dempster-Shafer evidence theory to calculate the confidence degree, and the feature weight coefficient is dynamically adjusted according to the signal-to-noise ratio and the interference characteristic to obtain a comprehensive interference feature.

[0067] Specifically, the spatial covariance matrix is calculated based on the dual-path interference signal monitoring data. The complex monitoring data is expressed in matrix form and divided into frames according to time windows, so that each frame contains a sufficient number of array observation samples to support robust spatial statistical calculations. On this basis, the spatial covariance matrix is constructed by calculating the covariance of all array elements in each time window, which reflects the correlation between any two array elements in the current time period and reveals the spatial structure of the interference signal and the energy distribution of the signal subspace. The spatial covariance matrix is decomposed to extract a set of eigenvalues and their corresponding eigenvectors in descending order of energy. The largest part of these eigenvalues corresponds to the subspace occupied by the main components of the interference signal, while the remaining smaller eigenvalues are considered as background noise or unstructured energy. To determine the number of potential interference sources in the current wireless environment, the minimum description length criterion is used to analyze the eigenvalue sequence. This criterion is a structure determination method based on information theory, which maximizes the sample data interpretation ability while simplifying the model parameters as much as possible. Therefore, by calculating the compression cost corresponding to the number of spatial subspaces and the remaining energy distribution under different assumptions, the decomposition point that minimizes the overall information description length is determined, thereby judging the actual number of interference sources in the environment and taking it as the key spatial domain interference feature. At the same time of spatial analysis, the original signal data from each array element in the dual array is subjected to time-frequency transformation to extract its dynamic spectral features. The S-transform processing method is used to process the time series data of each array element, which combines the segmentation characteristics of short-time Fourier transform and the scale adaptability of wavelet transform to provide clear frequency distribution details at different time scales. The output of S-transform is a two-dimensional matrix, where the horizontal axis is time and the vertical axis is frequency, and each element in the matrix reflects the energy intensity at the corresponding time and frequency. On this basis, a set of statistically representative feature indicators are extracted from the time-frequency distribution matrix of each array element, including spectral centroid to measure the concentration trend of signal frequency distribution, spectral bandwidth to reflect the expansion range of signal in frequency domain, spectral skewness to quantify the asymmetry degree of spectrum, spectral kurtosis to describe the sharpness of spectral shape, and spectral entropy to measure the chaos degree of spectral distribution. These statistical indicators can reflect the basic morphological features of interference signals and also provide important clues for interference type identification, constituting the core content of time-frequency domain interference features. After completing the spatial domain and time-frequency domain feature extraction, the unified processing before the fusion of the two types of features is performed, that is, the dimension normalization operation is performed on the feature data of different dimensions and different physical units to make all feature dimensions in a unified dimension space by using Z-score standardization or maximum and minimum normalization, forming a standardized multi-domain feature set that can be compared and fused. The standardized multi-domain features are combined with the actual output results of the dual-array detection system to realize the fusion decision at the feature level through the Dempster-Shafer evidence theory.In the specific operation, the spatial domain feature and the time-frequency domain feature are defined as independent sources respectively, and corresponding trust functions are constructed. An initial trust degree about "existence of interference" or "non-existence of interference" is given to each type of feature, and then the trust values are input into the D-S evidence combination rule to solve the conflict, synthesize the evidence and fuse the trust degrees, so as to obtain the fusion trust evaluation result as a whole. In order to adapt to different interference intensity, interference complexity and signal-to-noise ratio conditions, a dynamic feature weight adjustment mechanism is introduced. According to the overall energy level of the current environment signal, the noise distribution structure and the interference feature concentration, the contribution of each type of feature to the overall judgment is evaluated in real time, and the weight proportion of the spatial domain feature and the time-frequency domain feature in the final fusion result is dynamically adjusted, so that the spatial distribution ability is highlighted in the high signal-to-noise ratio scene, and the spectral structure stability is highlighted in the low signal-to-noise ratio condition, and finally the comprehensive interference feature is obtained.

[0068] In a specific embodiment, the process of performing step 105 can specifically include the following steps:

[0069] The generalized likelihood ratio test is performed on the comprehensive interference feature to obtain a test statistic, and the test statistic is compared with an initial threshold to obtain an interference signal existence judgment result;

[0070] Based on the interference signal existence judgment result, density peak clustering is performed on the comprehensive interference feature to obtain an interference type recognition result;

[0071] Based on the interference type recognition result, interference parameter estimation is performed to obtain interference source parameter information;

[0072] According to the interference source parameter information, the receiver filter parameter matrix and the detection threshold vector are alternately optimized by iterative calculation to obtain an optimal system parameter combination;

[0073] Based on the optimal system parameter combination, the receiver filter configuration and the interference detection decision threshold of the dual-array cooperative detection system are dynamically adjusted to complete the anti-interference detection processing of multiple types of interference signals in a complex electromagnetic environment.

