Satellite communication authentication integrated method and device

Abstract

The application provides a satellite communication authentication integrated method and device, which comprises the following steps: obtaining a data frame signal of a satellite to be authenticated, inputting the data frame signal into a trained neural network model to form a 3D sample, and calculating the distance between the output feature vector and the feature vector of each legal satellite; determining the launch source authentication identity of the satellite to be authenticated by using the distance; comparing the launch source authentication identity with the launch source identity carried by the signal of the satellite to be authenticated, and if the two are consistent, determining that the satellite to be authenticated is authenticated successfully, which can not only accurately identify and authenticate the identity of the legal satellite signal launch source, but also identify the signal launched by an illegal device. Then, the data frame signal is transmitted to an upper computer for further processing. The application can realize authentication of the satellite signal from two dimensions of time and space, simplifies the authentication process, guarantees the reliability of satellite communication, effectively improves the accuracy of satellite signal launch source identity authentication, and reduces the authentication delay.

Description

Technical Field

[0001] This invention belongs to the field of physical layer identity authentication technology for communication satellites, and specifically relates to an integrated method and device for satellite communication authentication. Background Technology

[0002] Satellite internet is a technology based on satellite communication for internet access. It can form a large-scale network using a certain number of satellites, covering the globe and building a satellite system with real-time information processing capabilities. It is a new type of network capable of providing broadband internet access and other communication services to ground and air terminals, characterized by wide coverage, low latency, broadband speed, and low cost. However, the open environment, the presence of both friendly and adversarial elements, and the weak signal strength caused by long communication distances and rapid environmental changes make satellite communication vulnerable to numerous security threats. The most significant potential security risks are satellite spoofing attacks and replay attacks. Identification and authentication of satellite signal transmitters are effective means to address these security issues. Satellite identification can effectively combat attacks, protect messages, and maintain message integrity. Radio Frequency Fingerprinting (RFF) is a technology for device identification based on transmitter-related signal characteristics. Its basic idea is to extract radio frequency features such as carrier frequency offset, IQ imbalance, and nonlinear offset of radio frequency components from the transmitted signal, and establish a mapping from these features to device identity, achieving unique identification of the transmitting device. Radio frequency fingerprinting technology has advantages such as low communication and computing overhead, non-intrusiveness, and continuity, and has become one of the key technologies for wireless device identification.

[0003] Radio frequency fingerprinting (RF fingerprinting), a radio identification technology based on the characteristics of radio signals, has been widely used in satellite node identification. The existence of RF fingerprints stems from defects within a certain margin of error caused by shortcomings in the manufacturing process of satellite node RF transceivers. These defects are unique, unforgeable, and environmentally independent, resulting in different characteristics in the RF signals transmitted by different communication satellites. RF fingerprinting technology generates a unique identifier for satellite nodes by collecting and analyzing the RF signals transmitted by satellites, used for the identification and authentication of satellite signal transmission sources. Compared to traditional satellite authentication methods, such as cryptographic authentication, radio frequency (RF) fingerprinting technology offers significant advantages: First, RF fingerprinting can authenticate the signal transmitter without third-party security measures, whereas traditional cryptographic authentication requires at least a secure server for identity storage or a secure physical channel for transmitting authentication keys. Second, RF fingerprinting provides greater concealment, as it relies on the characteristics of the satellite's radio frequency circuitry itself, eliminating the need for additional data transmission. This reduces energy consumption by satellite nodes, prevents attackers from interfering with specific authentication processes, and enhances the concealment of the identification process. Third, RF fingerprinting offers strong real-time performance. If the intermediate IQ samples can be transmitted to the authentication processor simultaneously with digital communication, authentication can be completed directly upon receiving the authentication signal, without sending additional data or control frames or adding extra authentication overhead to the data payload. Currently, RF authentication schemes for satellites are under extensive research.

[0004] Currently, researchers have proposed a physical layer identification scheme, PAST-AI, for Iridium satellite signal transmitters. PAST-AI identifies Iridium satellites based on the spatial characteristics of IQ imbalance measured from raw IQ samples. It extracts raw IQ samples from Iridium satellite alarm messages and converts them into a set of grayscale images to demonstrate the satellite's unique IQ imbalance characteristics. This transforms the satellite identification problem into a typical image pattern recognition problem of grayscale images generated by satellite signals, which is solved by a deep convolutional network based on an 18-layer residual network. Although this model has demonstrated potential in satellite identification, its recognition accuracy is positively correlated with the number of IQ samples per unit grayscale image, indicating significant room for improvement in both accuracy and time. Furthermore, this model currently exists only in technical research and has not yet been studied or attempted for practical application or the design of supporting equipment.

