Parameter measurement method for frequency modulation monitoring antenna

Abstract

The invention relates to the technical field of signal parameter measurement, and provides a parameter measurement method for a frequency modulation monitoring antenna, which realizes self-adaptive monitoring and accurate measurement of a complex electromagnetic environment through multi-dimensional data acquisition, intelligent analysis and dynamic optimization, and deeply integrates environment perception, signal processing and parameter optimization. A complete technical chain is formed, so that the parameter measurement method can have self-learning and self-adjusting capabilities, can cope with environmental changes and interference challenges in real time, remarkably improves the monitoring efficiency and the measurement precision, and provides a systematic solution for the application of a frequency modulation monitoring antenna in a complex scene.

Description

Technical Field

[0001] This application relates to the technical field of signal parameter measurement, and particularly to a parameter measurement method for a frequency modulation monitoring antenna. Background Art

[0002] In the field of frequency modulation monitoring, traditional antenna parameter measurement methods mainly rely on single-dimensional data acquisition and analysis. In early technologies, fixed monitoring frequencies and beam directions were manually set to collect signals in the electromagnetic environment, and simple signal processing algorithms were used to analyze basic parameters such as signal strength and frequency. With the development of technology, sensor networks have been gradually introduced to achieve automated acquisition of electromagnetic data, and algorithms such as adaptive filtering have been adopted to suppress some interference.

[0003] In recent years, with the rise of machine learning technology, some solutions have tried to perform preliminary pattern recognition on electromagnetic signals through convolutional neural networks to distinguish common interference types, and at the same time use optimization means such as genetic algorithms to adjust the scanning frequency of the antenna, which has improved the monitoring efficiency to a certain extent. In terms of environmental factors, some studies have also considered temperature and humidity data, and corrected the influence of the environment on signal propagation through empirical formulas or simple linear models. In addition, some solutions have begun to use spatial geometric models to describe the influence of reflectors on signals, and ray tracing algorithms to simulate the multipath propagation path of signals. Most of them have not solved the problem of how to generate signals based on multi-source heterogeneous data to adapt to the adaptive optimization of complex monitoring environments and monitoring parameters. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, this application provides a parameter measurement method for a frequency modulation monitoring antenna. The method includes: obtaining interference data of the monitoring area in real time, where the interference data includes environmental electromagnetic data, temperature and humidity data, and reflector coordinates, performing pattern recognition on the interference data through a convolutional neural network to output interference features, where the interference features include interference types and spatio-temporal intensity distributions, and constructing a multi-objective optimization function to process the interference features to generate a test signal; Injecting the test signal into the frequency modulation monitoring antenna, while receiving and monitoring the backscattered signal, obtaining the scattering parameters, energy distribution, and phase delay of the backscattered signal in real time, and comprehensively judging the signal error between the test signal and the backscattered signal based on the scattering parameters, energy distribution, and phase delay, and dynamically compensating the backscattered signal according to the signal error; Optimizing the beam direction and scanning frequency of the monitoring area according to the compensated backscattered signal, re-obtaining the interference data of the monitoring area based on the optimized beam direction and scanning frequency of the monitoring area, updating the interference features at the same time, and updating the parameter combination of the test signal through a genetic algorithm.

[0005] As an optional implementation manner, the generation strategy of the test signal includes: Construct a generative adversarial network. The generator processes interference features to output candidate parameter combinations of the test signal, and the discriminator evaluates the candidate parameter combinations of the test signal through a multi-objective optimization function; Configure a physical constraint layer according to the properties of the frequency modulation monitoring antenna, and screen the candidate parameter combinations of the test signal according to the physical constraint layer; Optimize the candidate parameter combinations of the test signal through the simulated annealing algorithm. After multiple iterative trainings of the generator and the discriminator, generate the test signal. The parameter combinations of the test signal include the frequency step size, modulation index, and duty cycle.

[0006] As an optional implementation manner, the output sub-strategy of the interference features includes: Obtain interference data including environmental electromagnetic data, temperature and humidity data, and reflector coordinates in the monitoring area in real time, and align the interference data according to the time stamp; Convert the environmental electromagnetic data into a time-frequency map matrix through short-time Fourier transform, process the temperature and humidity data through Kriging interpolation to generate a temperature and humidity field distribution, and construct a spatial topology of the reflector coordinates through the octree algorithm; Construct a three-branch convolutional neural network including an electromagnetic branch, an environmental branch, and a spatial branch. Extract the features of the interference data through the three-branch convolutional neural network, and perform weighted fusion through the attention mechanism to generate interference features including interference types and spatio-temporal intensity distributions; Compare and analyze the interference features with the historical feature data to determine whether to adjust the parameters of the three-branch convolutional neural network and the attention weights.

[0007] As an optional implementation manner, the electromagnetic branch identifies the interference patterns of the time-frequency map matrix through a residual network, the environmental branch extracts the gradient changes of the temperature and humidity field distribution through a two-dimensional convolutional layer, and the spatial branch analyzes the spatial topology through a graph convolutional network and extracts the path complexity of multipath reflection.

[0008] As an optional implementation manner, the construction sub-strategy of the multi-objective optimization function includes: Regard interference suppression, antenna matching, and propagation efficiency as optimization objectives, and dynamically adjust the objective weights of the optimization objectives according to the interference type through a Bayesian network; Construct a multi-objective optimization function by weighted summation of the optimization objectives and the objective weights.

[0009] As an optional implementation manner, the compensation strategy for the backscattered signal includes: Extract the dominant component of the signal error, and determine the compensation direction of the backscattered signal according to the dominant component of the signal error; Based on the compensation direction of the backscattered signal, coordinately adjust the parameter combination of the test signal to compensate for the backscattered signal; Real-time monitor the signal quality of the backscattered signal after compensation. The signal quality includes the convergence degree and energy concentration of the scattering parameters, and judge according to the signal quality to update the adjustment step of the parameter combination of the test signal and switch the compensation direction of the backscattered signal.

[0010] As an alternative implementation, the sub-strategy for judging the signal error between the test signal and the backscattered signal includes: Inject the test signal into the frequency modulation monitoring antenna, receive the backscattered signal, and obtain the scattering parameters, energy distribution and phase delay of the backscattered signal in real time; Perform cross-correlation analysis on the scattering parameters, energy distribution and phase delay of the backscattered signal to obtain a correlation matrix, and determine the error weights of the scattering parameters, energy distribution and phase delay according to the correlation matrix; Judge the signal error between the test signal and the backscattered signal by weighted summation according to the error weights.

[0011] As an alternative implementation, scan the monitoring area based on the field strength probe, perform three-dimensional spatial sampling on the backscattered signal, obtain the energy distribution of the backscattered signal by acquiring the energy value of each sampling point; compare the carrier phases of the test signal and the backscattered signal through coherent detection technology to obtain the phase delay of the backscattered signal.

[0012] As an alternative implementation, the optimization strategy for the beam direction of the monitoring area includes: Map the energy distribution and phase delay of the compensated backscattered signal to the direction-of-arrival space through Fourier transform, construct an energy density map, and identify the energy concentration area through the peak detection algorithm and mark it as the target signal direction; Combine the temperature and humidity field distribution of the monitoring area and the spatial topology of the reflector to evaluate the propagation feasibility of the beam direction; Through the reinforcement learning algorithm, use the received power of the target signal and the interference energy as the reward function to dynamically optimize the beam direction of the monitoring area.

[0013] As an alternative implementation, the optimization strategy for the scanning frequency of the monitoring area includes: Divide the scattering parameters of the compensated backscattered signal into multiple frequency bands according to the frequency, and combine the monitoring area after optimizing the beam direction to analyze the signal attenuation and interference intensity distribution of each frequency band in the monitoring area to identify the scanning frequency band and frequency step size; Obtain the spectrum occupancy status of the monitoring area in real time through spectrum sensing technology, combine the interference type to judge the spectrum change trend, and mark the risk frequency band and idle frequency band; Determine the constraint conditions by combining the scanning frequency band, frequency step size, and spectrum change trend, and optimize the scanning frequency of the monitoring area through the genetic algorithm.

[0014] Compared with the prior art, the beneficial effects of this application are as follows: Through multi-dimensional data acquisition, intelligent analysis, and dynamic optimization, the adaptive monitoring and accurate measurement of complex electromagnetic environments are realized. This method deeply integrates environmental perception, signal processing, and parameter optimization to form a complete technical chain, enabling the parameter measurement method to have the capabilities of self-learning and self-adjustment, being able to respond to environmental changes and interference challenges in real time, significantly improving the monitoring efficiency and measurement accuracy, and providing a systematic solution for the application of frequency modulation monitoring antennas in complex scenarios.

