Deep learning-based aviation aircraft satellite communication fault diagnosis and repair method and system

Abstract

The invention relates to an aviation aircraft satellite communication fault diagnosis and repair method and system based on deep learning, and belongs to the technical field of satellite fault diagnosis. The method comprises the following steps: acquiring a communication satellite topology network, acquiring observation parameters of the communication satellite topology network through a dual-mode heterogeneous receiving terminal, and performing data archiving to generate a constellation measurement data set; obtaining abnormal node positioning information through a satellite group behavior perception model according to the constellation measurement data set and the communication satellite topology network; obtaining an inter-satellite communication abnormal mode through an abnormal mode reconstruction model according to the abnormal node positioning information; and performing fault repair according to the inter-satellite communication abnormal mode. And aviation aircraft satellite communication fault diagnosis and repair based on deep learning are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of satellite fault diagnosis, and particularly relates to a method and system for satellite communication fault diagnosis and repair of an aircraft based on deep learning. Background Art

[0002] With the continuous development of global informatization, digitization, and aerospace technology, traditional single satellite communication systems have been difficult to meet the increasingly complex application scenarios and high-efficiency service requirements. In this context, the construction of multi-satellite communication systems has become one of the core directions of modern aerospace technology and plays an irreplaceable role in many important fields. However, with the significant increase in the number of satellites, the probability of multiple satellites failing simultaneously also increases. How to accurately and timely diagnose the fault types of multiple satellites has become an important challenge in satellite communication fault repair.

[0003] Traditional fault detection methods usually rely on the fault diagnosis of a single satellite. However, in a multi-satellite system, this method often cannot effectively handle the situation where multiple satellites fail simultaneously, easily leading to misdiagnosis or missed diagnosis. In addition, due to the general scarcity of fault sample data in satellite systems and the strong concealment and diversity of fault types, historical operation data cannot provide sufficient information to support accurate fault classification and location. This makes it impossible to identify the fault source and locate the fault type in a timely manner in practical applications, further increasing the complexity of satellite communication fault repair. Therefore, how to accurately identify the unknown fault types of multi-satellite communication is an urgent problem to be solved. Summary of the Invention

[0004] To solve the above problems existing in the prior art, the present invention provides a method and system for satellite communication fault diagnosis and repair of an aircraft based on deep learning.

[0005] The object of the present invention can be achieved by the following technical solutions:

[0006] A method for satellite communication fault diagnosis and repair of an aircraft based on deep learning includes:

[0007] Obtain the communication satellite topology network, collect the observation parameters of the communication satellite topology network through a dual-mode heterogeneous receiving terminal and perform data archiving to generate a constellation measurement data set;

[0008] Obtain the abnormal node location information through a star cluster behavior perception model according to the constellation measurement data set and the communication satellite topology network;

[0009] Obtain the inter-satellite communication abnormal mode through an abnormal mode reconstruction model according to the abnormal node location information; perform fault repair according to the inter-satellite communication abnormal mode.

[0010] Preferably, the constellation behavior perception model includes:

[0011] Obtain the initial network topology, positioning parameters, and candidate abnormal nodes through preliminary fault detection based on the constellation measurement data set and the communication satellite topology network;

[0012] Obtain the abnormal node positioning information through back substitution detection based on the initial network topology, the candidate abnormal nodes, and the positioning parameters.

[0013] Preferably, the preliminary fault detection specifically includes:

[0014] Obtain the measurement data and predicted pseudorange in the measurement data set, and calculate the pseudorange deviation amount through residual calculation based on the measurement data and the predicted pseudorange;

[0015] Obtain the normalized pseudorange deviation amount through normalizing the pseudorange deviation amount;

[0016] Obtain the maximum normalized pseudorange deviation amount through sorting based on the normalized pseudorange deviation amount, and mark the satellite corresponding to the maximum normalized pseudorange deviation amount as a candidate abnormal node;

[0017] Calculate and sort the maximum correlation coefficient through correlation calculation based on the candidate abnormal nodes; preset a correlation threshold. If the maximum correlation coefficient is greater than the correlation threshold, the satellite corresponding to the maximum correlation is a candidate abnormal node; if the maximum correlation is not greater than the correlation threshold, the satellite corresponding to the maximum correlation is a normal node;

[0018] Obtain the initial network topology by excluding the candidate abnormal nodes from the communication satellite topology network; obtain the positioning parameters based on the initial network topology.