[0005] Patent CN116224377B discloses a satellite handover authentication method for low-Earth orbit satellite networks. This method uses a cyclic shift sequence code with a low autocorrelation peak as the authentication code; the cyclically shifted authentication code represents the authentication information to be transmitted; the authentication information undergoes phase modulation processing using multiple code shift keying combinations; and the phase-modulated authentication information is added to a fixed position in the satellite intermediate frequency signal navigation message. This method can effectively solve the problem of spoofing interference during civilian GNSS signal broadcasting, improve the authentication information transmission rate at the transmitter, and increase the accuracy of information decoding at the user receiver. However, this method requires significant modifications to the satellite transmitter, necessitating changes to the entire upper-layer protocol to achieve satellite handover authentication. The inability to modify the inherent components of currently launched satellites affects the practicality of the method.

[0006] Patent CN116056078B discloses a high-speed terminal security authentication method based on trajectory prediction in a space-ground integrated scenario. This method utilizes temporary identity information generated by the high-speed terminal to protect its true identity, and generates new temporary identity information when switching to a new satellite access point, ensuring the security of each authentication process. The method leverages the predictable trajectory of satellites along their orbits to calculate the trajectory key for the current journey and generate a switching credential. When switching to another satellite access point, the switching process can be completed using only the new satellite access point and the high-speed terminal. While this method can achieve switching authentication based on cryptography and certain signal characteristics, its fundamental security guarantee still relies on cryptographic operations such as hashing and chaotic mapping, resulting in high latency and computational complexity.

[0007] Based on the above analysis, the problems and shortcomings of the existing technology are as follows:

[0008] (1) Existing radio frequency identification models for satellite signal transmitters rely on a large number of training samples, requiring the equipment to have certain computing and storage capabilities. Moreover, the developed models can only be used for signal identification, that is, they can only classify the signal to be identified into a certain category of known signals, and cannot identify the signal as an unknown transmitter, that is, they cannot identify illegal transmitters, and thus cannot achieve the authentication of the signal.

[0009] (2) Most existing radio frequency authentication methods for satellite signal transmitters based on radio frequency authentication require modification of the satellite transmission signal structure. On the one hand, the signal used for authentication occupies the bandwidth of satellite communication, leading to a decrease in satellite communication performance; on the other hand, most existing satellite communication systems are deployed in stages, and it is difficult to change the signal structure of satellites launched earlier to ensure compatibility with the authentication scheme.

[0010] (3) Most of the existing radio frequency authentication devices for satellite signal transmitters based on radio frequency authentication have not improved the radio frequency authentication method. Since the IQ samples required for model operation need to allocate a certain bus bandwidth to the receiving device, it affects the normal data communication of satellite signals and cannot achieve real-time radio frequency authentication of satellite signals based on hardware devices.

[0011] The difficulty in solving the above problems and defects is as follows:

[0012] (1) The amount of data invested in radio frequency identification will affect the recognition accuracy of the recognition model. Directly reducing the dimensionality of training data and model input data will directly lead to a decrease in recognition accuracy. Radio frequency identification methods rely on neural network algorithms and cannot infer the possibility that data labels do not appear in the model.

[0013] (2) Most radio frequency identification and authentication uses radio frequency features that need to be estimated based on pilots. Existing satellite communication protocols have their own pilot signals, and the protocols are not open. It is not possible to directly use the pilot symbols set by the protocol. The signal structure must be changed to achieve radio frequency feature extraction.

[0014] (3) The current development of satellite communication is still in the exploratory stage. Most satellite signal transceivers focus more on efficient satellite signal reception and ignore the security threats faced by satellite communication. In order to achieve cost control, they only use general embedded processors and a small amount of DDR, which cannot bear the data transmission bandwidth required for radio frequency certification. Summary of the Invention

[0015] To address the aforementioned problems in the existing technology, this invention provides an integrated satellite communication authentication method. The technical problem to be solved by this invention is achieved through the following technical solution:

[0016] In a first aspect, the present invention provides an integrated satellite communication authentication method comprising:

[0017] S100: Acquire the data frame signal of the satellite to be authenticated, and generate a 3D sample using the data frame signal; wherein, the data frame signal carries the identity of the satellite's transmission source;

[0018] S200, input the 3D sample of the satellite to be certified into the trained neural network model to output the feature vector of the satellite to be certified, and calculate the distance between the feature vector and the feature vector of each legitimate satellite.

[0019] S300: The source authentication identity of the satellite to be authenticated is determined by using the distance between the feature vector of the satellite to be authenticated and the feature vector of each legitimate satellite.