[0015] Multi-source data fusion constructs a multi-dimensional feature space to comprehensively represent the physical characteristics, spatial distribution, and environmental correlations of interference, providing a complete information basis for subsequent analysis. The deep feature extraction ability of the convolutional neural network automatically identifies complex interference patterns, avoiding the limitations of manual feature engineering, and improving the accuracy of interference type judgment and the positioning accuracy of spatio-temporal intensity distribution. The multi-objective optimization function makes the generated test signal more suitable for the actual monitoring scenario through dynamic weight adjustment.

[0016] Simultaneously monitor the scattering parameters, energy distribution, and phase delay of the backscattered signal to construct a complete signal feature vector, comprehensively evaluate the signal transmission state, avoid the one-sidedness of single-parameter evaluation, comprehensively judge the signal error between the test signal and the backscattered signal, and dynamically compensate the backscattered signal according to the signal error to effectively suppress environmental interference and improve the signal quality and stability. Description of the Drawings

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Among them: Figure 1 It is the flowchart of the parameter measurement method for the frequency modulation monitoring antenna provided by the embodiment of this application; Figure 2 It is the output sub-strategy diagram of the interference characteristics of the parameter measurement method for the frequency modulation monitoring antenna provided by the embodiment of this application; Figure 3 It is the optimization strategy diagram of the beam pointing of the monitoring area of the parameter measurement method for the frequency modulation monitoring antenna provided by the embodiment of this application. Detailed Embodiments

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings of the specification. Apparently, the described embodiments are only a part rather than all of the embodiments of this application.

[0019] Embodiment As Figure 1 shown, the present application provides a method flow chart of a parameter measurement method for a frequency modulation monitoring antenna. The method includes: S1. Obtain interference data of the monitoring area in real time. The interference data includes environmental electromagnetic data, temperature and humidity data, and reflector coordinates. Perform pattern recognition on the interference data through a convolutional neural network to output interference features. The interference features include interference types and spatio-temporal intensity distributions, and construct a multi-objective optimization function to process the interference features to generate a test signal.

[0020] As Figure 2 shown, the output sub-strategy of the interference features includes: Obtain interference data including environmental electromagnetic data, temperature and humidity data, and reflector coordinates in the monitoring area in real time, and align the interference data according to the time stamp; Convert the environmental electromagnetic data into a time-frequency map matrix through short-time Fourier transform, generate a temperature and humidity field distribution by processing the temperature and humidity data through Kriging interpolation method, and construct a spatial topology of the reflector coordinates through an octree algorithm; Construct a three-branch convolutional neural network including an electromagnetic branch, an environmental branch, and a spatial branch. Extract the features of the interference data through the three-branch convolutional neural network, and perform weighted fusion through an attention mechanism to generate interference features including interference types and spatio-temporal intensity distributions; Compare and analyze the interference features with historical feature data to determine whether to adjust the parameters and attention weights of the three-branch convolutional neural network.

[0021] There are differences in the environmental electromagnetic data, temperature and humidity data, and reflector coordinates of the monitoring area in terms of data type, acquisition frequency, and format, which need to be uniformly processed to provide a standardized input for subsequent feature extraction. At the same time, the impacts of different data on signal propagation and interference are correlated in the time dimension, and aligning by timestamp can accurately reflect the dynamic changes of interference characteristics. The environmental electromagnetic data is obtained through a spectrum analyzer to acquire the signal intensity distribution in different frequency bands, the temperature and humidity data is obtained through temperature and humidity sensors, and the monitoring area is scanned in real time by lidar to obtain the three-dimensional coordinate data of the reflector. Then, through the timestamp matching algorithm, the environmental electromagnetic data, temperature and humidity data, and reflector coordinates at the same time point are integrated into interference data. Thus, the standardized processing and time synchronization of interference data are achieved, avoiding analysis deviations caused by data asynchronization. The unified data format and structure provide a good foundation for the efficient processing of the subsequent three-branch convolutional neural network, improving the accuracy and efficiency of interference feature extraction. The preprocessed and fused interference data is used as the input of the three-branch convolutional neural network to ensure that the three-branch convolutional neural network can process data features in different dimensions in parallel, laying a foundation for the accurate extraction of interference features.

[0022] The environmental electromagnetic data has time-varying characteristics. Among them, impulse interference shows sudden spikes in the time domain and has wide-spectrum characteristics in the frequency domain. Single time-domain or frequency-domain analysis cannot comprehensively capture interference features. The continuous environmental electromagnetic data is segmented into multiple short time periods of fixed length, and a Hanning window function is applied to each period to reduce spectral leakage. Then, the fast Fourier transform is performed on each windowed period to convert the time-domain signal into a frequency-domain representation, obtaining the spectral distribution within that period. The spectra of all periods are arranged in chronological order to form a two-dimensional matrix, where the horizontal axis represents time and the vertical axis represents frequency, and the matrix element value is the signal intensity at the corresponding time and frequency points. Thus, the environmental electromagnetic data is converted into a visual time-frequency map matrix. The time-frequency map matrix intuitively shows the dynamic change characteristics of electromagnetic interference. Preferably, narrowband interference appears as a continuous high-energy line at a fixed frequency in the time-frequency map, while impulse interference appears as a wide-spectrum burst within a short time, which enables the subsequent neural network to more accurately identify different types of interference patterns.

[0023] The temperature and humidity sensors in the monitoring area are usually discretely distributed. Since the signal propagation is affected by the temperature and humidity changes in the continuous space, it is necessary to convert the discrete sensor data into a continuous temperature and humidity field distribution through an interpolation method to quantify the impact of environmental factors on signal propagation. Deploy multiple temperature and humidity sensors in the monitoring area according to a certain grid density, regularly obtain the temperature and humidity values at each point, and analyze the spatial correlation between the data of each sensor, that is, the temperature and humidity values of points closer in distance are more similar, and the correlation decreases with increasing distance. Based on the spatial correlation, for any point without a sensor installed in the monitoring area, calculate the estimated temperature and humidity value of this point according to the known sensor data around it and its spatial position relationship. Through iterative calculations, finally generate the temperature and humidity field distribution of the entire monitoring area. The Kriging interpolation method fully considers the spatial correlation of temperature and humidity data. Compared with simple linear interpolation methods, it can more accurately restore the distribution of the actual temperature and humidity field. Especially in areas with sparse sensor distributions, it can also provide relatively reliable estimated values.

[0024] The reflectors in the monitoring area are irregularly distributed and numerous. The reflectors include buildings, metal structures, etc. A data structure is needed to represent their spatial relationships for subsequent analysis of the multipath propagation effect of signals. The octree simplifies the complex spatial scene into a hierarchical structure by recursively dividing the three-dimensional space, which is convenient for quickly querying and analyzing the topological relationships between reflectors. Define the entire monitoring area as the initial cubic space, and then recursively divide the cube into eight equal sub-cubes until the number of reflectors contained in each sub-cube is less than a preset number threshold. Each cube node contains a pointer to the sub-cube and a list of reflectors within the cube. In this way, the reflectors in the three-dimensional space are organized into a hierarchical tree structure, and each node represents a specific spatial region. Based on the octree structure, analyze the connection relationships between adjacent spatial regions and the reflector distribution density, and extract characteristic parameters reflecting the spatial topological characteristics, including the degree of reflector clustering and spatial connectivity, etc. The octree structure greatly improves the efficiency of spatial query and analysis, enabling quick positioning and analysis of the spatial characteristics of specific regions when dealing with a large amount of reflector data. At the same time, this hierarchical structure is also convenient for subsequent graph convolutional networks to extract features such as the path complexity of multipath reflections.

[0025] Environmental electromagnetic data, temperature and humidity data, and the information characteristics contained in reflector coordinates are different. It is difficult for a single network structure to effectively extract their respective key features. The three-branch convolutional neural network performs feature extraction for data characteristics through different network architectures. Combining the attention mechanism can further highlight important features and achieve effective fusion of multi-dimensional features. A three-branch convolutional neural network including an electromagnetic branch, an environmental branch, and a spatial branch is constructed. Among them, the electromagnetic branch performs multi-layer convolution operations on the time-frequency map matrix through a residual network, and solves the problem of gradient disappearance caused by the increase in network depth through residual connections, and extracts interference patterns in the time-frequency map matrix, including features such as spectral spikes of narrowband interference and continuous spectral distributions of broadband interference. The environmental branch processes the temperature and humidity field distribution through a two-dimensional convolutional layer, extracts the gradient change features of the temperature and humidity field distribution through the convolutional kernel, captures the spatial change trends of temperature and humidity, and quantifies the impact of environmental factors on signal propagation. The spatial branch takes the spatial topological structure of the reflector as the input through a graph convolutional network, analyzes the connection relationship between reflectors through information transfer and aggregation between nodes, and extracts the path complexity of multipath reflection.

