A method and system for identifying the identity of an unmanned aerial vehicle

Abstract

The present application provides a method and system for unmanned aerial vehicle (UAV) identity recognition. Specifically, it receives the flight behavior data stream from the UAV, processes the flight behavior data stream using time series analysis and Bayesian networks to obtain a multi-dimensional behavior trajectory model, compares the multi-dimensional behavior trajectory model with a preset identity behavior library through a graph matching algorithm to identify the behavior identifier that matches the multi-dimensional behavior trajectory model, and uses a spatio-temporal consistency verification mechanism and a random forest algorithm to cross-verify the relevance between the behavior identifier and the pre-stored UAV identity information to determine the identity of the UAV. The present application improves the accuracy and efficiency of UAV identity recognition.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of data analysis and identity recognition of unmanned aerial vehicle flight behaviors, and particularly to a method and system for unmanned aerial vehicle identity recognition. Background Art

[0002] With the wide application of unmanned aerial vehicles in fields such as logistics distribution, agricultural monitoring, security patrol, and environmental monitoring, ensuring the safety and legality of unmanned aerial vehicle operations has become crucial. Unmanned aerial vehicle identity recognition not only needs to process the position change information and flight attitude adjustment records in the unmanned aerial vehicle flight behavior data stream, but also needs to conduct comprehensive analysis by combining multi-dimensional behavior characteristics to ensure the accurate confirmation of the unmanned aerial vehicle identity. In addition, in the face of complex and changeable flight environments and dynamic flight modes, traditional methods based on single identifiers or static path matching can no longer meet the requirements of modern unmanned aerial vehicle management. Therefore, it is particularly necessary to develop a method that can effectively process a large amount of flight behavior data and use advanced data analysis techniques to achieve high-precision identity recognition.

[0003] Currently, unmanned aerial vehicle identity recognition mainly relies on the following methods: First, RFID- or GPS-based identity verification confirms the identity by sending a unique identifier through a device installed on the unmanned aerial vehicle. This method is simple and direct, but is easily affected by signal interference or device failures. Second, visual detection based on image recognition uses a camera to capture the appearance characteristics of the unmanned aerial vehicle and compares them with the images in the database. Although intuitive, it performs poorly in complex environments such as low light and occlusion. Third, traditional trajectory matching based on the flight path identifies the identity by analyzing the historical flight path of the unmanned aerial vehicle and comparing it with a preset standard path. However, this method is difficult to cope with dynamic flight modes and complex flight environments. Finally, some studies have tried to use machine learning algorithms to classify the behavior data of unmanned aerial vehicles, but are often limited to specific types of flight behaviors and lack the construction of a comprehensive multi-dimensional behavior trajectory model and a systematic cross-validation mechanism.

[0004] Although existing solutions can meet certain identity recognition requirements under specific conditions, they generally have the following defects: First, RFID- or GPS-based identity verification methods are easily affected by signal interference or device failures, resulting in recognition failures or misjudgments; second, image recognition-based methods perform poorly in complex environments such as low light and occlusion, restricting their application scope; third, traditional trajectory matching methods only rely on historical path data and lack the analysis of the dynamic characteristics of unmanned aerial vehicle flight behaviors, and cannot effectively cope with changing flight modes and complex flight environments; finally, existing machine learning algorithms are difficult to achieve high-precision identity recognition when processing large-scale flight behavior data, especially in the presence of noise and partial data loss, and the accuracy drops significantly. Summary of the Invention

[0005] An embodiment of the present application provides a method and system for identifying the identity of an unmanned aerial vehicle, aiming to solve the problems of low accuracy and poor efficiency in identifying the identity of unmanned aerial vehicles in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for identifying the identity of an unmanned aerial vehicle, including:

[0007] Receiving a flight behavior data stream from the unmanned aerial vehicle, where the flight behavior data stream includes position change information and flight attitude adjustment records;

[0008] Processing the flight behavior data stream by using time series analysis and Bayesian network to obtain a multi-dimensional behavior trajectory model;

[0009] Comparing the multi-dimensional behavior trajectory model with a preset identity behavior library through a graph matching algorithm to identify a behavior identifier that matches the multi-dimensional behavior trajectory model;

[0010] Using a spatio-temporal consistency verification mechanism and a random forest algorithm to cross-verify the relevance between the behavior identifier and the pre-stored unmanned aerial vehicle identity information to determine the identity of the unmanned aerial vehicle.

[0011] Optionally, the step of comparing the multi-dimensional behavior trajectory model with a preset identity behavior library through a graph matching algorithm to identify a behavior identifier that matches the multi-dimensional behavior trajectory model includes:

[0012] Using the principal component analysis method and the stochastic neighborhood embedding method to perform dimensionality reduction processing on the flight behavior data stream, and performing anomaly detection according to an autoencoder to obtain a flight behavior data set;

[0013] Based on the flight behavior data set, constructing a behavior graph of the multi-dimensional behavior trajectory model, and constructing a corresponding standard behavior graph for each unmanned aerial vehicle in the preset identity behavior library;

[0014] Using a graph matching algorithm to compare the behavior graph with the standard behavior graph to calculate a similarity score, and based on the similarity score, preliminarily screening candidate matching objects from the multi-dimensional behavior trajectory model and the preset identity behavior library, and using a probabilistic graph model to evaluate the candidate matching objects to obtain an optimized similarity score list;

[0015] According to a set similarity threshold, combining Bayesian decision theory, and using spatio-temporal context information, screening out behavior identifiers from the optimized similarity score list.

[0016] Optionally, using the graph matching algorithm, compare the behavior graph with the standard behavior graph to calculate a similarity score. Based on the similarity score, preliminarily screen candidate matching objects from the multi-dimensional behavior trajectory model and the preset identity behavior library, and use the probabilistic graph model to evaluate the candidate matching objects to obtain an optimized similarity score list, including:

[0017] Using the graph matching algorithm, compare the behavior graph with the standard behavior graph to generate an initial similarity score;

[0018] Using the machine learning algorithm, set a dynamic threshold. Based on the dynamic threshold, screen the target behavior graph from the initial similarity score, and use the clustering algorithm to determine the behavior identifier of the target behavior graph, and use the behavior identifier as the candidate matching identifier;

[0019] Using the probabilistic graph model, evaluate the occurrence probability of the candidate matching identifier under different conditions, and combine the time series characteristics of the UAV flight behavior and the influence of key environmental factors. Adopt the variational inference statistical method to adjust the initial similarity score of the candidate matching identifier to generate an optimized similarity score;

[0020] Using the ranking learning algorithm, rank the optimized similarity scores to obtain an optimized similarity score list.

[0021] Optionally, using the probabilistic graph model, evaluate the occurrence probability of the candidate matching identifier under different conditions, and combine the time series characteristics of the UAV flight behavior and the influence of key environmental factors. Adopt the variational inference statistical method to adjust the initial similarity score of the candidate matching identifier to generate an optimized similarity score, including:

[0022] Using the probabilistic graph model, evaluate the occurrence probability of the candidate matching identifier under different conditions to obtain an initial probability distribution;

[0023] According to the initial probability distribution, combine the time series analysis technology to model the time evolution law of the candidate matching identifier to obtain an intermediate probability distribution;

[0024] Collect and analyze the data of key environmental factors. Based on the data of the key environmental factors, adjust the intermediate probability distribution to generate a target probability distribution. The key environmental factors include wind speed, temperature, and humidity;

[0025] Apply the variational inference statistical method. Based on the target probability distribution, evaluate the probability distribution of the candidate matching identifier, and adjust the initial similarity score of the candidate matching identifier to generate an optimized similarity score.