[0074] Specifically, the generalized likelihood ratio test is used to calculate the comprehensive interference characteristics, and the initial judgment of whether there is a significant interference signal in the current wireless environment is realized. The test method calculates the test statistic by ratio operation of the joint probability model of the actual observed signal under the interference hypothesis and the no interference hypothesis. The statistic can quantify the deviation between the observed data and the background noise, and is compared with the initial decision threshold preset in the system. When the statistical value exceeds the threshold, it is considered that the interference signal exists at the current time, otherwise it is considered that there is no interference or low intensity interference background. The process is continuously performed on the time-frequency window, thereby forming a time axis interference existence discrimination sequence for the whole channel state. The density peak clustering operation is performed on the feature sample set confirmed to exist interference. The method does not depend on sample label information, but calculates the local density of each data point and the minimum distance between it and the point with higher density, constructs a density-distance two-dimensional space image, automatically finds the representative clustering center, and adsorbs the remaining samples to the corresponding center according to the density gradient, forming a natural division result of the interference type. The clustering algorithm is suitable for the case that the interference samples present non-spherical and non-uniform distribution, and can effectively identify different types of interference signal structures such as narrowband, wideband, sweep, frequency hopping, pulse and continuous carrier, providing category constraint conditions and structure boundaries for the next step of parameter estimation. After completing the type recognition, the customized parameter estimation module is called for different interference types to analyze and calculate the key physical quantities. For narrowband interference signals, the center frequency, signal strength and modulation mode are estimated; for sweep interference, the start frequency, sweep rate and periodic characteristics are extracted; and for pulse interference, the repetition frequency, pulse width, interval and envelope structure are estimated. The parameter estimation method can be realized by combining frequency domain analysis, matching pursuit, spectrum fitting and time series modeling, etc., to ensure a high description of the behavior characteristics of the interference source, and then the above estimated values are transmitted as inputs to the system control module for guiding the configuration reconstruction of the receiving channel. According to the obtained interference source parameter information, the filter parameter matrix and the detection threshold vector of the receiver are alternately optimized through iteration to realize the optimal anti-interference configuration of the whole system in the current interference scene. The specific method is to initialize the filter parameters and the decision threshold, and define the anti-interference performance objective function as the maximization of the minimum detection probability in the full frequency range, that is, the maximization of the worst performance to ensure that the system still has basic detection ability under the most unfavorable conditions. In the optimization process, the bisection search strategy is used to continuously approach the acceptable performance boundary. In each iteration, the current detection threshold is fixed, the filter parameters are linearly expanded to form a solvable sub-problem, and the filter parameters are updated by solving the sub-problem through optimization algorithms such as interior point method. Then the threshold vector is optimized again under the new filter parameters, and the cycle is continued until the search precision meets the condition, and the most suitable system parameter combination for the current environment is obtained.Based on the above optimization results, the key parameters in the dual-array cooperative detection architecture are dynamically adjusted, wherein the passive detection array needs to reset the center frequency point and passband range of the front-end filter, adjust the amplifier gain to match the signal strength variation, and reconstruct the frequency domain sampling rate to adapt to the bandwidth characteristics; the active detection array updates the beam pointing, phase control and beam width setting in real time according to the parameter results, thereby improving the focusing ability of the interference signal in a specific direction and effectively suppressing the sidelobe interference energy in the non-target direction; in the decision module, the detection decision threshold is adjusted synchronously to adapt to the performance difference of different interference types in the frequency spectrum distribution and power distribution, so as to ensure that the system can still maintain the dual goals of sensitive response and low misjudgment when facing frequent changes in the interference state.

