System and Method for Drone RemoteID Detection Based on Bluetooth and Wi-Fi

Abstract

The present invention relates to a Bluetooth and Wi-Fi-based drone remote identification detection system and method, wherein the system comprises: a data collection unit that collects a remote identification signal, a Bluetooth signal, and a Wi-Fi signal transmitted from the drone using a Bluetooth module and a Wi-Fi detection module, and generates drone flight data by integrating the remote identification signal, the Bluetooth signal, and the Wi-Fi signal; a data analysis unit that detects the drone by analyzing the Bluetooth signal and Wi-Fi signal collected by the data collection unit, classifies the type of the drone, detects abnormal flight patterns, and predicts the drone's flight path; a data visualization unit that provides the collected remote identification number and the results analyzed by the data analysis unit to a user; and a data storage unit that stores the Bluetooth signal, Wi-Fi signal, and remote identification signal collected by the data collection unit, and stores training data and no-fly zone information.

Description

Technology Field

[0001] The present invention relates to a Bluetooth and Wi-Fi-based drone detection system and method, and more specifically, to a Bluetooth and Wi-Fi-based drone remote identification detection system and method for detecting the location and flight status of a drone using Bluetooth and Wi-Fi signals, and for detecting abnormal flight patterns and predicting the flight path of a drone by analyzing remote identification information. Background Technology

[0002] Key technologies for conventional drone detection and identification include radar-based detection methods and methods utilizing RF (Radio Frequency) scanners. While these conventional technologies can detect the presence of drones to some extent, they face various limitations due to the recent increase in unauthorized drone operations and the diversification of drone materials and flight characteristics.

[0003] Radar detection is a technology that determines location and speed by analyzing radar cross-sections (RCS) reflected from aircraft such as drones; however, there are problems such as the difficulty in detecting small drones with small reflective surfaces or those made of plastic or synthetic materials due to weak reflected signals, the occurrence of blind spots for low-altitude aircraft caused by surrounding environments like terrain, buildings, and trees, and the fact that radars operated at fixed angles not only have limited detection performance for drones flying at low altitudes but also incur high costs for equipment construction and maintenance.

[0004] On the other hand, while the method using an RF scanner can track the drone's location and operational information by capturing radio frequencies transmitted by the drone, it has disadvantages such as confusion in detection if false signals are generated through RF spoofing, the scanner's reception performance being neutralized or degraded by RF jamming, and difficulty in frequency separation and tracking when multiple frequencies are received simultaneously.

[0005] Furthermore, in accordance with the U.S. Federal Aviation Administration (FAA) Remote ID Rule and the European Aviation Safety Agency (EASA) UAS Remote Identification Regulation, the installation of Remote ID capabilities is becoming mandatory for drones exceeding a certain range in the United States and Europe. This technology enables effective drone management and operational regulation by allowing drones to autonomously transmit location and identification information in real time during flight, as defined in the ASTM International F3411-22a standard.

[0006] However, since drones imported into the country are often introduced with this Remote ID function disabled, it is difficult to effectively detect and identify unauthorized drone operations. In particular, most drones recently purchased in the domestic market do not utilize the FAA and EASA standard Remote ID, making it virtually impossible to achieve immediate detection through remote identification and systematic analysis using AI and machine learning.

[0007] As such, conventional technologies centered on radar detection and RF scanners show limitations in the efficient detection and tracking of unauthorized drones, and because the RemoteID function of domestic drones is disabled, it is difficult to collect and analyze relevant information in real time. Therefore, there is a need for a method to enable RemoteID suitable for the domestic environment, as well as an automated drone detection and identification system that incorporates AI and machine learning analysis technologies based on this.

[0008] The matters described above as background technology are intended only to enhance understanding of the background of the present invention and should not be construed as an acknowledgment that they constitute prior art already known to those skilled in the art. Prior art literature

[0009] Korean Registered Patent Publication No. 10-2128590 (June 24, 2020) Korean Published Patent Publication No. 10-2019-0134720 (December 4, 2019) The problem to be solved

[0010] Accordingly, the present invention aims to provide a drone detection system and method capable of maintaining safety by effectively collecting and analyzing Remote ID data transmitted from a drone based on Bluetooth and Wi-Fi signals to resolve the aforementioned conventional problems, thereby accurately tracking the real-time location, altitude, and speed of the drone and rapidly detecting abnormal flight patterns.

[0011] In addition, another objective of the present invention is to enable safer and more efficient aviation management in domestic and international drone operating environments by presenting a drone detection system and method that solves the aforementioned problems by classifying drone identification signals and signals from surrounding IoT devices through AI and machine learning-based data analysis, identifying abnormal drone flight in advance to provide warnings, and predicting the drone's flight path to be linked with an air traffic control system. means of solving the problem

[0012] A Bluetooth and Wi-Fi-based drone remote identification detection system (1) for achieving the above-mentioned purpose, which receives and analyzes Bluetooth signals and Wi-Fi signals including a remote identification signal (Remote ID) transmitted from a drone to detect the drone, detect abnormal flight patterns of the drone, and predict a flight path, wherein the system comprises: a data collection unit (100) that collects the remote identification signal, Bluetooth signal, and Wi-Fi signal transmitted from the drone using a Bluetooth module and a Wi-Fi detection module, and generates drone flight data by synthesizing the remote identification signal, Bluetooth signal, and Wi-Fi signal; a data analysis unit (200) that detects the drone by analyzing the Bluetooth signal and Wi-Fi signal collected by the data collection unit (100), classifies the type of the drone, detects abnormal flight patterns, and predicts the flight path of the drone; a data visualization unit (300) that provides the collected remote identification number and the results analyzed by the data analysis unit (200) to a user; and a data storage unit (400) that stores the Bluetooth signal, Wi-Fi signal, and remote identification signal collected by the data collection unit (100), and stores training data and flight prohibition zone information. It is possible.

[0013] In one embodiment according to the present invention, the remote identification signal is data transmitted from the drone, and the remote identification signal is characterized by including at least one of a drone identification number, a drone location, a drone altitude, a drone flight speed, a drone direction of travel, an ID of the drone pilot, a flight status, and a time at which the remote identification signal was transmitted. The data collection unit (100) may be characterized by including a Bluetooth signal collection unit (110) that detects and receives a Bluetooth signal transmitted from the drone using a Bluetooth detection module, a Wi-Fi signal collection unit that detects and receives a Wi-Fi signal transmitted from the drone using a Wi-Fi detection module, and a data generation unit that integrates the Bluetooth signal and the Wi-Fi signal to generate drone flight data.

[0014] In one embodiment according to the present invention, the Bluetooth signal is characterized by including at least one of the remote identification signal, a Bluetooth unique identifier (UUID), a received signal strength (RSSI), a frequency, a transmission interval, and an AUI, and the Wi-Fi signal may be characterized by including at least one of the remote identification number, a drone network (SSID), and a hardware unique identifier (MAC).

[0015] In one embodiment according to the present invention, the data analysis unit (200) may be characterized by including: a data preprocessing unit (210) that performs preprocessing to refine and normalize the Bluetooth signal and Wi-Fi signal collected by the data collection unit (100) before analyzing the Bluetooth signal and Wi-Fi signal; a data learning unit (220) that learns a machine learning model using training data and drone flight data; and a data inference unit (230) that classifies the drone type based on the drone flight data, detects abnormal drones, and predicts the drone flight path using the machine learning model learned in the data learning unit (220).

[0016] In one embodiment according to the present invention, the data preprocessing unit (210) is characterized by performing preprocessing through noise removal, missing value processing, and normalization processes in the Bluetooth signal and Wi-Fi signal, and the data learning unit (220) is characterized by performing supervised learning using normal flight data of an authorized drone and the drone flight data among the learning data, and may be characterized by performing unsupervised learning using unauthorized drone data and abnormal flight data among the learning data and the drone flight data.