[0019] Preferably, the back substitution detection includes:

[0020] Calculate the global detection statistic through global statistic calculation based on the candidate abnormal nodes and the positioning parameters;

[0021] Preset a global threshold, and obtain the final seat without candidate abnormal nodes through recovery judgment based on the global detection statistic and the global threshold;

[0022] The recovery judgment includes judging whether the global detection statistic is less than the global threshold. If so, the candidate abnormal node is restored to a normal node; if not, it remains a candidate abnormal node;

[0023] Obtain the abnormal node positioning information through least squares estimation based on the final seat without candidate abnormal nodes.

[0024] Preferably, the abnormal mode reconstruction model includes:

[0025] Obtain the telemetry data of the faulty satellite, and obtain the latent space representation through the compression model according to the telemetry data of the faulty satellite;

[0026] Probability model;

[0027] Obtain the inter-satellite communication anomaly pattern through the scoring model according to the in-distribution likelihood ratio.

[0028] Preferably, the compression model includes:

[0029] Reduce the dimension of the telemetry data of the faulty satellite to obtain the dimension-reduced telemetry data of the faulty satellite;

[0030] Obtain the latent space representation through the low-dimensional representation optimization model according to the dimension-reduced telemetry data of the faulty satellite.

[0031] Preferably, the low-dimensional representation optimization model includes:

[0032] Obtain the initial low-dimensional representation of the sample by minimizing the reconstruction loss function according to the dimension-reduced telemetry data of the faulty satellite;

[0033] The reconstruction loss function is expressed as:

[0034]

[0035] where L rec represents the reconstruction loss function, N represents the number of dimension-reduced telemetry data of the faulty satellite, i represents the i-th dimension-reduced telemetry data of the faulty satellite, x i represents the dimension reduction of the telemetry data of the faulty satellite, Attention(x′ i , ai) represents the attention mechanism, x′ i represents the data after passing through the reconstruction loss function, a i represents the attention weight, λ is the regularization parameter, j represents the j-th weight, M represents the number of weights, w j represents the weight;

[0036] Obtain the latent space representation by calculating according to the initial low-dimensional representation of the sample through the contrast loss function.

[0037] Preferably, the probability model includes:

[0038] Obtain the component type probability through the artificial neural network according to the latent space representation;

[0039] Obtain the complete mixture component parameters through maximization calculation according to the component type probability, and the complete mixture component parameters include weight parameter, mean vector parameter, and covariance matrix parameter;

[0040] The in-distribution likelihood ratio is obtained through probability calculation according to the complete mixture component parameters and the latent space representation;

[0041] The probability calculation expression is:

[0042]

[0043] where P(x) is the in-distribution likelihood ratio, c represents the c-th complete mixture component parameter, C represents the number of complete mixture component parameters, is the weight, |∑ c | represents the determinant of the covariance matrix of the c-th complete mixture component, tr represents the trace operation, x represents the latent space representation, μ c is the mean vector, and T represents the transpose calculation.

[0044] Preferably, the scoring model includes:

[0045] The out-of-distribution likelihood value is obtained through scoring calculation according to the in-distribution likelihood ratio;

[0046] The scoring calculation expression is:

[0047]

[0048] where Score(x out ) is the out-of-distribution likelihood value, c represents the c-th complete mixture component parameter, C represents the number of complete mixture component parameters, is the weight, |∑ c | represents the determinant of the covariance matrix of the c-th complete mixture component, x in represents the latent space representation, μ c is the mean vector, T represents the transpose calculation, and the out-of-distribution likelihood value is obtained through negative logarithm calculation of the in-distribution likelihood ratio;

[0049] A preset scoring threshold is set, and the inter-satellite communication abnormal mode is obtained through scoring judgment according to the out-of-distribution likelihood value and the scoring threshold;

[0050] The scoring judgment includes judging whether the out-of-distribution likelihood value is greater than the scoring threshold. If so, the inter-satellite communication abnormal mode is of an unknown type; if not, the inter-satellite communication abnormal mode is of a known type.

