System for Detecting and Countering Drone RF Signals and its Method

Abstract

A drone RF signal detection and response system according to the present invention is a system for detecting and responding to radio frequency (RF) signals of a drone, comprising: a signal collection unit that receives a broadband RF signal and converts it into digital I / Q data; a conversion unit that generates data representing time-frequency characteristics from the digital I / Q data; an artificial intelligence-based signal detection unit that detects a specific area corresponding to a drone signal based on the time-frequency data and generates a detection result as a trigger signal; a control unit that selectively performs synchronization and demodulation on the digital I / Q data in the section corresponding to the trigger signal; a decoding unit that decodes the selectively demodulated signal to extract drone-related information; and a jamming unit that generates a jamming signal based on the detected area information and the decoded information, wherein whether or not to perform decoding is controlled according to the detection result of the artificial intelligence-based signal detection unit.

Description

Technology Field

[0001] The present invention relates to a technology for detecting and responding to unmanned aerial systems (UAS), and more specifically, to a drone RF signal detection and smart jamming technology that converts a drone's radio frequency (RF) signal into visualization data in the time-frequency domain, detects the drone signal through deep learning-based pattern recognition, selectively demodulates the detected signal to extract identification information of the drone, and selectively generates and transmits a jamming signal for a specific frequency and time interval based on the information. Background Technology

[0003] With the rapid expansion of unmanned aerial systems (UAS), particularly drones, security threats to critical facilities such as airports, military facilities, power plants, and urban areas are increasing, and consequently, the importance of drone detection and response technologies is being significantly highlighted.

[0004] Conventional drone detection technologies are primarily based on radar, image recognition, or RF signal analysis; among these, RF signal-based detection methods are utilized as effective methods because they can directly analyze communication signals between the drone and the controller.

[0005] However, conventional RF signal-based detection technology relies primarily on analyzing the energy distribution or spectral characteristics of specific frequency bands. Consequently, it fails to accurately reflect signal characteristics that change rapidly over time, such as frequency hopping (FHSS), and suffers from reduced detection accuracy due to the difficulty of distinguishing it from noise or other radio signals.

[0006] Furthermore, the method of decoding the entire range of large amounts of I / Q data collected in broadband RF environments induces excessive computational load, limiting real-time processing; in particular, there are issues of processing delay and performance degradation in environments where multiple wireless signals are mixed.

[0007] Meanwhile, conventional jamming technology generally uses a method of transmitting interference signals across an entire specific frequency band, which causes problems such as interference not only to the target drone but also to adjacent wireless communication systems, and is also inefficient in terms of power consumption.

[0008] Furthermore, conventional technology has a structure in which signal detection, decoding, and jamming are performed independently of each other, so detection results are not efficiently reflected in the decoding or jamming stage, which has a limitation in that the overall response efficiency of the system is reduced.

[0009] Therefore, a new drone response technology is required that can detect drone RF signals more accurately, enable real-time processing while maintaining computational efficiency, and provide an effective response while minimizing interference with surrounding communications. Prior art literature

[0011] Registered Patent No. 10-2194734 Anti-drone system and method using GPS spoofing deception attack The problem to be solved

[0012] The present invention relates to a technology for effectively detecting and identifying drone signals in a broadband RF signal environment and performing a response thereto, and aims to solve the following problems.

[0013] First, this study aims to resolve the problem of false detection caused by the difficulty in distinguishing between conventional frequency-based signal analysis methods and noise or other radio signals. This is because conventional frequency-based signal analysis methods rely on the energy distribution of specific frequency bands or simple spectral characteristics, and thus fail to accurately reflect the characteristics of drone signals that change rapidly over time, such as frequency hopping (FHSS).

[0014] Second, since the conventional method of decoding the entire range of a large amount of I / Q data entering a broadband RF environment causes an excessive computational load that degrades real-time processing performance, we aim to improve computational efficiency by providing a structure that can selectively decode only the signals of interest.

