A signal interference strategy generation method and device based on deep learning, and a medium
By generating radar signal jamming strategies using deep learning technology, the problem of radar signal systems being unable to determine signal attenuation in a timely manner under rainfall conditions is solved, and dynamic jamming strategy generation and precise jamming effect are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING HAIGE SHENZHOU COMM TECH
- Filing Date
- 2026-05-08
- Publication Date
- 2026-06-05
Smart Images
Figure CN122151009A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of signal interference technology, and in particular to a method, device and medium for generating signal interference strategies based on deep learning. Background Technology
[0002] In radar signal systems, the electromagnetic spectrum in the environment is typically controlled through reconnaissance and jamming. The propagation of electromagnetic signals is affected by the humidity of the natural environment, and rainfall is one of the main natural factors affecting humidity. In the microwave and millimeter-wave bands, signal attenuation caused by rainfall can reach 20 dB or even higher, which can seriously affect reconnaissance range, positioning accuracy, and jamming effectiveness.
[0003] In existing technologies, traditional systems primarily rely on weather radar for rainfall prediction. However, in cases of rainfall beyond visual range, it's difficult to promptly determine signal attenuation across different propagation methods. For radar signal interference links, OISAC links can access real-time telemetry data such as received signal level, transmitted power, link length, operating frequency, and polarization, eliminating the need for dedicated sensor hardware. Existing non-link systems typically set corresponding signal compensation power to compensate for rain attenuation of the interference signal. However, when setting interference parameters, it's impossible to dynamically analyze the link status of both the interfering and affected parties based on the actual loss of the interference signal's path through rain. This prevents proactively utilizing the attenuation differences of rainfall across different frequency bands to dynamically formulate corresponding signal interference strategies based on the differences in link signal attenuation between the two parties. Therefore, a method for generating dynamic signal interference strategies based on OISAC links, tailored to rainfall environments, is urgently needed to address these technical problems. Summary of the Invention
[0004] This application provides a signal interference strategy generation method, device, and medium based on deep learning, which solves the technical problem that existing signal interference strategies cannot meet the dynamic requirements of interference signal links in rainy environments.
[0005] In a first aspect, embodiments of this application provide a signal interference strategy generation method based on deep learning, characterized in that the method includes: acquiring reference level sampling data and real-time level sampling data of each communication link, and calculating the initial path rain attenuation value corresponding to the reference level sampling data and real-time level sampling data; based on the initial path rain attenuation value, filtering the wet antenna attenuation component of the real-time level sampling data by digital filtering to determine the pure path rain attenuation value of each communication link; inverting the pure path rain attenuation value through a rain attenuation power-law model to obtain the path average rainfall intensity corresponding to each communication link; performing spatial interpolation optimization on a preset interference area according to the path average rainfall intensity to determine a two-dimensional rainfall intensity distribution map of the interference area; calculating the vertical distribution parameter of the real-time atmospheric refractive index, and calculating the signal propagation time delay corresponding to the ray trajectory of the interference target based on the vertical distribution parameter to obtain the corrected coordinates of the interference target; and determining the interference strategy corresponding to the interference target by calculating the link interference parameters and predicting the interference window according to the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target.
[0006] In one implementation of this application, calculating the initial path rain attenuation value corresponding to the reference level sampling data and the real-time level sampling data specifically includes: calculating the total additional attenuation of the links corresponding to the reference level sampling data and the real-time level sampling data; obtaining the corresponding non-rain attenuation of each communication link, and subtracting the non-rain attenuation from the total additional attenuation to obtain the initial path rain attenuation value.
[0007] In one implementation of this application, based on the initial path rain attenuation value, the wet antenna attenuation component of the real-time level sampling data is filtered by digital filtering to determine the pure path rain attenuation value of each communication link. Specifically, this includes: extracting the attenuation change corresponding to the real-time level sampling data to determine the dynamic attenuation characteristics of each communication link; and based on the dynamic attenuation characteristics, distinguishing between the wet antenna attenuation component and path rain attenuation in the initial path rain attenuation value by digital filtering, and filtering the wet antenna attenuation component in the initial path rain attenuation value to determine the pure path rain attenuation value.
[0008] In one implementation of this application, the pure path rain attenuation value is inverted using a rain attenuation power-law model to obtain the path-average rain intensity corresponding to each communication link. Specifically, this includes: obtaining the operating frequency and polarization mode of each communication link, and matching the corresponding empirical coefficients in the rain attenuation power-law model based on the operating frequency and polarization mode; inputting the pure path rain attenuation value into the rain attenuation power-law model according to the empirical coefficients, and inverting to solve for the path-average rain intensity corresponding to each communication link.