[0017] In one embodiment according to the present invention, the data inference unit (230) may be characterized by including: a drone type classification unit (231) that determines the presence or absence of a drone and classifies the drone type based on the drone flight data using a machine learning model that has completed training; an abnormal drone detection unit (232) that detects an abnormal flight pattern of the drone based on the drone flight data using a machine learning model that has completed training and determines whether the drone has entered a no-fly zone; and a drone movement path prediction unit (233) that detects the current location of a drone where the abnormal flight pattern was not detected by the abnormal drone detection unit (232) using a machine learning model that has completed training and predicts the movement path of the drone based on the current location of the drone.

[0018] In one embodiment according to the present invention, the drone type classification unit (231) may be characterized by including a drone identification module that identifies whether the drone is a drone or a non-drone using a manufacturer identification number among the drone flight data, or identifies whether the drone is a drone or a non-drone using a drone network (SSID) among the drone flight data, and a drone type classification module that classifies the drone type into commercial and military types by analyzing and clustering the drone flight data using a clustering algorithm.

[0019] In one embodiment according to the present invention, the abnormal drone detection unit (232) may be characterized by including an abnormal flight pattern detection module that detects an abnormal flight pattern by analyzing the flight pattern of the drone based on the drone flight data using a clustering algorithm, and an intrusion determination module that extracts GPS data of the drone from the drone flight data using a machine learning model and determines whether the drone has entered a no-fly zone by analyzing the drone GPS data.

[0020] In one embodiment according to the present invention, the intrusion determination module may be characterized by extracting the drone position, drone altitude, drone flight speed, and drone direction of travel of the drone flight data as the drone GPS data, and if the coordinates of the drone GPS data enter within a preset distance of the no-fly zone radius, determining that the drone has intruded into the no-fly zone and generating a first intrusion warning information.

[0021] In one embodiment according to the present invention, the drone movement path prediction unit (233) may be characterized by including a current location detection module that extracts the current location of a drone in which the abnormal flight pattern is not detected in the abnormal drone detection unit (232) using a distance estimation method based on received signal strength (RSSI) and a location calculation method based on TDOA, and a movement path prediction module that predicts the movement path of the drone using an LSTM-based time series prediction model based on the extracted current location of the drone and determines whether the predicted movement path of the drone intrudes into the no-fly zone.

[0022] A Bluetooth and Wi-Fi-based drone detection method that receives and analyzes Bluetooth signals and Wi-Fi signals, including a remote identification signal (Remote ID) transmitted from a drone, to detect a drone, detect abnormal flight patterns of the drone, and predict a flight path, wherein the method comprises: a data collection step in which a data collection unit (100) collects a remote identification signal, a Bluetooth signal, and a Wi-Fi signal transmitted from the drone using a Bluetooth module and a Wi-Fi detection module, and generates drone flight data by synthesizing the remote identification signal, the Bluetooth signal, and the Wi-Fi signal; a data analysis step in which a data analysis unit (200) analyzes the Bluetooth signal and Wi-Fi signal collected by the data collection unit (100) to detect the drone, classifies the type of the drone, detects abnormal flight patterns, and predicts a flight path of the drone; a data visualization step in which a data visualization unit (300) provides the collected remote identification number and the results analyzed by the data analysis unit (200) to a user; and a data storage step in which a data storage unit (400) stores the Bluetooth signal, Wi-Fi signal, and remote identification signal collected by the data collection unit (100), and stores training data and flight prohibition zone information. It can be characterized as a feature. Effects of the invention

[0023] According to the Bluetooth and Wi-Fi based drone remote identification detection system (1) and the method described above, the following effects are achieved.

[0024] First, the system and method according to the present invention have the effect of accurately tracking the position, altitude, speed, etc. of a drone even in a complex environment where multiple drones are flying simultaneously by collecting and analyzing Remote ID signals transmitted from the drone and Bluetooth and Wi-Fi signals in real time.

[0025] Second, the system and method according to the present invention have the effect of distinguishing between drone signals and signals generated from general IoT devices through AI and machine learning algorithms, and automatically classifying the type of drone to quickly identify unauthorized drone operation or abnormal flight.

[0026] Third, the system and method according to the present invention detect no-fly zones and abnormal flight patterns in real time, generate a warning when a drone approaches a dangerous area or when abnormal signals such as GPS spoofing occur, and, if necessary, enable an immediate response by linking with an air traffic control system.

[0027] Fourth, the system and method according to the present invention provide a path prediction function utilizing Received Signal Strength Indicator (RSSI) and Time Difference of Arrival (TDOA), thereby identifying and predicting the future movement path of a drone in advance, which creates a safe drone operation environment and has the effect of preventing drone-related accidents or illegal activities.

[0028] Fifth, the system and method according to the present invention have the advantage of being able to operate for a long time by utilizing the low-power characteristics of Bluetooth and Wi-Fi technologies.

[0029] The effects obtainable from the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art to which the present invention belongs from the description below. Brief explanation of the drawing

[0030] FIG. 1 is a configuration diagram of a Bluetooth and Wi-Fi-based drone remote identification detection system (1) according to an embodiment of the present invention. FIG. 2 is a configuration diagram of a data collection unit (100) according to an embodiment of the present invention. FIG. 3 is a configuration diagram of a data analysis unit (200) according to an embodiment of the present invention. FIG. 4 is a configuration diagram of a data inference unit (230) according to an embodiment of the present invention. FIG. 5 is a flowchart of a Bluetooth and Wi-Fi-based drone detection method according to an embodiment of the present invention. FIG. 6 is a flowchart of a Bluetooth and Wi-Fi-based drone detection method according to another embodiment of the present invention. Specific details for implementing the invention

[0031] With reference to the attached drawings, a Bluetooth and Wi-Fi-based drone remote identification detection system (1) and method according to embodiments of the present invention will be described in detail. Since the present invention is subject to various modifications and may take various forms, specific embodiments are illustrated in the drawings and described in detail in the text. However, this is not intended to limit the present invention to specific disclosed forms, and it should be understood that it includes all modifications, equivalents, and substitutions that fall within the spirit and scope of the present invention. Similar reference numerals have been used for similar components in the description of each drawing. In the attached drawings, the dimensions of the structures are shown enlarged from the actual size to ensure clarity of the present invention, or reduced from the actual size to understand the schematic configuration.

[0032] Additionally, terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms. These terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be named the second component, and similarly, the second component may be named the first component. Meanwhile, unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the present invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0033] The present invention relates to a Bluetooth and Wi-Fi-based drone detection system (1) and a method, and more specifically, to a Bluetooth and Wi-Fi-based drone remote identification detection system (1) and a method for detecting the location and flight status of a drone using Bluetooth and Wi-Fi signals, and detecting abnormal flight patterns and predicting the movement path of a drone by analyzing remote identification information.

[0034] Conventional technology, centered on radar detection and RF scanner methods, has limitations in the efficient detection and tracking of unauthorized drones. Furthermore, due to the deactivation of the RemoteID function of domestic drones, it is difficult to collect and analyze relevant information in real time. Therefore, the system according to the present invention aims to solve the above-mentioned problems.

[0035] Accordingly, the present invention aims to provide a drone detection system (1) and a method that can maintain safety by effectively collecting and analyzing remote identification (Remote ID) data transmitted from a drone based on Bluetooth and Wi-Fi signals to resolve the aforementioned conventional problems, thereby accurately tracking the real-time location, altitude, speed, etc. of the drone and rapidly detecting abnormal flight patterns.