[0015] Third, since conventional jamming technology primarily utilizes a method of indiscriminately transmitting interference signals across an entire specific frequency band, it causes problems by inducing interference with adjacent wireless communication systems other than the target drone. Therefore, this invention aims to minimize unnecessary radio interference by performing selective jamming synchronized with the temporal and frequency characteristics of the drone signal.

[0016] Fourth, since conventional technology has a structure in which signal detection, decoding, and jamming are performed independently of each other, there is a problem of reduced processing efficiency of the entire system. Therefore, we aim to improve the processing efficiency and response accuracy of the entire system by providing an interconnected structure that controls whether to perform decoding based on the AI-based signal detection result and performs jamming based on the decoding result.

[0017] Fifth, unlike conventional response methods that do not reflect the characteristics of drone models and communication protocols, we aim to enable more precise and effective drone response by applying jamming methods differentially based on decoded drone identification information.

[0018] The problems solved by the present invention are not limited to those mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0020] A drone RF signal detection and response system according to the present invention is a system for detecting and responding to radio frequency (RF) signals of a drone, comprising: a signal collection unit that receives a broadband RF signal and converts it into digital I / Q data; a conversion unit that generates data representing time-frequency characteristics from the digital I / Q data; an artificial intelligence-based signal detection unit that detects a specific area corresponding to a drone signal based on the time-frequency data and generates a detection result as a trigger signal; a control unit that selectively performs synchronization and demodulation on the digital I / Q data in the section corresponding to the trigger signal; a decoding unit that decodes the selectively demodulated signal to extract drone-related information; and a jamming unit that generates a jamming signal based on the detected area information and the decoded information, wherein whether or not to perform decoding is controlled according to the detection result of the artificial intelligence-based signal detection unit.

[0021] Preferably, the decoding unit is characterized by performing decoding only on the region detected by the artificial intelligence-based signal detection unit.

[0022] Preferably, the time-frequency data is characterized by including a spectrogram generated by a short-time Fourier transform (STFT).

[0023] Preferably, the artificial intelligence-based signal detection unit is characterized by detecting a drone signal area using an object detection model on the spectrogram.

[0024] Preferably, the object detection model is characterized by including a deep learning-based model.

[0025] Preferably, the jamming unit is characterized by generating a jamming signal synchronized with the frequency and time intervals of the drone signal.

[0026] Preferably, the jamming unit is characterized by generating a jamming signal synchronized with the time of the temporal burst of the drone signal.

[0027] Preferably, the decoding unit is characterized by extracting identification information, location information, or communication protocol information of the drone.

[0028] Preferably, the jamming unit is characterized by varying the jamming method according to the decoded communication protocol information.

[0029] In addition, the drone RF signal detection and response method according to the present invention comprises, in the method for detecting and responding to a drone RF signal, a step of collecting a broadband RF signal and converting it into digital I / Q data; a step of generating time-frequency data from the digital I / Q data; a step of detecting a drone signal using artificial intelligence based on the time-frequency data to detect a specific time-frequency range; a step of selectively performing decoding on the digital I / Q data corresponding to the range only if the detected range exists; a step of extracting identification information or communication characteristics of the drone from the decoding result; and a step of generating a jamming signal synchronized with the time interval where the drone signal exists, based on the extracted information and the detected time-frequency range. Effects of the invention

[0031] According to the present invention, the following effects can be obtained.

[0032] First, by adopting a structure that controls whether to perform decoding based on the results of AI-based pattern recognition, decoding is selectively performed only on signals of interest, unlike the conventional method that performs decoding on the entire broadband RF signal in bulk. This results in a significant reduction in computational load and an improvement in real-time processing performance.

[0033] Second, by converting raw I / Q data into spectrogram images in the time-frequency domain and analyzing them using an object detection method, it is possible to more accurately identify abnormal and dynamic signal patterns, such as frequency hopping (FHSS), unlike conventional methods that rely on simple frequency component analysis, thereby improving detection precision.

[0034] Third, unlike conventional systems where each function is performed independently, the overall processing flow is organically combined through a structure in which AI-based signal detection, selective decoding, and jamming control are interconnected, thereby improving system processing efficiency and response speed.