[0020] S400, compare the source authentication identity of the satellite to be authenticated with the source authentication identity carried in the data frame signal of the satellite to be authenticated. If the two are consistent, it is determined that the satellite to be authenticated has been successfully authenticated.

[0021] The S500 transmits the data frame signal of the successfully authenticated satellite to the host computer for further processing.

[0022] Secondly, the present invention provides an integrated satellite communication authentication device comprising:

[0023] The radio frequency receiving and processing module is configured to acquire data frame signals from the satellite to be authenticated; wherein the data frame signals carry the source identity of the satellite to be authenticated.

[0024] The fingerprint generation and authentication module is configured to generate a 3D sample using the data frame signal of the satellite to be authenticated, and input it into a trained neural network model to output the feature vector of the satellite to be authenticated, and calculate the distance between the feature vector and the feature vector of each legitimate satellite; using the distance between the feature vector of the satellite to be authenticated and the feature vector of each legitimate satellite, the transmitter authentication identity of the satellite to be authenticated is determined; if the transmitter authentication identity of the satellite to be authenticated is compared with the transmitter identity carried in the data frame signal of the satellite to be authenticated, and the two match, the authentication of the satellite to be authenticated is determined to be successful.

[0025] The data interface transmits the data frame signal of the successfully authenticated satellite to the host computer for further processing.

[0026] Beneficial effects:

[0027] 1. In training the neural network model, this invention extracts the feature vector of each legitimate satellite and performs identity authentication by comparing the distance between the feature vector of the satellite signal to be authenticated and the feature vector of the legitimate satellite. This not only accurately identifies and authenticates the identity of the legitimate satellite signal transmitter, but also identifies the signals emitted by illegal devices.

[0028] 2. This invention does not require structural changes to existing satellite transmission signals during the authentication process. By utilizing the unbalanced IQ characteristics of satellite signals, a 3D sample map is created from the IQ samples, enabling authentication of satellite signals from both temporal and spatial dimensions. This simplifies the authentication process, ensures the reliability of satellite communication, effectively improves the accuracy of satellite signal source identification, and reduces authentication latency. Furthermore, this method can be applied to various satellite communication protocols, is simple and easy to implement, and has high practical and application value.

[0029] 3. This invention designs an integrated satellite communication authentication device that can perform radio frequency (RF) authentication while the device is transmitting and receiving RF signals from communication satellites. The authentication data and communication data are transmitted through two essentially independent data paths to ensure no interference. This invention enables real-time signal authentication while normal digital communication is being conducted, solving the long-standing problem that RF authentication cannot be directly applied to commercial signal transceivers.

[0030] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0031] Figure 1 This is a flowchart of an integrated satellite communication authentication method provided by the present invention;

[0032] Figure 2 This is a schematic diagram of an integrated satellite communication authentication method provided by the present invention;

[0033] Figure 3 This is a structural diagram of an integrated satellite communication authentication device provided by the present invention;

[0034] Figure 4 This is an experimental result diagram provided in an embodiment of the present invention. Detailed Implementation

[0035] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0036] In satellite communication scenarios, user data undergoes protocol processing by the sender and ultimately reaches the receiver after passing through the transmitter, wireless link, and receiver. The physical layer processing between the transmitter and receiver affects the wireless signal, causing effects such as signal fading, Doppler shift, multipath fading, and signal distortion. In satellite communication, unavoidable errors exist in the manufacturing of radio frequency (RF) chipsets. These errors cause the wireless signal to exhibit specific radiation characteristics, which are ultimately reflected in the transmitted wireless signal. Theoretically, these radiation characteristics can be used to identify signal source nodes in satellite communication networks. For convenience, these radiation characteristics will be referred to as feature vectors below.

[0037] This invention does not employ conventional signal analysis to extract specific features as the basis for radio frequency identification and authentication. Instead, it directly uses the intermediate state of the signal during the communication process to process the signal.

[0038] Combination Figure 1 and Figure 2 This invention provides an integrated satellite communication authentication method, comprising:

[0039] S100: Acquire the data frame signal of the satellite to be authenticated, and generate a 3D sample using the data frame signal; wherein, the data frame signal carries the identity of the satellite's transmission source;

[0040] This step involves creating 3D samples from the data frame signal in the same way that 3D samples are constructed from the authentication dataset when training a neural network model.

[0041] S200, input the 3D sample of the satellite to be certified into the trained neural network model to output the feature vector of the satellite to be certified, and calculate the distance between the feature vector and the feature vector of each legitimate satellite.

[0042] S300: The source authentication identity of the satellite to be authenticated is determined by using the distance between the feature vector of the satellite to be authenticated and the feature vector of each legitimate satellite.

