A method and system for detecting an active phased array radar based on SDR

Abstract

This application provides an active phased array radar detection method and system based on SDR. By acquiring historical spectrum data, it uses HMM / particle filtering to predict the next-hop frequency band and generate a pre-interference template. The fast channel transmits directionally in the predicted sub-band according to the template, while the slow channel performs system identification, protocol analysis, DOA and micro-Doppler analysis, and backfills waveforms and beams. For multiple targets, a three-dimensional joint scheduling system of "time slot-frequency band-beam" is constructed, allocating frequencies according to threat level and inherently constraining frequency bands / sectors, power, sidelobes, etc. The system includes an antenna array, SDR, prediction, parallel processing, scheduling, threat assessment, compliance management, and an industrial control computer, achieving low-latency, accurate, and compliant suppression of fast frequency hopping and concurrent targets.

Description

Technical Field

[0001] This application relates to the field of radar, and in particular to a detection method and system for an active phased array radar based on SDR. Background Technology

[0002] With the widespread application of low-altitude small unmanned aerial vehicle (UAV) technology, especially in civilian fields such as express delivery, agricultural monitoring, and environmental protection, traditional UAV detection and countermeasure systems are facing increasing challenges. Small UAVs are characterized by their small size, low flight altitude, and high speed, and some UAVs employ multi-band and multi-modulation communication technologies. This makes it difficult for traditional radar systems and spectrum detection methods to effectively detect and counter them in low-altitude environments. Especially in complex electromagnetic environments, the frequency band jumping and the complexity of communication systems further increase the difficulty of detection.

[0003] Currently, most existing drone detection and countermeasure technologies rely on single-band radar or fixed-band radar systems, typically employing traditional electromagnetic spectrum detection methods. These technologies have several limitations: First, insufficient frequency coverage makes them unable to address issues such as target frequency jumps or spectrum changes, especially when drones jump frequencies due to improper operation or environmental factors, making it difficult for existing systems to quickly identify and respond. Second, most existing systems use a serial processing approach, identifying the target before countermeasures, which results in significant reaction delays and unsatisfactory countermeasure effects when facing frequency jumps and complex modulation methods. Finally, in multi-target scenarios, when multiple drones simultaneously use the same or similar frequency bands, existing systems are prone to resource conflicts or accidental attacks, and fail to adequately schedule resources in the three dimensions of time, frequency, and space, making it impossible to accurately counter each target.

[0004] Traditional technological approaches typically rely on single-band solutions or costly, complex hardware to enhance system detection capabilities. However, these methods fail to provide sufficient flexibility and adaptability, particularly when faced with issues such as frequency hopping by drones, improper operation (e.g., incorrect frequency switching, signal interference), and complex modulation methods, where technological bottlenecks remain.

[0005] Therefore, there is an urgent need for a drone detection and countermeasure solution that can overcome the limitations of existing technologies, effectively address drone frequency hopping, improve system response speed, and achieve precise resource scheduling in multi-target scenarios to ensure maximum countermeasure effectiveness, especially in civilian drone applications to address problems caused by improper operation or misoperation. Summary of the Invention

[0006] The purpose of this application is to provide an SDR-based active phased array radar detection method that can effectively cope with frequency band hopping of multiple UAVs.

[0007] According to one aspect of this application, an active phased array radar detection method based on SDR is provided, comprising the following steps:

[0008] When the UAV is in fast frequency hopping mode, the S100 collects historical spectrum data of the target UAV.

[0009] Based on the historical spectrum data, S200 predicts the frequency hopping band of the target UAV at the next moment through a hidden Markov model or particle filter algorithm, and generates a corresponding pre-interference template.

[0010] The S300 starts fast-channel processing threads and slow-channel processing threads to execute in parallel, wherein:

[0011] The fast-channel processing thread controls the phased array antenna array to transmit a first interference signal to the predicted frequency hopping band based on the pre-interference template.

[0012] The slow channel processing thread performs detailed analysis on the spectrum data, including system identification, protocol analysis, angle of arrival estimation, and micro-Doppler analysis, and generates accurate interference waveform templates and beam weights;

[0013] When multiple target drones exist, the S400 constructs a resource scheduling model; based on the threat level and frequency hopping prediction results of each target drone, it dynamically allocates time slots, frequency bands and beam resources to each target drone.

[0014] The scheduling model described in S500 has built-in compliance constraints, including protection bands, protection sectors, power limits, beam isolation, and sidelobe limits, to ensure that countermeasures operations comply with the system's compliance requirements.