[0009] In one implementation of this application, spatial interpolation optimization is performed on a preset interference area based on the path-averaged rainfall intensity to determine a two-dimensional rainfall intensity distribution map of the interference area. Specifically, this includes: acquiring the DEM terrain corresponding to the interference area and discretizing the DEM terrain into corresponding grid cells; wherein the information of each grid cell includes: grid cell coordinates and the set of rainfall intensities for each grid cell; constructing a feature vector corresponding to each grid cell based on the path-averaged rainfall intensity and the geometric features of each communication link and grid cell; wherein the geometric features include: the shortest distance from the center of the grid cell to the projection of each link, the link crossing the grid state, and the azimuth angle of the grid relative to the start and end points of the link; inputting the feature vector into a preset random forest model and training it until the model converges to obtain the rainfall intensity parameters corresponding to the grid cell; interpolating the rainfall intensity parameters until the discrete error is extended to all grid cells to determine the rainfall intensity error surface of the interference area; and integrating the rainfall intensity error surface into the rainfall intensity parameters to determine the two-dimensional rainfall intensity distribution map.
[0010] In one implementation of this application, the signal propagation time delay corresponding to the ray trajectory of the interfering target is calculated based on the vertical distribution parameters to obtain the corrected coordinates of the interfering target. Specifically, this includes: obtaining the initial coordinates of the interfering target and the coordinates of the reconnaissance station, and calculating the ray azimuth angle based on the initial coordinates and the reconnaissance station coordinates; calculating the ray path from the reconnaissance station coordinates to the interfering target through numerical integration based on the ray azimuth angle; calculating the propagation time delay of the interfering signal on the ray path, and calculating the arrival time difference from each reconnaissance station coordinate to the interfering target based on the propagation time delay; correcting the arrival time difference measurement value through the arrival time difference, determining the arrival time difference correction amount, and locating the arrival time difference correction amount to obtain the corrected coordinates of the interfering target.
[0011] In one implementation of this application, based on a two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, an interference strategy corresponding to the interference target is determined through link interference parameter calculation and interference window prediction. Specifically, this includes: determining the original signal-to-noise ratio (SNR) of the link corresponding to the interference target based on the corrected coordinates of the interference target using electromagnetic propagation simulation; inputting the original SNR into a rain attenuation power-law model to obtain the current SNR of the link corresponding to the interference target; extracting the rain area path from the transmission point of the interference signal to the receiving point of the link corresponding to the interference target, and calculating the additional rain attenuation loss corresponding to the rain area path; defining the susceptibility index of the link corresponding to the interference target based on the additional rain attenuation loss; predicting the rain area movement trend on the two-dimensional rainfall intensity distribution map to obtain the link movement trend corresponding to the interference target; updating the susceptibility index based on the link movement trend, and marking the current time window as an interference time window when the susceptibility index exceeds a preset interference threshold; and determining the interference strategy by matching the corresponding interference signal parameters according to the interference time window.
[0012] In one implementation of this application, after determining the interference strategy corresponding to the interference target based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, through link interference parameter calculation and interference window prediction, the method further includes: executing the interference strategy to obtain the interference signal parameters of the link corresponding to the interference target, and evaluating the interference effect of the interference signal parameters to determine the interference success rate; adjusting the interference threshold and empirical coefficient according to the interference success rate to obtain the updated interference strategy.
[0013] Secondly, embodiments of this application also provide a signal interference strategy generation device based on deep learning, characterized in that the device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to implement a signal interference strategy generation method based on deep learning.
[0014] Thirdly, embodiments of this application also provide a non-volatile computer storage medium for generating signal interference strategies based on deep learning, which stores computer-executable instructions, characterized in that the computer-executable instructions, when executed, are capable of implementing a signal interference strategy generation method based on deep learning.
[0015] This application provides a signal interference strategy generation method, device, and medium based on deep learning. Based on an OISAC link, it accesses data on the received level, transmitted power, link length, operating frequency, and polarization of multiple existing communication links. By distinguishing between rain attenuation and wet antenna attenuation, waveguide correction, and rain attenuation model analysis, it obtains a two-dimensional rainfall intensity distribution map and corrected coordinates of the interference target. Based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, it analyzes interference parameters and timing, predicts the interference strategy of the target link, and solves the technical problem in existing technologies where signal interference strategies cannot meet the dynamic requirements of interference signal links under rainfall conditions. It also enables the analysis of attenuation differences in different frequency bands caused by rainfall, improving the interference efficiency of interference signals under rainfall conditions. Attached Figure Description
[0016] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart of a signal interference strategy generation method based on deep learning is provided for embodiments of this application; Figure 2 This is a schematic diagram of the internal structure of a signal interference strategy generation device based on deep learning, provided in an embodiment of this application. Detailed Implementation
[0017] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0018] This application provides a signal interference strategy generation method, device, and medium based on deep learning. Based on an OISAC link, it accesses data on the received level, transmitted power, link length, operating frequency, and polarization of multiple existing communication links. By distinguishing between rain attenuation and wet antenna attenuation, waveguide correction, and rain attenuation model analysis, it obtains a two-dimensional rainfall intensity distribution map and corrected coordinates of the interference target. Based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, it analyzes interference parameters and timing, predicts the interference strategy of the target link, and solves the technical problem in existing technologies where signal interference strategies cannot meet the dynamic requirements of interference signal links under rainfall conditions. It also enables the analysis of attenuation differences in different frequency bands caused by rainfall, improving the interference efficiency of interference signals under rainfall conditions.