[0075] In a specific embodiment, the process of obtaining the optimal system parameter combination by iteratively calculating and alternately optimizing the receiver filtering parameter matrix and the detection threshold vector according to the interference source parameter information can specifically include the following steps:

[0076] The interference source parameter information is used to initialize the receiver filtering parameter matrix and the detection threshold vector, and the maximum minimum interference detection probability of the full frequency band is set as the optimization objective function;

[0077] The optimization objective function is transformed into an equivalent form by introducing auxiliary variables, and the upper and lower bounds of the bisection search and the precision requirement are set to obtain the initialized bisection search interval;

[0078] In each iteration, the target detection probability is set as the midpoint value in the initialized bisection search interval, the detection threshold vector is fixed, the receiver filtering parameter matrix is first-order Taylor expanded at the current point to obtain a convex approximation problem;

[0079] The interior point method is used to solve the convex approximation problem, the receiver filtering parameter matrix is updated, and the detection threshold vector is optimized under the condition of fixing the updated receiver filtering parameter matrix to obtain the system parameter combination of the current iteration;

[0080] It is judged whether the interference detection performance of the system parameter combination of the current iteration in the full frequency band meets the target detection probability, if yes, the lower bound is updated, otherwise the upper bound is updated to obtain a new search interval;

[0081] When the length of the new search interval is less than the precision requirement, the iteration is stopped, and the optimal system parameter combination is output.

[0082] Specifically, the system initialization configuration framework is constructed based on the interference source parameter information, wherein the interference source parameter information includes the center frequency, signal strength, spectral bandwidth, occurrence period, duration, directivity distribution, and signal-to-noise ratio characteristics of the interference signal, and through the structured extraction and induction of the information, an explicit optimization direction is provided for the current working state of the receiver. The initial setting of the filter parameter matrix of the receiver needs to refer to the main energy distribution position of the interference signal, select an appropriate passband range in the frequency domain, set the amplification coefficient matching the interference power in the gain control, and give the element cooperative adjustment capability in the phase control; the initial value of the detection threshold vector is set to a relatively neutral low reference level according to the maximum strength of the interference signal and the ambient noise distribution, so as to ensure that the system has sufficient detection sensitivity in the early stage of iteration. In the optimization stage, the objective function is to maximize the minimum interference detection probability in the full frequency band, which has the characteristic of "maximizing the minimum value" and belongs to a typical fractional structure optimization problem, which is difficult to solve directly in the original form. Therefore, an auxiliary variable is introduced as the lower bound of the optimization target, and the original nonlinear objective function is reconstructed into a linear equivalent form under the constraint condition, and at the same time the upper and lower bounds of the bisection search are set, wherein the upper bound can be estimated by the theoretical optimal detection probability, and the lower bound is given by the actual detection performance of the system in the non-optimized state, and a small enough precision threshold is set between the two as the iteration termination standard, and the initial bisection search interval is obtained. In each iteration process, the system takes the midpoint value of the current upper and lower bounds as the target detection probability, and expands the optimization operation in this round. The specific process is as follows: first, fix the current detection threshold vector unchanged, and focus on optimizing the receiver filter parameter matrix, but since the influence of the matrix on the objective function has nonlinear properties, direct solution will lead to uncontrollable calculation, therefore, the first-order Taylor expansion is performed on the objective function at the current point to linearize it, and a local convex optimization problem is constructed based on this, which has analytical structure and solvability, and the interior point method and other convex optimization algorithms are used for efficient solution to obtain the optimal filter parameter update value under the current detection threshold. Under the condition of keeping the just updated filter parameter matrix unchanged, the detection threshold vector is optimized and adjusted again, and the target is to make the system more sensitive to the interference signal and the false alarm rate controllable under the current filtering state. In this stage, the false alarm probability back-propagation mechanism is used to update the decision threshold, that is, a reasonable decision threshold distribution is mapped from the expected detection probability of the current system through a statistical model, and the parameters are adjusted according to the preset performance requirements. The updated filter parameter matrix and threshold vector jointly constitute the system parameter combination formed in this round of iteration.The interference detection performance of the parameter combination in the full frequency range is evaluated, the detection probability at each frequency point is calculated through actual observation or simulation, and the minimum value is taken as the extreme performance index of the current combination. The performance index is compared with the target detection probability. If it is equal to or greater than the target probability, it means that the current combination has high robustness, so it is accepted as a solution and the lower bound of the search is updated to the current midpoint value. Otherwise, if it does not meet the performance requirement, the upper bound is compressed to the current midpoint value, forming a new search interval with a smaller range. This process is repeated continuously, and each iteration further compresses the uncertainty in the parameter space, continuously updates the filter parameters and detection threshold of the receiving channel, and always optimizes the detection ability under the most unfavorable conditions. When the search interval length is less than the preset precision threshold, the system determines that the convergence condition is met, and the iteration process is terminated. The filter parameter matrix and detection threshold vector corresponding to the current system parameters are output as the final optimal system parameter combination.