[0043] S400, if the source authentication identity of the satellite to be authenticated is compared with the source authentication identity carried in the data frame signal of the satellite to be authenticated, and if the two are consistent, then the satellite to be authenticated is determined to be successfully authenticated.

[0044] refer to Figure 2 This step utilizes a radio frequency antenna to receive satellite signals, amplifies, filters, and mixes them to obtain a mixed signal. This mixed signal is then digitally sampled to obtain the baseband signal, i.e., the data frame signal. It is divided into two data paths. One path is used for demodulation and decoding, error control, and physical layer processing to obtain the transmitter identity of the satellite to be authenticated. In the other communication path, the data frame signal is sampled and mapped to a grayscale image. These grayscale images are stacked to create a 3D sample, which is then used by a neural network model to identify the 3D sample and obtain the transmitter authentication identity of the satellite to be authenticated. The transmitter authentication identity and the actual transmitter identity are compared. If they match, authentication is successful; otherwise, it fails. Successful authentication allows the signal to be sent to the host computer for further processing. Otherwise, signal processing is terminated.

[0045] The S500 continues to process data frame signals from satellites that have successfully passed authentication.

[0046] This step uses a digital interface to transmit data frame signals to the host computer. This data interface can support common digital communication protocols such as USB and UART and provide a hardware interface, enabling data transmission between the device and the host computer.

[0047] As an optional embodiment of the present invention, the training process of the trained neural network model includes:

[0048] S000 collects communication signal data from all legitimate satellites in the satellite system and acquires two in-phase quadrature IQ sampling data corresponding to their data frames;

[0049] S001, the IQ sampling data is normalized, a two-dimensional IQ sampling grayscale image is generated, and multiple IQ sampling grayscale images are stacked to generate a 3D sample;

[0050] As an optional embodiment of the present invention, S001 includes:

[0051] S001a, the IQ sampled data of the data frame is normalized to obtain the normalized IQ sampled data. The normalization formula is as follows:

[0052]

[0053] Where x represents the IQ sampled data, mean represents the average of the IQ sampled data, and std represents the variance of the IQ sampled data. For IQ sampling data after normalization;

[0054] S001b: Using the two paths of the IQ sampling data after normalization as the horizontal and vertical axes respectively, a constellation diagram is drawn, and the range of the constellation diagram is divided into a d×d grid. The number of sampling points in each grid is counted, and a grayscale image is made according to its distribution frequency.

[0055] S001c, stack the grayscale images from top to bottom in groups of n to form a d×d×n 3D sample.

[0056] This invention directly utilizes IQ sampling of signals for fingerprint construction and authentication, thus preserving the features caused by hardware defects most completely. The data source used is baseband IQ sampling after down-conversion via FPGA, which can be represented as follows:

[0057] S = [[I 0, Q0],[I 1, Q1],…,[I i Q i ],…,[I N Q N ]]

[0058] In the formula, S represents a digital sampling sequence of a satellite communication signal, where each tuple [I o Q o ] represents the i-th orthogonal in-phase sampling point in the sequence, and N is the length of the sampling sequence. Different lengths will affect the size of the fingerprint database and the accuracy of recognition.

[0059] The obtained digital sampled sequence S of the signal is first normalized to obtain S′, and then S′ is divided into n disjoint subsequences S′. j In general, n∈{2,3,4,5}. In each subsequence, each element [Ii Q i ] with its I i and Q i The x and y axes are mapped to a sampling point in a two-dimensional coordinate system, and the entire subsequence is ultimately mapped to a scatter plot G. j Divide the complete scatter plot into m*n sub-blocks and map the scatter plot to a grayscale image M. j The mapping method is as follows:

[0060]

[0061] In the formula, ρ i,j This indicates the number of sampling points that fall within the corresponding block.

[0062] Multiple grayscale images M j The samples are stacked to form a 3D sample corresponding to the satellite communication signal, which serves as the satellite's large radio frequency fingerprint. This 3D sample can be represented as:

[0063] Sata(S) = [M1, M2, ..., M n ]

[0064] Before official use, for the 66 satellite launch nodes of the Iridium system, communication signals from each satellite were continuously collected offline using open-source software such as gr-iridium and a proposed integrated satellite communication authentication device to construct an RFID fingerprint database. For each satellite, the sampling sequence length was set to N = 13 * 10-1. 6 The number of subsequences is n=5, and two hundred standard 3D samples (Data(S)) are stored. j As a standard fingerprint, where j is the satellite number, the authentication dataset DB is finally constructed. RF .