[0015] In one specific embodiment, in the frequency hopping prediction step, the hidden Markov model or particle filter algorithm dynamically updates the frequency band hopping probability matrix by analyzing the target's historical frequency band hopping data in real time, and predicts the most likely frequency band for the next moment based on the matrix.

[0016] In one specific embodiment, in the parallel processing step, the response time of the fast channel processing thread is less than 50 milliseconds, and the slow channel processing thread fills back the generated accurate interference waveform template and beam weights to the fast channel processing thread for updating the pre-interference template for the next cycle.

[0017] In one specific embodiment, the threat level assessment in the resource scheduling step includes:

[0018] The threat score for each target drone is dynamically calculated using a threat assessment model based on one or more factors, including the size, type, speed, distance from the protected sector, and frequency hopping speed.

[0019] In one specific embodiment, the implementation of the compliance constraints includes:

[0020] A beamforming algorithm is used to control the transmit beam of the phased array antenna array, ensuring that the interference between multiple beams transmitted simultaneously is below a preset threshold.

[0021] By setting a protection radius and exclusion list, interference signals can be avoided from being emitted within the protected frequency band and protected sector;

[0022] Real-time monitoring of transmission power and equipment temperature, and dynamic adjustment of power distribution to ensure safe and stable operation of the equipment.

[0023] In one specific embodiment, the joint scheduling, by considering intrinsic compliance constraints, ensures that the system resource scheduling process complies with compliance requirements such as protected frequency bands, protected sectors, power, and thermal management, thereby avoiding interference to non-target areas.

[0024] In one specific embodiment, the three-dimensional resource scheduling further includes the following steps:

[0025] S401: Through time slot-level scheduling, countermeasure time slots are dynamically allocated based on the threat level and priority of the target, with higher priority targets allocated more time slot resources;

[0026] S402: Dynamically allocate frequency band resources based on the target's frequency band and frequency hopping prediction results to avoid conflicts when multiple targets use the same frequency band;

[0027] S403: Based on the target's spatial location and trajectory, dynamically adjust the beam pointing and gain to ensure that each target can be accurately jammed;

[0028] S405: Automatically excludes protected frequency bands and protected sectors to prevent interference signals from affecting non-target or sensitive areas;

[0029] S406: Based on the power output and heat dissipation capacity of the equipment, limit the power of interference signals during scheduling to avoid overload or overheating;

[0030] S407: By controlling beam spacing and sidelobe leakage, it ensures that the jamming signal will not affect other targets or areas, thus avoiding accidental damage.

[0031] According to another aspect of this application, an active phased array radar detection system based on SDR is provided, including a phased array antenna array, a radio SDR platform, and a frequency hopping prediction module; the phased array antenna array is used to receive and transmit electromagnetic signals; the radio SDR platform, connected to the antenna array, is used to sample and demodulate the received signals; the frequency hopping prediction module is used to run a hidden Markov model or a particle filter algorithm to predict the frequency hopping band of the target UAV.

[0032] The parallel processing module includes a fast-channel processing unit, a slow-channel processing unit, a resource scheduling module, a threat assessment module, a compliance constraint management module, and an industrial control computer. The fast-channel processing unit transmits a first interference signal based on the frequency hopping prediction result; the slow-channel processing unit performs fine signal analysis and generates precise interference parameters; the resource scheduling module is used for resource scheduling in multi-target scenarios; the threat assessment module is used to calculate the threat level of each target UAV in real time; the compliance constraint management module is used to ensure that the scheduling process complies with various constraints; and the industrial control computer is used to run all the above modules.

[0033] In one specific embodiment, the fast channel processing unit and the slow channel processing unit interact with each other through shared memory, and the fast channel processing unit can directly access the control interface of the phased array antenna array.

[0034] In one specific embodiment, the resource scheduling module employs a combination of time slot allocation algorithm, frequency band allocation algorithm, and beam allocation algorithm to achieve joint optimization scheduling of three-dimensional resources in time, frequency, and space.

[0035] Therefore, this application presents an active phased array radar detection method and system based on SDR, which integrates software radio and active phased array radar. Utilizing SDR to provide wideband sampling and real-time demodulation, the system can flexibly adapt to different frequency bands and cope with frequency hopping and complex modulation schemes. It predicts the frequency hopping of target UAVs using a hidden Markov model or particle filter algorithm and generates corresponding pre-jamming templates. Simultaneously, through parallel processing of fast and slow channels, the fast channel quickly identifies the target's frequency band and modulation scheme, initiating jamming in advance based on the prediction results, while the slow channel performs detailed analysis of the target, including system identification, micro-Doppler analysis, and angle of arrival estimation, ensuring the accuracy of the countermeasure signal. Furthermore, through joint scheduling, in multi-target scenarios, time slots, frequency bands, and beam resources are dynamically allocated according to the target's threat level, frequency hopping prediction, and spatial location. This effectively solves the problems of insufficient frequency band coverage, excessive time delay, and difficulty in multi-target resource scheduling in traditional systems, achieving efficient detection and precise countermeasures against low-altitude small UAVs. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0037] Figure 1 This is a flowchart of an active phased array radar detection method based on SDR;

[0038] Figure 2 This is a module relationship diagram of an SDR-based active phased array radar detection system.