[0026] Optionally, using the spatio-temporal consistency verification mechanism and the random forest algorithm to cross-verify the relevance between the behavior identifier and the pre-stored UAV identity information to determine the identity of the UAV, including:

[0027] Using the spatio-temporal consistency verification mechanism, based on the time and space information extracted from the multi-dimensional behavior trajectory model, verify the behavior identifier to obtain a candidate behavior identifier;

[0028] Using the random forest algorithm, perform cross-validation processing on the candidate behavior identifier and the pre-stored UAV identity information to obtain a cross-validation result;

[0029] Based on a set matching degree threshold, screen out candidate behavior identifiers higher than the set matching degree threshold from the cross-validation results to generate a candidate list;

[0030] Combining the candidate behavior identifier and the cross-validation result, select a target behavior identifier from the candidate list to confirm the identity of the UAV.

[0031] Optionally, using the random forest algorithm to perform cross-validation processing on the candidate behavior identifier and the pre-stored UAV identity information to obtain a cross-validation result, including:

[0032] Construct a random forest algorithm model according to the candidate behavior identifier;

[0033] Pair the candidate behavior identifier with the pre-stored UAV identity information to generate a data set;

[0034] Use the random forest algorithm model to evaluate the data set to calculate a matching probability value;

[0035] Summarize the matching probability values to obtain a cross-validation result.

[0036] Optionally, using time series analysis and Bayesian network to process the flight behavior data stream to obtain a multi-dimensional behavior trajectory model, including:

[0037] Using the time series analysis method, decompose the time and space information in the flight behavior data stream to extract the behavior patterns of the UAV in different time periods, and generate a key time feature set;

[0038] Using the Bayesian network model combined with the key time feature set to generate a multi-dimensional behavior trajectory model.

[0039] In a second aspect, an embodiment of the present application provides a UAV identity recognition system, including:

[0040] A receiving module, configured to receive a flight behavior data stream from a drone, where the flight behavior data stream includes position change information and flight attitude adjustment records;

[0041] A processing module, configured to process the flight behavior data stream by using time series analysis and Bayesian network to obtain a multi-dimensional behavior trajectory model;

[0042] A comparison module, configured to compare the multi-dimensional behavior trajectory model with a preset identity behavior library through a graph matching algorithm to identify a behavior identifier matching the multi-dimensional behavior trajectory model;

[0043] A verification module, configured to cross-verify the relevance between the behavior identifier and pre-stored drone identity information by using a spatio-temporal consistency verification mechanism and a random forest algorithm to determine the identity of the drone.

[0044] In a third aspect, an embodiment of the present application provides a computing device, including a processor and a memory, where a computer program is stored in the memory, and the processor is configured to run the computer program to execute any one of the methods for identifying a drone identity in the first aspect.

[0045] In a fourth aspect, an embodiment of the present application provides a computer storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the methods for identifying a drone identity in any one of the first aspect are implemented.

[0046] In the embodiment of the present application, a flight behavior data stream from a drone is received, where the flight behavior data stream includes position change information and flight attitude adjustment records; the flight behavior data stream is processed by using time series analysis and Bayesian network to obtain a multi-dimensional behavior trajectory model; the multi-dimensional behavior trajectory model is compared with a preset identity behavior library through a graph matching algorithm to identify a behavior identifier matching the multi-dimensional behavior trajectory model; the relevance between the behavior identifier and pre-stored drone identity information is cross-verified by using a spatio-temporal consistency verification mechanism and a random forest algorithm to determine the identity of the drone.

[0047] The technical solution of the present application has the following beneficial effects:

[0048] This application comprehensively utilizes time series analysis, Bayesian networks, graph matching algorithms, spatio-temporal consistency verification mechanisms, and random forest algorithms. This method can accurately identify the behavior patterns of drones in complex environments and effectively screen out the behavior identifiers that match the preset identity behavior library, thereby achieving the accurate confirmation of the identity of drones. By adopting multi-level data processing and verification technologies, including the verification of spatio-temporal consistency and the cross-validation of machine learning algorithms, the security and reliability in the process of drone identity recognition are ensured. Even in the presence of interference or partial data loss, the system can still maintain a high level of accuracy. This method is not only applicable to the identity recognition of drones in static environments but also can handle dynamically changing flight environments. By deeply analyzing the time series characteristics of the flight behavior data stream, the behavior characteristics and their evolution laws of the drone at different time periods can be captured, enhancing the adaptability of the system. Using advanced machine learning algorithms (such as random forests) for fast and effective cross-validation significantly improves the speed and efficiency of drone identity recognition. Compared with the traditional one-by-one comparison method, this method can greatly shorten the recognition time and meet the requirements of real-time monitoring. This method is flexibly designed and easy to integrate new data analysis technologies and algorithm models, and can be adjusted and optimized according to the needs of specific application scenarios. In addition, with the continuous enrichment and improvement of the identity behavior library, the recognition ability of the system will be further enhanced, showing good scalability.

[0049] Furthermore, the multi-dimensional behavior trajectory model is compared with the preset identity behavior library through a graph matching algorithm to identify the matching behavior identifiers. First, the principal component analysis (PCA) and stochastic neighbor embedding (SNE) methods are used to reduce the dimension of the flight behavior data stream, and anomaly detection is performed based on the autoencoder to generate a clean flight behavior data set. Then, according to this data set, the behavior graph of the multi-dimensional behavior trajectory model is constructed, and a corresponding standard behavior graph is constructed for each drone in the identity behavior library. Next, the graph matching algorithm is used to calculate the similarity score between the behavior graph and the standard behavior graph, initially screening out candidate matching objects, and the probabilistic graph model is further used to evaluate these candidate objects to generate an optimized similarity score list. Finally, combined with the set similarity threshold and Bayesian decision theory, the spatio-temporal context information is used to screen out the final behavior identifiers from the optimized similarity score list.

[0050] This method achieves high-precision UAV identity recognition through a series of advanced data analysis techniques. First, dimensionality reduction is performed by PCA and SNE and combined with the anomaly detection of the autoencoder to ensure the quality of the input data and reduce the influence of noise and outliers. Second, a behavior graph is constructed and the similarity score is calculated through a graph matching algorithm, enabling the system to accurately find the standard behavior graph that is closest to the behavior pattern of the target UAV. In addition, a probabilistic graph model is used to evaluate the candidate matching objects, improving the reliability of the recognition results. Finally, combining Bayesian decision theory and spatio-temporal context information, the most compliant behavior identifiers are further screened out, significantly enhancing the accuracy and robustness of identity recognition. This method not only enhances the adaptability and anti-interference ability of the system but also enables efficient and accurate UAV identity confirmation in complex environments.

[0051] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0053] Figure 1 It is a flowchart of a UAV identity recognition method provided by an embodiment of the present application;

[0054] Figure 2 It is a schematic structural diagram of a UAV identity recognition system provided by an embodiment of the present application;

[0055] Figure 3 It is a schematic structural diagram of a computing device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0057] In some processes described in the specification, claims, and the above-mentioned drawings of this application, a number of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent a sequence, and do not limit that "first" and "second" are of different types.