[0065] S002, the 3D samples are combined with the IDs of legitimate satellites to form a labeled authentication dataset, and then the authentication dataset is divided into a training set and a verification set;

[0066] S003, the training set is input into a preset neural network model, and the backpropagation algorithm is used to train it to obtain a trained neural network model.

[0067] Based on transfer learning, the 3D ResNet18 algorithm trained on the Kinetics dataset was used to train the input data to obtain a neural network model that can identify legitimate satellites.

[0068] As an optional embodiment of the present invention, S003 includes:

[0069] S003a, Initialize the training parameters of the preset neural network model, the training parameters including: maximum number of iterations, learning rate, weights and bias parameters;

[0070] S003b, each 3D sample in the training set is input into a preset neural network model to obtain the predicted probability p of the signal source identity corresponding to each legitimate satellite. i and probability distribution q i ;

[0071] S003c, based on the predicted probability p i and the probability distribution q i The cross-entropy loss of the neural network model is calculated using the cross-entropy loss function, which is expressed by the formula:

[0072]

[0073] Where i is the ID of the corresponding legitimate satellite, i = 1, 2, ..., u, and u is the number of legitimate satellites;

[0074] S003d, use the exponential decay method to adjust the learning rate of the neural network model;

[0075] S003e, Repeat steps S003b to S003d until the maximum number of iterations is reached or the loss function converges, to obtain the trained neural network model.

[0076] As an optional embodiment of the present invention, S200 includes:

[0077] S210, input the 3D sample of the satellite to be certified into the trained neural network model to obtain the feature vector f of the satellite to be certified. t ;

[0078] S220, Calculate the feature vector F t With the eigenvector F of each legitimate satellite i Euclidean distance o i cosine similarity to the included angle i , is represented as:

[0079]

[0080] Where i∈1,2,…,u, u represents the number of legal satellites;

[0081] S230, according to the Euclidean distance o i Similarity to the cosine of the included angle (cos) i Calculate the eigenvector F t With the eigenvector F of each legitimate satellite i distance di , is represented as:

[0082]

[0083] As an optional embodiment of the present invention, S300 includes:

[0084] S310, calculate the distance between the feature vector of the satellite to be certified and the feature vector of each legitimate satellite;

[0085] S320, if there is only one distance between the feature vector of the satellite to be certified and the feature vectors of all legitimate satellites that is not greater than the judgment threshold, then find the feature vector of the legitimate satellite with the smallest distance to the feature vector of the satellite to be certified, and determine the identity of the legitimate satellite corresponding to the feature vector as the launch source certification identity of the satellite to be certified.

[0086] S330, if there is no or multiple distances between the feature vector of the satellite to be certified and the feature vectors of all legal satellites that are not greater than the judgment threshold, then the satellite to be certified is initially certified as an illegal satellite.

[0087] As an optional embodiment of the present invention, the judgment threshold is obtained through the following steps:

[0088] S004, The training set is re-input into the trained neural network model to calculate the feature vector corresponding to each legitimate satellite, as follows:

[0089]

[0090] Among them, F t The training set is the dataset of feature vectors output by the trained neural network model for each legal satellite correctly classified, where t = 1, 2, ..., m, and m is the number of data points in the dataset of correctly classified satellites.

[0091] S005, the verification set is input into the trained neural network model to obtain the judgment threshold for judging whether the satellite is legal or illegal.

[0092] As an optional embodiment of the present invention, S005 includes:

[0093] S005a, Based on the feature vector output by the neural network model, set the discrimination threshold θ for distinguishing between legal and illegal satellites. i ;

[0094] S005b, The validation set is input into the trained neural network model, and different discrimination thresholds θ are used. i The accuracy rate;

[0095] S005c, select the discrimination threshold with the highest accuracy as the judgment threshold θ.

[0096] Building upon this, an illegal satellite identification layer is added before the last layer of the neural network model. In this layer, the feature vector F of each legal satellite is used to identify illegal satellites. i distance d i The legality and illegality judgment threshold θ is used to authenticate the source identity of the satellite signal to be authenticated, and finally constitute an authentication model that can identify illegal sources.

[0097] As an optional embodiment of the present invention, S400 includes:

[0098] S410, the data frame signal of the satellite to be authenticated is demodulated and decoded to obtain the transmitter identity X of the satellite to be authenticated. a ;

[0099] S420, authenticate the identity of the launch source of the satellite to be authenticated. i The launch source identity X of the satellite to be certified a If the two are consistent, then the satellite to be certified is confirmed to have been successfully certified.