[0019] The technical solutions proposed in the embodiments of this application will be described in detail below with reference to the accompanying drawings.
[0020] Figure 1 This document provides a flowchart of a deep learning-based signal interference strategy generation method as an embodiment of this application. Figure 1 As shown in the figure, the signal interference strategy generation method based on deep learning provided in this application embodiment specifically includes the following steps: Step 101: Obtain the reference level sampling data and real-time level sampling data of each communication link, and calculate the initial path rain attenuation value corresponding to the reference level sampling data and real-time level sampling data.
[0021] For example, Opportunity-Aware Communication (OISAC) can access data such as the received level, transmitted power, link length, operating frequency, and polarization of communication links in real time through multi-protocol and link gateways. The millimeter-wave ISAC signal processing flow does not rely on dedicated weather radar. It uses microwave or millimeter-wave communication links that have been deployed or originally existed in the interference area as a distributed sensor network to acquire data in order to calculate the initial path rain attenuation value and provide a data basis for rain attenuation analysis of its own interference signals.
[0022] Specifically, the calculation of the initial path rain attenuation value corresponding to the reference level sampling data and the real-time level sampling data includes: calculating the total additional attenuation of the links corresponding to the reference level sampling data and the real-time level sampling data; obtaining the corresponding non-rain attenuation of each communication link, and subtracting the non-rain attenuation from the total additional attenuation to obtain the initial path rain attenuation value.
[0023] In one embodiment, firstly, Opportunity-Aware Communication (OISAC) polls the gateway devices of microwave links in the current area at a set frequency according to the communication links, and obtains telemetry data packets for each link through a custom protocol. Each data packet contains: link ID, current received level, nominal transmit power, link length, operating frequency, and polarization mode (horizontal).
[0024] Then, the automatic weather station periodically reports environmental data via a bus, such as temperature 25.3℃, humidity 78%, and air pressure 1013.2 hPa. Its central station can also be equipped with a raindrop spectrometer to obtain a histogram of raindrop diameter distribution and load a digital elevation model (DEM) of the area from a local server.
[0025] Finally, clear-sky received signal level data (no rain, no fog) is collected over a continuous period. The median received signal level within each 10-minute statistical window is calculated and stored as a historical sequence. The baseline data is updated using an exponentially weighted moving average to obtain the reference signal level sampling data and real-time signal level sampling data for this link.
[0026] The total additional attenuation can be obtained by subtracting the reference level sampling data from the real-time level sampling data. The water vapor absorption attenuation calculated based on the current temperature, humidity, air pressure and water vapor density model (which can be determined by looking up a table) and the fixed atmospheric gas absorption (the same data as above) are then subtracted from the total additional attenuation to obtain the preliminary rain attenuation.
[0027] Step 102: Based on the initial path rain attenuation value, the wet antenna attenuation component of the real-time level sampling data is filtered by digital filtering to determine the pure path rain attenuation value of each communication link.
[0028] For example, wet antenna attenuation (WAA) is one of the uncertainties in the rainfall process of microwave links. It analyzes the short-term dynamic characteristics of signal attenuation changes through high-frequency sampling data to distinguish wet antenna attenuation from path rain attenuation, thereby eliminating the wet antenna attenuation component in the path rain attenuation value.
[0029] Specifically, based on the initial path rain attenuation value, the wet antenna attenuation component of the real-time level sampling data is filtered by digital filtering to determine the pure path rain attenuation value of each communication link. This includes: extracting the attenuation change corresponding to the real-time level sampling data to determine the dynamic attenuation characteristics of each communication link; and based on the dynamic attenuation characteristics, distinguishing between the wet antenna attenuation component and path rain attenuation in the initial path rain attenuation value by digital filtering, and filtering the wet antenna attenuation component in the initial path rain attenuation value to determine the pure path rain attenuation value.
[0030] In one embodiment, the received level of the link is sampled at a frequency of 1 Hz. When rainfall begins, a short-term rapid change in the received level is detected. For example, in a certain time window, the level drops sharply from -52.3 dBm to -55.1 dBm and then rises back to -52.5 dBm within 0.5 seconds. This change is consistent with the wet antenna attenuation characteristics caused by water accumulation on the radome surface. In contrast, the path rain attenuation changes relatively smoothly and does not exhibit similar abrupt changes.
[0031] The path rain attenuation is obtained by extracting the slow-varying component from the original attenuation sequence using a low-pass filter. The difference between the original attenuation and the slow-varying component is taken as the WAA component, and the WAA component is filtered out to obtain the pure path rain attenuation. The wet antenna attenuation and pure path rain attenuation of the link are then labeled.
