Drone Communication System Threat Signal Detection Method, System, Device and Medium

Abstract

The present invention provides a method for detecting threat signals in a UAV communication system, belonging to the field of UAV signal detection, including: obtaining real-time traffic data of the UAV communication system; the real-time traffic data includes multiple fields of different links; inputting the real-time traffic data into a pre-trained natural language processing model BERT, and through the encoder of BERT, performing feature representation learning on the context information of multiple fields in the real-time traffic data to obtain the implicit vector features of the timing information in different links, and the implicit vector features are the segmented semantic representations of the data in different links; mapping the implicit vector features to obtain the complete real-time traffic data, and analyzing the complete real-time traffic data to obtain the threat signals in the data. The present invention can obtain the complete real-time traffic data through the pre-trained natural language processing model BERT, and then effectively obtain the threat signals with attacks from the complete real-time traffic data.

Description

Technical Field

[0001] The present invention belongs to the technical field of UAV signal detection, and particularly relates to a method, system, device and medium for detecting threat signals in a UAV communication system. Background Art

[0002] During the operation of a UAV, it often encounters signal threat problems. Since the UAV system has heterogeneous communication protocol types, threat early warning for the entire link needs to consider the traffic of multiple protocol types. The data characteristics of the communication link of the UAV (the communication link is the physical channel between two nodes in the network) generally include instantaneous characteristics (parameters, protocol types, carrier amplitude, phase and frequency, etc.), statistical characteristics (signal distance, cumulant and cyclostationarity, etc.) and transform characteristics (Fourier transform, wavelet transform, S transform, etc.). By analyzing the above characteristics, it is then determined whether there are threat signals entering according to the analysis results.

[0003] However, in the existing UAV signal acquisition process, most of the important characteristics are omitted. When detecting attack signals based on the above characteristics, due to the lack of characteristics, the effective attacks cannot be effectively identified, so it is very difficult to make early warning detection. Summary of the Invention

[0004] In order to solve the problem that in the existing UAV signal acquisition process, most of the important characteristics are omitted, and thus the effective attacks cannot be effectively identified, the present invention provides a method, system, device and medium for detecting threat signals in a UAV communication system.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] A method for detecting threat signals in a UAV communication system includes the following steps:

[0007] Obtain the real-time traffic data of the UAV communication system; the real-time traffic data contains multiple fields of different links;

[0008] Input the real-time traffic data into the pre-trained natural language processing model BERT, and perform feature representation learning on the context information of multiple fields in the real-time traffic data through the encoder of the pre-trained natural language processing model BERT to obtain the implicit vector features of the timing information in different links, and the implicit vector features are the segmented semantic representations of the data in different links;

[0009] Map the implicit vector features to obtain the complete real-time traffic data, and analyze the complete real-time traffic data to obtain the threat signals in the data.

[0010] Preferably, it further includes training the natural language processing model BERT to obtain a pre-trained natural language processing model BERT, specifically including the following steps:

[0011] Obtain the segment data of different links in the historical traffic data, set a mask in the segment data, and cover a continuous segment representation through the mask to construct a training set;

[0012] Input the data in the training set into the natural language processing model BERT, and encode the historical segment data containing the mask through the encoder of the natural language processing model BERT to obtain the corresponding hidden layer feature values, and obtain the pre-trained natural language processing model BERT.

[0013] Preferably, the historical traffic data includes multiple segment data of different links, and the segment data of different links is represented as , where represents the data of a certain segment, and n represents the number of segments; is the mask, i represents the i-th segment.

[0014] Preferably, the historical segment data of different links includes time-domain signal data and frequency-domain signal data.

[0015] Preferably, map the implicit vector features through the fully connected layer of the convolutional neural network to obtain the complete real-time traffic data.

[0016] Preferably, both the historical traffic data and the real-time traffic data are data containing multiple protocol types.

[0017] The present invention also provides a threat signal detection system for an unmanned aerial vehicle communication system, including:

[0018] A data acquisition module for acquiring the real-time traffic data of the unmanned aerial vehicle communication system; the real-time traffic data includes multiple fields of different links;

[0019] A data processing module for inputting the real-time traffic data into the pre-trained natural language processing model BERT, and performing feature representation learning on the context information of multiple fields in the real-time traffic data through the encoder of the pre-trained natural language processing model BERT to obtain the implicit vector features of the time series information in different links, and the implicit vector features are the segmented semantic representations of the data in different links;

[0020] A signal extraction module for mapping the implicit vector features to obtain the complete real-time traffic data, and analyzing the complete real-time traffic data to obtain the threat signals in the data.