[0100] This invention provides an integrated satellite communication authentication device comprising:

[0101] The radio frequency receiving and processing module is configured to acquire data frame signals from the satellite to be authenticated and create a 3D sample from them; wherein the data frame signals carry the source identity of the satellite to be authenticated.

[0102] This step utilizes the RF antenna in the RF receiving and processing module to receive the RF signal from the satellite to be authenticated. The RF processing and control module amplifies, filters, mixes, performs digital-to-analog conversion, and performs digital intermediate frequency (IF) processing on the RF signal. This down-converts the RF signal to IF and baseband signals, converts the analog signal to baseband signals, and ultimately achieves digital sampling, transforming the RF electrical signal into a digitally sampleable baseband signal. The data frame signal is obtained by sampling this baseband signal. The RF processing and control module of this invention includes an RF processing chip composed of amplifiers, filters, mixers, and analog-to-digital converters, as well as an FPGA for digital IF and RF control. The signal processed by the FPGA is directly connected to the physical layer control module, and simultaneously, via hardwired, a new data copy is created and directly connected to the fingerprint generation and authentication module, providing a data source for the fingerprint generation and authentication module.

[0103] The radio frequency receiving and processing module distributes the radio frequency signal to the physical layer control module in two paths, thereby enabling real-time synchronization of communication and authentication.

[0104] This invention can utilize a physical layer control module to decode baseband signals at the physical layer and extract the identity identifier of the signal, that is, to extract the transmitter identity X from the data frame signal of the satellite to be authenticated. a The physical layer control module can convert data frame signals into binary bit streams, complete physical layer decoding, obtain the control information contained in the physical layer frame header, acquire the link layer frame, and directly extract the source identity identifier of the signal from the link layer frame, and present the identifier to the fingerprint generation and authentication module.

[0105] The physical layer control module in this step consists of an embedded processor and other necessary components to support the processor in performing the required operations. These necessary components include DDR memory, a clock, and flash memory. The physical layer control module needs to connect to the fingerprint generation and authentication module, providing it with the necessary data to perform its functions. This data includes, in addition to completing routine physical layer processing, extracting the identity declared by the signal using the physical layer payload. This identity can be a link layer identity, and in certain cases, a network layer identity.

[0106] The fingerprint generation and authentication module is configured to generate a 3D sample using the data frame signal of the satellite to be authenticated, and input it into a trained neural network model to output the feature vector of the satellite to be authenticated, and calculate the distance between the feature vector and the feature vector of each legitimate satellite; using the distance between the feature vector of the satellite to be authenticated and the feature vector of each legitimate satellite, the transmitter authentication identity of the satellite to be authenticated is determined; if the transmitter authentication identity of the satellite to be authenticated is compared with the transmitter identity carried in the data frame signal of the satellite to be authenticated, and the two match, the authentication of the satellite to be authenticated is determined to be successful.

[0107] The fingerprint generation and authentication module in the device of this invention is used to generate radio frequency fingerprints of satellite signals, identify radio frequency signals, and authenticate the transmitter source radio frequency, and can generate the transmitter source authentication identity X of the satellite to be authenticated. i And authenticate the launch source identity X of the satellite to be authenticated. i The launch source identity X of the satellite to be certified a The fingerprint generation and authentication module compares and generates the fingerprint recognition result of the satellite to be authenticated, thus completing the legitimate authentication of the satellite. The fingerprint generation and authentication module consists of a high-speed processor and other necessary components to support the processor in performing the required operations, including DDR, clock, and flash memory. The fingerprint generation and authentication module can independently perform radio frequency identification and authentication for a specific satellite signal. The fingerprint generation and authentication module, along with the radio frequency receiving and processing module, provides the data required for authentication, including physical layer IQ sampling and upper-layer device identity. The fingerprint generation and authentication module provides an authentication feedback interface connected to the FPGA, allowing the FPGA to control subsequent radio frequency processing of the signal.

[0108] The data interface transmits the data frame signal of the successfully authenticated satellite to the host computer for further processing.

[0109] refer to Figure 3 This invention provides an integrated satellite communication authentication device. The device is connected to an active antenna RST740 and set to the working frequency band of Iridium satellite, f=1.622GHz. The device monitors the wireless signals in this frequency band in real time. After amplification, filtering, digital-to-analog conversion, down-conversion and other operations on the wireless signals in this frequency band, the device finally forms a digital sampling sequence of the signal.

[0110] This invention provides an integrated satellite communication authentication device that can perform physical layer decoding, control, and authentication of satellite signals. The upper-layer protocol can be implemented by a host computer, and the device connects to the host computer via protocols such as USB or UART. The device can obtain link-layer frames and authentication results through wireless signals in the channel and present them to the upper-layer protocol using the aforementioned protocol.