[0032] Step 103: Using the rain attenuation power-law model, the pure path rain attenuation value is inverted to obtain the path average rain intensity corresponding to each communication link.
[0033] For example, the standard rain attenuation power-law model can be used to invert the path-average rainfall intensity, calculate the path-average rainfall intensity of each communication link, and identify the corresponding link start and end point coordinates.
[0034] Specifically, the pure path rain attenuation value is inverted using the rain attenuation power law model to obtain the path average rain intensity corresponding to each communication link. This includes: obtaining the operating frequency and polarization mode of each communication link, and matching the corresponding empirical coefficients in the rain attenuation power law model based on the operating frequency and polarization mode; inputting the pure path rain attenuation value into the rain attenuation power law model according to the empirical coefficients, and inverting to solve for the path average rain intensity corresponding to each communication link.
[0035] In one embodiment, the corresponding empirical coefficients can be obtained by looking up a table based on the link's operating frequency and polarization using the power-law model, according to ITU-R Recommendation P.838 (Standard Raindrop Spectrum Model Coefficients).
[0036] Then, the obtained pure path rainfall attenuation is substituted into the power-law model constructed with the corresponding empirical coefficients, and the average rainfall intensity of the link path is obtained by inversion solution.
[0037] Step 104: Based on the average rainfall intensity along the path, perform spatial interpolation optimization on the preset interference area to determine the two-dimensional rainfall intensity distribution map of the interference area.
[0038] For example, since rainfall conditions are dynamic, interference strategies will also change accordingly. In order to simultaneously analyze the attenuation changes of both the interference signal and the signal being interfered with, this application constructs a two-dimensional rainfall intensity distribution map of the interference area based on the path-averaged rainfall intensity of multiple microwave links, providing a data foundation for the dynamic analysis of rainfall attenuation on both sides.
[0039] Specifically, based on the path-averaged rainfall intensity, spatial interpolation optimization is performed on the preset interference area to determine the two-dimensional rainfall intensity distribution map of the interference area. This includes: acquiring the DEM terrain corresponding to the interference area and discretizing the DEM terrain into corresponding grid cells; wherein the information of the grid cell includes: grid cell coordinates and the set of rainfall intensity of the grid cell; constructing feature vectors corresponding to each grid cell based on the path-averaged rainfall intensity and the geometric features of each communication link and the grid cell; wherein the geometric features include: the shortest distance from the center of the grid cell to the projection of each link, the link crossing the grid state, and the azimuth angle of the grid relative to the start and end points of the link; inputting the feature vectors into a preset random forest model and training until the model converges to obtain the rainfall intensity parameters corresponding to the grid cell; interpolating the rainfall intensity parameters until the discrete error is extended to all grid cells to determine the rainfall intensity error surface of the interference area; integrating the rainfall intensity error surface into the rainfall intensity parameters to determine the two-dimensional rainfall intensity distribution map.
[0040] In one embodiment, firstly, the entire interference region is discretized into... Each grid cell, For each sample, the sample label is set to the center coordinates of the grid cell, and the rain gauge parameters corresponding to the grid are obtained (if none are available, other available measurement methods can be selected) as supervision data for model training.
[0041] Furthermore, for grids without supervised data (i.e., ground truth), the preliminary inversion results of tomography algorithms can be used as pseudo-labels, and the processed grids can also be used for training.
[0042] Then, for each communication link, its geometric features relative to the grid are calculated to construct the feature vector corresponding to the grid. The geometric features include: the shortest distance from the center of the grid cell to the projection of each link, the link crossing the grid state, and the azimuth angle of the grid relative to the start and end points of the link.
[0043] For the link-based rainfall intensity feature, which is the path-average rainfall intensity of each link, the above data is input into the random forest model. For grids with observed ground truth values, the loss function assigns higher weights, and vice versa (for grids with pseudo-labels).
[0044] Finally, the trained machine learning model is used to predict the rainfall intensity of all grids, and the prediction error is calculated for each grid containing the true value of the rain gauge.
[0045] By using the Kriging interpolation method, the discrete error is extended to the entire grid to obtain the rainfall intensity error surface, and the rainfall intensity error surface is integrated into the rainfall intensity parameters to determine the two-dimensional rainfall intensity distribution map.
[0046] Step 105: Calculate the vertical distribution parameters of the real-time atmospheric refractive index, and based on the vertical distribution parameters, calculate the signal propagation time delay corresponding to the ray trajectory of the interfering target to obtain the corrected coordinates of the interfering target.
[0047] For example, when the location coordinates of multiple reconnaissance stations are determined, the atmospheric waveguide-corrected true target coordinates are calculated based on the target signal arrival time difference (TDOA) measurements obtained from each reconnaissance station, so that the calculated TDOA correction value matches the measurement value, thereby reducing the positioning error of distant interfering targets.