[0026] Taking the features extracted by the three branches as inputs, the multi-head attention mechanism automatically assigns weights according to the importance of different features for interference recognition, and performs weighted fusion to generate interference features including interference types and spatio-temporal intensity distributions. The three-branch convolutional neural network performs feature extraction for different data characteristics, fully leveraging the advantages of each network architecture. Compared with a single network structure, it can extract interference features more comprehensively and accurately. Concatenate the features output by the three branches into total features, process the total features through a fully connected network to output the weight values of the three-branch network, perform normalization processing on the output weight values to ensure that the sum of all weights is 1, multiply the features of each branch by the corresponding weight values, and then sum to obtain the final interference features. Then, map the fused interference features to interference types and spatio-temporal intensity distributions through a fully connected layer. Among them, the interference type is the interference pattern extracted in the electromagnetic branch, and the spatio-temporal intensity distribution includes time-series intensity and spatial-region intensity. The attention mechanism further enhances the weights of key features, effectively suppresses redundant information, improves the quality and distinguishability of interference features. The generated interference features contain rich interference information, which is used as the input of the multi-objective optimization function, providing a key basis for constructing the objective function and generating test signal parameters.

[0027] The interference environment in the monitoring area is dynamically changing, including the movement of interference sources and the fluctuations of environmental temperature and humidity. Fixed network parameters cannot adapt to these changes. By comparing the current interference characteristics with historical feature data and dynamically adjusting the parameters and attention weights of the three-branch convolutional neural network, the three-branch convolutional neural network can continuously maintain the ability to accurately extract interference characteristics. Compare the currently generated interference characteristics with historical feature data, calculate the similarity between features through the cosine similarity algorithm, and set a similarity threshold. If the average similarity between the current interference characteristics and historical feature data is less than the similarity threshold, it indicates that the interference environment has changed significantly. At this time, trigger the backpropagation algorithm, use the current interference data and expected output as training samples to fine-tune the parameters of the three-branch convolutional neural network, and at the same time update the weight parameters of the attention mechanism to enhance the response ability to newly emerging interference characteristics. Thus, the adaptive adjustment of the three-branch convolutional neural network to the dynamic interference environment is realized, which can timely capture the changes of interference characteristics, avoid the failure of feature extraction caused by environmental changes, ensure the accuracy and stability of the interference feature output, and the calibrated interference features can better reflect the current actual interference situation, providing more accurate data support for the subsequent construction of the multi-objective optimization function and helping to generate more effective test signal parameters.

[0028] Furthermore, the electromagnetic branch identifies the interference pattern of the time-frequency map matrix through the residual network, the environmental branch extracts the gradient change of the temperature and humidity field distribution through the two-dimensional convolutional layer, and the spatial branch analyzes the spatial topology through the graph convolutional network and extracts the path complexity of multipath reflection.

[0029] The time-frequency map matrix contains a large amount of distribution information of electromagnetic signals in the time and frequency dimensions. It is difficult for traditional convolutional networks to effectively extract deep features when increasing the number of layers, and the problem of gradient disappearance is likely to occur. However, through the residual connection structure, the residual network can mine complex interference patterns without losing information. Here, a residual network is constructed to identify the time-frequency map matrix. A convolutional layer is set at the front end of the residual network for preliminary feature extraction. Each residual block includes a convolutional layer and a skip connection structure, which are used to learn the detailed features of the interference pattern. Local features in the time-frequency map matrix are captured through convolutional operations, where local features include spectral spikes and time-domain pulses. The shallow features and deep features are fused using the skip connection to enhance the expression ability of the residual network for interference patterns. Preferably, for narrowband interference, the residual network can learn the continuous signal intensity change at a fixed frequency through the residual block, while for impulse interference, the burst energy characteristics in the time domain can be identified.

[0030] At the end of the residual network, a global average pooling layer is used to compress the local features of the time-frequency map matrix into a one-dimensional vector, and then the classification results of interference patterns are output through a fully connected layer. The classification results of interference patterns include narrowband interference, broadband interference, pulse interference, etc. Compared with ordinary convolutional networks, residual networks can extract interference features at deeper levels, avoid gradient disappearance, improve the recognition accuracy of complex time-varying interference, and can automatically learn the feature combinations of different interference patterns without the need for manual design of feature extraction rules. The output interference patterns, as the core part of interference features, provide key inputs for the multi-objective optimization function, such as guiding the selection of test signal frequencies to avoid interference frequency bands.

[0031] The gradient changes in the temperature and humidity field distribution directly affect the signal propagation path and quality. Among them, the signal refraction caused by the temperature gradient and the signal attenuation caused by the humidity difference need to capture the temperature and humidity change trends in the spatial dimension through a two-dimensional convolutional layer to quantify the impact of environmental factors on the signal. The temperature and humidity field distribution generated by Kriging interpolation is sliced into multiple two-dimensional matrices according to the height dimension. Each matrix represents the temperature and humidity distribution of a certain height plane and serves as the input of the two-dimensional convolutional layer. The receptive field is gradually expanded through convolutional kernels of different sizes, and the local gradient change features of the temperature and humidity field are extracted through the convolutional layer. The local gradient change features include the temperature gradient direction and the humidity mutation region, and the activation function is used to enhance the non-linear expression of the features. The outputs of each convolutional layer are concatenated in channels, and then the most significant gradient features are extracted through the global maximum pooling layer, and the output is the feature vector of the temperature and humidity field gradient, including parameters such as the temperature gradient intensity and the humidity change rate. The two-dimensional convolutional layer can efficiently capture the spatial change patterns of the temperature and humidity field, convert the impact of environmental factors on the signal into a quantifiable feature vector, and improve the feature extraction efficiency.

[0032] The spatial topological structure of the reflector determines the signal multipath propagation path. Traditional methods are difficult to handle irregular spatial relationships. Through information transmission between nodes, the graph convolutional network can effectively analyze the connection relationship of reflectors, extract the path complexity of multipath reflection, and the path complexity includes the number of paths and the number of reflections. The spatial topology of the reflector generated by the octree algorithm is transformed into a graph structure. Each reflector is used as a node, and the node features include position coordinates and reflection coefficients. The edges between nodes represent the spatial proximity relationship between reflectors. The weights of the edges are set according to the distance and the signal reflection probability. By aggregating the feature information of adjacent nodes, the node state is updated. Through multi-layer graph convolutional operations, the node features gradually fuse the global spatial information. A fully connected layer is set at the end of the graph convolutional network to map the node features to the path complexity of multipath reflection, including the average path length, the distribution of the number of reflections, and the direction of the strongest reflection path, etc. The graph convolutional network breaks through the limitations of traditional grid data, can accurately analyze the multipath effect in the irregular spatial topology, reduce the path complexity evaluation error, and the output path complexity of multipath reflection is used to optimize the parameters of the test signal. Preferably, the modulation index is adjusted to reduce the inter-symbol interference caused by multipath interference and assist in beam pointing optimization to avoid the direction of the strong reflection path.

[0033] The construction sub-strategies of the multi-objective optimization function include: Taking interference suppression, antenna matching, and propagation efficiency as optimization objectives, and dynamically adjusting the objective weights of the optimization objectives according to the interference type through the Bayesian network; Construct a multi-objective optimization function by weighted summation of the optimization objectives and the objective weights.

[0034] During the frequency modulation monitoring process, the interference signal will affect the measurement accuracy. The matching degree between the test signal and the antenna determines the signal transmission efficiency, and complex environmental factors will affect the signal propagation effect. Therefore, it is necessary to set optimization objectives from three dimensions: interference suppression, antenna matching, and propagation efficiency to improve the overall monitoring performance. The interference suppression objective is used to reduce the impact of the interference signal on the test signal, reduce the overlap of the interference signal and the test signal in the frequency spectrum, and improve the signal-to-noise ratio of the received signal. Antenna matching is to ensure that the parameters of the test signal, such as frequency and polarization mode, match the frequency response and polarization characteristics of the frequency modulation monitoring antenna, and maximize the receiving efficiency of the antenna for the test signal. Propagation efficiency is to consider environmental factors such as temperature, humidity, and reflectors in the monitoring area, optimize the parameters of the test signal, reduce attenuation and multipath effects during signal propagation, and improve the integrity and accuracy of the signal reaching the receiving end. Thus, the direction and focus of optimization are clarified, providing a clear framework for subsequent weight allocation and objective function construction, enabling the generation of test signals to comprehensively consider various factors, using the three core objectives as the basis, guiding the Bayesian network for weight allocation, and then affecting the construction of the multi-objective optimization function and the generation strategy of test signal parameters.