[0051] An aircraft satellite communication fault diagnosis and repair system based on deep learning includes a data acquisition module, a fault detection module, and a type diagnosis module:

[0052] The data acquisition module is used to obtain the communication satellite topology network, collect the observation parameters of the communication satellite topology network through a dual-mode heterogeneous receiving terminal, and perform data archiving to generate a constellation measurement data set;

[0053] The fault detection module is used to obtain the abnormal node positioning information through a satellite group behavior perception model according to the constellation measurement data set and the communication satellite topology network;

[0054] The type diagnosis module is used to obtain the inter-satellite communication abnormal mode through an abnormal mode reconstruction model according to the abnormal node positioning information; and perform fault repair according to the inter-satellite communication abnormal mode.

[0055] The beneficial effects of the present invention are as follows:

[0056] (1) Through the satellite group behavior perception model and deep learning technology, the position of the faulty satellite can be accurately identified from satellite measurement data, and the determination of the abnormal node positioning information is further optimized through back substitution detection, improving the accuracy and reliability of fault detection.

[0057] (2) Through the low-dimensional characterization of the telemetry data of the faulty satellite and probability model calculation, the system can efficiently diagnose the inter-satellite communication abnormal mode. Combining the optimization of the contrast loss function of deep learning and the component type probability analysis, the system can determine the fault type in a short time, improving the efficiency and scalability of diagnosis, and providing timely decision support for subsequent repair work.

[0058] (3) Through the out-of-distribution likelihood value calculated by the scoring model, the system can detect unknown faults. Combining the preset scoring threshold, it can identify unknown obstacles and take corresponding repair measures, improving the accuracy and efficiency of satellite communication fault repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] For the convenience of those skilled in the art to understand, the present invention will be further described below with reference to the accompanying drawings.

[0060] Figure 1 It is a schematic flowchart of a method for diagnosing and repairing satellite communication faults of an aircraft based on deep learning according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0061] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, with reference to the accompanying drawings and preferred embodiments, describe in detail the specific embodiments, structures, features and their effects according to the present invention.

[0062] Please refer to Figure 1 , a method for diagnosing and repairing satellite communication faults of an aircraft based on deep learning, including:

[0063] Obtain the communication satellite topology network, collect the observation parameters of the communication satellite topology network through a dual-mode heterogeneous receiving terminal, and perform data archiving to generate a constellation measurement data set;

[0064] Obtain the abnormal node positioning information according to the constellation measurement data set and the communication satellite topology network through a satellite group behavior perception model;

[0065] Obtain the inter-satellite communication abnormal mode according to the abnormal node positioning information through an abnormal mode reconstruction model; perform fault repair according to the inter-satellite communication abnormal mode.

[0066] Specifically, the dual-mode heterogeneous receiving terminal includes GPS L1 and BDS B1; the constellation measurement data includes pseudorange; the constellation represents a satellite network, and the satellites in the satellite network cooperate with each other according to specific orbits and configurations.

[0067] In this embodiment, the constellation consists of 24 satellites, which are distributed in 6 orbital planes, with 4 satellites in each plane, and the angular interval between the satellites is 60 degrees to ensure global coverage. All satellites are located in the medium Earth orbit and have a relative distance of 3600 kilometers. Each satellite is responsible for specific tasks, including functions such as global navigation and positioning, communication, weather monitoring, disaster warning, and precise time synchronization.

[0068] Specifically, the satellite group behavior perception model includes:

[0069] S201: Obtain the initial network topology, positioning parameters, and candidate abnormal nodes through preliminary fault detection according to the constellation measurement data set and the communication satellite topology network;

[0070] S202: Obtain the abnormal node positioning information through back substitution detection according to the initial network topology, the candidate abnormal nodes, and the positioning parameters.