[0058] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present application.

[0059] Figure 1 The following is a flowchart of a method for identifying the identity of a drone provided for an embodiment of the present application, as Figure 1 shown, the method includes:

[0060] Step 101: Receive the flight behavior data stream from the drone, where the flight behavior data stream includes position change information and flight attitude adjustment records;

[0061] In this step, the flight behavior data stream is a series of continuous data points, including the position change information of the drone (such as longitude, latitude, altitude, etc.) and the flight attitude adjustment records (such as pitch angle, yaw angle, roll angle). These data are used to describe the movement trajectory and its attitude change of the drone within a specific time period, and are the basis for subsequent analysis.

[0062] In actual operation, the system receives the flight behavior data sent by the drone in real time through a wireless communication module. These data are first stored in a buffer, and then preprocessed to remove noise and outliers. The preprocessed data will be used to construct a multi-dimensional behavior trajectory model.

[0063] For example, in an intelligent logistics distribution scenario, multiple drones are responsible for the task of delivering packages from a warehouse to a customer's address. Each drone is equipped with a high-precision GPS module and an inertial measurement unit, and can send its position and attitude data to the central control system in real time. For example, when a drone takes off from a warehouse, it will send a set of data containing the current position coordinates (latitude, longitude, altitude) and attitude angles (pitch angle, yaw angle, roll angle) to the system every second. These data not only help the system monitor the flight status of the drone, but also provide basic data for subsequent behavior analysis.

[0064] Step 102: Use time series analysis and Bayesian networks to process the flight behavior data stream and obtain a multi-dimensional behavior trajectory model;

[0065] In this step, time series analysis is a statistical method used to extract features in the time dimension, such as trends, seasonality, and random fluctuations. Bayesian networks are a type of probabilistic graphical model used to represent the dependencies between variables. Combining these two methods can refine a multi-dimensional behavior trajectory model that reflects the behavior patterns of drones from flight behavior data.

[0066] In actual operation, first, use time series analysis techniques to decompose the flight behavior data stream and identify the trend, seasonality, and random fluctuation components. Then, based on the extracted time feature set, use Bayesian networks for modeling to learn the probability distribution of the drone's behavior pattern and its evolution over time, and finally generate a multi-dimensional behavior trajectory model.

[0067] For example, continuing with the above intelligent logistics distribution case, the central control system performs time series analysis on the received position and attitude data. The system first denoises and smooths the data, and then uses an autoregressive integrated moving average model to identify the behavior patterns of the drone in different time periods, such as takeoff, cruising, turning, hovering, etc. Next, the system uses Bayesian networks for modeling, trains the model based on historical data, and predicts the future behavior trend of the drone. For example, during a package delivery process, the system finds that a certain drone frequently turns and hovers in a certain area, which may be due to the presence of obstacles or traffic control in that area. The system then adjusts the flight path of the drone accordingly to ensure safe and efficient task completion.

[0068] Step 103: Use a graph matching algorithm to compare the multi-dimensional behavior trajectory model with a preset identity behavior library to identify the behavior identifier that matches the multi-dimensional behavior trajectory model;

[0069] In this step, the graph matching algorithm is a technique for comparing the similarity between two graph structures. Here, it is used to compare the behavior graph of the UAV with the standard behavior graph in the preset identity behavior library. By calculating the similarity score, the standard behavior graphs that match the behavior pattern of the target UAV can be preliminarily screened out, thereby identifying the candidate behavior identifiers.

[0070] In actual operation, based on the multi-dimensional behavior trajectory model generated in the previous step, the system constructs the corresponding behavior graph and matches it with the standard behavior graph in the identity behavior library. The graph matching algorithm is used to calculate the similarity score between the two, and the candidate matching objects are screened according to the set threshold. This process helps to narrow down the potential identity range and improve the efficiency of subsequent verification.

[0071] For example, in the intelligent logistics distribution scenario, when it is found that a certain UAV deviates from the preset path, the system will immediately trigger the graph matching algorithm and compare the behavior graph of this UAV with all the standard behavior graphs in the identity behavior library. Suppose the detected behavior pattern is highly similar to that of a UAV responsible for distribution in a specific area, and the system lists this UAV as a suspected target. Specifically, the system will extract a series of behavior nodes (such as takeoff, cruise, turn, landing, etc.) of this UAV in the past hour and match them with the standard behavior graph in the identity behavior library. If the similarity score exceeds the preset threshold, it is considered that this UAV highly matches a known UAV, preparing for further identity confirmation.

[0072] Step 104: Use the spatio-temporal consistency verification mechanism and the random forest algorithm to cross-verify the relevance between the behavior identifier and the pre-stored UAV identity information, and determine the identity of the UAV.

[0073] In this step, the spatio-temporal consistency verification mechanism is used to verify the continuity and rationality of the UAV behavior in time and space. The random forest algorithm is an ensemble learning method applicable to classification and regression tasks. By cross-verifying the relevance between the behavior identifier and the pre-stored UAV identity information, the true identity of the UAV can be determined.

[0074] In actual operation, first, use the spatio-temporal consistency verification mechanism to check whether the time and space attributes of the candidate behavior identifier are reasonable. Then, apply the random forest algorithm to cross-verify these identifiers, evaluate the association strength between them and the pre-stored identity information, and finally determine the identity of the UAV. This method improves the accuracy and reliability of identification.

[0075] For example, in the scenario of intelligent logistics distribution, the suspected drones selected by the graph matching algorithm need to have their identities further confirmed. The system verifies whether their behaviors conform to the normal operation logic through a spatio-temporal consistency verification mechanism. For example, the system checks whether the drones fly along the expected path within a specific time period and whether they have reasonable stopping and turning behaviors. Next, the system uses the random forest algorithm to cross-verify these behavior identifiers and evaluates the matching degree of each behavior identifier with the known drone identity information through the trained random forest model. Suppose the system finds that the behavior identifiers of a certain drone highly match those of the drones responsible for distribution in a specific area and its behaviors are also completely consistent spatio-temporally. The system finally confirms the identity of the drone and updates its task status. This not only ensures the safe execution of the task but also enhances the anti-interference ability of the system.

[0076] Through the collaborative work of the above four steps, this method realizes efficient and accurate drone identity recognition. First, it receives and processes the flight behavior data stream of the drones to ensure the integrity and accuracy of the input data; then, it uses time series analysis and Bayesian networks to construct a multi-dimensional behavior trajectory model to lay a foundation for identity recognition; then, it preliminarily screens out candidate behavior identifiers through the graph matching algorithm to narrow down the identity scope; finally, it combines the spatio-temporal consistency verification mechanism and the random forest algorithm for the final identity confirmation. The whole process not only improves the accuracy and robustness of the recognition but also shows good adaptability and anti-interference ability in complex environments and is applicable to various actual application scenarios. Especially in intelligent logistics distribution, this method effectively improves the safety and efficiency of drone operation and ensures the smooth execution of the task.