[0035] Fourth, by performing jamming synchronized not only with the frequency of the drone signal but also with the timing of the temporal burst, it is possible to minimize unnecessary radio wave emissions, improve energy efficiency, and reduce interference with adjacent wireless communication systems.

[0036] Fifth, by applying jamming strategies differentially based on drone model and communication protocol information extracted through decoding, it is possible to respond more precisely and effectively to target drones, unlike simple physical signal blocking methods.

[0037] Sixth, by performing the data storage path and signal analysis path in parallel through a Dual-Path structure, real-time analysis is possible without data loss, and the stability and reliability of the system are improved.

[0038] The effects of the present invention are not limited to those mentioned above, and other unmentioned problems will be clearly understood by those skilled in the art from the description below. Brief explanation of the drawing

[0040] FIG. 1 is a block diagram showing the overall architecture and data processing flow of an SDR-based drone detection and smart jamming system according to one embodiment of the present invention. FIG. 2 is a diagram illustrating the STFT transformation and spectrogram generation process of an I / Q signal according to one embodiment of the present invention. FIG. 3 is a diagram showing the result of detecting a drone signal on a spectrogram according to one embodiment of the present invention. FIG. 4 is a diagram showing the drone signal decoding result according to one embodiment of the present invention. FIG. 5 is a diagram showing the process of performing jamming based on drone signal detection according to one embodiment of the present invention. FIG. 6 is a diagram showing the dynamic updating of a danger zone according to drone movement in accordance with an embodiment of the present invention. FIG. 7 is a flowchart showing the overall operational flow of drone detection and response according to one embodiment of the present invention. Specific details for implementing the invention

[0041] Further objects, features, and advantages of the present invention can be more clearly understood from the following detailed description and the accompanying drawings.

[0042] Before providing a detailed description of the present invention, it should be understood that the present invention is capable of various modifications and may have various embodiments, and that the examples described below and illustrated in the drawings are not intended to limit the present invention to specific embodiments, but rather include all modifications, equivalents, and substitutions that fall within the spirit and scope of the present invention.

[0043] When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. On the other hand, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between.

[0044] The terms used in this specification are used merely to describe specific embodiments and are not intended to limit the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0045] Furthermore, in the description referring to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof are omitted. In describing the present invention, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the present invention, such detailed description is omitted.

[0046] This system features a controlled, interconnected structure rather than a simple parallel connection, enabling it to operate on high-performance edge computing hardware such as NVIDIA Jetson Orin or RK3588, and performs GPU and NPU acceleration through a dedicated software framework like AirStack. Collected data is transmitted without latency between the CPU and GPU via a Zero Copy memory architecture, maximizing real-time analysis performance.

[0048] FIG. 1 is a block diagram schematically showing the overall architecture and data processing flow of an SDR-based drone detection and smart jamming system according to one embodiment of the present invention.

[0049] Referring to FIG. 1, the system of the present invention may be configured to include an antenna unit (110), an SDR front-end (120), a PCIe interface (130), a main processor (140), a system memory (150), a data acquisition engine (210), a circular buffer (220), a storage controller (230), an STFT acceleration module (310), a YOLO pattern recognition engine (320), a synchronization and demodulation controller (330), a multi-protocol decoding engine (340), a metadata extractor (410), an RF fingerprinting module (420), a user interface (430), and a smart jamming module (440).

[0050] The antenna unit (110) receives a broadband RF signal from an external space, and the SDR front-end (120) can convert the received RF signal into digital I / Q data. The PCIe interface (130) can provide a data transmission path that transmits the digital I / Q data to the main processor (140) at high speed. The system memory (150) is implemented as a Zero-Copy based shared memory structure, so that the collected Raw I / Q data can be utilized in parallel without delay in multiple processing paths.

[0051] As illustrated in FIG. 1, the system of the present invention adopts a Dual-Path architecture to process input Raw I / Q data in parallel through two different paths. The first path (Path A) is a real-time storage and management path and may include a data collection engine (210), a circular buffer (220), and a storage controller (230). The data collection engine (210) receives and sorts stream data flowing in through a PCIe interface (130), the circular buffer (220) can temporarily store a large amount of I / Q data without loss, and the storage controller (230) can write the data to a storage device such as NVMe to enable post-analysis or evidence acquisition.