[0036] In addition, another objective of the present invention is to enable safer and more efficient aviation management in domestic and international drone operation environments by presenting a drone detection system (1) and method that, in order to solve the above-mentioned problems, classifies drone identification signals and signals from surrounding IoT devices through AI and machine learning-based data analysis, identifies abnormal flight of the drone in advance and provides a warning, and predicts the flight path of the drone and links with an air traffic control system.

[0037] FIG. 1 is a configuration diagram of a Bluetooth and Wi-Fi-based drone remote identification detection system (1) according to an embodiment of the present invention.

[0038] As illustrated in FIG. 1, a Bluetooth and Wi-Fi-based drone remote identification detection system (1) according to an embodiment of the present invention is characterized by including a data collection unit (100), a data analysis unit (200), a data visualization unit (300), a data storage unit (400), and a data transmission unit (500).

[0039] More specifically, in a Bluetooth and Wi-Fi-based drone remote identification detection system (1) that receives and analyzes Bluetooth signals and Wi-Fi signals including a remote identification signal (Remote ID) transmitted from a drone to detect the drone, detect abnormal flight patterns of the drone, and predict flight paths, the system is characterized by including: a data collection unit (100) that uses a Bluetooth module and a Wi-Fi detection module to collect a remote identification signal, a Bluetooth signal, and a Wi-Fi signal transmitted from the drone to detect the drone and generate drone flight data by synthesizing the remote identification signal, the Bluetooth signal, and the Wi-Fi signal; a data analysis unit (200) that analyzes the Bluetooth signal and Wi-Fi signal collected by the data collection unit (100) to classify the type of the drone, detect abnormal flight patterns, and predict flight paths of the drone; a data visualization unit (300) that provides the collected remote identification number and the results analyzed by the data analysis unit (200) to a user; and a data storage unit (400) that stores the Bluetooth signal, Wi-Fi signal, and remote identification signal collected by the data collection unit (100), and stores training data and flight prohibition zone information.

[0040] The system according to the present invention is characterized by further including a data transmission unit (500) that transmits the drone type classification result, abnormal flight pattern detection result, and drone movement path prediction result performed by the data analysis unit (200) to an air traffic system.

[0041] In one embodiment of the present invention, the data visualization unit (300) is characterized by including a screen and a speaker. More specifically, it is characterized by providing to the user detailed information on remote identification numbers, Bluetooth signals, and Wi-Fi signals collected by the data collection unit (100), and drone type classification results, abnormal flight pattern detection results, and drone flight path prediction results analyzed and inferred by the data analysis unit (200). The data visualization unit (300) is characterized by providing at least one of a map, a graph, and alarm information through the screen, and is characterized by providing a voice alarm through the speaker.

[0042] Here, the Remote ID is unique identification information transmitted by a drone to identify itself, characterized by having a fixed data format in accordance with international standards (such as ASTM F3411). Through this, the system according to the present invention allows external verification of the drone's identification number, real-time location, and flight status, which can be utilized to identify unauthorized drones or manage flight safety. Furthermore, the Remote ID includes information such as the pilot's location or flight history, enabling air traffic control or public institutions to transparently monitor drone operations.

[0043] More specifically, the remote identification number according to the present invention is data transmitted by the drone along with a Bluetooth signal or a Wi-Fi signal, and is characterized by including at least one of a drone identification number, a drone location, a drone altitude, a drone flight speed, a drone direction of travel, a drone pilot ID, a flight status, and a time stamp at which the remote identification signal was transmitted. Additionally, the remote identification number is characterized by further including the pilot location.

[0044] Specifically, the drone identification number is characterized as being at least one of the drone's unique ID (Serial Number) or flight registration number (equipment serial number); the drone location is characterized as including the latitude and longitude of the drone measured via GPS; the pilot location refers to the location of the drone pilot's ground control device; and the flight status is characterized as including at least one of the drone flight start time, end time, current status (takeoff, landing), and remaining battery level.

[0045] In one embodiment of the present invention, the learning data among the data stored in the data storage unit (400) is characterized by including authorized drone data, normal flight data, unauthorized drone data, and abnormal flight data.

[0046] In addition, the data storage unit (400) stores the drone manufacturer DB described later.

[0048] FIG. 2 is a configuration diagram of a data collection unit (100) according to an embodiment of the present invention.

[0049] As illustrated in FIG. 2, the data collection unit (100) according to the present invention is characterized by including a Bluetooth signal collection unit (110), a Wi-Fi signal collection unit (120), and a data generation unit (130).

[0050] More specifically, the data collection unit (100) according to the present invention is characterized by including a Bluetooth signal collection unit (110) that detects and receives a Bluetooth signal transmitted from the drone using a Bluetooth detection module, a Wi-Fi signal collection unit (120) that detects and receives a Wi-Fi signal transmitted from the drone using a Wi-Fi detection module, and a data generation unit (130) that integrates the Bluetooth signal and the Wi-Fi signal to generate drone flight data.

[0051] In one embodiment of the present invention, the Bluetooth signal collection unit (110) is characterized by detecting and identifying the drone by using the Bluetooth module to detect a Bluetooth signal including a remote identification signal (Remote ID) transmitted from the drone. The Bluetooth module is characterized as being a Bluetooth 5.0 or higher detection module, and the Bluetooth signal collection unit (110) is characterized by collecting the Bluetooth signal of the drone from a range of 1 km or more using the Bluetooth module.

[0052] Herein, the Bluetooth signal is characterized by including at least one of the remote identification signal, Bluetooth Unique Identifier (UUID), Received Signal Strength Indicator (RSSI), frequency, transmission interval, and Manufacturer's Unique Identifier (OUI). The Manufacturer's Unique Identifier (OUI) refers to a standard manufacturer's unique identifier assigned by the IEEE.

[0053] The Bluetooth signal according to the present invention is characterized as being data in the form of a beacon, and the Bluetooth signal collection unit (110) is characterized by determining that the target transmitting the Bluetooth signal is a drone if the manufacturer identification number (OUI) among the Bluetooth signals is a manufacturer registered in the drone manufacturer DB.

[0054] In one embodiment of the present invention, the Wi-Fi signal collection unit (120) is characterized by detecting and identifying the drone by using the Bluetooth module to detect a Wi-Fi signal including a remote identification signal (Remote ID) transmitted from the drone. The Wi-Fi module is characterized as being at least one of Wi-Fi Direct or Wi-Fi Aware, and the Wi-Fi signal collection unit (120) is characterized by collecting the Wi-Fi signal of the drone within a range of 2 to 5 km using the Wi-Fi module.

[0055] Herein, the Wi-Fi signal is characterized by including at least one of the remote identification number, the drone network (SSID), and the hardware unique identifier (MAC). The hardware unique identifier (MAC) is characterized by including the manufacturer identification number (OUI) and the equipment serial number.

[0056] The Wi-Fi signal collection unit (120) according to the present invention is characterized by determining that the target transmitting the Wi-Fi signal is a drone if the manufacturer identification number (OUI) among the Wi-Fi signals is a manufacturer registered in the drone manufacturer DB.

[0057] Additionally, the data collection unit (100) is characterized by having at least one or more. Multiple data collection units (100) can detect and identify a single drone, or multiple data collection units (100) can detect and identify multiple drones.

[0058] The data generation unit (130) according to the present invention is characterized by generating drone flight data by combining the Bluetooth signal collected by the Bluetooth signal collection unit (110) and the Wi-Fi signal collected by the Wi-Fi signal collection unit (120) when the drone is detected and identified by the Bluetooth signal collection unit (110) and the Wi-Fi signal collected by the Wi-Fi signal collection unit (120). Meanwhile, since the ranges in which the Bluetooth module and the Wi-Fi module detect and receive signals are different, the data collection unit (100) may collect only the Bluetooth signal or only the Wi-Fi signal. Accordingly, the drone flight data is characterized by including all of the remote identification number, the Bluetooth signal, and the Wi-Fi signal, or consisting only of the remote identification number and the Bluetooth signal, or consisting only of the remote identification number and the Wi-Fi signal.