[0035] Different types of interference affect parameter monitoring in different ways and degrees. Among them, narrowband interference mainly affects the spectral purity of the signal, while broadband pulse interference will damage the time-domain characteristics of the signal. Therefore, it is necessary to dynamically adjust the weights of each optimization objective according to the interference type to achieve more effective optimization. A mapping relationship table between interference types and target weights is pre-constructed. Preferably, for narrowband interference, the target weight of interference suppression is set to 0.6, the target weight of antenna matching is 0.3, and the target weight of propagation efficiency is 0.1. For broadband pulse interference, the target weights are adjusted to 0.3, 0.2, and 0.5 respectively. According to the real-time interference characteristics, combined with historical data and prior probabilities, the Bayesian inference algorithm dynamically calculates and updates the target weights of each optimization objective. Preferably, when it is detected that the interference type changes from narrowband interference to broadband pulse interference, the Bayesian network automatically adjusts the target weights and enhances the target weight of propagation efficiency to cope with problems such as multipath effects brought by broadband pulse interference. Thus, the multi-objective optimization function can adaptively adjust the optimization focus according to the change of interference type. Compared with the fixed weight scheme, it can optimize more precisely for different interference scenarios. The dynamically adjusted target weights directly participate in the calculation of the multi-objective optimization function, determining the importance of each optimization objective in the multi-objective optimization function, thereby affecting the generation direction of the parameter combination of the test signal.

[0036] The dynamic weights need to be combined with the core optimization objectives to form a quantitative evaluation criterion for judging the quality of parameter combinations of different test signals, so as to guide the generation of test signals. The multi-objective optimization function is constructed by weighted summation. The three objectives of interference suppression, antenna matching, and propagation efficiency are multiplied by their corresponding target weights respectively, and then added to obtain the target function value. At the same time, physical constraint conditions are built into the multi-objective optimization function. According to the hardware characteristics such as the working frequency band, bandwidth, and maximum transmission power of the frequency modulation monitoring antenna, as well as the physical laws of electromagnetic propagation, that is, the relationship between signal frequency and wavelength, to limit the value range of test signal parameters. Preferably, it is stipulated that the frequency step size must be within the working frequency band of the antenna and meet certain frequency resolution requirements, and the modulation index cannot exceed the maximum non-linear distortion range that the antenna can handle, etc. The constructed multi-objective optimization function can comprehensively consider multiple optimization objectives and physical constraints, and intuitively evaluate the quality of parameter combinations of different test signals through the quantitative target function value, providing a clear optimization direction and screening criterion for the generation of test signals. The multi-objective optimization function is used as the basis for the discriminator to evaluate the candidate parameter combinations of the test signal. During the test signal generation process, it guides the generator and discriminator to perform iterative training and parameter optimization until the parameter combination of the test signal that meets the requirements is generated.

[0037] The generation strategy of the test signal includes: Build a generative adversarial network. The generator processes interference features to output candidate parameter combinations of the test signal, and the discriminator evaluates the candidate parameter combinations of the test signal through a multi-objective optimization function; Configure a physical constraint layer according to the properties of the frequency modulation monitoring antenna, and screen the candidate parameter combinations of the test signal according to the physical constraint layer; Optimize the candidate parameter combinations of the test signal through the simulated annealing algorithm. After multiple iterative trainings of the generator and the discriminator, a test signal is generated. The parameter combinations of the test signal include the frequency step size, the modulation index, and the duty cycle.

[0038] There are numerous parameter combinations for the test signal. Traditional exhaustive methods or random search methods are inefficient and difficult to find the optimal solution. The generative adversarial network has powerful generative and learning capabilities and can quickly search for potential high-quality parameter combinations in the high-dimensional parameter space. Build a generative adversarial network, which includes a generator and a discriminator. Design a multi-layer fully connected neural network as the generator. The generator learns the mapping relationship between interference features and the parameter combinations of the test signal to generate candidate parameter combinations, namely the frequency step size, the modulation index, and the duty cycle. The discriminator also uses a multi-layer fully connected neural network. The discriminator evaluates by comparing the candidate parameter combinations output by the generator with the real parameter combinations according to the multi-objective optimization function and outputs a probability value to characterize the degree to which the candidate parameter combination is close to the optimal solution.

[0039] Optimize the generator and the discriminator through adversarial training. The generator attempts to generate parameter combinations that can deceive the discriminator, while the discriminator continuously improves its ability to distinguish the generated parameter combinations. During the training process, the generator adjusts its own parameters according to the feedback of the discriminator and gradually generates better candidate parameter combinations. Through the powerful generative ability of the generative adversarial network, a large number of high-quality candidate parameter combinations of the test signal can be generated in a short time. Compared with traditional search methods, the efficiency and coverage of parameter search are greatly improved, providing a rich candidate resource for subsequent screening of the optimal parameter combination. The generated candidate parameter combinations, as preliminary results, enter the physical constraint layer for screening, eliminating combinations that do not meet the actual physical conditions and retaining feasible parameter combinations for the next optimization.

[0040] The candidate parameter combinations generated by the generator do not conform to the physical characteristics and electromagnetic propagation laws of the FM monitoring antenna, including the frequency exceeding the operating frequency band of the antenna and the modulation index being too large resulting in signal distortion, etc. Therefore, it is necessary to screen the candidate parameter combinations through the physical constraint layer to ensure that the generated test signals can operate effectively in actual devices; according to the properties of the FM monitoring antenna, including technical parameters and the basic principles of electromagnetic propagation, physical constraint conditions are set, where the technical parameters include the settings of the operating frequency band, bandwidth, and maximum modulation index, and the basic principles of electromagnetic propagation include the relationship between signal frequency and wavelength and the limitations of duty cycle and pulse repetition frequency, etc.; each candidate parameter combination output by the generator is compared with the physical constraint conditions one by one. For the frequency step size, check whether it is within the operating frequency band of the antenna. For the modulation index, judge whether it exceeds the maximum non-linear distortion range that the antenna can withstand. For the duty cycle, ensure that it meets the timing requirements of signal transmission. Eliminate the parameter combinations that do not meet any constraint conditions and retain the candidate parameter combinations that meet all physical constraints for optimization; through the filtering of the physical constraint layer, the infeasible parameter combinations are excluded, ensuring that the parameter combinations used in the subsequent optimization process are feasible in actual applications, reducing the invalid calculations and engineering implementation risks of test signal generation. The screened candidate parameter combinations are used as effective solutions and are globally optimized through the simulated annealing algorithm to obtain the optimal parameter combination of the test signal.

[0041] Although the candidate parameter combinations screened by the physical constraint layer are feasible, they are not necessarily the optimal solutions. The simulated annealing algorithm has the ability of global search and can search for the globally optimal or approximately globally optimal parameter combinations in the feasible solution space, so as to generate the test signal with the best performance. The screened candidate parameter combinations are used as the initial solutions of the simulated annealing algorithm, and algorithm parameters such as the initial temperature, temperature decrease rate, and number of iterations are set. In each iteration, a neighborhood solution is generated based on the current solution. The neighborhood solution can be obtained by making small random adjustments to the parameters. Calculate the objective function values of the current solution and the neighborhood solution. If the objective function value of the neighborhood solution is better, accept the neighborhood solution as the new current solution. If the neighborhood solution is worse, accept the neighborhood solution with a certain probability, and this probability decreases as the temperature decreases. By continuously iterating and gradually reducing the temperature, the algorithm converges to the globally optimal or approximately globally optimal solution. When the preset number of iterations is reached or the temperature drops to a certain extent, the algorithm stops, and the finally obtained optimal solution is used as the parameter combination of the test signal to generate the final test signal. The simulated annealing algorithm can perform global search in the complex parameter space, effectively avoid falling into the local optimal solution, and can find a better parameter combination of the test signal compared with other local search algorithms, significantly improving the performance of the test signal, thereby improving the parameter measurement accuracy and reliability of the frequency modulation monitoring antenna. The generated test signal is injected into the frequency modulation monitoring antenna to drive subsequent processes such as backscattered signal reception, signal error judgment, and dynamic compensation, and is the key input signal for the entire frequency modulation monitoring method.

[0042] S2. Inject the test signal into the frequency modulation monitoring antenna, and at the same time receive and monitor the backscattered signal, obtain the scattering parameters, energy distribution, and phase delay of the backscattered signal in real time, and comprehensively judge the signal error between the test signal and the backscattered signal based on the scattering parameters, energy distribution, and phase delay, and perform dynamic compensation on the backscattered signal according to the signal error.

[0043] The judgment sub-strategy for the signal error between the test signal and the backscattered signal includes: Inject the test signal into the frequency modulation monitoring antenna, and receive the backscattered signal, and obtain the scattering parameters, energy distribution, and phase delay of the backscattered signal in real time; Perform cross-correlation analysis on the scattering parameters, energy distribution, and phase delay of the backscattered signal to obtain a correlation matrix, and determine the error weights of the scattering parameters, energy distribution, and phase delay according to the correlation matrix; Judge the signal error between the test signal and the backscattered signal by weighted summation according to the error weights.