[0111] This device utilizes open-source tools such as gr-iridium to perform frame detection, frame preamble localization, frame demodulation, and frame decoding on the real-time acquired temporal sampling sequence to obtain a sample sequence of valid frames. Following the same steps as in Method 1, the real-time acquired satellite communication signals are then processed into 3D samples. These 3D samples are then input into a trained 3D convolutional model, and the output of the model is the identified device identity.

[0112] Figure 4 The experimental data presented in this invention demonstrates that the integrated satellite communication authentication method and device described in this example can guarantee performance gains compared to known features. Currently, the only available Iridium satellite identification method based on 3D ResNet achieves an accuracy of 93% with a single data sample using 4200 RRC stage IQ samples (i.e., identification waiting time less than 9 seconds). This invention can improve accuracy while reducing identification waiting time.

[0113] This device unpacks the satellite signal data carried by the Iridium satellite frame and extracts the signal number, comparing it with a fingerprint. When the identification result is an illegal transmission source, the signal is directly identified as being transmitted by an illegal node, and subsequent signal delivery is terminated. When the identification result is a specific satellite, it is compared with the satellite ID claimed by the signal source. If the two are the same, authentication is successful, and data frame signals from that satellite continue to be received; otherwise, authentication fails, and data frame signals from that satellite are stopped, thus achieving integrated satellite communication authentication processing.

[0114] It is worth noting that the terms "first" and "second" in this invention are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0115] Although this application has been described herein in conjunction with various embodiments, those skilled in the art will understand and implement other variations of the disclosed embodiments by reviewing the accompanying drawings, the disclosure, and the appended claims in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality.

[0116] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A satellite communication authentication integrated method, characterized in that, include: S100: Acquire the data frame signal of the satellite to be certified, and generate a 3D sample using the data frame signal; wherein, the data frame signal carries the identity of the transmitter of the satellite to be certified; wherein, generating a 3D sample using the data frame signal includes extracting two in-phase orthogonal IQ sampling data from the data frame signal of the satellite to be certified, normalizing the two in-phase orthogonal IQ sampling data, generating a grayscale image, and finally stacking them to form a 3D sample; S200, input the 3D sample of the satellite to be certified into the trained neural network model to output the feature vector of the satellite to be certified, and calculate the distance between the feature vector and the feature vector of each legitimate satellite. S300: The source authentication identity of the satellite to be authenticated is determined by using the distance between the feature vector of the satellite to be authenticated and the feature vector of each legitimate satellite. S400, compare the source authentication identity of the satellite to be authenticated with the source authentication identity carried in the data frame signal of the satellite to be authenticated. If the two are consistent, it is determined that the satellite to be authenticated has been successfully authenticated. S500 transmits the data frame signal of the successfully authenticated satellite to the host computer for further processing; S200 includes: S210, Input the 3D sample of the satellite to be certified into the trained neural network model to obtain the feature vector of the satellite to be certified. ; S220, Calculate the feature vector With the feature vector of each legitimate satellite European distance Similarity with the cosine of the included angle , is represented as: , Represents the number of legal satellites; S230, according to the Euclidean distance Similarity with the cosine of the included angle Calculate the eigenvectors With the feature vector of each legitimate satellite distance , is represented as: 。 2. The integrated satellite communication authentication method according to claim 1, characterized in that, The training process of the trained neural network model includes: S000 collects communication signal data from all legitimate satellites in the satellite system and acquires two in-phase quadrature IQ sampling data corresponding to their data frames; S001, the IQ sampling data is normalized, a two-dimensional IQ sampling grayscale image is generated, and multiple IQ sampling grayscale images are stacked to generate a 3D sample; S002, the 3D samples are combined with the IDs of legitimate satellites to form a labeled authentication dataset, and then the authentication dataset is divided into a training set and a verification set; S003, the training set is input into a preset neural network model, and the backpropagation algorithm is used to train it to obtain a trained neural network model.

3. The integrated satellite communication authentication method according to claim 2, characterized in that, S001 includes: S001a, the IQ sampled data of the data frame is normalized to obtain the normalized IQ sampled data. The normalization formula is as follows: in, For IQ sampling data, The average value of the IQ sampled data. The variance of the IQ sampling data, For IQ sampling data after normalization; S001b, using the two paths of the entire normalized IQ sampling data as the horizontal and vertical axes respectively, plots a constellation diagram, and divides the constellation diagram range into... The grid is used to count the number of sampling points in each grid and to create a grayscale image based on their distribution frequency. S001c, The grayscale image is rotated from top to bottom... Grouped into sets of one, stacked together 3D samples.