[0060] FIG. 3 is a configuration diagram of a data analysis unit (200) according to an embodiment of the present invention.

[0061] As illustrated in FIG. 3, the data analysis unit (200) according to the present invention is characterized by including a data preprocessing unit (210), a data learning unit (220), and a data inference unit (230).

[0062] More specifically, in one embodiment of the present invention, the data analysis unit (200) is characterized by including a data preprocessing unit (210) that performs preprocessing to refine and normalize the Bluetooth signal and Wi-Fi signal collected by the data collection unit (100) before analyzing the Bluetooth signal and Wi-Fi signal; a data learning unit (220) that learns a machine learning model using training data and drone flight data; and a data inference unit (230) that classifies the drone type based on the drone flight data, detects abnormal drones, and predicts the drone flight path using the machine learning model learned in the data learning unit (220).

[0063] The data preprocessing unit (210) according to the present invention is a component that performs preprocessing to refine and normalize drone flight data generated by integrating Bluetooth signals and Wi-Fi signals transmitted from the drone in the data collection unit (100). More specifically, the data preprocessing unit (210) is characterized by performing preprocessing through noise removal, missing value processing, and normalization processes in the Bluetooth signals and Wi-Fi signals.

[0064] The noise filtering process of the data preprocessing unit (210) according to the present invention is a process of filtering signal interference and noise from the drone flight data. The noise filtering process is a process of removing signal interference; when the data collection unit (100) simultaneously collects signals generated from the drone and IoT devices such as smartphones and tablets, only the signals generated from the drone must be selected. The data preprocessing unit (210) is characterized by using RSSI-based filtering (Received Signal Strength Indicator) technology to consider weak signals received at a distance greater than a certain distance as noise and remove them. For example, since the Wi-Fi signal of a smartphone, which is not directly related to the drone, is weaker than the Wi-Fi signal of the drone, the Wi-Fi signal of the smartphone is considered as noise and removed.

[0065] The missing data handling process of the data preprocessing unit (210) according to the present invention is characterized by a process of predicting and compensating for the current location using the drone's previous location and flight speed when GPS data, such as drone location information of the remote identification signal of the drone flight data, is temporarily lost. The data preprocessing unit (210) corrects the location of the drone based on TDOA (Time Difference of Arrival) technology, and specifically, predicts the location by comparing the time when a plurality of data collection units (100) receive the drone signal. For example, if the drone's GPS coordinates are missing, the drone's location can be calculated and corrected using TDOA technology.

[0066] The normalization process of the data preprocessing according to the present invention is characterized by normalizing the mean and standard deviation of drone flight data, such as received signal strength (RSSI), drone altitude, and drone speed, using the Z-Score Normalization method. For example, if the received signal strength (RSSI) values ​​collected by the data collection unit (100) are measured to vary from -30dBm to -90dBm, there is a problem that it is difficult for a machine learning model to learn if the range of the data is too large. Therefore, the received signal strength (RSSI) values ​​must be normalized to a range of 0 to 1 to provide a uniform learning environment. The data preprocessing unit (210) performs the normalization process, and since the normalized drone flight data has all input values ​​within the same range, there is an advantage that learning and analysis become easier in the data learning unit (220) and data inference unit (230) described later.

[0067] The data learning unit (220) according to the present invention is a component that learns a machine learning model using training data and drone flight data in the data storage unit (400). More specifically, it is characterized by performing supervised learning using normal flight data of an authorized drone and the preprocessed drone flight data among the training data, and by performing unsupervised learning using unauthorized drone data and abnormal flight data among the training data and the preprocessed drone flight data.

[0068] The above data learning unit (220) is a component that performs both supervised learning and unsupervised learning. The above data learning unit (220) is characterized by learning the machine learning model through supervised learning, which enables the distinction between normal drones and unauthorized drones by learning normal flight data and drone flight data of authorized drones. Additionally, the above data learning unit (220) is characterized by learning the machine learning model through unsupervised learning, which enables the detection of abnormal flights by learning unauthorized drones, abnormal flight data, and drone flight data.

[0069] In addition, the data learning unit (220) according to the present invention is characterized by continuously collecting new Bluetooth signals and Wi-Fi signals from the data collection unit (100), and continuously updating the machine learning model and improving performance by retraining the machine learning model when new Bluetooth signals and Wi-Fi signals are collected.

[0071] FIG. 4 is a configuration diagram of a data inference unit (230) according to an embodiment of the present invention.

[0072] The data inference unit (230) according to the present invention is characterized by being a component that classifies drone types based on drone flight data, detects abnormal drones, and predicts drone movement paths using a machine learning model that has been trained in the data learning unit (220).

[0073] As illustrated in FIG. 4, the data inference unit (230) according to the present invention is characterized by including a drone type classification unit (231), an abnormal drone detection unit (232), and a drone movement path prediction unit (233).

[0074] More specifically, the data inference unit (230) is characterized by including a drone type classification unit (231) that determines the presence or absence of a drone and classifies the drone type based on the drone flight data using a machine learning model that has completed training, an abnormal drone detection unit (232) that detects an abnormal flight pattern of the drone based on the drone flight data using a machine learning model that has completed training and determines whether the drone has entered a no-fly zone, and a drone movement path prediction unit (233) that detects the current location of a drone where the abnormal flight pattern was not detected by the abnormal drone detection unit (232) using a machine learning model that has completed training and predicts the movement path of the drone based on the current location of the drone.

[0075] In one embodiment of the present invention, the drone type classification is a component that determines the presence or absence of a drone based on the drone flight data and classifies the drone type.

[0076] More specifically, the drone type classification unit (231) is characterized by including a drone identification module that identifies whether the drone is a drone or a non-drone using a manufacturer identification number among the drone flight data, or identifies whether the drone is a drone or a non-drone using a drone network (SSID) among the drone flight data, and a drone type classification module that classifies the drone type into commercial and military types by analyzing and clustering the drone flight data using a clustering algorithm.

[0077] In one embodiment of the present invention, the drone identification module is a component that analyzes drone flight data through a machine learning model to distinguish whether a target sending a Bluetooth signal or a Wi-Fi signal is a drone or a non-drone. The method of distinguishing whether a drone or a non-drone in the drone identification module includes a method of distinguishing and identifying using the manufacturer identification number (OUI) among the drone flight data and a method of distinguishing and identifying using the drone network (SSID) among the drone flight data.

[0078] The method of distinguishing and identifying the above drone flight data using an OUI (Ownership of the Manufacturer) is characterized by determining that the target transmitting the Bluetooth signal or Wi-Fi signal is a drone if the OUI assigned by the IEEE in the above drone flight data is a manufacturer registered in the drone manufacturer database. For example, the drone identification module according to the present invention is characterized by determining that the target transmitting the Bluetooth signal or Wi-Fi signal is a drone if the OUI is a drone manufacturer such as DJI, Parrot, or Autel, and determining that the target transmitting the Bluetooth signal or Wi-Fi signal is a non-drone, including smartphones, tablets, smartwatches, vehicle Bluetooth devices, and other IoT devices, if the OUI is a manufacturer not registered in the drone manufacturer database such as Raspberry Pi, Apple iPhone, or Samsung.