[0048] Specifically, based on the vertical distribution parameters, the signal propagation time delay corresponding to the ray trajectory of the interfering target is calculated to obtain the corrected coordinates of the interfering target. This includes: obtaining the initial coordinates of the interfering target and the coordinates of the reconnaissance station, and calculating the ray azimuth angle based on the initial coordinates and the reconnaissance station coordinates; calculating the ray path from the reconnaissance station coordinates to the interfering target through numerical integration based on the ray azimuth angle; calculating the propagation time delay of the interfering signal on the ray path, and calculating the arrival time difference from each reconnaissance station coordinate to the interfering target based on the propagation time delay; correcting the arrival time difference measurement value through the arrival time difference, determining the arrival time difference correction amount, and locating the arrival time difference correction amount to obtain the corrected coordinates of the interfering target.
[0049] In one embodiment, the atmosphere is first discretized along the height direction. In a spherically symmetric layered medium, the ray path satisfies Snell's law. Numerical integration is performed using arc length or height parameters, or ray tracing in a Cartesian coordinate system can be used to solve for the actual propagation time from the reconnaissance station to the jamming target.
[0050] In atmospheric waveguide environments, there may be multiple paths (multiple elevation angles corresponding to the same endpoint). It is necessary to select a reasonable path, and in general, the path with the shortest propagation time or the main energy path should be selected.
[0051] Then, for any two reconnaissance stations, a time delay correction is defined between them. This correction represents the deviation of the TDOA change caused by atmospheric waveguides from the linear assumption. The time difference of arrival correction is determined by correcting the time difference of arrival measurement value for the time difference of arrival.
[0052] Finally, the TDOA positioning parameters are corrected, the corrected TDOA measurement value is determined, and the time difference of arrival correction is used for time difference of arrival positioning to obtain the corrected coordinates of the interfering target.
[0053] For example, the TDOA values of the radiation signals from sea surface targets measured by three reconnaissance stations are first calculated using the traditional straight-line TDOA to obtain the initial coordinates of the candidate targets.
[0054] Then, ray tracing was performed on each reconnaissance station using a spherical layered atmospheric model with a layer thickness of 50m and the refractive index determined by the profile of the vertical distribution parameters.
[0055] Starting from reconnaissance station 1, the integration begins with an azimuth of 78.3° pointing towards the candidate target and an initial elevation angle of 0.5°. When the integration reaches a target height of 15m, the horizontal deviation at the endpoint is 320m. After adjusting the elevation angle to 0.65° using the target shooting method, the deviation is less than 10m. At the same time, the propagation delay is calculated. Similarly, the propagation delays of reconnaissance stations 2 and 3 can be obtained.
[0056] After calculating the linear time delay, the time delay correction is obtained. The original TDOA measurement value is corrected using the correction. Then, the new target coordinates are calculated. After N iterations, the calculation converges to obtain the corresponding interference target coordinates and error distance. This error distance should meet the preset index.
[0057] Step 106: Based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, determine the interference strategy corresponding to the interference target through link interference parameter calculation and interference window prediction.
[0058] For example, based on OISAC real-time inversion of rainfall fields, the parameter settings of the interference signal and the time window for the interference signal to be emitted under the current rainfall environment are determined by calculating the link interference parameters and predicting the interference window. This application can dynamically optimize the output strategy of the interference signal under the interference rainfall environment by predicting the decision parameters of the interference signal and the interfered signal under each link of both parties.
[0059] Specifically, based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, the interference strategy corresponding to the interference target is determined through link interference parameter calculation and interference window prediction. This includes: determining the original signal-to-noise ratio (SNR) of the link corresponding to the interference target based on the corrected coordinates of the interference target through electromagnetic propagation simulation; inputting the original SNR into a rain attenuation power-law model to obtain the current SNR of the link corresponding to the interference target; extracting the rain area path from the transmission point of the interference signal to the receiving point of the link corresponding to the interference target, and calculating the additional rain attenuation loss corresponding to the rain area path; defining the susceptibility index of the link corresponding to the interference target based on the additional rain attenuation loss; predicting the rain area movement trend on the two-dimensional rainfall intensity distribution map to obtain the link movement trend corresponding to the interference target; updating the susceptibility index based on the link movement trend, and marking the current time window as the interference time window when the susceptibility index exceeds a preset interference threshold; and matching the corresponding interference signal parameters according to the interference time window to determine the interference strategy.
[0060] In one embodiment, a two-dimensional rainfall intensity distribution map of the current moment is obtained, with each one-kilometer square grid point marked with a rainfall intensity value, and the coordinates of the receiving station and transmitting station of the long-distance communication link have been obtained through correction.
[0061] First, based on the corrected coordinates of the interfered communication link, the original signal-to-noise ratio (SNR) of the link under clear-sky conditions is calculated through electromagnetic propagation simulation. Free-space loss and atmospheric absorption are calculated to obtain the received power. Given a low noise level, the difference between these two losses yields the corresponding original SNR.