[0077] To solve the problems of insufficient data dimensionality reduction and anomaly detection in the existing methods and further improve the accuracy and robustness of drone identity recognition, in some embodiments, in step 103, through the graph matching algorithm, comparing the multi-dimensional behavior trajectory model with a preset identity behavior library to identify the behavior identifiers that match the multi-dimensional behavior trajectory model includes:

[0078] Using the principal component analysis method and the stochastic neighbor embedding method, perform dimensionality reduction processing on the flight behavior data stream, and perform anomaly detection based on the autoencoder to obtain a flight behavior data set; based on the flight behavior data set, construct a behavior map of the multi-dimensional behavior trajectory model, and construct a corresponding standard behavior map for each unmanned aircraft in the preset identity behavior library; use the graph matching algorithm to compare the behavior map with the standard behavior map to calculate a similarity score, and based on the similarity score, preliminarily screen candidate matching objects from the multi-dimensional behavior trajectory model and the preset identity behavior library, and use the probabilistic graph model to evaluate the candidate matching objects to obtain an optimized similarity score list; according to the set similarity threshold, combined with the Bayesian decision theory, use the spatio-temporal context information to screen out the behavior identifiers from the optimized similarity score list.

[0079] In this embodiment, principal component analysis is a statistical method used to reduce the data dimension while retaining as much information as possible. Stochastic neighbor embedding is a non-linear dimensionality reduction technique, especially suitable for the visualization and clustering of high-dimensional data. The autoencoder is a neural network structure that can automatically learn the feature representation of data and is used for anomaly detection. Through these methods, noise and outliers can be effectively removed, and a high-quality flight behavior data set can be generated. The behavior map and the standard behavior map constructed based on these data can more accurately reflect the true behavior patterns of unmanned aircraft.

[0080] In the embodiment of the present application, first, use PCA and SNE to perform dimensionality reduction processing on the flight behavior data stream, extract the main features and remove redundant information. Then, apply the autoencoder for anomaly detection, identify and filter out abnormal data points to ensure the quality of the data set. Then, construct a behavior map of the multi-dimensional behavior trajectory model based on the processed data set, and establish a standard behavior map for each unmanned aircraft in the identity behavior library. Next, use the graph matching algorithm to calculate the similarity score between the two, and preliminarily screen out candidate matching objects. Subsequently, use the probabilistic graph model to evaluate these candidate objects to generate an optimized similarity score list. Finally, combined with the set similarity threshold and the Bayesian decision theory, use the spatio-temporal context information to screen out the final behavior identifiers from the optimized score list.

[0081] The following is a specific embodiment:

[0082] In a smart city surveillance system, multiple drones are deployed for urban patrol and emergency response tasks. To ensure the safety and legality of drone operations, the system needs to monitor the behavior of drones in real time and perform identity verification. First, the system receives the flight behavior data stream from each drone, including location change information (such as longitude, latitude, altitude) and flight attitude adjustment records (such as pitch angle, yaw angle, roll angle). To improve data quality, the system uses PCA and SNE to perform dimensionality reduction on this data, extract key features, and remove redundant information. Then, an autoencoder is applied for anomaly detection to identify and filter out abnormal data points, such as sudden altitude changes or unreasonable attitude adjustments, ensuring the integrity and accuracy of the data set.

[0083] Based on the processed data set, the system constructs a behavior map of the multi-dimensional behavior trajectory model of each drone and establishes a standard behavior map for each drone in the identity behavior library. During a daily patrol, a certain drone deviated from the preset path. The system immediately triggers the graph matching algorithm, compares the behavior map of this drone with all the standard behavior maps in the identity behavior library, and calculates the similarity score. Suppose the system finds that the behavior pattern of this drone is highly similar to that of a drone responsible for patrolling a specific area and lists it as a suspected target.

[0084] To further confirm the identity, the system uses a probabilistic graphical model to evaluate the relevance between the suspected target and other potential matching objects, generating an optimized list of similarity scores. Combining the set similarity threshold and Bayesian decision theory, the system uses spatio-temporal context information (such as the historical flight path of the drone and the current environmental conditions) to screen out the final behavior identification from the optimized score list. Suppose the system finally confirms that this drone is indeed the one responsible for patrolling the specific area and updates its task status, ensuring the safe execution of the task. This method not only improves the accuracy and robustness of drone identity recognition but also demonstrates good adaptability and anti-interference ability in complex environments.

[0085] Optionally, the use of the graph matching algorithm in step 103 to compare the behavior map with the standard behavior map to calculate the similarity score, based on the similarity score, preliminarily screening candidate matching objects from the multi-dimensional behavior trajectory model and the preset identity behavior library, and using a probabilistic graphical model to evaluate the candidate matching objects to obtain an optimized list of similarity scores further includes:

[0086] Using a graph matching algorithm, compare the behavior graph with the standard behavior graph to generate an initial similarity score; using a machine learning algorithm, set a dynamic threshold, based on the dynamic threshold, screen the target behavior graph from the initial similarity score, and use a clustering algorithm to determine the behavior identifier of the target behavior graph, and use the behavior identifier as a candidate matching identifier; using a probabilistic graph model, evaluate the occurrence probability of the candidate matching identifier under different conditions, and combine the time series characteristics of the UAV flight behavior and the influence of key environmental factors, and adopt a variational inference statistical method to adjust the initial similarity score of the candidate matching identifier to generate an optimized similarity score; using a ranking learning algorithm, rank the optimized similarity scores to obtain a list of optimized similarity scores. Optionally, the step of using a probabilistic graph model to evaluate the occurrence probability of the candidate matching identifier under different conditions, and combining the time series characteristics of the UAV flight behavior and the influence of key environmental factors, and adopting a variational inference statistical method to adjust the initial similarity score of the candidate matching identifier to generate an optimized similarity score includes: using a probabilistic graph model to evaluate the occurrence probability of the candidate matching identifier under different conditions to obtain an initial probability distribution; according to the initial probability distribution, combining time series analysis techniques, model the time evolution law of the candidate matching identifier to obtain an intermediate probability distribution; collect and analyze data on key environmental factors, and based on the data of the key environmental factors, adjust the intermediate probability distribution to generate a target probability distribution, where the key environmental factors include wind speed, temperature, and humidity; apply a variational inference statistical method, based on the target probability distribution, evaluate the probability distribution of the candidate matching identifier, and adjust the initial similarity score of the candidate matching identifier to generate an optimized similarity score.

[0087] In this embodiment, the graph matching algorithm is used to compare the similarity between two graph structures, which here refers to comparing the behavior graph of the UAV with the standard behavior graph in the preset identity behavior library. The machine learning algorithm is used to set a dynamic threshold, and automatically adjust the threshold according to historical data to adapt to the changing environment. The clustering algorithm is used to identify behavior identifiers with similar characteristics to form a set of candidate matching identifiers. The probabilistic graph model is used to evaluate the occurrence probability of the candidate matching identifier under different conditions, combine the influence of time series analysis techniques and key environmental factors (such as wind speed, temperature, humidity), and adopt a variational inference statistical method to adjust the initial similarity score to generate an optimized similarity score. The ranking learning algorithm is used to rank the optimized similarity scores to obtain the final score list.