[0079] The method of distinguishing and identifying using the drone network (SSID) among the above drone flight data is a method of distinguishing whether the object transmitting the Wi-Fi signal is a drone or a non-drone by utilizing the characteristic that drones generally use a unique SSID format specific to each manufacturer to communicate with control software. The above drone identification module is characterized by determining the object transmitting the Wi-Fi signal as a drone if it is a unique drone manufacturer network using the drone network (SSID) of the Wi-Fi signal among the drone flight data, and determining the object transmitting the Wi-Fi signal as a non-drone if it is a network (SSID) having a format other than that of a drone manufacturer network. Examples of networks (SSID) having a format other than that of a drone manufacturer network include SSID=Home_WiFi, SSID=Free_Cafe_WiFi, etc.

[0080] In addition to the two methods described above, the drone identification module can identify whether a target transmitting the Bluetooth signal or Wi-Fi signal is a drone or a non-drone by using a machine learning model to learn and classify signal patterns of drones and non-drones. More specifically, the drone identification module is characterized by using at least one of the Received Signal Strength Indicator (RSSI), Bluetooth signal transmission interval, frequency band, and Manufacturer Identification Number (OUI) from the drone flight data as an input value to the drone identification algorithm, determining it to be a drone if the signal characteristics are a beacon signal and have a specific frequency pattern, and determining it to be a non-drone if the signal characteristics are a general Wi-Fi AP and have an irregular signal pattern. Here, the drone identification algorithm is characterized by being at least one of Random Forest, Support Vector Machine (SVM), and Convolutional Neural Network (CNN).

[0081] In one embodiment of the present invention, the drone type classification module is characterized by analyzing and clustering the drone flight data using a clustering algorithm to classify the drone type into commercial and military types. The drone type classification module is characterized by classifying the type of the identified drone when the drone identification module identifies a target transmitting a Bluetooth signal or a Wi-Fi signal as a drone.

[0082] More specifically, the drone type classification module uses received signal strength (RSSI), drone flight speed, drone altitude, drone location, and frequency band among the drone flight data as input values ​​for the clustering algorithm, and the clustering algorithm clusters the input values ​​to group them into identical drone data and classifies the drone type into commercial and military use through analysis.

[0083] Here, the clustering algorithm is characterized as being at least one of K-Means, DBSCAN, and GMM (Gaussian Mixture Model).

[0084] The K-Means algorithm is an algorithm that automatically classifies data from K predefined groups. The drone type classification module of the present invention is characterized by using a clustering algorithm that vectorizes the signal characteristics of the drone flight data and automatically classifies it into K types (commercial, military). When the drone type classification module uses the K-Means algorithm among the clustering algorithms, it is characterized by classifying the drone into commercial and military use by using Received Signal Strength Indicator (RSSI), drone flight speed, drone altitude, drone location, and frequency band as input values ​​for the K-Means algorithm.

[0085] The DBSCAN (Density-Based Spatial Clustering) algorithm is an algorithm that uses a density-based clustering technique. The drone type classification module of the present invention is characterized by analyzing the drone flight data through the DBSCAN algorithm to output the flight pattern of the drone, and classifying the drone into commercial and drone types based on the flight pattern. For example, if the drone flight pattern follows a fixed path based on GPS location or shows a pattern of moving at low altitude without abnormal speed changes, the drone is classified as a commercial drone. If the drone flight pattern shows rapid acceleration and high-speed movement in a specific area or a pattern of moving at high altitude, the drone is classified as a military drone.

[0086] The Gaussian Mixture Model (GMM) algorithm is an algorithm utilizing a probability-based clustering technique. The drone type classification module of the present invention is characterized by classifying drone types by analyzing the continuity of signal patterns in drone flight data. Specifically, it is characterized by assuming that the drone flight data follows multiple Gaussian distributions and classifying the data by calculating the probability that each data point belongs to a specific cluster.

[0087] In addition to the method of classifying the type of the drone using the clustering algorithm described above, the drone type classification module according to the present invention can classify the type of the drone using remote identification signals, Bluetooth signals, and Wi-Fi signals collected by the data collection unit (100).

[0088] A remote identification signal (Remote ID) conforming to the ASTM F3411 standard includes information such as a drone identification number, drone location, drone altitude, and drone flight speed. The drone type classification module is characterized by classifying the type of the drone by learning the flight pattern of the drone from the remote identification number using a machine learning model. Generally, military drones tend to use specific security protocols (encrypted remote identification signals) or exhibit flight patterns such as high-speed movement and maintaining abnormal altitude. Accordingly, the drone type classification module according to the present invention is characterized by classifying the drone as a military drone when it transmits an encrypted remote identification signal or exhibits flight patterns such as high-speed movement and maintaining abnormal altitude, and classifying the drone as a commercial drone when it transmits an unencrypted remote identification signal.

[0089] In addition, the drone type classification module is characterized by classifying the drone as a military drone when the drone transmits encrypted Bluetooth signals or beacon data, and the drone type classification module is characterized by classifying the drone as a military drone when the drone transmits encrypted Wi-Fi signals, i.e., encrypted drone network (SSID).

[0090] The abnormal drone detection unit (232) according to the present invention is a component that detects abnormal flight patterns of a drone based on the drone flight data using a machine learning model that has completed training, and determines whether the drone has entered a no-fly zone.

[0091] More specifically, the above abnormal drone detection unit (232) is characterized by including an abnormal flight pattern detection module and an intrusion determination module, wherein the abnormal flight pattern detection module detects an abnormal flight pattern by analyzing the flight pattern of the drone based on the drone flight data using a clustering algorithm. The intrusion determination module is characterized by extracting the drone's GPS data from the drone flight data using a machine learning model and analyzing the drone GPS data to determine whether the drone has intruded into a no-fly zone.

[0092] In one embodiment of the present invention, the abnormal flight pattern detection module is a component that detects an abnormal flight pattern by analyzing the flight pattern of the drone using a clustering algorithm. Herein, the clustering algorithm is characterized as being at least one of K-Means and DBSCAN. In addition, the abnormal flight pattern detection module can detect the abnormal flight pattern using a machine learning model in addition to the clustering algorithm, and the machine learning model is characterized as being a Long Short-Term Memory (LSTM).

[0093] And the above abnormal flight pattern refers to a drone that deviates from a normal flight pattern and attempts to enter a no-fly zone, and also refers to a flight pattern of a drone that exhibits sudden changes in speed, rapid changes in direction, irregular changes in altitude, etc.

[0094] When the abnormal flight pattern detection module according to the present invention detects the abnormal flight pattern using the K-Means algorithm, the abnormal flight pattern detection module is characterized by predetermining K clusters and using the K-Means algorithm to group drone flight data with similar flight patterns together. More specifically, the abnormal flight pattern detection module is characterized by using the K-Means algorithm to cluster normal flight data stored in the data storage unit (400) to create a normal flight cluster (cluster 0), and then comparing drone flight data collected from the data collection unit (100) with the corresponding cluster; if it is within the normal range, it is classified as a normal flight cluster, and if it is outside, it is detected as an abnormal flight pattern, and the abnormal flight pattern data is classified as an abnormal flight cluster (cluster 1).

[0095] Specifically, the abnormal flight pattern detection module establishes a normal flight cluster based on normal flight data (speed, altitude, direction, etc.), collects drone flight data in real time, assigns the collected drone flight data to K-Means clusters, and determines whether it belongs to the normal flight cluster. If the collected drone flight data is located more than a preset distance away from the normal flight cluster or does not belong to any preset cluster, it is determined to be an abnormal flight pattern.

[0096] The abnormal flight pattern detection module according to the present invention is characterized by detecting abnormal flight patterns using the DBSCAN algorithm, wherein drone flight data densely clustered in similar locations is determined to be normal flight, and drone flight data scattered individually is determined to be an abnormal flight pattern.