[0062] The original signal-to-noise ratio (SNR) is input into the rain attenuation power-law model to obtain the SNR under the current environment. The eight grid points traversed by the link path are read from the rainfall intensity distribution map, and the rainfall intensity value of each grid point is determined. Using a preset rain attenuation coefficient, the rain intensity is first calculated to the power of the coefficient for each grid point, then multiplied by a fixed coefficient to obtain the unit distance attenuation for that grid point, and then multiplied by the path length. The attenuation values of the eight grid points are summed to obtain the total path rain attenuation, and this attenuation value is subtracted from the original SNR to obtain the current SNR.
[0063] Then, the rain path between our jamming signal transmission point and the enemy's link receiving point is extracted to obtain the additional rain attenuation loss of the jamming signal on the rain-penetrating path. The jamming intensity to be applied is set, and the current signal-to-noise ratio (SNR) is subtracted from the jamming intensity to obtain the equivalent SNR at the enemy's receiver after jamming. The vulnerability index is obtained by subtracting the equivalent SNR after jamming from the clear-sky SNR and then dividing by the clear-sky SNR.
[0064] Furthermore, the movement trend of the rain zone is predicted based on the two-dimensional rainfall intensity distribution map. The movement of the rain clusters is estimated by continuously observing the most recent six frames of the rainfall map (each frame spaced one minute apart). The prediction for the next twenty minutes indicates that the receiving station of the other link will enter an area with higher, unchanged, or lower rainfall intensity, as well as changes in the path of the interfering signal through the rain. Based on this movement trend, the susceptibility index is recalculated every five minutes. When the predicted value exceeds the threshold, the current time window is marked as the interference time window.
[0065] Finally, the interference strategy is determined by matching the interference signal parameters according to the interference time window. During the window period, it is recommended to use the same frequency as the other party, and the interference pattern is noise amplitude modulation.
[0066] It is important to note that the determination of interference power is based on the power required to suppress the link under clear sky conditions and the additional rain attenuation loss of the corresponding rain-penetrating path of the interference signal, and the rain-penetrating path should be shortened as much as possible.
[0067] Furthermore, after determining the interference strategy corresponding to the interference target based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, through link interference parameter calculation and interference window prediction, the method also includes: executing the interference strategy to obtain the interference signal parameters of the link corresponding to the interference target, and evaluating the interference effect of the interference signal parameters to determine the interference success rate; adjusting the interference threshold and empirical coefficient according to the interference success rate to obtain the updated interference strategy.
[0068] In one embodiment, following the jamming strategy generated in Embodiment 1, the UAV jamming platform is controlled via the eastward interface to jam the enemy's communication link within a marked jamming time window. After the jamming ends, the reconnaissance receiver collects the signal parameters after jamming and evaluates the jamming effect.
[0069] Before jamming, the baseline signal parameters of the enemy link are recorded, including signal-to-noise ratio, bit error rate, and communication status. After jamming begins, the reconnaissance receiver continuously monitors the signal changes of the same link.
[0070] If a significant decrease in signal-to-noise ratio, a significant increase in bit error rate, the occurrence of retransmission requests, or communication interruptions are detected, the interference is deemed effective; otherwise, the interference is deemed ineffective or partially effective. The interference success rate is calculated based on the degree of difference in signal parameters before and after the interference.
[0071] Based on the interference success rate and measured deviation, adjust the corresponding interference threshold. If the interference success rate is higher than expected and the interference triggering time is earlier, it indicates that the current threshold setting is too low. In this case, the interference threshold should be appropriately increased to avoid premature triggering. If the interference success rate is lower than expected and the interference triggering time is later or not triggered, it indicates that the current threshold setting is too high. In this case, the interference threshold should be appropriately decreased to ensure that interference is initiated in a timely manner at a favorable time.
[0072] For the rain attenuation power-law model coefficients under the current operating frequency and polarization, adjustments are made based on the ratio of measured additional attenuation to predicted attenuation. If the measured attenuation is greater than the predicted attenuation, the coefficients are increased proportionally; if the measured attenuation is less than the predicted attenuation, the coefficients are decreased proportionally. The adjusted coefficients are stored in an empirical database for subsequent link calculations under the same conditions.
[0073] The updated interference strategy is stored as empirical rules, including parameters such as interference threshold and rain attenuation model coefficients. In subsequent approximate scenario analysis, the updated parameters can be directly called, and the feedback adjustment completes the optimization logic loop through the westward interface.
[0074] In one embodiment, after determining the interference time window and interference frequency, it is necessary to calculate the minimum power value required to perform the interference to prevent premature exposure.
[0075] First, determine the standard jamming power required to suppress the target link under clear-sky conditions. Then, based on the two-dimensional rainfall intensity distribution map, extract the rain zone path traversed from the proposed jammer deployment location to the target link receiving point, and calculate the attenuation of electromagnetic waves at each rain zone grid point segment by segment. The minimum required jamming power can be obtained by multiplying the clear-sky standard power by a factor corresponding to the total additional attenuation in decibels.