[0083] The wireless anti-interference signal detection method in the embodiment of the application is described above, and the wireless anti-interference signal detection device in the embodiment of the application is described below. Please refer to Figure 2 An embodiment of the wireless anti-interference signal detection device in the embodiment of the application includes:

[0084] The acquisition module 201 is configured to acquire multi-band interference signal data by a multi-scene wireless signal receiving device, perform time-frequency domain feature extraction and preprocessing on the multi-band interference signal data, and obtain a standardized interference feature vector.

[0085] The enhancement module 202 is configured to perform interference feature enhancement processing on the standardized interference feature vector to obtain an interference feature sample set.

[0086] The construction module 203 is configured to construct a dual-array cooperative detection system including a passive detection array and an active detection array according to the interference feature sample set, and perform real-time monitoring on a wireless environment by the dual-array cooperative detection system to obtain dual-path interference signal monitoring data.

[0087] The fusion module 204 is configured to perform spatial domain and time-frequency domain dual feature extraction and fusion processing on the dual-path interference signal monitoring data to obtain comprehensive interference features.

[0088] The adjustment module 205 is configured to dynamically adjust the receiver filter parameter matrix and the detection threshold vector of the dual-array cooperative detection system according to the comprehensive interference features.

[0089] Through the synergistic cooperation of the above various components, based on multi-scene multi-band interference signal collection and time-frequency domain feature extraction, the method comprehensively captures the feature performance of various interference sources in different environments, uses the Wasserstein generative adversarial network for interference feature enhancement processing, effectively solves the problem of insufficient actual interference signal samples, expands the training data set scale to 10 times the original data, and significantly enhances the learning ability of the model to various interference features. A dual-array cooperative detection system combining passive detection arrays and active detection arrays is adopted, and the spatial diversity technology is used to greatly improve the interference detection capability of the system in complex electromagnetic environments, which is especially suitable for complex scenes with multiple interference sources coexisting. Combined with spatial domain and time-frequency domain dual feature extraction and fusion processing mechanism, the feature differences of interference signals in different domains are fully utilized to improve the recognition ability of the system to weak interference and complex interference patterns. Based on the hierarchical interference signal detection model, interference signal detection, interference type recognition and interference parameter estimation are completed in turn to realize accurate recognition and parameter acquisition of different types of interference signals. Through iterative calculation, the receiver filter parameters and detection threshold are dynamically adjusted, so that the system can adaptively cope with the changes of different interference environments. The interference signal generation-disappearance process tracking model effectively captures the random appearance and disappearance characteristics of interference signals, provides the system with interference behavior prediction ability, and enhances the perception and response ability to burst interference. Through the two cooperative working modes of synchronous scanning and complementary scanning, the communication spectrum is fully covered, and the monitoring range and spectrum perception ability of the system are greatly expanded.

[0090] The above Figure 2 The wireless anti-interference signal detection device in the embodiment of the application is described in detail from the perspective of modular functional entities, and the wireless anti-interference signal detection device in the embodiment of the application is described in detail from the perspective of hardware processing.

[0091] Figure 3is a structural schematic diagram of a wireless anti-interference signal detection device provided by an embodiment of the present application. The wireless anti-interference signal detection device 300 can have great differences due to different configurations or performances, and can include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, one or more storage media 330 (for example, one or more mass storage device ends) storing application programs 333 or data 332. The memory 320 and the storage medium 330 can be temporary storage or persistent storage. The programs stored in the storage medium 330 can include one or more modules (not shown in the figure), and each module can include a series of instruction operations in the wireless anti-interference signal detection device 300. Further, the processor 310 can be configured to communicate with the storage medium 330 and execute a series of instruction operations in the storage medium 330 on the wireless anti-interference signal detection device 300 to realize the steps of the wireless anti-interference signal detection method described above.

[0092] The wireless anti-interference signal detection device 300 can also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that the wireless anti-interference signal detection device 300 can also include other components, and the components shown in the figure are not exhaustive. Figure 3 The structure of the wireless anti-interference signal detection device shown in the figure does not constitute a limitation on the wireless anti-interference signal detection device provided by the present application, and can include more or fewer components than shown in the figure, or combine certain components, or different component arrangements.

[0093] The present application also provides a computer readable storage medium, which can be a non-volatile computer readable storage medium, and can also be a volatile computer readable storage medium. The computer readable storage medium has instructions stored therein, and when the instructions are run on a computer, the computer executes the steps of the wireless anti-interference signal detection method.