[0039] Explanation of icon numbers:

[0040] 510. Phased array antenna array; 520. Radio SDR platform; 530. Frequency hopping prediction module; 540. Parallel processing module; 541. Fast channel processing unit; 542. Slow channel processing unit; 550. Resource scheduling module; 560. Threat assessment module; 570. Compliance constraint management module; 210. Industrial control computer; 100. An active phased array radar detection system based on SDR. Detailed Implementation

[0041] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings. Preferred embodiments of this application are shown in the drawings. However, this application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of this application.

[0042] It should be noted that when a component is said to be "fixed to" another component, it can be directly attached to the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0043] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0044] Please refer to Figure 1 - Figure 2 This application provides an SDR-based active phased array radar detection method, comprising the following steps:

[0045] When the UAV is in fast frequency hopping mode, the S100 collects historical spectrum data of the target UAV.

[0046] Based on the historical spectrum data, S200 predicts the frequency hopping band of the target UAV at the next moment through a hidden Markov model or particle filter algorithm, and generates a corresponding pre-interference template.

[0047] The S300 starts fast-channel processing threads and slow-channel processing threads to execute in parallel, wherein:

[0048] The fast-channel processing thread controls the phased array antenna array to transmit a first interference signal to the predicted frequency hopping band based on the pre-interference template.

[0049] The slow channel processing thread performs detailed analysis on the spectrum data, including system identification, protocol analysis, angle of arrival estimation, and micro-Doppler analysis, and generates accurate interference waveform templates and beam weights;

[0050] When multiple target drones exist, the S400 constructs a resource scheduling model; based on the threat level and frequency hopping prediction results of each target drone, it dynamically allocates time slots, frequency bands and beam resources to each target drone.

[0051] The scheduling model described in S500 has built-in compliance constraints, including protection bands, protection sectors, power limits, beam isolation, and sidelobe limits, to ensure that countermeasures operations comply with the system's compliance requirements.

[0052] In one specific embodiment, in the frequency hopping prediction step, the hidden Markov model or particle filter algorithm dynamically updates the frequency band hopping probability matrix by analyzing the target's historical frequency band hopping data in real time, and predicts the most likely frequency band for the next moment based on the matrix.

[0053] In one specific embodiment, in the parallel processing step, the response time of the fast channel processing thread is less than 50 milliseconds, and the slow channel processing thread fills back the generated accurate interference waveform template and beam weights to the fast channel processing thread for updating the pre-interference template for the next cycle.

[0054] In one specific embodiment, the threat level assessment in the resource scheduling step includes:

[0055] The threat score for each target drone is dynamically calculated using a threat assessment model based on one or more factors, including the size, type, speed, distance from the protected sector, and frequency hopping speed.

[0056] In one specific embodiment, the implementation of the compliance constraints includes:

[0057] A beamforming algorithm is used to control the transmit beam of the phased array antenna array, ensuring that the interference between multiple beams transmitted simultaneously is below a preset threshold.

[0058] By setting a protection radius and exclusion list, interference signals can be avoided from being emitted within the protected frequency band and protected sector;

[0059] Real-time monitoring of transmission power and equipment temperature, and dynamic adjustment of power distribution to ensure safe and stable operation of the equipment.

[0060] In one specific embodiment, the joint scheduling, by considering intrinsic compliance constraints, ensures that the system resource scheduling process complies with compliance requirements such as protected frequency bands, protected sectors, power, and thermal management, thereby avoiding interference to non-target areas.

[0061] In one specific embodiment, the three-dimensional resource scheduling further includes the following steps:

[0062] S401: Through time slot-level scheduling, countermeasure time slots are dynamically allocated based on the threat level and priority of the target, with higher priority targets allocated more time slot resources;

[0063] S402: Dynamically allocate frequency band resources based on the target's frequency band and frequency hopping prediction results to avoid conflicts when multiple targets use the same frequency band;

[0064] S403: Based on the target's spatial location and trajectory, dynamically adjust the beam pointing and gain to ensure that each target can be accurately jammed;

[0065] S405: Automatically excludes protected frequency bands and protected sectors to prevent interference signals from affecting non-target or sensitive areas;

[0066] S406: Based on the power output and heat dissipation capacity of the equipment, limit the power of interference signals during scheduling to avoid overload or overheating;

[0067] S407: By controlling beam spacing and sidelobe leakage, it ensures that the jamming signal will not affect other targets or areas, thus avoiding accidental damage.