[0021] The present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory, and the processor executes the computer program to implement the steps of any one of the methods for detecting threat signals of the drone communication system.

[0022] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is loaded by a processor, it can execute the steps of any one of the methods for detecting threat signals of the drone communication system.

[0023] The method for detecting threat signals of the drone communication system provided by the present invention has the following beneficial effects:

[0024] In the present invention, the encoder of the pre-trained natural language processing model BERT is used to perform feature representation learning on the context information of multiple fields in the real-time traffic data, so as to realize the semantic representation of the full-link traffic. Through learning, all the temporal information of multiple fields in different links can be comprehensively understood to obtain the specific meaning of the entire link, and then the implicit vector features of the temporal information in different links can be obtained. Then, the implicit vector features are mapped to obtain the complete real-time traffic data, and threat signals with attacks can be effectively obtained from the complete real-time traffic data. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention and their design schemes, the accompanying drawings required for this embodiment will be briefly introduced below. The accompanying drawings in the following description are only partial embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0026] Figure 1 It is a flowchart of the method for detecting threat signals of the drone communication system proposed in Embodiment 1 of the present invention;

[0027] Figure 2 It is a threat signal detection framework based on BERT proposed in Embodiment 1 of the present invention;

[0028] Figure 3 It is a flowchart of the system for detecting threat signals of the drone communication system proposed in Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] In order to enable those skilled in the art to better understand the technical solutions of the present invention and implement them, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and cannot be used to limit the protection scope of the present invention.

[0030] Embodiment 1

[0031] Communication signals and natural languages share the same characteristic, that is, they are both time-series signals. The information that appears in different time periods is not the same, and only by integrating all the time-series information (context information) can the specific meaning be understood. Based on this, the present invention proposes a method for detecting threat signals in a UAV communication system. In order to comprehensively retain as much information as possible in the signals, both time-domain signals and frequency-domain signals are used as the input of the detection model in the present invention.

[0032] As Figure 1 shown, the method for detecting threat signals in the UAV communication system proposed by the present invention includes the following steps:

[0033] S1. Obtain the real-time traffic data of the UAV communication system; the real-time traffic data contains multiple fields of different links, and at the same time, the real-time traffic data is also data containing multiple protocol types.

[0034] S2. Input the real-time traffic data into the pre-trained natural language processing model BERT, and through the encoder of the pre-trained natural language processing model BERT, perform feature representation learning on the context information of multiple fields in the real-time traffic data to obtain the hidden vector features of the time-series information in different links. The hidden vector features are the segmented semantic representations of the data in different links.

[0035] S3. Map the hidden vector features to obtain the complete real-time traffic data, and analyze the complete real-time traffic data to obtain the threat signals in the data.

[0036] Specifically, the hidden vector features are mapped through the fully connected layer of the convolutional neural network to obtain the complete real-time traffic data.

[0037] This embodiment can perform early warning analysis based on the segment representation of the communication signals in different links in the complete real-time traffic data. If there are components in the analysis result where the traffic does not meet the predetermined requirements (i.e., is abnormal), an early warning prompt is made.

[0038] In this embodiment, the pre-trained natural language processing model BERT is the natural language processing model BERT after training. The training of the model specifically includes the following steps:

[0039] Obtain the segment data of different links in the historical traffic data, set a mask in the segment data, and cover a continuous segment representation through the mask to construct a training set. Among them, the historical segment data of different links includes time-domain signal data and frequency-domain signal data. At the same time, the historical traffic data is data containing multiple protocol types.

[0040] Input the data in the training set into the natural language processing model BERT. Encode the historical segment data with masks through the encoder of the natural language processing model BERT to obtain the corresponding hidden layer feature values, and obtain the pre-trained natural language processing model BERT.

[0041] The specific model framework is as Figure 2 shown. In this embodiment, the historical traffic data includes multiple segment data of different links, and the segment data of different links are represented as , where represents the data of a certain segment, and n represents the number of segments; is the mask, i represents the i-th segment. The mask masks a continuous segment of wireless signal segments, allowing the entire model to judge the specific content of the signal at that place. That is, through the encoder in the figure, the masked signal in the context environment is predicted . The final obtained complete hidden layer feature values are such as O1 - O5. In this way, the model can well learn the internal relationship between signals and can well extract useful features, and then these features can be fine-tuned and applied to threat detection in the link.

[0042] Based on the same inventive concept, the present invention also provides a threat signal detection system for an unmanned aerial vehicle communication system, as Figure 3 shown, including a data acquisition module 101, a data processing module 102, and a signal extraction module 103.