[0052] Meanwhile, the second path (Path B) is an AI-based analysis and protocol identification path and may include an STFT acceleration module (310), a YOLO pattern recognition engine (320), a synchronization and demodulation controller (330), and a multi-protocol decoding engine (340). Specifically, the STFT acceleration module (310) can rapidly convert input raw I / Q data into a spectrogram in the time-frequency domain, and the YOLO pattern recognition engine (320) can detect feature patterns corresponding to drone signals on the spectrogram using an object detection method to generate bounding box and class information. Such AI detection results can be transmitted to the synchronization and demodulation controller (330) as an AI Detection Trigger.

[0053] A feature of the present invention is that the result of AI-based pattern recognition is not utilized merely as detection information, but functions as a control signal that directly controls whether subsequent decoding is performed. Therefore, unlike conventional methods that perform decoding on the entire broadband RF signal collectively, the invention has a structure in which decoding is performed selectively only on specific sections identified as signals of interest by AI.

[0054] The above Dual-Path structure is configured to execute the data storage path and the signal analysis path in parallel independently of each other, thereby preventing analysis delays from leading to data loss and enabling the simultaneous assurance of real-time performance and data integrity.

[0055] According to one embodiment of the present invention, an artificial intelligence-based pattern recognition engine (320) is configured to generate a trigger that controls whether to perform decoding based on a specific time-frequency range detected on a spectrogram, and a synchronization and demodulation controller (330) may be controlled to selectively perform decoding only on Raw I / Q data of the interval corresponding to the trigger.

[0056] Therefore, unlike conventional methods that perform batch decoding of the entire band, it has a structure in which decoding is performed selectively only on the signal region of interest. Due to this selective decoding structure, there is no need to decode the entire broadband RF signal, which significantly reduces computational load and enables real-time processing even in environments with high-speed frequency hopping signals.

[0057] In particular, stable processing performance can be maintained without bottlenecks even in environments where multiple drone signals are mixed. Through this structure, the present invention structurally resolves the computational bottleneck problem that inevitably occurs in broadband signal environments, and provides a system architecture capable of real-time response beyond simple processing speed improvement.

[0058] The synchronization and demodulation controller (330) can extract Raw I / Q data of a corresponding specific time interval based on detection location information provided by the pattern recognition engine (320), perform synchronization signal verification and demodulation control, and transmit it to the multi-protocol decoding engine (340).

[0059] The multi-protocol decoding engine (340) can perform a decoding algorithm corresponding to a drone-specific protocol, such as DJI DroneID, to restore the drone's serial number, model, location information, etc. That is, the present invention can secure computational efficiency and real-time capability by performing decoding only on the signal of interest section detected first by artificial intelligence, rather than indiscriminately decoding the entire broadband signal.

[0060] The output of the decoding engine (340) is transmitted to the metadata extractor (410) to organize information such as the drone's SN, model, and location into human-readable text or structured data. Additionally, an RF fingerprinting module (420) may be used as needed, and the RF fingerprinting module (420) can perform object identification within the same model using the 1D-ResNet Identity Re-ID method, etc.

[0061] The user interface (430) can provide the operator with visualized analysis results, spectrogram images, detected objects, decoding information, etc., in the form of a dashboard.

[0062] Additionally, the smart jamming module (440) analyzes the occupied frequency band and the timing of the temporal burst of the drone signal based on the decoding result and detected pattern information, and can selectively generate and transmit jamming signals in accordance with the frequency and time interval where the drone signal actually exists. Accordingly, unlike the conventional method of indiscriminately transmitting jamming signals across the entire band, it is possible to effectively neutralize the control signals or video transmission signals of the target drone while reducing interference with other wireless communications such as Wi-Fi and Bluetooth in adjacent bands. By performing such time-synchronized jamming, unnecessary radio wave emissions can be minimized, energy efficiency improved, and interference with adjacent wireless communication systems can be reduced.