[0068] According to another aspect of this application, an active phased array radar detection system based on SDR is provided, including a phased array antenna array, a radio SDR platform, and a frequency hopping prediction module; the phased array antenna array is used to receive and transmit electromagnetic signals; the radio SDR platform, connected to the antenna array, is used to sample and demodulate the received signals; the frequency hopping prediction module is used to run a hidden Markov model or particle filter algorithm to predict the frequency hopping band of the target UAV; the parallel processing module includes a fast channel processing unit, a slow channel processing unit, a resource scheduling module, a threat assessment module, a compliance constraint management module, and an industrial control computer. The fast channel processing unit transmits a first interference signal based on the frequency hopping prediction result; the slow channel processing unit performs fine signal analysis and generates accurate interference parameters; the resource scheduling module is used to perform resource scheduling in multi-target scenarios; the threat assessment module is used to calculate the threat level of each target UAV in real time; the compliance constraint management module is used to ensure that the scheduling process complies with various constraints; the industrial control computer is used to run the above modules.

[0069] In one specific embodiment, the fast channel processing unit and the slow channel processing unit interact with each other through shared memory, and the fast channel processing unit can directly access the control interface of the phased array antenna array.

[0070] In one specific embodiment, the resource scheduling module employs a combination of time slot allocation algorithm, frequency band allocation algorithm, and beam allocation algorithm to achieve joint optimization scheduling of three-dimensional resources in time, frequency, and space.

[0071] See Figure 2 An active phased array radar detection system 100 based on SDR includes: a phased array antenna array 510 disposed on the upper part of a cabinet, with array elements deployed facing a second direction Y; a radio SDR platform 520 electrically coupled to the radio frequency transceiver port of 510; a frequency hopping prediction module 530; a parallel processing module 540 containing a fast channel processing unit 541 and a slow channel processing unit 542; a resource scheduling module 550; a threat assessment module 560; a compliance constraint management module 570; and an industrial control computer 210. Preferably, the industrial control computer 210 communicates with each module via a backplane bus, which can be either Ethernet or PCIe, and its orthographic projection covers the control plane of each module. This paragraph provides a common background for the embodiments.

[0072] Further, see Figure 1 In one implementation, the radio SDR platform writes broadband I / Q data from the phased array antenna array into a circular buffer managed by the industrial control computer with a fixed frame length, and marks the timestamp of the acquisition time and the current local oscillator frequency in the frame header of each frame (S100). The purpose is to provide "alignable, traceable, and replayable" data boundaries for all subsequent algorithms, so that frequency hopping prediction, short-time template matching, and beamforming can all operate on a unified time scale without cross-frame misalignment. The direct effect of doing so is to reduce the frequency band determination and phase estimation errors caused by data misalignment, thereby reducing the probability of fast channel false triggering or slow channel backfill mismatch.

[0073] Subsequently, the frequency hopping prediction module statistically analyzes the power spectrum changes over time in the near-time window and establishes the state transition relationship of the target frequency band, outputting the candidate frequency band for the next moment and the matching pre-interference template (S200). The purpose is to transform the problem of "where to shoot" into a small set selection, so that the fast channel does not need to blindly scan the entire bandwidth but directly loads a small number of high-confidence templates, thereby compressing the first-mover delay to the millisecond level. Its effect is to significantly improve the hit rate of the predicted subband and reduce invalid transmissions.

[0074] Next, the parallel processing module simultaneously activates two processing branches (S300): the fast channel processing unit immediately commands the phased array antenna array to transmit the first jamming signal on the predicted subband via the array weight control word issued by the industrial control computer according to the pre-interference template. The purpose is to occupy the possible working subband and suppress the target receiver with directional energy before the target completes the next hop. The effect is to change the serial link of "identification before action" to a parallel link of "prediction first, identification correction". At the same time, the slow channel processing unit performs fine analysis on the signal system, protocol consistency, angle of arrival and micro-Doppler within the same time window and generates a more accurate jamming waveform and beam weight, and fills the fast channel with an effective timestamp. The purpose is to correct the coarse matching of the fast channel with high-resolution estimation and replace the first-hand template with parameters that better fit the target characteristics in the next cycle. The effect is that the closed loop converges to a higher jamming gain and a lower sidelobe leakage cycle by cycle.