[0094] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, system and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0095] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or the entire or part of the technical solutions that essentially contribute to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0096] The above description and the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for detecting wireless anti-interference signals, characterized in that, include: Multi-band interference signal data is collected by a multi-scenario wireless signal receiving device, and time-frequency domain feature extraction and preprocessing are performed on the multi-band interference signal data to obtain a standardized interference feature vector. The standardized interference feature vectors are subjected to interference feature enhancement processing to obtain an interference feature sample set; A dual-array cooperative detection system, comprising a passive detection array and an active detection array, is constructed based on the interference feature sample set. This system is used to monitor the wireless environment in real time, obtaining dual-channel interference signal monitoring data. Specifically, the system includes: tracking the generation and disappearance of interference signals based on the interference feature sample set to obtain an interference state transition probability matrix; configuring a passive detection array with M receiving antennas and an active detection array with N array elements according to the interference state transition probability matrix to obtain the dual-array cooperative detection system; setting the receiving parameters of the passive detection array, including low-noise amplifier gain, mixer parameters, and filter bandwidth; setting the digital phased array unit and beamforming parameters of the active detection array; establishing a high-speed data transmission link between the passive and active detection arrays, employing both synchronous scanning and complementary scanning cooperative working modes to obtain a dual-array cooperative monitoring mechanism; and using this mechanism to collect interference signals in the wireless environment in real time, recording the omnidirectional interference data received by the passive detection array and the directional interference data acquired by the active detection array to obtain dual-channel interference signal monitoring data. The dual-channel interference signal monitoring data are subjected to spatial domain and time-frequency domain dual feature extraction and fusion processing to obtain comprehensive interference features; The receiver filtering parameter matrix and detection threshold vector of the dual-array cooperative detection system are dynamically adjusted based on the comprehensive interference characteristics.

2. The wireless anti-interference signal detection method according to claim 1, characterized in that, The process involves collecting multi-band interference signal data using a multi-scenario wireless signal receiving device, extracting and preprocessing time-frequency domain features from the multi-band interference signal data to obtain a standardized interference feature vector, including: Multi-band interference signal data were obtained by performing full-spectrum scanning sampling in various electromagnetic environments using a multi-scenario wireless signal receiving device. Time-domain filtering and frequency-domain filtering are performed on the multi-band interference signal data to obtain a preliminarily purified interference signal; The interference signal that has been initially purified is subjected to amplitude normalization and time synchronization calibration to obtain standardized time-domain interference data; The standardized time-domain interference data is subjected to a short-time Fourier transform to obtain the time-frequency domain interference feature distribution; The spectral envelope, center frequency offset, spectral bandwidth, and time-frequency distribution pattern are extracted from the time-frequency domain interference feature distribution to obtain a multi-dimensional feature set, and a standardized interference feature vector is constructed based on the multi-dimensional feature set.

3. The wireless anti-interference signal detection method according to claim 1, characterized in that, The standardization interference feature vector is subjected to interference feature enhancement processing to obtain an interference feature sample set, including: Perform statistical distribution analysis on the standardized interference feature vector to obtain interference feature distribution parameters, and generate a random noise vector with a standard normal distribution based on the interference feature distribution parameters; The random noise vector of the standard normal distribution is input into the Wasserstein generative adversarial network. The generator network in the Wasserstein generative adversarial network performs a nonlinear transformation mapping to map the random noise vector of the standard normal distribution to the interference feature space, thereby obtaining candidate interference feature samples. The Wasserstein distance and gradient penalty value between the candidate interference feature sample and the standardized interference feature vector are calculated by the discriminator network in the Wasserstein generative adversarial network to obtain the sample authenticity score. Based on the sample authenticity score, high-quality synthetic interference feature samples are selected and retained, and the synthetic interference feature samples are merged with the standardized interference feature vector to obtain the interference feature sample set.