[0063] The smart jamming module (440) includes a jamming control unit (441), the jamming control unit analyzes the characteristics of the jamming target signal based on detection and decoding results and controls the generation and transmission of the jamming signal.

[0064] According to one embodiment of the present invention, the smart jamming module (440) can generate a jamming signal by using the frequency band of the drone signal as well as the timing of the occurrence of a temporal burst as a synchronization standard. Specifically, by analyzing the periodic or non-periodic burst pattern of the drone communication signal and generating a pulsed or narrowband jamming signal synchronized with the time interval in which the burst occurs, effective communication disruption can be performed with minimal power.

[0065] In addition, according to one embodiment of the present invention, the smart jamming module (440) can apply a jamming strategy differentially based on the drone model and communication protocol information extracted through the decoding engine (340). For example, it can be operated in a manner that degrades video quality by applying high-intensity intermittent jamming that induces communication disconnection to the drone's control channel and continuous low-power jamming to the video transmission channel.

[0066] According to one embodiment of the present invention, artificial intelligence-based pattern recognition uses a spectrogram image generated by applying a Short-Time Fourier Transform (STFT) to Raw I / Q data as input and is configured to recognize signal patterns on an object basis in a two-dimensional feature space including time and frequency axes. Unlike FFT-based signal processing that simply analyzes frequency components, this provides a high-dimensional feature recognition structure that simultaneously reflects the temporal change characteristics and frequency distribution of the signal.

[0067] As shown in FIG. 1, the main processor (140) can control the data flow between each of the components and perform computation and control functions of the entire system, and can be configured so that the storage path and the analysis path are organically linked around the system memory (150). Accordingly, the present invention can perform the collection of broadband RF signals, real-time visualization, AI-based pattern detection, selective protocol demodulation, metadata extraction, and selective smart jamming within a single integrated architecture.

[0068] In particular, the present invention has the effect of minimizing unnecessary radio wave emissions and improving energy efficiency, while simultaneously reducing interference to adjacent wireless systems, by performing jamming by synchronizing with the actual time interval in which the drone signal exists.

[0069] Accordingly, the present invention provides an integrated structure in which AI-based signal detection, selective decoding, and time-frequency synchronous jamming are interconnected, which is a technical feature distinct from a parallel combination of simple signal analysis technology and jamming technology.

[0071] FIG. 2 is a diagram showing a spectrogram and a spectrum graph representing the time-frequency analysis results of an SDR-based drone signal according to an embodiment of the present invention.

[0072] Referring to FIG. 2, a spectrogram (520) in the time-frequency domain is shown at the top, and a frequency spectrum (510) for a specific time interval is shown at the bottom.

[0073] Specifically, the spectrogram (520) is a result generated by a Short-Time Fourier Transform (STFT) acceleration module and can be represented as two-dimensional image data including time axis, frequency axis, and color information representing signal strength. On the spectrogram, the drone signal appears as a repetitive burst form or an energy pattern concentrated in a specific band, which has a characteristic shape that distinguishes it from general noise or other wireless signals.

[0074] In addition, the spectrum graph (510) at the bottom represents the signal strength by frequency based on the average or maximum value for a certain time interval (e.g., 100 ms), allowing for the quantitative verification of the frequency band and power distribution occupied by the drone signal. In the spectrum, a sharp peak appears in a specific frequency band, which can be used to estimate the center frequency and bandwidth of the drone communication link.

[0075] Such spectrograms (520) and spectrum graphs (510) are subsequently used as input data for a YOLO-based pattern recognition engine and can function as feature data for drone signal detection. That is, the present invention is configured to identify drone signals more accurately by applying a time-frequency image-based object recognition structure, going beyond simple RF signal analysis. Unlike FFT-based signal processing methods that simply analyze the magnitude of frequency components, this adopts a structure that recognizes patterns in the time-frequency domain in the form of images, thereby enabling more precise identification of abnormal changes and frequency hopping characteristics of the signal.