[0075] When there are multiple targets, the resource scheduling module constructs a three-dimensional resource pool of "time slot - frequency band - beam" and combines the threat assessment results and frequency hopping prediction probability to allocate countermeasure time slots, occupied frequency bands and beam directions (S400) to each target. The purpose is to ensure that important targets are prioritized and that each target does not interfere with each other under limited hardware concurrency capabilities. The effect is to maintain a stable and executable scheduling table even in dense scenarios.

[0076] Finally, the compliance constraint management module injects the protected frequency band, protected sector, power limit, beam isolation and sidelobe limit as hard boundaries into the scheduler (S500). The purpose is to make regulatory and safety constraints a "feasible set" in advance. The effect is that any instruction issued to the phased array antenna array and fast channel processing unit naturally meets the regulations and equipment safety requirements without the need for post-correction.

[0077] It should be noted that in the S200 "prediction and template generation" step of the above embodiment, the frequency hopping prediction module 530 can operate along two complementary paths: one is the Hidden Markov Model (HMM) path, and the other is the Particle Filter (PF) path. Both only output the "set of candidate frequency bands for the next time step, including the center frequency, bandwidth, and their confidence order," and derive a pre-interference template based on this, which is then loaded by the fast-channel processing unit 541. At the same time, the probability distribution is transmitted to the resource scheduling module 550.

[0078] As a priori weight on the spectrum side, the purpose is to obtain forward-looking capability for fast frequency hopping with reasonable computational overhead, express uncertainty in probabilistic form, and allow subsequent scheduling and transmission paths to use this uncertainty information to make conservative or aggressive strategy choices. Its direct effect is that the fast channel can complete template readiness and transmission triggering in milliseconds without blind scanning, thereby transforming the serial bottleneck of "recognition and then action" into a parallel closed loop of "predictive action + recognition and correction".

[0079] In the Markov Model (HMM) path:

[0080] The frequency hopping prediction module discretizes the "operating frequency band" into several states (e.g., dividing the spectrum into N sub-bands with equal or adaptive width, where the selection of N matches the effective bandwidth of the radio SDR platform and the target hopping granularity, and the value is not a unique constraint). It uses the "power spectrum and system indication characteristics" as observations, and maintains the initial probability vector, state transition matrix, and observation likelihood parameters online. Whenever the radio SDR platform writes a new frame of timestamped power spectrum slices at 520, the frequency hopping prediction module calculates the observation likelihood of each state based on the sub-band energy, cyclic spectrum intensity, or pilot consistency index of that frame. It then performs small-step updates to the state transition matrix and observation model using online EM or incremental Baum-Welch. Subsequently, it gives the "most likely state at the next moment" and its adjacent states using forward-backward probability or Viterbi one-step extrapolation. The model generates a pre-interference template for each candidate state (the template includes the center frequency, template bandwidth, expected duration, and optional modulation placeholders) for high-probability states. The mechanism involves using the indirectly observable "real operating frequency band" as a hidden state and the "measured time-frequency energy and system clues" as observations. It utilizes the "temporal correlation of state transitions" to constrain the reachability of hops and employs online learning to follow scene changes, thus consistently providing a probability ranking for the next hop even with limited samples and noise. The engineering effect is that when the target frequency hopping statistics drift slowly or exhibit a phased pattern, the model can maintain prediction convergence with low computational cost. The fast channel can prioritize loading the templates corresponding to the first K states to achieve a higher hit rate, while the resource scheduling module 550 can directly use this probability vector as a priori for frequency band allocation, reducing attempts at co-frequency conflict.

[0081] In the particle filter (PF) path:

[0082] To avoid quantization errors caused by overly coarse discrete binning, the frequency hopping prediction module models and maintains a set of weighted particles (each particle represents a candidate center frequency and its weight; the number of particles is set according to computing power and latency targets). Upon arrival of each frame, the particles undergo a small random walk according to the state transition model or are extrapolated based on the identified frequency drift velocity. Then, the observation likelihood is calculated based on the subband energy, cyclic spectrum, or synchronization head consistency of the current frame, and the particle weights are updated. Subsequently, systematic or hierarchical resampling is performed to suppress the monopoly of a few high-weight particles and maintain particle diversity. Finally, several highest-weighted particles are output as candidate center frequencies for the next time step, and a pre-interference template is generated for each center frequency (the template bandwidth can be determined based on the particle weights). The mechanism is to approximate the posterior distribution with the sample set, allowing the center frequency to slide in the continuous domain and providing multi-peak coverage for sudden transitions. The observation likelihood is directly derived from the time-frequency slices of the radio SDR platform 520 and the system clues of the parallel processing module 540. Resampling ensures that "a small number of particles misfollowing" will not cause the entire distribution to collapse. Its engineering effect is that under the conditions of fine-grained frequency shift, drift rate change or multi-peak uncertainty, it can still output Top-K candidates that cover the real next hop. The fast channel can be pre-set with two to three templates in parallel to offset the uncertainty, while the resource scheduling module 550 can use the normalized weight of the particle set as an "occupancy probability map" to guide the collision avoidance allocation of the frequency band.