[0088] In the embodiments of the present application, first, a graph matching algorithm is used to compare the behavior graph with the standard behavior graph to generate an initial similarity score. Then, a machine learning algorithm is used to set a dynamic threshold, and based on this threshold, the target behavior graphs are screened out from the initial similarity scores, and the behavior identifiers of these graphs are determined through a clustering algorithm as candidate matching identifiers. Next, a probabilistic graph model is used to evaluate the occurrence probabilities of the candidate matching identifiers under different conditions, and a time series analysis technique is combined to model its time evolution law, and data on key environmental factors are collected to adjust the intermediate probability distribution to generate a target probability distribution. The variational inference statistical method is applied to evaluate the probability distribution of the candidate matching identifiers based on the target probability distribution, adjust the initial similarity score, and generate an optimized similarity score. Finally, a ranking learning algorithm is used to rank the optimized similarity scores to obtain a list of optimized similarity scores.

[0089] The following is a specific embodiment:

[0090] In an intelligent agricultural monitoring system, multiple drones are deployed for farmland monitoring and crop spraying tasks. To ensure the safety and legality of drone operations, the system needs to monitor the behavior of drones in real time and perform identity verification.

[0091] First, the system receives the flight behavior data stream from each drone, including position change information (such as longitude, latitude, altitude) and flight attitude adjustment records (such as pitch angle, yaw angle, roll angle). The system uses a graph matching algorithm to compare the behavior graph of each drone with the standard behavior graph in the identity behavior library to generate an initial similarity score. For example, in a daily monitoring task, a certain drone deviates from the preset path, and the system triggers the graph matching algorithm and finds that its behavior pattern is highly similar to that of a drone responsible for monitoring a specific area, generating a relatively high initial similarity score.

[0092] Next, the system uses a machine learning algorithm to set a dynamic threshold, and based on this threshold, the target behavior graphs are screened out from the initial similarity scores, and the behavior identifiers of these graphs are determined through a clustering algorithm as candidate matching identifiers. Suppose the system screens out two suspected drones and lists them as candidate matching identifiers.

[0093] To further confirm the identity, the system uses a probabilistic graph model to evaluate the occurrence probabilities of the candidate matching identifiers under different conditions. The system first establishes an initial probability distribution according to the historical data and the time series characteristics of the current flight behavior. Then, the system collects data on key environmental factors, such as wind speed, temperature, and humidity, and adjusts the intermediate probability distribution based on these data to generate a target probability distribution. For example, suppose the current wind speed is relatively high, and the system predicts that this may affect the flight trajectory of the drone, thereby adjusting its probability distribution.

[0094] Apply variational inference statistical methods to evaluate the probability distribution of candidate matching identities based on the target probability distribution, and adjust their initial similarity scores to generate optimized similarity scores. For example, the system discovers that the probability distribution of one of the candidate drones is significantly higher than that of the other, indicating that it is more likely to be the target drone. Finally, the system uses a ranking learning algorithm to rank the optimized similarity scores to obtain a list of optimized similarity scores. Suppose the system finally confirms that this drone is indeed the one responsible for monitoring a specific area and updates its task status, ensuring the safe execution of the task.

[0095] To further improve the accuracy and reliability of drone identity recognition, in some embodiments, the step of using the spatio-temporal consistency verification mechanism and the random forest algorithm to cross-verify the relevance between the behavior identity and the pre-stored drone identity information to determine the identity of the drone includes:

[0096] Use the spatio-temporal consistency verification mechanism to verify the behavior identity based on the time and space information extracted from the multi-dimensional behavior trajectory model to obtain candidate behavior identities; use the random forest algorithm to perform cross-verification processing on the candidate behavior identities and the pre-stored drone identity information to obtain cross-verification results; based on a set matching degree threshold, screen out candidate behavior identities higher than the set matching degree threshold from the cross-verification results to generate a candidate list; combine the candidate behavior identities and the cross-verification results, and select the target behavior identity from the candidate list to confirm the identity of the drone. Optionally, the step of using the random forest algorithm to perform cross-verification processing on the candidate behavior identities and the pre-stored drone identity information to obtain cross-verification results includes: constructing a random forest algorithm model according to the candidate behavior identities; pairing the candidate behavior identities with the pre-stored drone identity information to generate a data set; using the random forest algorithm model to evaluate the data set to calculate the matching probability value; summarizing the matching probability values to obtain cross-verification results.

[0097] In this embodiment, the spatio-temporal consistency verification mechanism is a method for verifying the consistency of data in the time and space dimensions. It verifies the authenticity and legality by analyzing the time series and geographical location information in the behavior identity. The random forest algorithm is an ensemble learning method that makes predictions by constructing multiple decision trees and summarizes their results to obtain more accurate and reliable classification or regression results. In this scenario, the random forest algorithm is used to evaluate the matching probability value between the candidate behavior identity and the pre-stored drone identity information.

[0098] In the embodiments of the present application, first, a spatio-temporal consistency verification mechanism is adopted to verify the authenticity of behavior identifiers based on the time and space information extracted from the multi-dimensional behavior trajectory model, and eligible candidate behavior identifiers are screened out. Next, a model is constructed using the random forest algorithm, and the candidate behavior identifiers are paired with the pre-stored UAV identity information to form a data set. Then, the random forest algorithm is used to evaluate this data set, and the matching probability values between each candidate behavior identifier and the UAV identity information are calculated. These matching probability values are aggregated to form a cross-validation result. Based on a pre-set matching degree threshold, candidate behavior identifiers with a high matching degree are screened out from these results to generate a candidate list. Finally, considering the candidate behavior identifiers and their corresponding cross-validation results comprehensively, the most likely target behavior identifier is selected from them to confirm the identity of the UAV.

[0099] The following is a specific embodiment:

[0100] In an urban logistics distribution system, UAVs are widely used for the transportation of express packages. To ensure the safety and legality of each UAV's mission execution, the system needs to monitor the behavior of UAVs in real time and confirm their identities.

[0101] The system first collects various data during the flight of the UAV, including the flight path (spatial information) and flight time (time information). Through the spatio-temporal consistency verification mechanism, the system verifies these behavior identifiers, such as checking whether a certain flight conforms to the expected route and schedule, thereby obtaining a set of candidate behavior identifiers.

[0102] Next, the system uses the random forest algorithm to train a model based on historical data to evaluate the correlation between candidate behavior identifiers and the pre-stored UAV identity information. Suppose there is a group of UAVs that have participated in multiple distribution tasks. The system pairs the historical behavior data of these UAVs with the current mission behavior identifiers to create a data set with rich features. Using the random forest algorithm, the system evaluates this data set and calculates the matching probability values between each candidate behavior identifier and the identity information of a specific UAV.

[0103] Based on the set matching degree threshold, the system screens out those candidate behavior identifiers whose matching probability values exceed the threshold to form a candidate list. Considering the complexity of the actual operating environment, the system not only relies on the matching probability values but also combines other factors (such as weather conditions, flight area restrictions, etc.) to select the final target behavior identifier from the candidate list to confirm the identity of the UAV.

[0104] To further improve the analysis accuracy of UAV flight behavior data and ensure that the multi-dimensional behavior trajectory model can accurately reflect the behavior patterns of UAVs, in some embodiments, in step 102, using time series analysis and Bayesian network to process the flight behavior data stream to obtain a multi-dimensional behavior trajectory model, including:

[0105] Using the time series analysis method, decompose and process the time and space information in the flight behavior data stream to extract the behavior patterns of the UAV in different time periods and generate a key time feature set; using the Bayesian network model combined with the key time feature set to generate a multi-dimensional behavior trajectory model.