[0076] According to one embodiment of the present invention, artificial intelligence-based pattern recognition uses a spectrogram image generated by applying a Short-Time Fourier Transform (STFT) to Raw I / Q data as input and is configured to recognize signal patterns on an object basis in a two-dimensional feature space including time and frequency axes. Unlike FFT-based signal processing that simply analyzes frequency components, this provides a high-dimensional feature recognition structure that simultaneously reflects the temporal change characteristics and frequency distribution of the signal.

[0078] FIG. 3 is a diagram illustrating the results of artificial intelligence-based drone signal detection according to one embodiment of the present invention.

[0079] Referring to FIG. 3, areas identified as drone signals on the spectrogram (520) are indicated by a plurality of bounding boxes, and the probability of a drone's presence can be indicated for each area.

[0080] Specifically, the YOLO pattern recognition engine can detect characteristic signal patterns within a spectrogram image, recognize them as objects, and mark the corresponding area with a rectangular bounding box. Each bounding box is assigned a confidence score, such as “drone: 0.91” or “drone: 0.96,” which indicates the probability that the area is a drone signal.

[0081] In addition, as illustrated in FIG. 3, multiple drone signals may appear repeatedly in the time axis direction, and each signal may occur in the form of a burst with a constant period or irregular intervals. By recognizing these repeating patterns based on learning, the present invention can effectively detect not only a single signal but also multiple drone signals in an environment where they are mixed.

[0082] Meanwhile, the location information (time interval and frequency range) of each bounding box is used as an AI Detection Trigger (530), and can subsequently be used as a criterion for selectively extracting Raw I / Q data of the corresponding interval during the synchronization and demodulation control steps.

[0083] In other words, the present invention does not indiscriminately process the entire RF band, but rather selects a signal region of interest based on artificial intelligence and performs subsequent processing only on that region, thereby significantly reducing the amount of computation while ensuring real-time performance.

[0085] FIG. 4 is a diagram illustrating a Drone-ID payload structure as a result of drone signal decoding according to one embodiment of the present invention.

[0086] Referring to Fig. 4, drone communication data restored through a multi-protocol decoding engine can be displayed in the form of JSON or structured data.

[0087] Specifically, the decoded data may include packet length (pkt_len), version (version), sequence number (sequence_number), state information (state_info), aircraft serial number (serial_number), etc. Additionally, as drone location information, information such as latitude, longitude, altitude, and speed (v_north, v_east, v_up) may be included.

[0088] In addition, various operational information such as the relative position between the drone and the operator, app-based GPS information (app_lat, app_lon), home point (home position), and device type (device_type) may be included, and data integrity can be verified through CRC values ​​(crc-packet, crc-calculated).

[0089] Such Drone-ID data can be organized through a metadata extractor and displayed on a user interface or utilized as a control input for a smart jamming module. For example, jamming signals optimized for a specific drone can be generated based on the drone's identification information and communication characteristics.

[0090] In addition, unlike conventional methods that simply repeat received I / Q data or transmit jamming signals across an entire specific band, the present invention has a structure that determines the jamming target and jamming method based on the decoded drone's communication protocol and identification information.

[0091] In other words, the present invention can provide an integrated drone response system capable of surveillance, tracking, and response by going beyond simply detecting the presence of a drone signal and restoring the identification and location information of the actual drone.

[0093] FIG. 5 is a diagram illustrating a smart jamming execution scenario based on drone detection results according to an embodiment of the present invention.

[0094] Referring to FIG. 5, the system of the present invention may include a series of processes for performing selective jamming based on drone signal detection and decoding results.

[0095] Specifically, when a drone signal is detected in a specific frequency band and time interval during the artificial intelligence-based pattern recognition step (S510), the information can be transmitted to the jamming control unit (441) (S520). The jamming control unit (441) analyzes the frequency occupancy bandwidth of the detected drone signal and the timing of the temporal burst occurrence due to frequency hopping (FHSS) in real time, and can calculate parameters required for generating a jamming signal based on the analysis results.

[0096] At this time, the jamming control unit (441) can define the characteristics of the jamming target signal by analyzing information such as the center frequency, bandwidth, burst period, and signal strength of the drone signal (S530).