4. The integrated satellite communication authentication method according to claim 2, characterized in that, S003 includes: S003a, Initialize the training parameters of the preset neural network model, the training parameters including: maximum number of iterations, learning rate, weights and bias parameters; S003b, each 3D sample in the training set is input into a preset neural network model to obtain the predicted probability of the signal source identity corresponding to each legitimate satellite. and probability distribution ; S003c, based on predicted probability and the probability distribution The cross-entropy loss of the neural network model is calculated using the cross-entropy loss function, which is expressed by the formula: in, For the corresponding legitimate satellite ID, =1,2,..,u, where u is the number of legal satellites; S003d, use the exponential decay method to adjust the learning rate of the neural network model; S003e, Repeat steps S003b to S003d until the maximum number of iterations is reached or the loss function converges, to obtain the trained neural network model.

5. The integrated satellite communication authentication method according to claim 1, characterized in that, The S300 includes: S310, calculate the distance between the feature vector of the satellite to be certified and the feature vector of each legitimate satellite; S320, if there is only one distance between the feature vector of the satellite to be certified and the feature vectors of all legitimate satellites that is not greater than the judgment threshold, then find the feature vector of the legitimate satellite with the smallest distance to the feature vector of the satellite to be certified, and determine the identity of the legitimate satellite corresponding to the feature vector as the launch source certification identity of the satellite to be certified. S330, if there is no or multiple distances between the feature vector of the satellite to be certified and the feature vectors of all legal satellites that are not greater than the judgment threshold, then the satellite to be certified is initially certified as an illegal satellite.

6. The integrated satellite communication authentication method according to claim 5, characterized in that, The judgment threshold is obtained through the following steps: S004, re-input the training set into the trained neural network model to calculate the feature vector corresponding to each legitimate satellite, as follows: in, The training set is the dataset of feature vectors correctly classified for each legal satellite, output by the trained neural network model. , The number of data points in the dataset that correctly classifies this satellite; S005, input the validation set into the trained neural network model to obtain the threshold for judging whether a satellite is legal or illegal.

7. The integrated satellite communication authentication method according to claim 6, characterized in that, S005 includes: S005a, Based on the feature vector output by the neural network model, set a threshold for distinguishing between legal and illegal satellites. ; S005b, The validation set is input into the trained neural network model, and different discrimination thresholds are applied. The accuracy rate; S005c, Select the discrimination threshold with the highest accuracy as the judgment threshold. .

8. The integrated satellite communication authentication method according to claim 6, characterized in that, The S400 includes: S410, the source identity of the satellite to be authenticated is obtained by demodulating and decoding the data frame signal of the satellite to be authenticated. ; S420, authenticate the identity of the launch source of the satellite to be authenticated. The identity of the launch source of the satellite to be certified If the two are consistent, then the satellite to be certified is confirmed to have been successfully certified.

9. A satellite communication authentication integrated device, characterized in that, include: The radio frequency receiving and processing module is configured to acquire data frame signals from the satellite to be authenticated; wherein the data frame signals carry the source identity of the satellite to be authenticated. The fingerprint generation and authentication module is configured to generate 3D samples using the data frame signals of the satellite to be authenticated, and input them into a trained neural network model to output the feature vector of the satellite to be authenticated, and calculate the distance between the feature vector and the feature vector of each legitimate satellite; using the distance between the feature vector of the satellite to be authenticated and the feature vector of each legitimate satellite, the source authentication identity of the satellite to be authenticated is determined; if the source authentication identity of the satellite to be authenticated is compared with the source identity carried in the data frame signal of the satellite to be authenticated, and if they match, the satellite to be authenticated is determined to be successfully authenticated; wherein, generating 3D samples using data frame signals includes extracting two in-phase orthogonal IQ sampling data from the data frame signal of the satellite to be authenticated, normalizing the two in-phase orthogonal IQ sampling data, generating grayscale images, and finally stacking them to form 3D samples; The 3D samples are input into the trained neural network model to output the feature vector of the satellite to be certified, and the distance between this feature vector and the feature vector of each legitimate satellite is calculated, including: S210, Input the 3D sample of the satellite to be certified into the trained neural network model to obtain the feature vector of the satellite to be certified. ; S220, Calculate the feature vector With the feature vector of each legitimate satellite European distance Similarity with the cosine of the included angle , is represented as: , Represents the number of legal satellites; S230, according to the Euclidean distance Similarity with the cosine of the included angle Calculate the eigenvectors With the feature vector of each legitimate satellite distance , is represented as: ； The data interface transmits the data frame signal of the successfully authenticated satellite to the host computer for further processing.

Images