[0097] More specifically, the above abnormal flight pattern detection module clusters drone flight data where normal flight frequently occurs through past drone flight data stored in the data storage unit (400). Since normal flight tends to be concentrated or repeated in specific areas, such as over a park, drone flight data with high density in a specific area can be determined as being located in a normal flight or normal dense area. The above abnormal flight pattern detection module clusters normal flight data using past drone flight data, and then extracts a remote identification signal including drone location, drone altitude, drone flight speed, and drone direction of travel from the collected drone flight data to determine whether the current drone flight pattern belongs to a normal dense area. The abnormal flight pattern detection module is characterized by determining it as normal flight if the current drone flight pattern belongs to a normal dense area, and determining it as an abnormal flight pattern if the current drone flight pattern deviates from the normal dense area and is located in an isolated area or a no-fly zone.

[0098] When the abnormal flight pattern detection module according to the present invention detects an abnormal flight pattern using a Long Short-Term Memory (LSTM), it means detecting the abnormal flight pattern using time-series data. More specifically, it is characterized by detecting abnormal rapid changes as abnormal flight patterns by learning changes in drone flight speed, drone altitude, and drone position from the past to the present. Here, the abnormal rapid changes are characterized by including rapid ascent, rapid descent, rapid speed changes, etc.

[0099] Specifically, the abnormal flight pattern detection module is characterized by converting the normal flight data into time-series data to train the LSTM, and then inputting the collected drone flight data in real time to predict the next change. It is characterized by comparing the normal next drone position, drone flight speed, and drone altitude predicted by the LSTM with the collected drone flight data, and determining it as an abnormal flight if the error is greater than or equal to a preset value. Here, the preset error value is characterized as being 5.

[0100] Furthermore, the abnormal flight pattern detection module is characterized by detecting not only the abnormal flight pattern but also GPS spoofing, signal manipulation, and camouflage. GPS spoofing refers to an attack in which a drone receives a fake GPS signal and transmits manipulated location information. To prevent this, the abnormal flight pattern detection module according to the present invention is characterized by analyzing the difference from the actual location by comparing multiple drone flight data, or by determining whether GPS manipulation has occurred using TDOA-based signal analysis technology. Additionally, the abnormal flight pattern detection module is characterized by detecting signal manipulation and camouflage in which the drone remote control signal is interfered with externally (jamming) or abnormal signals are generated at unauthorized frequencies. To detect signal manipulation and camouflage, the abnormal flight pattern detection module according to the present invention is characterized by analyzing frequency patterns among the drone flight data using a machine learning model to classify normal frequencies and abnormal frequencies, and determining the abnormal frequencies as signal manipulation and camouflage.

[0101] In one embodiment of the present invention, the intrusion determination module is characterized by using a machine learning model to extract GPS data of the drone from the drone flight data and analyzing the drone GPS data to determine whether the drone has intruded into a no-fly zone.

[0102] More specifically, the intrusion determination module extracts the drone position, drone altitude, drone flight speed, and drone direction of travel of the drone flight data as the drone GPS data, and if the coordinates of the drone GPS data enter within a preset distance of the no-fly zone radius, it determines that the drone has intruded into the no-fly zone and generates a first intrusion warning information. At this time, the no-fly zone is characterized by including airports, military institutions, military bases, government buildings, etc., and the preset distance is characterized by being 5 km.

[0103] Here, the first intrusion warning is characterized by being transmitted to the air traffic system through the data transmission unit (500).

[0105] The movement path prediction unit according to the present invention is characterized by including a current location detection module and a movement path prediction module. More specifically, the current location detection module is characterized by extracting the current location of a drone in which the abnormal flight pattern is not detected in the abnormal drone detection unit (232) using at least one of a distance estimation method based on received signal strength (RSSI) and a location calculation method based on TDOA, and the movement path prediction module is characterized by predicting the movement path of the drone using an LSTM-based time series prediction model based on the extracted current location of the drone, and determining whether the predicted movement path of the drone intrudes into the no-fly zone.

[0106] In one embodiment of the present invention, the current location detection module is characterized by extracting the current location of a drone in which the abnormal flight pattern is not detected by using at least one of a distance estimation method based on received signal strength (RSSI) and a location calculation method based on time-delayed arrival (TDOA).

[0107] The above-described distance estimation method based on received signal strength (RSSI) is a method for estimating the distance between a transmitter (drone) and a receiver (data collection unit (100)) using the received signal strength. This method utilizes the characteristic that the distance is close when the received signal strength is strong and far when the received signal strength is weak, and is characterized by converting it into a distance using a path loss model. The path loss model typically uses a Log-Distance Path Loss Model (LDPLM), and the current position detection module is characterized by extracting the current position of the drone by calculating the distance based on the received signal strength through the following mathematical formula 1.

[0109] (Mathematical Formula 1)

[0110] (d : distance between transmitter (drone) and receiver (detection module) (m), P t : Transmitted power (power transmitted from the drone, dBm), P r : Received power (RSSI, dBm), n : Path loss index)

[0112] The above TDOA-based position calculation method is a method for accurately calculating the position of a drone by comparing the time at which the drone's signal arrives from a plurality of detection devices. Using this method, the current position detection module of the present invention is characterized by receiving a Bluetooth signal or a Wi-Fi signal from the drone at different times from a plurality of data collection units (100), and then using the difference in the time at which the signals arrive to extract the current position of the drone through the following mathematical formula 2 using a triangulation method.

[0114] (Mathematical Formula 2)

[0115] (d AB: Distance difference between detection modules A and B (m), c : Propagation speed, t A : Time when the signal was received by detection module A, t B : Time when the signal was received by detection module B)

[0117] The above propagation speed (c) is 3.0 × 10 8 It is characterized by being m / s, and the detection module refers to the data collection unit (100). The current location detection module can accurately extract the latitude and longitude of the drone by triangulation by comparing the reception times of the Bluetooth signal or Wi-Fi signal from the three data collection units (100). Additionally, the current location detection module can accurately extract the latitude, warning, and altitude of the drone by triangulation by comparing the reception times of the Bluetooth signal or Wi-Fi signal from the four data collection units (100).

[0118] In one embodiment of the present invention, the movement path prediction module predicts the movement path of the drone using an LSTM-based time series prediction model based on the current location of the extracted drone, and determines whether the predicted movement path of the drone intrudes into the no-fly zone.

[0119] More specifically, the above-described movement path prediction module is characterized by training an LSTM-based time series prediction model using the above-described normal flight data and drone flight data that has completed a preprocessing process in the above-described data preprocessing unit (210), thereby learning the change patterns of the drone flight speed, drone position (latitude and longitude), drone direction of travel, and drone altitude of the above-described drone flight data in chronological order, and based on this, predicting the expected flight speed, position (latitude and longitude), direction of travel, and altitude of the drone after a specific time in real time, thereby finally predicting the movement path of the above-described drone. The above-described specific time is preferably 1 second, but the above-described specific time is adjustable.

[0120] In addition, the movement path prediction module compares the extracted current location of the drone and the predicted movement path of the drone with no-fly zone information to determine whether there is a possibility that the drone will approach or intrude into the no-fly zone within a certain period of time. In this process, the coordinate data of the no-fly zone is compared with the drone's current location and predicted location to analyze whether intrusion has occurred. If the current location is within the no-fly zone, a second intrusion warning is generated, and the generated second intrusion warning is transmitted to the air traffic system through the data transmission unit (500). Furthermore, if the predicted location of the drone is within a radius of 3 km of the no-fly zone, a preliminary intrusion warning is generated, and the generated preliminary intrusion warning is transmitted to the air traffic system through the data transmission unit (500).

[0122] FIG. 5 is a flowchart of a Bluetooth and Wi-Fi-based drone detection method according to one embodiment of the present invention, and FIG. 6 is a flowchart of a Bluetooth and Wi-Fi-based drone detection method according to another embodiment of the present invention.