[0106] In this embodiment, time series analysis is a statistical technique used to identify patterns, trends, and periodicities in time series data. It can help understand the behavior patterns of UAVs in different time periods, such as takeoff, cruising, turning, landing, etc. The key time feature set includes useful information extracted from the original data, such as trend components, seasonal components, and random fluctuation components. The Bayesian network is a probabilistic graphical model used to represent the dependence relationships between variables. By combining the key time feature set, the Bayesian network can learn the probability distribution of UAV behavior patterns and the laws of their evolution over time, thereby generating a multi-dimensional behavior trajectory model that comprehensively describes the behavior of UAVs.

[0107] In the embodiments of the present application, first, use the time series analysis method to decompose and process the time and space information in the flight behavior data stream. Specifically, the system will decompose the original data into different components, such as long-term trends, seasonal variations, and short-term fluctuations. These components together constitute the key time feature set, reflecting the behavior patterns of the UAV in different time periods. Next, using the Bayesian network model combined with these key time feature sets, the system can learn the probability distribution of UAV behavior patterns and capture the laws of their changes over time. Finally, based on this information, a multi-dimensional behavior trajectory model is generated, which can comprehensively describe the behavior characteristics of the UAV and provide a basis for subsequent identity recognition.

[0108] The following is a specific embodiment:

[0109] In an intelligent agricultural monitoring system, multiple UAVs are deployed for farmland monitoring and crop spraying tasks. To ensure the safety and legality of UAV operations, the system needs to monitor the behavior of UAVs in real time and perform identity confirmation.

[0110] First, the system receives the flight behavior data stream from each drone, which includes position change information (such as longitude, latitude, and altitude) and flight attitude adjustment records (such as pitch angle, yaw angle, and roll angle). The system uses time series analysis methods to decompose and process this data. For example, in a daily monitoring task, a certain drone executed multiple round trips to cover the entire farmland area. The system first decomposes the flight path data into long-term trends (such as the overall flight direction), seasonal variations (such as changes in the daily flight time period), and short-term fluctuations (such as minor path deviations caused by wind speed changes), and extracts a key time feature set from them.

[0111] Next, the system uses the Bayesian network model to combine these key time feature sets to generate a multi-dimensional behavior trajectory model. Suppose the system discovers that the drone exhibits obvious cruising and hovering behavior patterns within a specific time period, which may be due to the presence of obstacles or areas that need to be monitored closely in this area. The Bayesian network establishes the probability distribution of the drone's behavior patterns by learning historical data and predicts its future behavior trends. For example, the system predicts that during the next flight, the drone may exhibit a similar behavior pattern in the same area again.

[0112] Based on the generated multi-dimensional behavior trajectory model, the system can not only better understand the current behavior pattern of the drone but also predict its future behavior trends, which helps to optimize scheduling and path planning. For example, in a crop spraying task, the system predicts based on the multi-dimensional behavior trajectory model that a certain farmland may require additional spraying operations and arranges the drone to go to that area in advance, improving the operation efficiency and accuracy.

[0113] This application takes into account the lack of accuracy and robustness of drone behavior recognition in complex environments, especially in dynamically changing flight patterns and complex flight environments, where traditional methods are difficult to provide high-precision identity recognition. Therefore, a new alternative solution is proposed, which includes:

[0114] Using the graph matching algorithm, compare the behavior graph with the standard behavior graph to calculate a similarity score. Based on the similarity score, preliminarily screen candidate matching objects from the multi-dimensional behavior trajectory model and the preset identity behavior library, and use the probabilistic graph model to evaluate the candidate matching objects to obtain an optimized similarity score list, including:

[0115] Using the graph matching algorithm, compare the behavior graph with the standard behavior graph to generate an initial similarity score, where the initial similarity score is calculated as follows:

[0116] ;

[0117] where, Represents the initial similarity score, Represents the time window length, Is the improved time decay factor, And Respectively represent the node set and edge set of the maximum common subgraph at time point The elements in these sets are the common parts between the two graphs and are obtained through graph matching algorithms. And Are respectively the importance weights of node And edge At time point The importance weight reflects the contribution of the node and edge to the overall similarity score. And Are respectively the attribute matching degrees of node And edge At time point The attribute matching degree measures the similarity of the attributes of the node and edge in the two graphs. And Are non - linear adjustment terms, And Respectively represent the node sets of the behavior graph and the standard behavior graph at time point The node sets are the collections of nodes in the two graphs. And Respectively represent the edge sets of the behavior graph and the standard behavior graph at time point The edge sets are the collections of edges in the two graphs.

[0118] The following gives a detailed explanation of each parameter:

[0119] Represents the initial similarity score, which is used to measure the similarity between the UAV behavior graph and the standard behavior graph.

[0120] Represents the time window length, which represents the time period considered. It is usually set according to the application scenario, such as per second, per minute, etc.

[0121] Represents the time decay factor, which reflects the influence of the behavior characteristics at different time points on the final score.

[0122] Can be defined by an exponential function (such as ), where Is the parameter that controls the decay rate.

[0123] And Respectively represent the node set and edge set of the maximum common subgraph at time point The elements in these sets are the common parts between the two graphs and are obtained through graph matching algorithms.

[0124] And Are respectively the importance weights of node And edge At time point The importance weight. The weight reflects the importance of a node or edge to the overall behavior pattern and can be preset according to domain knowledge or automatically adjusted through learning algorithms.

[0125] and are the attribute matching degrees of node and edge at time point respectively. The matching degree measures the similarity of the attributes of corresponding nodes or edges in two graphs and usually ranges from and is obtained by comparing the attributes of nodes or edges.

[0126] and are non - linear adjustment terms used to enhance or weaken the influence of the attribute matching degree. The specific non - linear function can be a power function (such as ), where and are parameters that control the intensity of the influence of the matching degree.

[0127] and represent the node sets of the behavior graph and the standard behavior graph at time point respectively.

[0128] and represent the edge sets of the behavior graph and the standard behavior graph at time point respectively.

[0129] The following introduces the reasons for each sub - design:

[0130] This part calculates the sum of the weighted attribute matching degrees of nodes and edges in the maximum common sub - graph. Through weighted and non - linear adjustment, it ensures that important nodes and edges contribute more to the score, while avoiding unimportant nodes and edges having too much influence on the score. The reason for weighted summation is to comprehensively consider the importance and matching degree of each node and edge, making the score more comprehensive and accurate.

[0131] This part calculates the total weights of nodes and edges in the behavior graph and the standard behavior graph and selects the larger one as the denominator. The purpose of doing this is to standardize the score and prevent the score from being too high or too low. The reason for using the maximum value is to consider that the behavior graph and the standard behavior graph may be asymmetric, and choosing the larger total weight can ensure the stability of the score.

[0132] Using machine learning algorithms, set a dynamic threshold. Based on the dynamic threshold, screen the target behavior map from the initial similarity scores, and use a clustering algorithm to determine the behavior identifiers of the target behavior map, and use the behavior identifiers as candidate matching identifiers;

[0133] Using a probabilistic graphical model, evaluate the occurrence probabilities of the candidate matching identifiers under different conditions, and combine the time series characteristics of the UAV flight behavior and the influence of key environmental factors. Adopt a variational inference statistical method to adjust the initial similarity scores of the candidate matching identifiers to generate optimized similarity scores, where the optimized similarity scores are calculated as follows:

[0134] ;

[0135] where, represents the optimized similarity score, represents the sigmoid function, is the adaptive learning rate, represents the initial similarity score, and are exponential adjustment parameters, represents the conditional probability, is the penalty term.