[0043] Among them, the data acquisition module 101 is used to acquire the real-time traffic data of the unmanned aerial vehicle communication system; the real-time traffic data includes multiple fields of different links.

[0044] The data processing module 102 is used to input the real-time traffic data into the pre-trained natural language processing model BERT, and perform feature representation learning on the context information of multiple fields in the real-time traffic data through the encoder of the pre-trained natural language processing model BERT to obtain the hidden vector features of the time series information in different links, and the hidden vector features are the segmented semantic representations of the data in different links.

[0045] The signal extraction module 103 is used to map the hidden vector features to obtain the complete real-time traffic data, and analyze the complete real-time traffic data to obtain the threat signals in the data.

[0046] Each module in the above threat signal detection system for an unmanned aerial vehicle communication system can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0047] The present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps in the embodiment of the method for detecting threat signals in a drone communication system. For the specific implementation method, reference can be made to the method embodiment, which will not be elaborated here.

[0048] Furthermore, the present invention also provides a non-transitory computer-readable storage medium containing instructions, on which a computer program is stored. For example, a memory containing instructions, and the above instructions can be executed by the processor of the computer device to complete the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc. When the computer program is executed by the processor, it can implement the steps in the embodiment of the method for detecting threat signals in a drone communication system. For the specific implementation method, reference can be made to the method embodiment, which will not be elaborated here.

[0049] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0050] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified function in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0051] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device, and the instruction device implements the specified function in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0052] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0053] It should be pointed out that the specific implementation methods described above can enable those skilled in the art to understand the invention more comprehensively, but do not limit the invention in any way. Therefore, although the invention has been described in detail in this specification and embodiments, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the protection scope of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved. Any simple change or equivalent replacement of the technical solution that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belongs to the protection scope of the present invention.

Claims

1. A method for detecting threat signals in a drone communication system, characterized in that: The following steps are involved: Acquire real-time traffic data of the UAV communication system; the real-time traffic data includes multiple fields of different links; The real-time traffic data is input into the pre-trained natural language processing model BERT, and the context information of multiple fields in the real-time traffic data is learned through the encoder of the pre-trained natural language processing model BERT, so as to obtain the implicit vector features of the timing information in different links, and the implicit vector features are the segmented semantic representations of the data in different links; The latent vector features are mapped to obtain complete real-time traffic data, and the complete real-time traffic data is analyzed to obtain threat signals in the data; specifically, the latent vector features are mapped through the fully connected layer of the convolutional neural network to obtain complete real-time traffic data; It also includes training the natural language processing model BERT to obtain a pre-trained natural language processing model BERT, which specifically includes the following steps: Obtaining fragment data of different links in historical traffic data, setting a mask in the fragment data, masking a continuous fragment representation by the mask, and constructing a training set; the historical fragment data of different links includes time domain signal data and frequency domain signal data; The data in the training set is input into the natural language processing model BERT, and the historical segment data containing the mask is encoded through the encoder of the natural language processing model BERT to obtain the corresponding hidden layer feature values, thereby obtaining a pre-trained natural language processing model BERT.

2. The method for detecting threat signals of a drone communication system according to claim 1, characterized in that: The historical traffic data includes multiple fragment data of different links, and the fragment data of different links are represented as ,in Represents the data of a certain fragment, and n represents the number of fragments; For the mask, i represents the i-th fragment.

3. The method for detecting threat signals of a drone communication system according to claim 1, characterized in that: Both the historical traffic data and the real-time traffic data contain data of multiple protocol types.

4. A system for implementing the method for detecting threat signals of a drone communication system according to any one of claims 1 to 3, characterized in that: include: A data acquisition module is used to obtain real-time traffic data of the UAV communication system; The real-time traffic data includes multiple fields of different links; A data processing module is used to input the real-time traffic data into a pre-trained natural language processing model BERT, and perform feature representation learning on the context information of multiple fields in the real-time traffic data through the encoder of the pre-trained natural language processing model BERT, so as to obtain implicit vector features of the timing information in different links, wherein the implicit vector features are segmented semantic representations of the data in different links; The signal extraction module is used to map the implicit vector features to obtain complete real-time traffic data, analyze the complete real-time traffic data, and obtain threat signals in the data.

5. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the drone communication system threat signal detection method according to any one of claims 1 to 3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is loaded by a processor, it can execute the steps of the threat signal detection method of the drone communication system described in any one of claims 1 to 3.

Images