[0097] The jamming control unit (441) determines whether the jamming target channel is a control channel or a video channel based on the decoded drone model, communication method, and protocol information, and can selectively perform a control channel jamming step (S550) or a video channel jamming step (S560) depending on the result of the determination.

[0098] Afterward, the smart jamming module (440) can selectively generate and transmit a jamming signal only in the frequency range and time slot where the drone signal actually exists, based on the analyzed characteristics (S540). For example, if the drone control signal is transmitted intermittently within a specific ISM band, a narrowband or pulsed jamming signal can be applied in accordance with the burst timing.

[0099] Additionally, according to one embodiment of the present invention, the drone model, communication method, and protocol information extracted through the multi-protocol decoding engine (340) are provided as control inputs to the smart jamming module (440), and the smart jamming module (440) may be configured to determine the jamming method, output strength, and application time interval based on the information.

[0100] In addition, the present invention can distinguish between a drone control channel and a video transmission channel and apply different jamming strategies to each. For example, strong intermittent jamming can be applied to the drone control channel to break the link (S550), and continuous low-power jamming can be applied to the video transmission channel to degrade the video quality (S560).

[0101] Unlike indiscriminate jamming across the entire frequency band, this selective jamming method is effective in ensuring high response efficiency against target drones while minimizing interference with adjacent wireless systems.

[0102] Unlike conventional methods that transmit jamming signals across an entire specific frequency band, this structure selectively performs jamming only on the time and frequency intervals actually occupied by the drone signal.

[0104] FIG. 6 is a diagram illustrating the concept of dynamic updating of a danger zone according to drone movement in accordance with an embodiment of the present invention.

[0105] Referring to FIG. 6, the system of the present invention can set and update danger zones in real time based on the location and movement status of the drone.

[0106] Specifically, the current direction and path of movement of the drone can be estimated using the drone's position coordinates and velocity vector obtained from the decoded Drone-ID data. Additionally, the rate of change in slope or the rate of change in movement can be calculated by calculating the amount of change in position at regular time intervals.

[0107] If the above rate of change exceeds a preset threshold (Rth), it can be determined that the drone is performing a sudden change in direction or speed, and accordingly, the danger zone can be expanded or redefined. Conversely, if the rate of change is low, it is determined to be in a stable movement state, and the danger zone can be reduced or maintained.

[0108] In addition, the present invention can update the danger zone at regular intervals by setting a periodic recalculation cycle, and in the case of multiple drones, can set a danger zone individually for each drone and integrate them to form an overall surveillance area.

[0109] Through this dynamic risk zone setting function, unlike simple surveillance systems based on fixed boundaries, active and predictable responses to the actual behavioral patterns of drones become possible.

[0111] FIG. 7 is a flowchart showing the overall operation flow of an SDR-based drone detection and response system according to one embodiment of the present invention.

[0112] Referring to FIG. 7, the system of the present invention may include a drone signal collection step (S710), a signal preprocessing and spectrogram generation step (S720), an artificial intelligence-based pattern detection step (S730), a drone detection signal determination step (S740), a synchronization and demodulation step (S750), a protocol decoding step (S760), a metadata extraction step (S770), and a response control step (S780).

[0113] First, in the drone signal collection step (S710), a broadband RF signal can be received through an antenna and an SDR front-end and converted into digital I / Q data.

[0114] Next, in the signal preprocessing step (S720), STFT operations can be performed to generate spectrogram data in the time-frequency domain.

[0115] Afterwards, in the artificial intelligence-based pattern detection step (S730), a YOLO-based object detection algorithm can be used to detect the area corresponding to the drone signal on the spectrogram and generate the detection result as a trigger signal.

[0116] In the drone signal detection judgment step (S740), if a drone signal is not detected, the process returns to step S710; if a drone signal is detected, the Raw I / Q data for the time interval corresponding to the detected drone signal is transmitted to the synchronization and demodulation step.

[0117] In the synchronization and demodulation step (S750), synchronization and demodulation processing can be performed on the extracted Raw I / Q data signal.