[0123] A Bluetooth and Wi-Fi-based drone detection method comprising receiving and analyzing Bluetooth signals and Wi-Fi signals including a remote identification signal (Remote ID) transmitted from a drone to detect a drone, detect abnormal flight patterns of a drone, and predict a flight path, wherein the method comprises: a data collection step in which a data collection unit (100) collects a remote identification signal, a Bluetooth signal, and a Wi-Fi signal transmitted from the drone using a Bluetooth module and a Wi-Fi detection module, and generates drone flight data by synthesizing the remote identification signal, the Bluetooth signal, and the Wi-Fi signal; a data analysis step in which a data analysis unit (200) analyzes the Bluetooth signal and Wi-Fi signal collected by the data collection unit (100) to detect the drone, classifies the type of the drone, detects abnormal flight patterns, and predicts the flight path of the drone; a data visualization step in which a data visualization unit (300) provides the collected remote identification number and the results analyzed by the data analysis unit (200) to a user; and a data storage step in which a data storage unit (400) stores the Bluetooth signal, Wi-Fi signal, and remote identification signal collected by the data collection unit (100), and stores training data and flight prohibition zone information. It is characterized by the fact that.

[0124] In addition, the method according to the present invention is characterized by further including the data transmission step in the data transmission unit (500).

[0125] For the sake of convenience of explanation, the above-described judgment system has been used as described below; however, the present invention is not simply limited thereto and can be implemented through various devices, systems, and terminals capable of performing similar operations. Data transmitted to the air traffic system regarding the drone type classification result, abnormal flight pattern detection result, and drone movement path prediction result performed in the data analysis step.

[0127] In one embodiment of the present invention, the remote identification signal is data transmitted from the drone, and the remote identification signal is characterized by including at least one of a drone identification number, a drone location, a drone altitude, a drone flight speed, a drone direction of travel, a drone pilot ID, a flight status, and a time at which the remote identification signal was transmitted.

[0128] The data collection step according to the present invention is characterized by including a Bluetooth signal collection step in which a Bluetooth signal transmitted from the drone is detected and received using a Bluetooth detection module in a Bluetooth signal collection unit (110), a Wi-Fi signal collection step in which a Wi-Fi signal transmitted from the drone is detected and received using a Wi-Fi detection module in a Wi-Fi signal collection unit (120), and a data generation step in which the Bluetooth signal and the Wi-Fi signal are integrated to generate drone flight data in a data generation unit (130).

[0129] Herein, the Bluetooth signal is characterized by including at least one of the remote identification signal, Bluetooth unique identifier (UUID), received signal strength (RSSI), frequency, transmission interval, and manufacturer identification number (OUI), and the Wi-Fi signal is characterized by including at least one of the remote identification number, drone network (SSID), and hardware unique identifier (MAC).

[0131] The data analysis step according to the present invention is characterized by comprising: a data preprocessing step in which a data preprocessing unit (210) performs preprocessing to refine and normalize the Bluetooth signal and Wi-Fi signal collected from the data collection unit (100) before analyzing the Bluetooth signal and Wi-Fi signal; a data learning step in which a machine learning model is trained using training data and drone flight data in a data learning unit (220); and a data inference step in which a machine learning model trained in the data learning unit (220) is used to classify drone types based on the drone flight data, detect abnormal drones, and predict drone flight paths in a data inference unit (230).

[0132] In one embodiment of the present invention, the data preprocessing step is characterized by performing preprocessing through noise removal, missing value processing, and normalization processes in the Bluetooth signal and Wi-Fi signal.

[0133] In one embodiment of the present invention, the data learning step is characterized by performing supervised learning using normal flight data of an authorized drone and drone flight data among the learning data, and by performing unsupervised learning using unauthorized drone data and abnormal flight data among the learning data and drone flight data.

[0135] In one embodiment of the present invention, the data inference step comprises: a drone type classification step that determines the presence or absence of a drone and classifies the drone type based on the drone flight data using a machine learning model that has been trained in a drone type classification unit (231); an abnormal drone detection step that detects an abnormal flight pattern of a drone based on the drone flight data using a machine learning model that has been trained in an abnormal drone detection unit (232) and determines whether the drone has entered a no-fly zone; and a drone movement path prediction step that detects the current location of a drone where the abnormal flight pattern was not detected in the abnormal drone detection unit (232) using a machine learning model that has been trained in a drone movement path prediction unit and predicts the movement path of the drone based on the current location of the drone.

[0136] Here, the drone type classification step is characterized by including a drone identification step in which the drone identification module identifies whether the drone is a drone or a non-drone using a manufacturer identification number among the drone flight data, or identifies whether the drone is a drone or a non-drone using a drone network (SSID) among the drone flight data, and a type classification step in which the drone type classification module analyzes and clusters the drone flight data using a clustering algorithm to classify the drone type into commercial and military types.

[0137] In addition, the abnormal drone detection step is characterized by including an abnormal flight pattern detection step that detects an abnormal flight pattern by analyzing the flight pattern of the drone based on the drone flight data using a clustering algorithm in an abnormal flight pattern detection module, and an intrusion determination step that extracts GPS data of the drone from the drone flight data using a machine learning model in an intrusion determination module and determines whether the drone has intruded into a no-fly zone by analyzing the drone GPS data.

[0138] More specifically, the intrusion determination step is characterized by extracting the drone location, drone altitude, drone flight speed, and drone direction of travel of the drone flight data as the drone GPS data, and if the coordinates of the drone GPS data enter within a preset distance of the radius of the no-fly zone, determining that the drone has intruded into the no-fly zone and generating a first intrusion warning information.

[0139] The above-mentioned drone movement path prediction step is characterized by including a current position detection step in which the current position of a drone that has not detected the abnormal flight pattern in the abnormal drone detection unit (232) is extracted using at least one of a distance estimation method based on received signal strength (RSSI) and a position calculation method based on TDOA in a current position detection module, and a movement path prediction step in which the movement path of the drone is predicted using an LSTM-based time series prediction model based on the extracted current position of the drone in a movement path prediction module, and a movement path prediction step in which the predicted movement path of the drone is determined whether it intrudes into the no-fly zone.

[0141] According to the Bluetooth and Wi-Fi based drone remote identification detection system (1) and the method described above, the following effects are achieved.

[0142] First, the system and method according to the present invention have the effect of accurately tracking the position, altitude, speed, etc. of a drone even in a complex environment where multiple drones are flying simultaneously by collecting and analyzing Remote ID signals transmitted from the drone and Bluetooth and Wi-Fi signals in real time.

[0143] Second, the system and method according to the present invention have the effect of distinguishing between drone signals and signals generated from general IoT devices through AI and machine learning algorithms, and automatically classifying the type of drone to quickly identify unauthorized drone operation or abnormal flight.

[0144] Third, the system and method according to the present invention detect no-fly zones and abnormal flight patterns in real time, generate a warning when a drone approaches a dangerous area or when abnormal signals such as GPS spoofing occur, and, if necessary, enable an immediate response by linking with an air traffic control system.

[0145] Fourth, the system and method according to the present invention provide a path prediction function utilizing Received Signal Strength Indicator (RSSI) and Time Difference of Arrival (TDOA), thereby identifying and predicting the future movement path of a drone in advance, which creates a safe drone operation environment and has the effect of preventing drone-related accidents or illegal activities.

[0146] Fifth, the system and method according to the present invention have the advantage of being able to operate for a long time by utilizing the low-power characteristics of Bluetooth and Wi-Fi technologies.

[0147] The effects obtainable from the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art to which the present invention belongs from the description below.