[0083] More preferably, in a preferred embodiment, the system automatically selects or combines two paths according to the scenario: when the spectrum environment is stable and the transitions are relatively regular, the Hidden Markov Model path is prioritized to obtain stable predictions with the lowest computing power; when an abnormal increase in frequency dispersion and transition amplitude or the presence of multiple peaks is detected, the particle filter path is switched or superimposed to obtain fine localization in the continuous domain; whether the two paths are used alone or superimposed, they only output the candidate frequency band set and confidence level without directly controlling the transmission. The actual first interference signal is still processed by the fast channel processing unit 541 on the deterministic path of "template-array weight-trigger", while the slow channel processing unit 542... In the next cycle, high-resolution system identification and angle of arrival estimation are used to correct the template and beam. The aim is to strictly limit the uncertainty of the prediction to the decision input level and keep the execution layer as a low-latency deterministic link. The effect is to gain the time advantage of "first move" while avoiding the risk of "model error being directly amplified to hardware transmission". In terms of parameter recommendations, the number of discrete subbands can be 64-256 to balance resolution and complexity, the number of particles can be 128-512 to balance coverage and latency, and Top-K is usually 2-3 to obtain sufficient redundancy without significantly increasing the fast channel loading time. The above values ​​are optional for engineering and do not constitute the only limitation.

[0084] Furthermore, the frequency hopping prediction module maintains a frequency band hopping probability matrix that is updated over time and performs online correction of the matrix based on new observations in each frame. Then, it calculates the most likely operating frequency band for the next moment from the matrix and outputs the corresponding center frequency and bandwidth to generate a pre-interference template. The purpose is to characterize non-stationary frequency hopping with a "data-driven time-varying transition relationship" instead of relying on fixed priors, so that the model can continue to closely approximate the real distribution when the scene changes. The direct effect is that the confidence interval of the next hop position is narrower and the template is filled faster.

[0085] Furthermore, in one implementation, the end-to-end response of the fast-channel processing unit is limited to less than 50 milliseconds. It completes the sub-phase weight writing and transmission triggering by directly reading the pre-interference template and control parameters from the shared memory, bypassing multi-level copying, and calling the register interface of the phased array control driver. The slow-channel processing unit calculates the "precise interference waveform and beam weight" and writes it back to the shared memory along with the "available time window". The fast channel switches to enable it in the specified time window. The purpose is to ensure the reliable landing of the "first move" with the shortest memory and control path and to ensure the phase consistency of parameter switching with timestamps. The effect is to meet the hard time limit on the one hand and avoid pointing deviation caused by cross-cycle phase misalignment on the other hand.

[0086] Furthermore, in one implementation, the threat assessment module integrates one or more pieces of information, such as the target size or type, frequency domain indicators of speed or acceleration, geometric distance from the protected sector, and number of transitions per unit time, output by the slow channel, to generate a single threat score, which is periodically sent to the resource scheduling module. The purpose is to directly map the "degree of hazard" to the scheduling priority using a single scalar. The effect is that resource allocation corresponds one-to-one with task risk, prioritizing the protection of time slots, frequency bands, and beams for high-risk targets.

[0087] Furthermore, in one implementation, the compliance constraint management module applies sidelobe and inter-beam correlation constraints to concurrent beams during beamforming calculation, directly eliminates protected frequency bands and protected sectors as infeasible sets during the scheduling phase, and monitors transmit power and equipment temperature in real time during the execution phase, reducing power or limiting duty cycle when a threshold is triggered. The purpose is to transform "regulatory boundaries and equipment safety" into hard constraints during the solution process and closed-loop constraints during execution. The effect is to ensure that the spectrum and space do not exceed the limits and the devices are in a safe range even under multi-beam concurrent and high duty cycle operation.