[0118] In the protocol decoding step (S760), drone identification information and operational data can be restored by applying a decoding algorithm suitable for the drone communication protocol.

[0119] In the metadata extraction step (S770), the decoding result can be organized into a structured data form and transmitted to a user interface or an external system.

[0120] Finally, in the response control step (S780), smart jamming or other countermeasures may be performed by considering the drone's location, frequency, communication characteristics, etc.

[0121] Based on this overall operational flow, the present invention can implement the entire process from the detection, identification, analysis, and response of drone signals into a single integrated system, thereby ensuring both real-time performance and efficiency.

[0123] The embodiments described in this specification and the accompanying drawings are merely illustrative of a part of the technical concept included in the present invention. Accordingly, since the embodiments disclosed in this specification are intended to explain, not limit, the technical concept of the present invention, it is obvious that the scope of the technical concept of the present invention is not limited by these embodiments. All variations and specific embodiments that can be easily deduced by a person skilled in the art within the scope of the technical concept included in the specification and drawings of the present invention should be interpreted as being included within the scope of the rights of the present invention. Explanation of the symbols

[0125] 110: Antenna 120: SDR Front-End 130: PCIe Interface 140: Main Processor 150: System Memory 210: Data Ingestion Engine 220: Circular Buffer Manager 230: Storage Controller 310: STFT Acceleration Module 320: Pattern Recognition Engine 330: Synchronization and Demodulation Controller 340: Protocol Decoding Engine 410: Metadata Extractor 420: RF Fingerprinting Module 430: User Interface 440: Smart Jamming Module 510: Spectrum Data 520: Spectrogram Image 530: Detection Region 610: Decoding Information

Claims

Claim 1 A system for detecting and responding to radio frequency (RF) signals of a drone comprises: a signal acquisition unit that receives a broadband RF signal and converts it into digital I / Q data; a conversion unit that generates time-frequency data in the form of a spectrogram image including a time axis and a frequency axis by applying a Short Time Fourier Transform (STFT) to the digital I / Q data; an artificial intelligence-based signal detection unit that detects a detection area corresponding to a drone signal using an object detection model based on the time-frequency data, and generates detection area information including a time interval, frequency range, and reliability information of the detection area as a trigger signal; a control unit that, only when the trigger signal is generated, selectively extracts a portion of the digital I / Q data corresponding to the time interval and frequency range of the detection area, and performs synchronization and demodulation on the selectively extracted portion; and a decoding unit that decodes the selectively demodulated signal to extract drone-related information including at least one of drone identification information, location information, or communication protocol information. A drone RF signal detection and response system comprising: a jamming unit that generates a jamming signal based on the detection area information and the drone-related information; wherein the jamming unit generates a jamming signal synchronized with the frequency and time intervals where the drone signal exists, based on the time interval and frequency range included in the detection area information. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 delete Claim 7 delete Claim 8 delete Claim 9 A system according to claim 1, wherein the jamming unit varies the jamming method according to the decoded communication protocol information. Claim 10 A method for detecting and responding to radio frequency (RF) signals of a drone comprises: receiving a broadband RF signal and converting it into digital I / Q data; applying a Short Time Fourier Transform (STFT) to the digital I / Q data to generate time-frequency data in the form of a spectrogram image including a time axis and a frequency axis; detecting a detection area corresponding to a drone signal using an object detection model based on the time-frequency data, and generating detection area information including a time interval, frequency range, and reliability information of the detection area as a trigger signal; selectively extracting a portion of the digital I / Q data corresponding to the time interval and frequency range of the detection area only when the trigger signal is generated; performing synchronization and demodulation on the selectively extracted portion; and decoding the selectively demodulated signal to extract drone-related information including at least one of drone identification information, location information, or communication protocol information. A method for detecting and responding to drone RF signals, comprising the step of generating a jamming signal based on the detection area information and the drone-related information, wherein the step of generating the jamming signal is characterized by generating a jamming signal synchronized with the frequency and time intervals where the drone signal exists, based on the time interval and frequency range included in the detection area information.

Images