[0044] Based on the field intensity probe scanning the monitoring area, the backscattered signal is sampled in three dimensions. The energy distribution of the backscattered signal is obtained by obtaining the energy value of each sampling point. The carrier phase of the test signal and the backscattered signal is compared through coherent detection technology to obtain the phase delay of the backscattered signal.

[0045] After the test signal is injected into the antenna, the core characteristic parameters of the backscattered signal need to be obtained to evaluate the distortion during signal transmission and reflection. The scattering parameters, energy distribution and phase delay respectively reflect the reflection loss, spatial distribution and time delay characteristics of the signal, and are the key basis for judging signal errors. The vector network analyzer is used to scan the non-passing frequency bands in segments to obtain the scattering parameters of the backscattered signal, covering the main working frequency band of the FM monitoring antenna. The attenuation value of the analyzer is automatically adjusted according to the signal strength to ensure that both weak and strong signals can be accurately acquired. The time domain gate calibration is triggered every 100 sets of data to eliminate the system errors caused by transmission lines and connectors.

[0046] The energy distribution of the backscattered signal reflects the propagation and reflection characteristics of the signal in space. The interference source can be located and the multipath reflection path can be identified through three-dimensional spatial sampling, providing a basis for beam optimization. The detection units are arranged in the field strength probe array to cover the monitoring space. A distributed architecture is adopted. Before each scan, a standard signal is emitted by the built-in calibration source. The amplitude and phase consistency of all probes are calibrated, and the monitoring area is divided along the height direction. A spiral scanning path is adopted, starting from the center of the bottom layer and expanding outward in the form of an Archimedean spiral. After completing one layer, it rises to the next layer. In the area where the signal fluctuates violently, the sampling points are automatically added to reduce the spacing, and the area where the signal fluctuates violently is close to the metal reflector. The standard spacing is maintained in the signal stable area. All probes are synchronized with the GPS clock to ensure that the energy data obtained at the same time corresponds to the same signal state.

[0047] The raw data of each sampling point is digitally down-converted to convert the RF signal into a baseband signal. Square-law detection is then performed to obtain instantaneous power. A sliding window of fixed length is used to average filter the power data to eliminate the influence of random noise and improve the stability of energy measurement. Kriging interpolation is used to fill data in sparse sampling areas to generate a continuous energy distribution. This enables energy sampling in the monitoring area, clearly captures the energy distribution formed by multipath reflection, and detects potential interference sources in advance. The acquired energy distribution is used to construct an energy density map, providing a spatial distribution basis for beam pointing optimization.

[0048] The phase delay of the backscattered signal contains information such as the signal propagation path length and the number of reflections. By accurately measuring the phase delay, the path complexity of multipath reflections can be calculated to assist in locating the position of the reflector. Carrier synchronization between the test signal and the received signal is achieved through a phase-locked loop circuit. The received signal is mixed with a locally generated carrier of the same frequency and phase to obtain a zero-intermediate-frequency signal. Then, the phase delay is directly obtained by measuring the phase of the zero-intermediate-frequency signal. Parallel coherent detection channels are set up to simultaneously measure the phases of the reflected signals in different directions in each coherent detection channel. The zero-intermediate-frequency signal is orthogonally sampled to obtain I / Q data. The phase is calculated through the arctangent function, and the phase ambiguity problem is solved through the polyphase phase-locked loop algorithm. By comparing the phase changes of adjacent sampling points, the phase value is automatically adjusted to change continuously. The phase value is processed by a Kalman filter to fuse historical measurement values and current observation values, suppress random noise interference, and improve the phase measurement accuracy. An auxiliary frequency is superimposed on the test signal, and the phase delays at each frequency point are measured separately. The phase gradient is obtained by differentiating the phase delays at different frequency points, which is used to judge the refraction and diffraction phenomena in the signal propagation path. Thus, high-precision phase delay measurement is achieved. The multi-frequency point analysis ability can effectively identify the phase distortion in the non-line-of-sight propagation scenario. The obtained phase delay is used for spatial Fourier transform to convert the time-domain phase information into the direction-of-arrival information, providing an angular-domain basis for beam optimization.

[0049] The acquisition of multi-dimensional parameters comprehensively covers the key characteristics of signal transmission, avoids error misjudgment caused by single-parameter analysis, provides a complete data basis for subsequent error evaluation, and the obtained parameters are directly used as the input for cross-correlation analysis to improve the accuracy of error weight calculation, thereby optimizing the comprehensive error judgment result.

[0050] The scattering parameters, energy distribution, and phase delay affect each other during signal transmission. That is, the reflection loss changes the energy distribution, and the multipath effect simultaneously causes phase delay and energy dispersion. Single-parameter analysis cannot reflect the essential reasons for signal distortion. Through cross-correlation analysis, the potential causal relationships between parameters can be explored to provide a basis for accurately evaluating the error weight. The obtained scattering parameters are stored in frequency segments, the energy distribution is constructed as a spatial grid matrix, where the x, y, and z coordinates correspond to the energy values, and the phase delay is arranged as a vector according to the frequency points. The three types of data are normalized to eliminate the dimension difference. The energy distribution and phase delay are processed through a spatio-temporal sliding window. Preferably, the time window is 500 ms, and the spatial window covers 10% of the area of the monitoring region. The correlation coefficient between the energy centroid offset and the phase delay change within the spatio-temporal sliding window is calculated to identify the corresponding relationship between the energy concentration region and the phase anomaly.

[0051] By comparing the frequency response curves of the test signal and the backscatter signal, analyze the influence of reflection loss on phase linearity, where the reflection loss is also the scattering parameter. That is, if a certain frequency band is too high and causes a phase mutation, it is marked as a strongly correlated frequency band. And use the Granger causality test algorithm to judge the causal directionality between parameters. If the change in energy distribution always precedes the phase delay fluctuation in the time series, it is determined that the energy distribution has a causal influence on the phase delay; integrate the above analysis results into a correlation matrix, and the matrix element value represents the correlation strength between any two parameters in a specific scenario. The value range of the correlation strength is 0 - 1. Preferably, when there is a large metal reflector in the monitoring area, the correlation strength between the energy distribution and the scattering parameter can reach 0.8; through multi-dimensional correlation analysis, reveal the complex causal chain of signal distortion, avoid misjudgment caused by split analysis of parameters, and can identify potential signal problems in advance compared with traditional single-parameter evaluation. The quantified correlation matrix is the core input for determining the error weight, which directly affects the accuracy of weight allocation.

[0052] Under different interference scenarios, the contribution degrees of the scattering parameter, energy distribution, and phase delay to the signal error are different. Among them, multipath interference focuses on energy distribution and phase delay, and impedance mismatch focuses on the scattering parameter. Dynamically determining the error weight can make the comprehensive error calculation more in line with the actual situation and avoid evaluation biases caused by fixed weights; set initial weights for the three types of parameters based on historical data. Preferably, the scattering parameter is 0.4, the energy distribution is 0.3, and the phase delay is 0.3 as the benchmark for adaptive adjustment; input the correlation matrix and the current error state into the recurrent neural network to output the adjusted error weight. Among them, the recurrent neural network continuously optimizes the weight mapping relationship through online learning, sets the weight adjustment range to prevent excessive weight fluctuations, and at the same time ensures that the sum of all weights is always 1. When the correlation matrix is abnormal, preferably, when the correlation strength of a certain parameter suddenly drops below 0.2, trigger the Bayesian inference mechanism and correct the error weight in combination with prior knowledge. After every 5 adjustments of the error weight, compare the error evaluation accuracy before and after the adjustment through simulation. If the error judgment accuracy rate drops by more than 5%, then trace back to the correlation analysis step to check whether there are abnormalities in the data or algorithm; the adaptive weight mechanism enables the error evaluation to quickly adapt to environmental changes, that is, the movement of interference sources and sudden weather changes, improves the judgment accuracy rate of the comprehensive error, and the determined error weight is directly used for the comprehensive error calculation to guide the selection of the compensation strategy for the backscatter signal, that is, the parameter with a high error weight is compensated first.

[0053] Combine the error weights of each parameter with the actual parameter deviation, and quantify the overall signal error between the test signal and the backscattered signal through weighted summation to provide a clear quantitative index for subsequent compensation; calculate the deviation between the actual value and the ideal value of each parameter, perform Z-score standardization to eliminate the influence of dimension, accumulate the standardized parameter deviations multiplied by the corresponding weights to obtain the signal error between the test signal and the backscattered signal, calculate the 95% confidence interval of the comprehensive error, if the error value exceeds this interval, it is marked as an abnormal error and a more refined compensation strategy is triggered; the unified quantitative evaluation standard makes the signal errors in different scenarios comparable, facilitating the rapid positioning of signal quality problems, the standardization process and confidence interval evaluation enhance the robustness of error judgment, reduce the interference of accidental factors, and the comprehensive error, as the core basis for compensation decision-making, triggers the dynamic compensation process of the backscattered signal, and its accuracy directly affects the effectiveness of the compensation strategy.