[0136] The following is a detailed explanation of each parameter:

[0137] represents the optimized similarity score, which is used to measure the matching degree between the adjusted UAV behavior map and the standard behavior map.

[0138] represents the sigmoid function, which is used to map any real value to the interval. The specific form is . This function helps to limit the score range and provides a smooth probability interpretation.

[0139] represents the adaptive learning rate, which is used to control the speed of score update. A higher value will make the score respond to new information faster, while a lower value is more conservative and makes the score change more smoothly. Usually, it is automatically adjusted by machine learning algorithms or set according to experience.

[0140] represents the initial similarity score, which is generated by the graph matching algorithm and represents the preliminary matching degree between the behavior map and the standard behavior map.

[0141] and Represents an exponential adjustment parameter used to enhance or weaken the initial similarity score and the conditional probability on the final score. These parameters can be adjusted according to specific application scenarios to optimize the model performance.

[0142] Represents the conditional probability, which indicates that given environmental factors the probability of the candidate match identifier appearing. This probability is usually estimated through a Bayesian network or other probability models, reflecting the likelihood of the candidate match identifier appearing under different conditions.

[0143] Represents a penalty term used to reduce the impact of certain unfavorable features on the score. The specific penalty term can be a function of the frequency of the candidate match identifier appearance or any negative influencing factors associated with environmental factors.

[0144] The design reasons for each item are introduced as follows:

[0145] This part makes an exponential adjustment to the initial similarity score By introducing the exponential adjustment parameter , the impact of the initial similarity score can be flexibly enhanced or weakened. When , a larger value will be further amplified; when , a smaller value will be relatively reduced. This design makes the score more sensitive to changes in the initial similarity.

[0146] The conditional probability reflects the likelihood of a certain candidate match identifier appearing under specific environmental conditions. By introducing this item, the score can be better adjusted in combination with actual environmental factors, improving the accuracy and robustness of recognition.

[0147] Is a penalty term used to reduce the impact of certain unfavorable features on the score. For example, if a certain candidate match identifier has shown abnormal behavior multiple times in the past, its score can be reduced by increasing its penalty term. This helps to filter out those behavior patterns that do not meet expectations and improve the reliability of the final recognition result.

[0148] This part comprehensively considers the impacts of the initial similarity score, conditional probability, and penalty term, and through an adaptive learning rate and the exponential adjustment parameter Further adjust the score. The index adjustment parameter is similar to , which is used to enhance or weaken the influence of the entire expression.

[0149] By multiplying each sub-item, it is to use the multiplication operation between and in the formula. This is because both reflect the matching degree between the behavior graph and the standard behavior graph, but with different focuses. focuses on the similarity in graph structure, while considers the influence of environmental factors. The multiplication operation can effectively combine the information of these two aspects, making the score more comprehensive and accurate. is subtracted from . This is to reduce the influence of those factors that are not conducive to scoring. In this way, those unreasonable candidate matching identifiers can be effectively filtered out, improving the reliability of the final score.

[0150] The following is a specific example:

[0151] In an intelligent logistics distribution system, multiple drones are deployed for package delivery tasks. Suppose there are two drones A and B, which perform different delivery tasks respectively. To ensure the safety and legality of drone operations, the system needs to monitor the behavior of drones in real time and conduct identity verification.

[0152] Generate the initial similarity score Suppose after comparing the behavior graph of drone A with the standard behavior graph in the identity behavior library, the following parameters are obtained:

[0153] Time window length , time decay factor , importance weights of nodes and edges and are 0.8 and 0.6 respectively, attribute matching degrees and are 0.9 and 0.7 respectively, non-linear adjustment term ;

[0154] According to the formula:

[0155] ;

[0156] It is calculated that .

[0157] Set the dynamic threshold and screen the target behavior graph,

[0158] Use the machine learning algorithm to set the dynamic threshold to 0.8. Since Greater than the threshold value of 0.8, the behavior graph of UAV A is recognized as the target behavior graph. Further, its behavior identifier is determined through a clustering algorithm as the candidate matching identifier.

[0159] Generate the optimized similarity score ,

[0160] Assume that the initial similarity score of UAV A is , conditional probability , penalty term , adaptive learning rate , exponential adjustment parameter and .

[0161] According to the formula:

[0162] ;

[0163] where is the sigmoid function. The calculation result is:

[0164] ;

[0165] It can be seen from the calculation results that the finally obtained optimized similarity score is 0.61. This means that according to the existing data and conditions, the behavior pattern of UAV A is highly consistent with the standard behavior graph. However, considering the influence of some uncertainties and environmental factors, its final score is slightly lower than the initial score. This indicates that UAV A is very likely to belong to the expected identity, but further verification is still required to ensure accuracy. This method not only improves the accuracy of identification but also enhances the robustness and adaptability of the system, especially suitable for application scenarios such as intelligent logistics distribution that require high efficiency and safety.

[0166] Figure 2 This is a schematic structural diagram of a UAV identity recognition system provided by an embodiment of the present application. As Figure 2 shown, the system includes:

[0167] A receiving module 21 for receiving the flight behavior data stream from the UAV, where the flight behavior data stream includes position change information and flight attitude adjustment records;

[0168] A processing module 22 for processing the flight behavior data stream using time series analysis and Bayesian network to obtain a multi-dimensional behavior trajectory model;

[0169] A comparison module 23 for comparing the multi-dimensional behavior trajectory model with a preset identity behavior library through a graph matching algorithm to identify the behavior identifier matching the multi-dimensional behavior trajectory model;

[0170] A verification module 24, configured to use a spatio-temporal consistency verification mechanism and a random forest algorithm to cross-verify the correlation between the behavior identifier and the pre-stored UAV identity information, and determine the identity of the UAV.

[0171] Figure 2 The described UAV identity recognition system can execute Figure 1 The UAV identity recognition method described in the illustrated embodiment, and its implementation principle and technical effects will not be elaborated. For the UAV identity recognition system in the above embodiment, the specific manners in which each module and unit perform operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0172] In a possible design, Figure 2 The UAV identity recognition system of the illustrated embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0173] The storage component 31 stores one or more computer instructions, where the one or more computer instructions are called and executed by the processing component 32.

[0174] The processing component 32 is used for the above Figure 1 The UAV identity recognition method of the illustrated embodiment.

[0175] Among them, the processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components for executing the above method.

[0176] The storage component 31 is configured to store various types of data to support operations on the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0177] Of course, the computing device may necessarily further include other components, such as input / output interfaces, display components, communication components, etc.

[0178] The input / output interface provides an interface between the processing component and the peripheral interface module, and the peripheral interface module may be an output device, an input device, etc.

[0179] The communication component is configured to facilitate communication between the computing device and other devices in a wired or wireless manner, etc.

[0180] Among them, the computing device may be a physical device or an elastic computing host provided by a cloud computing platform, etc. At this time, the computing device may refer to a cloud server, and the above-mentioned processing component, storage component, etc. may be basic server resources rented or purchased from a cloud computing platform.

[0181] The embodiment of the present application also provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it can implement the above Figure 1 a method for identifying the identity of an unmanned aerial vehicle shown in the embodiment.