[0076] For deployment locations, for each candidate location, calculate the rain penetration path length and total additional attenuation from that location to the target link receiver. Also, determine whether the location is at the edge or inside the rain zone. If a location minimizes the rain penetration path and additional attenuation for the interfering signal, and is located inside or near the edge of the rain zone, then that location is marked as the optimal deployment point. For aerial platforms (such as drones), consider the variations in rain penetration paths at different flight altitudes: flying above the rain zone avoids rain penetration attenuation but may lose rain cover; flying inside or below the rain zone utilizes rain cover but requires compensation for rain penetration attenuation. Adjustments can be made dynamically based on actual needs.
[0077] Furthermore, regarding the execution of the jamming strategy, after receiving the instruction, the execution unit first performs instruction parsing and verification. The UAV platform automatically plans its flight path based on the waypoint sequence in the instruction, and hovers or cruises after reaching the designated location; the ground mobile jamming station then moves to the designated coordinates and deploys its antenna.
[0078] At the start of the interference time window, the execution unit precisely adjusts the power amplifier to the required transmission power level according to the frequency and pattern set in the instruction, and begins transmitting the interference signal. The execution unit transmits its own status information in real time, including current location, actual transmission power, operating mode, equipment temperature, etc., to monitor whether it is executing the instruction correctly.
[0079] If new adjustment instructions are received during execution, the execution unit can dynamically adjust parameters without interrupting the task. After the interference ends, the execution unit automatically shuts down the transmitter and generates an execution log, which includes records of actual transmission duration, average power, antenna pointing, etc., providing a data basis for interference assessment.
[0080] The above are embodiments of the method proposed in this application. Based on the same inventive concept, embodiments of this application also provide a signal interference strategy generation device based on deep learning, the structure of which is as follows: Figure 2 As shown.
[0081] Figure 2 This is a schematic diagram of the internal structure of a signal interference strategy generation device based on deep learning, provided as an embodiment of this application. Figure 2 As shown, the device includes: at least one processor 201; and a memory 202 communicatively connected to the at least one processor; wherein the memory 202 stores instructions executable by the at least one processor, which are executed by the at least one processor 201 to enable the at least one processor 201 to implement a signal interference strategy generation method based on deep learning.
[0082] Some embodiments of this application provide corresponding to Figure 1 A non-volatile computer storage medium based on deep learning for generating signal interference strategies stores computer-executable instructions, which, when executed, can realize a signal interference strategy generation method based on deep learning.
[0083] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments for IoT devices and media are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.
[0084] The systems, media, and methods provided in this application are one-to-one correspondences. Therefore, the systems and media also have similar beneficial technical effects as their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the systems and media will not be repeated here.
[0085] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0086] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0087] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0088] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0089] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0090] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0091] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0092] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0093] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. A signal interference strategy generation method based on deep learning, characterized in that, The method includes: Acquire the reference level sampling data and real-time level sampling data of each communication link, and calculate the initial path rain attenuation value corresponding to the reference level sampling data and the real-time level sampling data; Based on the initial path rain attenuation value, the wet antenna attenuation component of the real-time level sampling data is filtered by digital filtering to determine the pure path rain attenuation value of each communication link. By using a power-law model of rain attenuation, the pure path rain attenuation value is inverted to obtain the path average rain intensity corresponding to each communication link; Based on the average rainfall intensity along the path, spatial interpolation optimization is performed on the preset interference area to determine a two-dimensional rainfall intensity distribution map of the interference area; Calculate the vertical distribution parameters of the real-time atmospheric refractive index, and based on the vertical distribution parameters, calculate the signal propagation time delay corresponding to the ray trajectory of the interfering target to obtain the corrected coordinates of the interfering target; Based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, the interference strategy corresponding to the interference target is determined through link interference parameter calculation and interference window prediction.
2. The signal interference strategy generation method based on deep learning according to claim 1, characterized in that, Calculating the initial path rain attenuation value corresponding to the reference level sampling data and the real-time level sampling data specifically includes: Calculate the total additional attenuation of the links corresponding to the reference level sampling data and the real-time level sampling data; Obtain the corresponding non-rain attenuation for each communication link, and subtract the non-rain attenuation from the total additional attenuation to obtain the initial path rain attenuation value.
3. The signal interference strategy generation method based on deep learning according to claim 1, characterized in that, Based on the initial path rain attenuation value, the wet antenna attenuation component of the real-time level sampling data is filtered by digital filtering to determine the pure path rain attenuation value of each communication link, specifically including: Extract the attenuation changes corresponding to the real-time level sampling data to determine the dynamic attenuation characteristics of each communication link; Based on the dynamic attenuation characteristics, the wet antenna attenuation component and path rain attenuation in the initial path rain attenuation value are distinguished by digital filtering, and the wet antenna attenuation component in the initial path rain attenuation value is filtered out to determine the pure path rain attenuation value.