[0054] The compensation strategy for the backscattered signal includes: Extract the dominant component of the signal error and determine the compensation direction of the backscattered signal according to the dominant component of the signal error; Based on the compensation direction of the backscattered signal, coordinately adjust the parameter combination of the test signal to compensate the backscattered signal; Real-time monitor the signal quality of the backscattered signal after compensation. The signal quality includes the convergence degree and energy concentration of the scattering parameters, and judge according to the signal quality to update the adjustment step size of the parameter combination of the test signal and switch the compensation direction of the backscattered signal.

[0055] Signal errors can be caused by a variety of factors, such as phase delay, uneven energy distribution, and scattering loss. It is necessary to identify the dominant error source and select a targeted compensation strategy to avoid blind adjustment. Perform principal component analysis on the comprehensive error, project the signal error to different eigenvector directions, and calculate the error contribution in each direction. For example, if the error contribution in the phase delay direction exceeds 50%, the phase error is determined to be the dominant component. If the error caused by uneven energy distribution accounts for the highest proportion, energy optimization is taken as the priority direction. A decision tree with three layers of decision nodes is established. The first layer determines the dominant error type, including scattering parameters, energy distribution, and phase delay. The second layer analyzes the error source, including multipath effects, antenna mismatch, and environment. Interference, the third layer outputs the compensation direction. If the dominant component of the signal error is phase delay and is caused by multipath reflection, the compensation direction is phase correction and beamforming. If the uneven energy distribution is caused by the deviation of the antenna radiation direction, the antenna pointing is adjusted and the signal power is optimized. For the case where the signal error components are close, that is, the phase error is 45% and the energy error is 40%, the fuzzy logic algorithm is used to comprehensively consider the historical compensation data and the current environmental parameters, and dynamically adjust the priority of the compensation direction; thereby targeting the root cause of the error, improving the matching degree of the compensation strategy, reducing the number of invalid adjustments and improving the compensation efficiency. The clear compensation direction directly guides the test signal parameter adjustment strategy and narrows the parameter search space.

[0056] The parameter combinations of the test signal are interrelated and affect the signal quality. It is necessary to coordinately adjust multiple parameters according to the compensation direction to achieve a comprehensive compensation effect. Predefine parameter adjustment rules corresponding to different compensation directions. For example, if the dominant error is phase delay, increase the modulation index first to optimize the phase linearity, and fine-tune the frequency step to avoid the phase-sensitive frequency band. If the phase is advanced, reduce the carrier frequency, and if it is lagging, increase the carrier frequency. In the energy compensation scenario, if the energy is unevenly distributed, adjust the phase weighting coefficient of the antenna array to change the beam pointing direction, and simultaneously increase the duty cycle to enhance the signal energy in the target area. If the energy concentration area is offset, re-plan the scanning path. After adjusting the parameters for the first time, a feasibility check is performed according to the above-mentioned physical constraint layer. If the parameters are out of range, the adjustment range is proportionally rolled back or the adjustment strategy is changed. The particle swarm optimization algorithm is used to search for the optimal combination of multiple parameters in parallel, and the parameter adjustment range is divided into multiple subspaces. Each particle searches independently and exchanges the optimal solution regularly to accelerate the convergence speed. The coordinated adjustment of multiple parameters avoids the degradation of other performances caused by the optimization of a single parameter and achieves a comprehensive improvement in signal quality. The measured data shows that compared with the adjustment of a single parameter, the comprehensive error is reduced. The adjusted test signal is used for a new round of signal acquisition, and its parameter combination directly affects the monitoring results of the signal quality after compensation.

[0057] Single compensation cannot completely eliminate errors, and environmental changes can also lead to the attenuation of the compensation effect. It is necessary to monitor the signal quality in real time and dynamically adjust the compensation strategy and parameter step size; continuously monitor the scattering parameters, calculate the Euclidean distance between two adjacent measurements. If the Euclidean distance is less than the set distance threshold, it is considered that the scattering parameters converge. Evaluate the energy concentration by calculating the entropy value of the energy distribution. The lower the entropy value, the more concentrated the energy. Set the entropy value threshold to judge whether the energy distribution meets the standard. At the same time, monitor the standard deviation of the phase delay to judge whether the phase is stable. Dynamically adjust the adjustment step size of the parameter combination according to the change of the signal quality. If the error decreases significantly after compensation, increase the step size to accelerate the optimization. If the error fluctuates or increases, decrease the step size for fine adjustment to avoid over-adjustment; if the signal quality does not improve after 3 consecutive adjustments, trigger a re-evaluation of the compensation direction. Compare the current error component with the historical compensation data. If the original compensation direction is invalid, switch to the sub-optimal compensation direction, that is, switch from phase correction to energy optimization, and record the failure cases for subsequent strategy optimization; thus making the compensation strategy of the backscatter signal adaptive and maintaining the signal quality stable in a dynamic environment. The compensated backscatter signal is used to optimize the beam pointing and scanning frequency of the monitoring area, promoting the continuous optimization of the entire measurement method.

[0058] S3. Optimize the beam pointing and scanning frequency of the monitoring area according to the compensated backscatter signal, re-acquire the interference data of the monitoring area based on the optimized beam pointing and scanning frequency of the monitoring area, update the interference characteristics at the same time, and update the parameter combination of the test signal through the genetic algorithm.

[0059] As Figure 3 shown, the optimization strategy of the beam pointing of the monitoring area includes: Map the energy distribution and phase delay of the compensated backscatter signal to the direction-of-arrival space through Fourier transform, construct an energy density map, and identify the energy concentration area through the peak detection algorithm and mark it as the target signal direction; Combine the temperature and humidity field distribution of the monitoring area and the spatial topology of the reflector to evaluate the propagation feasibility of the beam pointing; Through the reinforcement learning algorithm, use the received power of the target signal and the interference energy as the reward function to dynamically optimize the beam pointing of the monitoring area.

[0060] The compensated backscattered signal contains the mixed information of the target signal and the interference signal. It is necessary to locate the signal source direction through spatial mapping to provide an initial target for beam pointing. The energy distribution and phase delay of the compensated backscattered signal are transformed from the time domain signal to the direction-of-arrival (DOA) space by the fast Fourier transform algorithm to generate an energy density map. The energy density map is processed by multi-scale Gaussian smoothing to eliminate random noise interference. And through the non-maximum suppression algorithm, the maximum energy points are retained in the neighborhood, and at the same time, an energy threshold is set to filter out weak signal points. Preferably, the energy threshold is 1.5 times the global average energy. Finally, 3-5 potential target signal directions are identified. Transforming the time-domain and spatial-domain signals into the angle-domain expression can quickly locate the target signal direction, and the determined target signal direction is used as the initial reference for beam pointing, providing direction constraints for subsequent feasibility evaluation and dynamic optimization.

[0061] There may be strong reflectors or harsh environmental conditions in the target signal direction, including high-humidity areas. Direct pointing will cause multipath interference or signal attenuation. It is necessary to evaluate the pointing feasibility in combination with environmental factors. Retrieve the temperature and humidity field distribution of the monitoring area and the spatial topology of the reflector. Quickly retrieve the environmental characteristics in the target direction through the octree spatial index structure. If a high-humidity gradient area or a large metal reflector is detected, it is marked as a high-risk area. The judgment of the high-humidity gradient area is obtained by comparing the humidity change rate with the change rate threshold, and the judgment of the large metal reflector is obtained by comparing the volume of the reflector with the volume threshold.

[0062] Based on the ray-tracing algorithm, the signal propagation paths are simulated in parallel on the GPU cluster, and the reflection, refraction, and absorption losses of each path are calculated. If the predicted total loss is greater than the preset loss threshold, then this direction is determined to be infeasible. For the infeasible direction, a greedy search algorithm is used to find a feasible alternative direction in its neighborhood, preferentially choosing a path that passes through a low-loss medium and avoids large reflectors. Through environmental constraint evaluation, it is possible to avoid the beam pointing from falling into high-loss or interference areas, improve the signal reception stability. The filtered feasible directions are used as the search space for the reinforcement learning algorithm, narrowing the optimization range and improving the adjustment efficiency of the beam pointing.