[0088] Furthermore, in one implementation, the resource scheduling module writes the protection frequency band, protection sector, power and thermal management provided by the compliance constraint management module as hard constraints into the scheduling solver. When no solution is found, it first performs priority rollback on low threat score targets to release resources until a feasible solution is obtained. The purpose is to ensure that any issued solution is inherently compliant while maintaining the continuous executability of the system when resources are scarce. The effect is to avoid interference to non-target areas and to ensure stable operation under conflict conditions.

[0089] Furthermore, the resource scheduling module solves the problem in a three-layer sequence of "time-frequency-space": First, at the time layer, countermeasure time slots are allocated based on threat scores and historical supply conditions to avoid long-term starvation (S401). Then, at the frequency layer, non-conflicting frequency bands are allocated to each target based on frequency hopping prediction results, and rearrangement or relocation is performed when there is competition for the same frequency (S402). Finally, at the spatial layer, beam pointing and gain are calculated based on angle of arrival and motion trajectory, and adjustments are made by backtracking to the upper layer when beam isolation or sidelobe thresholds are not met (S403, S407). At the same time, protection bands and protection sectors are automatically excluded during the entire solution process, and power is reduced or duty cycle is shortened when power or heat dissipation is limited (S405, S406). The purpose is to quickly obtain a globally feasible near-optimal solution within a millisecond window through hierarchical decomposition. The effect is that it can still stably produce a scheduling table that meets both the goals of revenue and compliance when multiple targets are dense and the spectrum is tight.

[0090] Furthermore, the frequency hopping prediction module, parallel processing module, resource scheduling module, threat assessment module, and compliance constraint management module are deployed on the industrial control computer and exchange spectrum frames, templates, and beam weights through a shared memory queue. The fast channel processing unit directly accesses the control register of the phased array antenna array through a driver to complete the writing of the sub-phase weights and triggering of transmission. The slow channel processing unit writes the precise parameters of "valid for the next cycle" into the designated shared area and marks the effective window with a timestamp. The purpose is to shorten the path from decision to execution and reduce deterministic latency with "unified reference system + zero-copy path + direct control interface". The effect is to maintain low latency and high throughput even under high concurrency targets and high load scenarios.

[0091] Furthermore, the fast-channel processing unit and the slow-channel processing unit exchange templates, beam weights, and valid timestamps through a lock-free circular queue. The fast-channel processing unit has direct access to the phased array control registers. The purpose is to reduce context switching and kernel path length with zero copy and direct control, and to avoid cross-cycle parameter contamination with timestamp monotonicity constraints. The effect is faster end-to-end response, less jitter, and reduced risk of false launches.

[0092] Furthermore, the resource scheduling module integrates three types of processes—time slot allocation, frequency band allocation, and beam allocation—and executes them in cascade. After each layer is completed, the compliance constraint management module is called to perform projection clipping before entering the next layer. The purpose is to quickly converge to a globally feasible scheduling scheme within a given time window using a block-based solution approach. The effect is to achieve joint optimization in the three dimensions of time, frequency, and space and ensure that any output scheme naturally meets constraints such as protection band, protection sector, power, and sidelobes.

[0093] Therefore, this application presents an active phased array radar detection method and system based on SDR, which integrates software radio and active phased array radar. Utilizing SDR to provide wideband sampling and real-time demodulation, the system can flexibly adapt to different frequency bands and cope with frequency hopping and complex modulation schemes. It predicts the frequency hopping of target UAVs using a hidden Markov model or particle filter algorithm and generates corresponding pre-jamming templates. Simultaneously, through parallel processing of fast and slow channels, the fast channel quickly identifies the target's frequency band and modulation scheme, initiating jamming in advance based on the prediction results, while the slow channel performs detailed analysis of the target, including system identification, micro-Doppler analysis, and angle of arrival estimation, ensuring the accuracy of the countermeasure signal. Furthermore, through joint scheduling, in multi-target scenarios, time slots, frequency bands, and beam resources are dynamically allocated according to the target's threat level, frequency hopping prediction, and spatial location. This effectively solves the problems of insufficient frequency band coverage, excessive time delay, and difficulty in multi-target resource scheduling in traditional systems, achieving efficient detection and precise countermeasures against low-altitude small UAVs.

[0094] The embodiments described above are merely examples of several implementations of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the scope of protection of this application.