[0148] Although the technical concept of the present invention described above has been specifically described in preferred embodiments, it should be noted that the aforementioned embodiments are for illustrative purposes only and are not intended to be limiting. Furthermore, those skilled in the art will understand that various embodiments are possible within the scope of the technical concept of the present invention. Accordingly, the true scope of technical protection of the present invention should be determined by the technical concept of the appended claims. Explanation of the symbols

[0149] 1: Bluetooth and Wi-Fi based drone remote identification detection system 100 : Data Collection Unit 200 : Data Analysis Department 300 : Data Visualization Section 400 : Data storage unit 500 : Data transfer section 110: Bluetooth signal acquisition unit 120: Wi-Fi signal collection unit 130 : Data generation section 210 : Data Preprocessing Section 220 : Data learning section 230 : Data Inference Unit 231 : Drone Type Classification Section 232 : Abnormal Drone Detection Unit 233 : Drone Flight Path Prediction Unit

Claims

Claim 1 A Bluetooth and Wi-Fi-based drone remote identification detection system that receives and analyzes Bluetooth signals and Wi-Fi signals, including a remote identification signal (Remote ID) transmitted from a drone, to detect a drone, detect abnormal flight patterns of the drone, and predict a flight path, wherein the system comprises: a data collection unit that uses a Bluetooth module and a Wi-Fi detection module to collect a remote identification signal, a Bluetooth signal, and a Wi-Fi signal transmitted from the drone to detect the drone, and synthesizes the remote identification signal, the Bluetooth signal, and the Wi-Fi signal to generate drone flight data; a data analysis unit that analyzes the Bluetooth signal and Wi-Fi signal collected by the data collection unit to classify the type of the drone, detect abnormal flight patterns, and predict the flight path of the drone; a data visualization unit that provides the collected remote identification number and the results analyzed by the data analysis unit to a user; and a data storage unit that stores the Bluetooth signal, Wi-Fi signal, and remote identification signal collected by the data collection unit, and stores training data and flight prohibition zone information. Claim 2 A Bluetooth and Wi-Fi based drone remote identification detection system according to claim 1, wherein the remote identification signal is data transmitted from the drone, and the remote identification signal includes at least one of a drone identification number, drone location, drone altitude, drone flight speed, drone direction of travel, drone pilot ID, flight status, and the time at which the remote identification signal was transmitted; and wherein the data collection unit comprises: a Bluetooth signal collection unit that detects and receives a Bluetooth signal transmitted from the drone using a Bluetooth detection module; a Wi-Fi signal collection unit that detects and receives a Wi-Fi signal transmitted from the drone using a Wi-Fi detection module; and a data generation unit that integrates the Bluetooth signal and the Wi-Fi signal to generate drone flight data. Claim 3 A Bluetooth and Wi-Fi based drone remote identification detection system according to claim 2, wherein the Bluetooth signal comprises at least one of the remote identification signal, Bluetooth unique identifier (UUID), received signal strength (RSSI), frequency, transmission interval, and manufacturer identification number (OUI), and the Wi-Fi signal comprises at least one of the remote identification number, drone network (SSID), and hardware unique identifier (MAC). Claim 4 A Bluetooth and Wi-Fi based drone remote identification detection system according to claim 1, wherein the data analysis unit comprises: a data preprocessing unit that performs preprocessing to refine and normalize the Bluetooth signal and Wi-Fi signal collected by the data collection unit before analyzing the Bluetooth signal and Wi-Fi signal collected by the data collection unit; a data learning unit that learns a machine learning model using training data and drone flight data; and a data inference unit that classifies drone types based on the drone flight data, detects abnormal drones, and predicts drone flight paths using the machine learning model learned by the data learning unit. Claim 5 A Bluetooth and Wi-Fi based drone remote identification detection system according to claim 4, wherein the data preprocessing unit performs preprocessing through noise removal, missing value processing, and normalization processes in the Bluetooth signal and Wi-Fi signal, and the data learning unit performs supervised learning using normal flight data of an authorized drone and the drone flight data among the learning data, and performs unsupervised learning using unauthorized drone data and abnormal flight data among the learning data and the drone flight data. Claim 6 In claim 4, the data inference unit comprises: a drone type classification unit that determines the presence or absence of a drone based on the drone flight data using a machine learning model that has completed training and classifies the drone type; an abnormal drone detection unit that detects abnormal flight patterns of the drone based on the drone flight data using a machine learning model that has completed training and determines whether the drone has entered a no-fly zone; and a drone movement path prediction unit that detects the current location of a drone where the abnormal flight pattern was not detected by the abnormal drone detection unit using a machine learning model that has completed training and predicts the movement path of the drone based on the current location of the drone. This characterizes a Bluetooth and Wi-Fi-based drone remote identification detection system. Claim 7 A Bluetooth and Wi-Fi based drone remote identification detection system according to claim 6, wherein the drone type classification unit comprises: a drone identification module that identifies whether a drone is a drone or a non-drone using a manufacturer identification number among the drone flight data, or identifies whether a drone is a drone or a non-drone using a drone network (SSID) among the drone flight data; and a drone type classification module that classifies the drone type into commercial and military use by analyzing and clustering the drone flight data using a clustering algorithm. Claim 8 In claim 6, the above-mentioned abnormal drone detection unit comprises: an abnormal flight pattern detection module that detects an abnormal flight pattern by analyzing the flight pattern of the drone based on the drone flight data using a clustering algorithm; and an intrusion determination module that extracts GPS data of the drone from the drone flight data using a machine learning model and determines whether the drone has intruded into a no-fly zone by analyzing the drone GPS data. This characterizes a Bluetooth and Wi-Fi-based drone remote identification detection system. Claim 9 A Bluetooth and Wi-Fi based drone remote identification detection system according to claim 8, wherein the intrusion determination module extracts the drone position, drone altitude, drone flight speed, and drone direction of travel of the drone flight data as the drone GPS data, and if the coordinates of the drone GPS data enter within a preset distance of the radius of the no-fly zone, determines that the drone has intruded into the no-fly zone and generates a first intrusion warning information. Claim 10 In claim 6, the drone movement path prediction unit comprises: a current location detection module that extracts the current location of a drone in which the abnormal flight pattern is not detected by the abnormal drone detection unit using at least one of a distance estimation method based on received signal strength (RSSI) and a location calculation method based on TDOA; and a movement path prediction module that predicts the movement path of the drone using an LSTM-based time series prediction model based on the extracted current location of the drone, and determines whether the predicted movement path of the drone intrudes into the no-fly zone; characterized in that it comprises a Bluetooth and Wi-Fi based drone remote identification detection system. Claim 11 A Bluetooth and Wi-Fi-based drone remote identification detection method comprising receiving and analyzing Bluetooth signals and Wi-Fi signals including a remote identification signal (Remote ID) transmitted from a drone to detect the drone, detect abnormal flight patterns of the drone, and predict a flight path, wherein the method comprises: a data collection step in which a data collection unit collects a remote identification signal, a Bluetooth signal, and a Wi-Fi signal transmitted from the drone using a Bluetooth module and a Wi-Fi detection module, and generates drone flight data by synthesizing the remote identification signal, the Bluetooth signal, and the Wi-Fi signal; a data analysis step in which a data analysis unit analyzes the Bluetooth signal and Wi-Fi signal collected by the data collection unit to detect the drone, classifies the type of the drone, detects abnormal flight patterns, and predicts a flight path of the drone; a data visualization step in which a data visualization unit provides the collected remote identification number and the results analyzed by the data analysis unit to a user; and a data storage step in which a data storage unit stores the Bluetooth signal, the Wi-Fi signal, and the remote identification signal collected by the data collection unit, and stores training data and flight prohibition zone information.

Images