[0063] Monitor the dynamic changes in the environment, namely the movement of interference sources and the fluctuations in environmental temperature and humidity. It is necessary to adjust the beam direction in real time to maximize the received power of the target signal and suppress interference. Through the deep deterministic policy gradient algorithm, the beam direction angle is regarded as a continuous action space. The beam direction angle includes the azimuth angle and the elevation angle. The ratio of the received power of the target signal to the interference energy is used as the reward function. An experience replay buffer is constructed. After every 50 beam adjustments are completed, samples are randomly selected for network parameter update. And through the soft update strategy of the target network, the target network parameters are updated every 100 steps. Action exploration is achieved through the OU noise process, increasing exploration randomness at the initial stage of training and gradually reducing the noise intensity as training progresses. Dynamic optimization enables the beam direction to track the target signal in real time and avoid interference. The optimized beam direction is directly applied to the monitoring antenna array, and at the same time, the re-acquisition of interference data is triggered, providing inputs in the new environment for subsequent scan frequency optimization and the update of the parameter combination of test signals.

[0064] The optimization strategy for the scan frequency of the monitoring area includes: Divide the scattering parameters of the compensated backscattered signal into multiple frequency bands according to frequency. Combine the monitoring area after beam direction optimization, and analyze the signal attenuation and interference intensity distribution of each frequency band in the monitoring area to identify the scan frequency band and the frequency step size. Obtain the spectrum occupancy status of the monitoring area in real time through spectrum sensing technology, combine the interference type to judge the spectrum change trend, and mark the risk frequency bands and idle frequency bands. Combine the scan frequency band, the frequency step size and the spectrum change trend to determine the constraint conditions, and optimize the scan frequency of the monitoring area through the genetic algorithm.

[0065] There are differences in the attenuation and interference intensity of the compensated backscattered signal in different frequency bands, and the signal propagation characteristics of the monitoring area change after beam direction optimization. It is necessary to combine the two for analysis to determine the efficient scan frequency band and the step size, and avoid ineffective scanning. Divide the scattering parameters of the backscattered signal into multiple continuous frequency bands according to frequency. Calculate the average value of the reflection coefficient and the transmission loss for each frequency band. At the same time, retrieve the monitoring area after beam direction optimization, simulate the propagation path of signals in different frequency bands in this monitoring area, based on the ray tracing algorithm, combine the coordinates of the reflectors and the temperature and humidity field distribution in the monitoring area, calculate the attenuation values of signals in each frequency band during propagation, and analyze the interference intensity distribution of each frequency band by statistically analyzing historical interference data, and mark the frequency bands with lower signal attenuation and weaker interference as candidate scan frequency bands.

[0066] Determine the step size according to the differences in signal characteristics within the candidate frequency bands. If the signal characteristics change drastically within the frequency band, that is, there are multiple interference spikes, a small step size is adopted. If the signal characteristics are stable, a large step size is adopted to ensure that both signal details can be captured and the scanning efficiency can be improved. Through comprehensive analysis of multiple factors, select a cost-effective scanning frequency band and frequency step size, which improves the scanning efficiency and reduces the signal distortion problem caused by improper frequency band selection. The preliminarily determined scanning frequency band and frequency step size, together with the results of the spectrum occupancy status analysis, constitute the basic constraint conditions for frequency optimization.

[0067] The spectrum usage situation in the monitoring area changes in real time. Relying solely on historical spectrum data cannot meet the requirements of dynamic monitoring. It is necessary to sense the spectrum occupancy status in real time and predict trends to avoid interference frequency bands and utilize idle resources. The compressed sensing technology is adopted to quickly obtain the broadband spectrum data of the monitoring area through sparse sampling, and the reconstruction algorithm is used to restore the complete spectrum. At the same time, the energy detection method and the feature recognition method are combined to distinguish different types of signals, including communication signals and interference signals, mark the boundaries and intensities of the occupied frequency bands, and analyze the distribution rules of the above-identified interference types, including narrowband interference and broadband interference, on the spectrum to obtain the spectrum change trend. The long short-term memory network is used to predict the occupancy probability of each frequency band within a certain period in the future based on historical spectrum data. The frequency bands with an occupancy probability greater than the first threshold are marked as risk frequency bands, and those less than the second threshold are marked as idle frequency bands.

[0068] Establish a visual spectrum map and update the annotation information of risk frequency bands and idle frequency bands in real time. When there is a rapid upward trend in the occupancy rate of a risk frequency band, trigger the early warning mechanism to prompt that the scanning strategy needs to be adjusted preferentially. Thus, the dynamic monitoring and forward-looking analysis of spectrum resources are realized, enabling the scanning frequency to avoid high-risk frequency bands to reduce the interference collision rate and effectively improving the anti-interference ability of the monitoring antenna. The spectrum occupancy status and trend prediction results, combined with the candidate scanning frequency band and frequency step size, form a complete frequency optimization constraint condition, providing an optimization basis for the genetic algorithm.

[0069] Under complex constraint conditions, traditional optimization algorithms are prone to falling into local optima. It is necessary to utilize the global search ability of genetic algorithms to quickly find the optimal scanning frequency combination. Encode the scanning start frequency, stop frequency, and frequency step size as chromosome genes, and ensure parameter accuracy through real-number encoding. Randomly generate an initial population consisting of 100 individuals, and perform feasibility verification according to the constraint conditions, eliminating individuals that exceed the antenna operating frequency band or violate the frequency step size limit. Construct a fitness function, taking the scanning efficiency, spectrum utilization rate, and interference suppression effect as the core evaluation indicators. Among them, the scanning efficiency is related to the number of scanning frequency bands and the frequency step size, the spectrum utilization rate refers to the degree of utilization of idle frequency bands, and the interference suppression effect refers to the degree of avoidance of risky frequency bands. Then calculate the fitness value of each individual through weighted summation. Preferably, the weight of the scanning efficiency is set to 0.4, the weight of the spectrum utilization rate is set to 0.3, and the weight of the interference suppression effect is set to 0.3.

[0070] Through selection, crossover, and mutation operations, iteratively update the population, retaining the individual with the highest fitness in each generation to accelerate the convergence of the genetic algorithm. Terminate the iteration when the fitness of the optimal individual does not improve significantly for multiple consecutive generations, and decode the optimal individual into the final scanning frequency. The genetic algorithm efficiently searches for the optimal scanning frequency combination under complex constraints, shortening the calculation time and improving the spectrum utilization rate. The optimized scanning frequency is applied to the monitoring antenna, and at the same time, it triggers the re-acquisition of interference data, promoting the entire monitoring process to enter a new round of dynamic optimization cycle.

[0071] The optimized beam pointing and scanning frequency change the way the monitoring antenna receives and acquires signals. In order to accurately reflect the interference situation in the current monitoring environment, it is necessary to re-acquire interference data. According to the optimized beam pointing, adjust the phase weighting coefficient or mechanical steering device of the monitoring antenna array to align the main beam of the antenna with the direction of the target signal. At the same time, according to the optimized scanning frequency setting, configure parameters such as the frequency range and frequency step size of the signal acquisition device, start the signal acquisition device, and perform continuous or timed data acquisition in the monitoring area. During the data acquisition process, simultaneously acquire interference data such as environmental electromagnetic data, temperature and humidity data, and reflector coordinates. To ensure the accuracy and consistency of the acquired interference data, perform time synchronization and calibration on different types of data, and make each sensor and signal acquisition device work under the same time reference. At the same time, perform calibration processing on the acquired data to eliminate the influence of device errors and environmental noise. The re-acquired interference data can more accurately reflect the actual situation in the current monitoring environment, providing a reliable data basis for subsequent interference feature updates and test signal parameter combination updates. The acquired interference data is directly used to update the interference features, providing a basis for the update of the test signal parameter combination.

[0072] Monitor the changes in the environment and signal propagation conditions over time. The newly acquired interference data will contain new interference patterns and characteristics, and the interference characteristics need to be updated in a timely manner. Preprocess the newly acquired interference data, including operations such as data cleaning, denoising, and normalization. For environmental electromagnetic data, remove high-frequency noise and low-frequency interference through filtering algorithms. For temperature and humidity data, perform outlier detection and processing, and perform normalization to bring them within the same numerical range. For reflector coordinates, perform spatial coordinate transformation and alignment to ensure data consistency. Then, extract features from the preprocessed interference data, fuse and update the extracted new features with the historical feature data to form the latest interference feature vector.

[0073] Evaluate and verify the updated interference characteristics to ensure their accuracy and reliability. Verify whether the new features can accurately identify and classify different types of interference by comparing with known interference types and historical feature data. At the same time, use methods such as cross-validation to evaluate the stability and generalization ability of the features to ensure their effectiveness in different monitoring environments. Timely updating of interference characteristics can adapt to changes in the monitoring environment and improve the ability to identify and process interference. The updated interference characteristics serve as an important basis for updating the parameter combinations of test signals, providing more accurate inputs for genetic algorithms and helping to generate more optimized parameter combinations of test signals.