[0182] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0183] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0184] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution or the part that contributes to the prior art can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0185] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than limiting them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A method for identifying a drone, characterized in that: include: Receiving a flight behavior data stream from the drone, the flight behavior data stream including position change information and flight attitude adjustment records; Using time series analysis and Bayesian networks, the flight behavior data stream is processed to obtain a multi-dimensional behavior trajectory model; By using a graph matching algorithm, the multidimensional behavior trajectory model is compared with a preset identity behavior library to identify a behavior identifier that matches the multidimensional behavior trajectory model; Using a spatiotemporal consistency verification mechanism and a random forest algorithm, cross-verify the association between the behavior identifier and the pre-stored drone identity information to determine the identity of the drone; The multi-dimensional behavior trajectory model is compared with a preset identity behavior library by using a graph matching algorithm to identify a behavior identifier that matches the multi-dimensional behavior trajectory model, including: The flight behavior data stream is subjected to dimensionality reduction processing by using a principal component analysis method and a random neighborhood embedding method, and anomaly detection is performed according to an autoencoder to obtain a flight behavior data set; Based on the flight behavior data set, a behavior map of the multi-dimensional behavior trajectory model is constructed, and a corresponding standard behavior map is constructed for each drone in the preset identity behavior library; Using a graph matching algorithm, the behavior graph is compared with the standard behavior graph to calculate a similarity score, based on the similarity score, candidate matching objects are preliminarily screened from the multidimensional behavior trajectory model and a preset identity behavior library, and the candidate matching objects are evaluated using a probabilistic graph model to obtain an optimized similarity score list; According to the set similarity threshold, combined with Bayesian decision theory, and using spatiotemporal context information, the behavior identifier is screened out from the optimized similarity score list.

2. The method according to claim 1, characterized in that The graph matching algorithm is used to compare the behavior graph with the standard behavior graph to calculate a similarity score, and based on the similarity score, candidate matching objects are preliminarily screened from the multidimensional behavior trajectory model and the preset identity behavior library, and the candidate matching objects are evaluated using a probabilistic graph model to obtain an optimized similarity score list, including: Using a graph matching algorithm, the behavior graph is compared with the standard behavior graph to generate an initial similarity score; Using a machine learning algorithm to set a dynamic threshold, based on the dynamic threshold, screening the target behavior map from the initial similarity score, and using a clustering algorithm to determine the behavior identifier of the target behavior map, and using the behavior identifier as a candidate matching identifier; The probability of occurrence of the candidate matching identifiers under different conditions is evaluated by using a probabilistic graphical model, and the initial similarity scores of the candidate matching identifiers are adjusted by using a variational inference statistical method in combination with the time series characteristics of the UAV flight behavior and the influence of key environmental factors to generate an optimized similarity score; The optimized similarity scores are sorted by using a ranking learning algorithm to obtain an optimized similarity score list.

3. The method according to claim 2, characterized in that The probability graph model is used to evaluate the occurrence probability of the candidate matching identifier under different conditions, and the variational inference statistical method is used to adjust the initial similarity score of the candidate matching identifier in combination with the time series characteristics of the UAV flight behavior and the influence of key environmental factors to generate an optimized similarity score, including: Using a probabilistic graphical model, the occurrence probability of the candidate matching identifier under different conditions is evaluated to obtain an initial probability distribution; Based on the initial probability distribution, combined with time series analysis technology, the time evolution law of the candidate matching identifier is modeled to obtain an intermediate probability distribution; Collecting and analyzing data of key environmental factors, and adjusting the intermediate probability distribution based on the data of the key environmental factors to generate a target probability distribution, wherein the key environmental factors include wind speed, temperature, and humidity; A variational inference statistical method is applied to evaluate the probability distribution of the candidate matching identifiers based on the target probability distribution, and the initial similarity scores of the candidate matching identifiers are adjusted to generate optimized similarity scores.

4. The method according to claim 1, characterized in that The use of the spatiotemporal consistency verification mechanism and the random forest algorithm to cross-verify the association between the behavior identifier and the pre-stored drone identity information to determine the drone's identity includes: Using a spatiotemporal consistency verification mechanism, based on the time and space information extracted from the multidimensional behavior trajectory model, the behavior identifier is verified to obtain a candidate behavior identifier; Using a random forest algorithm, cross-validation is performed on the candidate behavior identifier and the pre-stored drone identity information to obtain a cross-validation result; Based on a set matching degree threshold, candidate behavior identifiers having a matching degree higher than the set matching degree threshold are screened out from the cross-validation results to generate a candidate list; In combination with the candidate behavior identifiers and the cross-validation result, a target behavior identifier is selected from the candidate list to confirm the identity of the drone.

5. The method according to claim 4, characterized in that The random forest algorithm is used to cross-validate the candidate behavior identifier and the pre-stored drone identity information to obtain a cross-validation result, including: Constructing a random forest algorithm model according to the candidate behavior identifiers; Pairing the candidate behavior identifier with pre-stored drone identity information to generate a data set; Using the random forest algorithm model, the data set is evaluated to calculate a matching probability value; The matching probability values ​​are summarized to obtain a cross-validation result.

6. The method according to claim 1, characterized in that The flight behavior data stream is processed by using time series analysis and Bayesian network to obtain a multi-dimensional behavior trajectory model, including: Using a time series analysis method, the time and space information in the flight behavior data stream is decomposed to extract the behavior patterns of the UAV in different time periods and generate a key time feature set; A multi-dimensional behavior trajectory model is generated by combining the key time feature set with a Bayesian network model.

7. A drone identification system, characterized in that: include: A receiving module, used to receive a flight behavior data stream from the UAV, wherein the flight behavior data stream includes position change information and flight attitude adjustment records; A processing module, used to process the flight behavior data stream using time series analysis and Bayesian network to obtain a multi-dimensional behavior trajectory model; A comparison module, used to compare the multi-dimensional behavior trajectory model with a preset identity behavior library through a graph matching algorithm to identify a behavior identifier that matches the multi-dimensional behavior trajectory model; A verification module, used to cross-verify the association between the behavior identifier and the pre-stored drone identity information by using a spatiotemporal consistency verification mechanism and a random forest algorithm to determine the identity of the drone; The multi-dimensional behavior trajectory model is compared with a preset identity behavior library by using a graph matching algorithm to identify a behavior identifier that matches the multi-dimensional behavior trajectory model, including: The flight behavior data stream is subjected to dimensionality reduction processing by using a principal component analysis method and a random neighborhood embedding method, and anomaly detection is performed according to an autoencoder to obtain a flight behavior data set; Based on the flight behavior data set, a behavior map of the multi-dimensional behavior trajectory model is constructed, and a corresponding standard behavior map is constructed for each drone in the preset identity behavior library; Using a graph matching algorithm, the behavior graph is compared with the standard behavior graph to calculate a similarity score, based on the similarity score, candidate matching objects are preliminarily screened from the multidimensional behavior trajectory model and a preset identity behavior library, and the candidate matching objects are evaluated using a probabilistic graph model to obtain an optimized similarity score list; According to the set similarity threshold, combined with Bayesian decision theory, and using spatiotemporal context information, the behavior identifier is screened out from the optimized similarity score list.

8. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a drone identification method as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, a method for identifying a drone as described in any one of claims 1 to 6 is implemented.

Images