Claims

1. A method for detecting SDR-based active phased array radars, characterized in that, Includes the following steps: When the UAV is in fast frequency hopping mode, the S100 collects historical spectrum data of the target UAV. Based on the historical spectrum data, S200 predicts the frequency hopping band of the target UAV at the next moment through a hidden Markov model or particle filter algorithm, and generates a corresponding pre-interference template. The S300 starts fast-channel processing threads and slow-channel processing threads to execute in parallel, wherein: The fast channel processing thread controls the phased array antenna array to transmit the first interference signal to the predicted frequency hopping band based on the pre-interference template, and the response time of the fast channel processing thread is less than 50 milliseconds. The slow channel processing thread fills back the generated accurate interference waveform template and beam weights to the fast channel processing thread to update the pre-interference template for the next cycle. The slow channel processing thread performs detailed analysis on the spectrum data, including system identification, protocol analysis, angle of arrival estimation, and micro-Doppler analysis, and generates accurate interference waveform templates and beam weights; When multiple target drones exist, the S400 constructs a resource scheduling model; based on the threat level and frequency hopping prediction results of each target drone, it dynamically allocates time slots, frequency bands and beam resources to each target drone. The scheduling model described in S500 has built-in compliance constraints, including protection bands, protection sectors, power limits, beam isolation, and sidelobe limits, to ensure that countermeasures operations comply with the system's compliance requirements.

2. The SDR-based active phased array radar detection method of claim 1, wherein, In step s200, the hidden Markov model or particle filter algorithm dynamically updates the frequency band hopping probability matrix by analyzing the target's historical frequency band hopping data in real time, and predicts the frequency band hopping at the next moment based on the matrix.

3. The SDR-based active phased array radar detection method of claim 1, wherein, The threat level assessment in the S400 includes: The threat score for each target drone is dynamically calculated using a threat assessment model based on one or more factors, including the size, type, speed, distance from the protected sector, and frequency hopping speed.

4. The SDR-based active phased array radar detection method of claim 3, wherein, The implementation of the compliance constraints includes: A beamforming algorithm is used to control the transmit beam of the phased array antenna array, ensuring that the interference between multiple beams transmitted simultaneously is below a preset threshold. By setting a protection radius and exclusion list, interference signals can be avoided from being emitted within the protected frequency band and protected sector; Real-time monitoring of transmission power and equipment temperature, and dynamic adjustment of power distribution to ensure safe and stable operation of the equipment.

5. The SDR-based active phased array radar detection method of claim 1, wherein, The scheduling model, by taking into account built-in compliance constraints, ensures that the system resource scheduling process complies with the compliance requirements of protected frequency bands, protected sectors, power and thermal management, and avoids interference to non-target areas.

6. The SDR-based active phased array radar detection method of claim 1, wherein, The S400 further includes the following steps: S401: Through time slot-level scheduling, countermeasure time slots are dynamically allocated based on the threat level and priority of the target, with higher priority targets allocated more time slot resources; S402: Dynamically allocate frequency band resources based on the target's frequency band and frequency hopping prediction results to avoid conflicts when multiple targets use the same frequency band; S403: Based on the target's spatial location and trajectory, dynamically adjust the beam pointing and gain to ensure that each target can be accurately jammed; S405: Automatically excludes protected frequency bands and protected sectors to prevent interference signals from affecting non-target or sensitive areas; S406: Based on the power output and heat dissipation capacity of the equipment, limit the power of interference signals during scheduling to avoid overload or overheating; S407: By controlling beam spacing and sidelobe leakage, it ensures that the jamming signal will not affect other targets or areas, thus avoiding accidental damage.

7. A SDR-based active phased array radar detection system, characterized in that, The system is used to implement the SDR-based active phased array radar detection method according to any one of claims 1 to 6, and the system includes: A phased array antenna is used to receive and transmit electromagnetic signals; A radio SDR platform, connected to the antenna array, is used to sample and demodulate the received signals; The frequency hopping prediction module is used to run a hidden Markov model or particle filter algorithm to predict the frequency hopping band of the target UAV. The parallel processing module includes: The fast-channel processing unit transmits the first interference signal based on the frequency hopping prediction result; The slow channel processing unit performs fine signal analysis and generates accurate interference parameters. The resource scheduling module is used for resource scheduling in multi-objective scenarios; The threat assessment module is used to calculate the threat level of each target drone in real time. The compliance and constraint management module is used to ensure that the scheduling process complies with various constraints. The industrial control computer is used to run the above modules.

8. The SDR-based active phased array radar detection system according to claim 7, characterized in that, The fast channel processing unit and the slow channel processing unit interact with each other through shared memory, and the fast channel processing unit can directly access the control interface of the phased array antenna array.

9. The SDR-based active phased array radar detection system according to claim 7, characterized in that, The resource scheduling module employs a combination of time slot allocation algorithm, frequency band allocation algorithm, and beam allocation algorithm to achieve joint optimization scheduling of three-dimensional resources in time, frequency, and space.

Images