4. The wireless anti-interference signal detection method according to claim 1, characterized in that, The dual-domain and time-frequency domain feature extraction and fusion processing is performed on the dual-channel interference signal monitoring data to obtain comprehensive interference features, including: The spatial covariance matrix is ​​calculated based on the dual-channel interference signal monitoring data, and the spatial covariance matrix is ​​subjected to eigenvalue decomposition to obtain the eigenvalue sequence and the corresponding eigenvector set. The feature value sequence is analyzed based on the minimum description length criterion to determine the number of interference sources in the wireless environment and obtain spatial domain interference features. An S-transform is performed on each array element signal in the dual-channel interference signal monitoring data to generate a time-frequency distribution matrix. The spectral centroid, spectral bandwidth, spectral skewness, spectral kurtosis, and spectral entropy are extracted from the time-frequency distribution matrix to obtain the time-frequency domain interference characteristics. The spatial domain interference features and the time-frequency domain interference features are subjected to dimension normalization processing to obtain a standardized multi-domain feature set. The standardized multi-domain feature set and the dual-array detection results are used to calculate the fusion confidence level using Dempster-Shafer evidence theory, and the feature weight coefficients are dynamically adjusted according to the signal-to-noise ratio and interference characteristics to obtain the comprehensive interference features.

5. The wireless anti-interference signal detection method according to claim 4, characterized in that, The step of dynamically adjusting the receiver filtering parameter matrix and detection threshold vector of the dual-array cooperative detection system based on the comprehensive interference characteristics includes: The generalized likelihood ratio test is performed on the comprehensive interference features to obtain the test statistic. The test statistic is then compared with the initial threshold to obtain the result of the interference signal existence judgment. Based on the existence determination result of the interference signal, density peak clustering is performed on the comprehensive interference features to obtain the interference type identification result; Based on the interference type identification results, interference parameter estimation is performed to obtain interference source parameter information; Based on the interference source parameter information, the optimal system parameter combination is obtained by iteratively calculating and alternately optimizing the receiver filter parameter matrix and the detection threshold vector. Based on the optimal system parameter combination, the receiver filtering configuration and interference detection decision threshold of the dual-array cooperative detection system are dynamically adjusted to complete the anti-interference detection processing of multiple types of interference signals in complex electromagnetic environments.

6. The wireless anti-interference signal detection method according to claim 5, characterized in that, The step of obtaining the optimal system parameter combination by iteratively calculating and alternately optimizing the receiver filter parameter matrix and detection threshold vector based on the interference source parameter information includes: The interference source parameter information is used to initialize the receiver filter parameter matrix and detection threshold vector, and the objective function is set to maximize the minimum interference detection probability across the entire frequency band. The optimization objective function is transformed into an equivalent form by introducing auxiliary variables, and the upper and lower bounds and accuracy requirements of the binary search are set to obtain the initialized binary search interval. In each iteration, the target detection probability is set to the value of the midpoint of the initialized binary search interval, the detection threshold vector is fixed, and the receiver filter parameter matrix is ​​expanded in first order Taylor at the current point to obtain the convex approximation problem. The convex approximation problem is solved using the interior point method, the receiver filter parameter matrix is ​​updated, and the detection threshold vector is optimized under the condition of fixing the updated receiver filter parameter matrix to obtain the system parameter combination for the current iteration. Determine whether the interference detection performance of the current combination of system parameters across the entire frequency band meets the target detection probability. If it does, update the lower bound; otherwise, update the upper bound to obtain a new search interval. The iteration stops when the length of the new search interval is less than the accuracy requirement, and the optimal combination of system parameters is output.

7. A wireless anti-interference signal detection device, characterized in that, For performing the wireless anti-interference signal detection method as described in any one of claims 1-6, the wireless anti-interference signal detection device comprises: The acquisition module is used to acquire multi-band interference signal data through a multi-scenario wireless signal receiving device, and to perform time-frequency domain feature extraction and preprocessing on the multi-band interference signal data to obtain a standardized interference feature vector. The enhancement module is used to perform interference feature enhancement processing on the standardized interference feature vector to obtain an interference feature sample set; The construction module is used to construct a dual-array collaborative detection system containing a passive detection array and an active detection array based on the interference feature sample set, and to monitor the wireless environment in real time through the dual-array collaborative detection system to obtain dual-channel interference signal monitoring data. The fusion module is used to perform spatial domain and time-frequency domain dual feature extraction and fusion processing on the dual-channel interference signal monitoring data to obtain comprehensive interference features; The adjustment module is used to dynamically adjust the receiver filtering parameter matrix and detection threshold vector of the dual-array cooperative detection system according to the comprehensive interference characteristics.

8. A wireless anti-interference signal detection device, characterized in that, The wireless anti-interference signal detection device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the wireless anti-interference signal detection device to perform the wireless anti-interference signal detection method as described in any one of claims 1-6.

9. A computer-readable storage medium storing instructions thereon, characterized in that, When the instruction is executed by the processor, it implements the wireless anti-interference signal detection method as described in any one of claims 1-6.

Images