4. The signal interference strategy generation method based on deep learning according to claim 1, characterized in that, By inverting the pure path rain attenuation value using a rain attenuation power-law model, the path-average rainfall intensity corresponding to each communication link is obtained, specifically including: Obtain the operating frequency and polarization of each communication link, and match the corresponding empirical coefficients in the rain attenuation power law model based on the operating frequency and polarization. Based on the empirical coefficients, the pure path rain attenuation value is input into the rain attenuation power law model, and the path average rain intensity corresponding to each communication link is obtained by inversion solution.
5. The signal interference strategy generation method based on deep learning according to claim 1, characterized in that, Based on the average rainfall intensity along the path, spatial interpolation optimization is performed on the preset interference area to determine a two-dimensional rainfall intensity distribution map of the interference area, specifically including: Obtain the DEM terrain corresponding to the interference area, and discretize the DEM terrain into corresponding grid cells; wherein, the information of the grid cell includes: grid cell coordinates and the set of rainfall intensity of the grid cell; Based on the average rainfall intensity along the path and the geometric features of each communication link and the grid cell, a feature vector corresponding to each grid cell is constructed; wherein, the geometric features include: the shortest distance from the center of the grid cell to the projection of each link, the link crossing the grid state, and the azimuth angle of the grid relative to the start and end points of the link; The feature vector is input into a preset random forest model and trained until the model converges to obtain the rainfall intensity parameters corresponding to the grid cell. The rainfall intensity parameters are interpolated until the discrete error is extended to all grid cells to determine the rainfall intensity error surface of the interference area; The rainfall intensity error surface is integrated into the rainfall intensity parameter to determine the two-dimensional rainfall intensity distribution map.
6. The signal interference strategy generation method based on deep learning according to claim 1, characterized in that, Based on the vertical distribution parameters, the signal propagation time delay corresponding to the ray trajectory of the interfering target is calculated to obtain the corrected coordinates of the interfering target, specifically including: Obtain the initial coordinates of the jamming target and the coordinates of the reconnaissance station, and calculate the ray azimuth angle based on the initial coordinates and the coordinates of the reconnaissance station; Based on the ray azimuth angle, the ray path from the reconnaissance station coordinates to the interference target is obtained through numerical integration. Calculate the propagation time delay of the interference signal corresponding to the ray path, and based on the propagation time delay, calculate the arrival time difference from the coordinates of each reconnaissance station to the interference target; By correcting the arrival time difference measurement value using the arrival time difference, the arrival time difference correction amount is determined, and the arrival time difference correction amount is used for arrival time difference positioning to obtain the corrected coordinates of the interference target.
7. The signal interference strategy generation method based on deep learning according to claim 1, characterized in that, Based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, the interference strategy corresponding to the interference target is determined through link interference parameter calculation and interference window prediction, specifically including: Based on the corrected coordinates of the interference target, the original signal-to-noise ratio of the link corresponding to the interference target is determined through electromagnetic propagation simulation; The original signal-to-noise ratio is input into the rain attenuation power-law model to obtain the current signal-to-noise ratio of the link corresponding to the interference target; Extract the rain zone path from the interference signal transmission point to the corresponding link receiving point of the interference target, and calculate the additional rain attenuation loss corresponding to the rain zone path; Based on the additional rain attenuation loss, the interference susceptibility index of the link corresponding to the interference target is defined; The movement trend of the rain area is predicted based on the two-dimensional rainfall intensity distribution map to obtain the link movement trend corresponding to the interference target; Based on the link movement trend, the interference susceptibility index is updated, and when the interference susceptibility index exceeds a preset interference threshold, the current time window is marked as an interference time window. Based on the interference time window, the corresponding interference signal parameters are matched to determine the interference strategy.
8. The signal interference strategy generation method based on deep learning according to claim 1, characterized in that, After determining the interference strategy corresponding to the interference target based on the two-dimensional rainfall intensity distribution map and the corrected coordinates of the interference target, through link interference parameter calculation and interference window prediction, the method further includes: The interference strategy is executed to obtain the interference-post signal parameters of the link corresponding to the interference target, and the interference effect is evaluated on the interference-post signal parameters to determine the interference success rate. Based on the interference success rate, the interference threshold and empirical coefficient are adjusted to obtain the updated interference strategy.
9. A signal interference strategy generation device based on deep learning, characterized in that, The device includes: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, which are executed by the at least one processor to enable the at least one processor to implement a deep learning-based signal interference strategy generation method as described in any one of claims 1-8.
10. A non-volatile computer storage medium generated based on a deep learning-based signal interference strategy, storing computer-executable instructions, characterized in that... When the computer-executable instructions are executed, they are sufficient to implement the signal interference strategy generation method based on deep learning as described in any one of claims 1-8.