[0074] As the monitoring environment and interference characteristics change, the original parameter combinations of test signals are no longer optimal and need to be optimized through genetic algorithms to generate parameter combinations of test signals that are more suitable for the current environment. Encode the parameters of the test signal to form chromosomes. Through real-number encoding, map each parameter to a real-number interval, and randomly generate a certain number of initial chromosomes to form an initial population. Evaluate the quality of each chromosome through a fitness function. The design of the fitness function takes into account multiple factors, including interference suppression effect, antenna matching degree, and propagation efficiency. Preferably, the interference suppression effect can be used as the main goal. Evaluate its interference suppression ability by calculating the bit error rate or signal-to-noise ratio of the test signal in different interference scenarios. At the same time, consider the antenna matching degree and propagation efficiency, and evaluate its performance by calculating the radiation efficiency of the antenna and the coverage range of the signal in the monitoring area. Weight-sum these factors to obtain the fitness value of each chromosome.

[0075] Genetic operations are performed on the initial population, including selection, crossover, and mutation. In the selection operation, according to the fitness values of the chromosomes, methods such as roulette wheel selection or tournament selection are used to select chromosomes with higher fitness to enter the next generation. In the crossover operation, two chromosomes are randomly selected and gene exchange is carried out between them to generate new chromosomes. In the mutation operation, genes of the chromosomes are randomly mutated with a certain probability to introduce new gene diversity. The genetic operations are repeated, and the population is continuously iteratively optimized until the preset termination conditions are met. The termination conditions are reaching a certain number of iterations, the fitness value no longer increases, or other specific performance indicators are satisfied. In each iteration process, the optimal chromosome and its fitness value are recorded, and adjustments and optimizations are made as needed. When the termination conditions are met, the chromosome with the highest fitness is selected as the optimal solution, and it is decoded into the parameter combination of the test signal. These parameter combinations are the parameter combinations of the test signal optimized by the genetic algorithm and can be directly injected into the FM monitoring antenna.

[0076] Updating the parameter combination of the test signal through the genetic algorithm can enable the test signal to better adapt to the changes in the monitoring environment. Compared with traditional parameter optimization methods, the genetic algorithm has global search ability and strong robustness, and can find better solutions in the complex parameter space. The updated parameter combination of the test signal will be used to form a new test signal, which is injected into the FM monitoring antenna for a new round of monitoring and data acquisition, forming a closed-loop optimization process.

Claims

1. A parameter measurement method for a frequency modulation monitoring antenna, characterized in that Including: Obtain the interference data of the monitoring area in real time. The interference data includes environmental electromagnetic data, temperature and humidity data, and reflector coordinates. Perform pattern recognition on the interference data through a convolutional neural network to output interference features. The interference features include interference types and spatio-temporal intensity distributions, and construct a multi-objective optimization function to process the interference features to generate a test signal; Inject the test signal into the frequency modulation monitoring antenna, and at the same time receive and monitor the backscattered signal. Obtain the scattering parameters, energy distribution, and phase delay of the backscattered signal in real time, and comprehensively judge the signal error between the test signal and the backscattered signal based on the scattering parameters, energy distribution, and phase delay, and perform dynamic compensation on the backscattered signal according to the signal error; Optimize the beam pointing and scanning frequency of the monitoring area according to the compensated backscattered signal, re-obtain the interference data of the monitoring area based on the optimized beam pointing and scanning frequency of the monitoring area, update the interference features at the same time, and update the parameter combination of the test signal through a genetic algorithm.

2. The parameter measurement method for a frequency modulation monitoring antenna according to claim 1, characterized in that, The generation strategy of the test signal includes: Construct a generative adversarial network. The generator processes the interference features to output candidate parameter combinations of the test signal, and the discriminator evaluates the candidate parameter combinations of the test signal through a multi-objective optimization function; Configure a physical constraint layer according to the properties of the frequency modulation monitoring antenna, and screen the candidate parameter combinations of the test signal according to the physical constraint layer; Optimize the candidate parameter combinations of the test signal through a simulated annealing algorithm. After multiple iterative trainings of the generator and the discriminator, generate a test signal. The parameter combination of the test signal includes frequency step size, modulation index, and duty cycle.

3. The parameter measurement method for a frequency modulation monitoring antenna as described in claim 2, wherein The output sub-strategy of the interference features includes: Obtain the interference data including environmental electromagnetic data, temperature and humidity data, and reflector coordinates in the monitoring area in real time, and align the interference data according to the time stamp; Convert the environmental electromagnetic data into a time-frequency map matrix through short-time Fourier transform, process the temperature and humidity data through Kriging interpolation method to generate a temperature and humidity field distribution, and construct a spatial topology of the reflector coordinates through an octree algorithm; Construct a three-branch convolutional neural network including an electromagnetic branch, an environmental branch, and a spatial branch. Extract the features of the interference data through the three-branch convolutional neural network, and perform weighted fusion through an attention mechanism to generate interference features including interference types and spatio-temporal intensity distributions; Compare and analyze the interference features with the historical feature data to judge whether to adjust the parameters and attention weights of the three-branch convolutional neural network.

4. The parameter measurement method for a frequency modulation monitoring antenna according to claim 3, characterized in that, The electromagnetic branch identifies the interference pattern of the time-frequency map matrix through a residual network, the environmental branch extracts the gradient change of the temperature and humidity field distribution through a two-dimensional convolutional layer, and the spatial branch analyzes the spatial topology through a graph convolutional network and extracts the path complexity of multipath reflection.

5. The parameter measurement method for a frequency modulation monitoring antenna according to claim 4, characterized in that, The construction sub-strategy of the multi-objective optimization function includes: Take interference suppression, antenna matching, and propagation efficiency as optimization objectives, and dynamically adjust the objective weights of the optimization objectives according to the interference type through a Bayesian network; Construct a multi-objective optimization function by weighted summation of the optimization objectives and the objective weights.

6. The parameter measurement method for a frequency modulation monitoring antenna according to claim 5, characterized in that, The compensation strategy of the backscattered signal includes: Extract the dominant components of the signal error, and determine the compensation direction of the backscattered signal according to the dominant components of the signal error; Based on the compensation direction of the backscattered signal, coordinately adjust the parameter combination of the test signal to compensate the backscattered signal; Real-time monitor the signal quality of the compensated backscattered signal. The signal quality includes the convergence degree of scattering parameters and the energy concentration degree, and judge according to the signal quality to update the adjustment step size of the parameter combination of the test signal and switch the compensation direction of the backscattered signal.

7. The parameter measurement method for a frequency modulation monitoring antenna according to claim 6, characterized in that, The judgment sub-strategy for the signal error between the test signal and the backscattered signal includes: Inject the test signal into the frequency modulation monitoring antenna, receive the backscattered signal, and obtain the scattering parameters, energy distribution and phase delay of the backscattered signal in real time; Conduct cross-correlation analysis on the scattering parameters, energy distribution and phase delay of the backscattered signal to obtain a correlation matrix, and determine the error weights of the scattering parameters, energy distribution and phase delay according to the correlation matrix; Judge the signal error between the test signal and the backscattered signal by weighted summation according to the error weights.

8. The parameter measurement method for a frequency modulation monitoring antenna according to claim 7, characterized in that, Based on the field strength probe to scan the monitoring area, perform three-dimensional spatial sampling on the backscattered signal, obtain the energy distribution of the backscattered signal by acquiring the energy value of each sampling point; compare the carrier phases of the test signal and the backscattered signal through coherent detection technology to obtain the phase delay of the backscattered signal.

9. The parameter measurement method for a frequency modulation monitoring antenna according to claim 8, wherein, The optimization strategy for the beam direction of the monitoring area includes: Map the energy distribution and phase delay of the compensated backscattered signal to the direction-of-arrival space through Fourier transform, construct an energy density map, and identify the energy concentration area through the peak detection algorithm and mark it as the target signal direction; Combine the temperature and humidity field distribution of the monitoring area and the spatial topology of the reflector to evaluate the propagation feasibility of the beam direction; Through the reinforcement learning algorithm, use the received power of the target signal and the interference energy as the reward function to dynamically optimize the beam direction of the monitoring area.

10. The parameter measurement method for a frequency modulation monitoring antenna according to claim 9, characterized in that, The optimization strategy for the scanning frequency of the monitoring area includes: Divide the scattering parameters of the compensated backscattered signal into multiple frequency bands according to the frequency, and combine the monitoring area optimized by the beam direction to analyze the signal attenuation and interference intensity distribution of each frequency band in the monitoring area to identify the scanning frequency band and the frequency step size; Obtain the spectrum occupancy status of the monitoring area in real time through spectrum sensing technology, combine the interference type to judge the spectrum change trend, and mark the risk frequency band and the idle frequency band; Combine the scanning frequency band, the frequency step size and the spectrum change trend to determine the constraint conditions, and optimize the scanning frequency of the monitoring area through the genetic algorithm.

Images