[0071] Specifically, the preliminary fault detection specifically includes:

[0072] S201-1: Obtain the measurement data and predicted pseudorange in the measurement data set, and calculate the pseudorange deviation amount through residual calculation according to the measurement data and the predicted pseudorange;

[0073] S201-2: Obtain the normalized pseudorange deviation amount through normalization processing of the pseudorange deviation amount;

[0074] S201-3: Obtain the maximum normalized pseudorange deviation amount through sorting according to the normalized pseudorange deviation amount, and mark the satellite corresponding to the maximum normalized pseudorange deviation amount as a candidate abnormal node;

[0075] S201-4: Calculate and sort the maximum correlation coefficient based on the candidate abnormal node through correlation calculation; preset a correlation threshold. If the maximum correlation coefficient is greater than the correlation threshold, the satellite corresponding to the maximum correlation is the candidate abnormal node; if the maximum correlation is not greater than the correlation threshold, the satellite corresponding to the maximum correlation is a normal node;

[0076] S201-5: Obtain the initial network topology by excluding the candidate abnormal node from the communication satellite topology network; obtain the positioning parameters according to the initial network topology.

[0077] Specifically, the back substitution detection includes:

[0078] S202-1: Calculate the global detection statistic through global statistics based on the candidate abnormal node and the positioning parameters;

[0079] The calculation expression of the global statistic is:

[0080]

[0081] Among them, S is the global detection statistic, T represents the transpose operation, represents the positioning parameters, Δy is the pseudorange deviation, ρ represents the measurement data, l represents the distance between the receiver position and the satellite position, c represents the speed of light, δt s is the satellite clock error, δt u is the receiver clock error, ΔI is the ionospheric delay, and ΔT is the tropospheric delay;

[0082] S202-2: Preset a global threshold, and obtain the final seat without candidate abnormal nodes through recovery judgment based on the global detection statistic and the global threshold;

[0083] The recovery judgment includes judging whether the global detection statistic is less than the global threshold. If so, the candidate abnormal node is restored to a normal node; if not, it remains a candidate abnormal node;

[0084] S202-3: Obtain the abnormal node positioning information through least squares estimation based on the final seat without candidate abnormal nodes.

[0085] Specifically, the abnormal mode reconstruction model includes:

[0086] S301: Obtain the telemetry data of the faulty satellite, and obtain the potential space representation through a compression model based on the telemetry data of the faulty satellite;

[0087] S302: Probability model;

[0088] S303: Obtain the inter-satellite communication anomaly pattern through the scoring model according to the in-distribution likelihood ratio.

[0089] Specifically, the compression model includes:

[0090] S301-1: Obtain the telemetry dimensionality-reduced data of the faulty satellite by performing dimensionality reduction on the telemetry data of the faulty satellite;

[0091] S301-2: Obtain the latent space representation through the low-dimensional representation optimization model according to the telemetry dimensionality-reduced data of the faulty satellite.

[0092] Specifically, the low-dimensional representation optimization model includes:

[0093] Obtain the initial low-dimensional representation of the sample by minimizing the reconstruction loss function according to the telemetry dimensionality-reduced data of the faulty satellite;

[0094] The reconstruction loss function is expressed as:

[0095]

[0096] where L rec represents the reconstruction loss function, N represents the number of telemetry dimensionality-reduced data of the faulty satellite, i represents the i-th telemetry dimensionality-reduced data of the faulty satellite, x i represents the telemetry dimensionality reduction of the faulty satellite, Attention(x′ i , ai) represents the attention mechanism, x′ i represents the data after passing through the reconstruction loss function, a i represents the attention weight, λ is the regularization parameter, j represents the j-th weight, M represents the number of weights, and w j represents the weight;

[0097] Obtain the latent space representation by calculating according to the initial low-dimensional representation of the sample through the contrast loss function.

[0098] Specifically, the probability model includes:

[0099] S302-1: Obtain the component type probability through an artificial neural network according to the latent space representation;

[0100] S302-2: Obtain the complete mixture component parameters by maximizing the calculation according to the component type probability;

[0101] S302-3: Obtain the in-distribution likelihood ratio through probability calculation according to the complete mixture component parameters and the latent space representation;

[0102] The probability calculation expression is:

[0103]

[0104] Among them, P(x) is the in-distribution likelihood ratio, c represents the c-th complete mixture component parameter, C represents the number of complete mixture component parameters, is the weight, |Σ c | represents the determinant of the covariance matrix of the c-th complete mixture component, tr represents the trace operation, x represents the latent space representation, μ c is the mean vector, T represents the transpose calculation;

[0105] Specifically, the complete mixture component parameters include weight parameters, mean vector parameters, and covariance matrix parameters.

[0106] In this embodiment, the artificial neural network includes three hidden layers. The first hidden layer has 128 neurons; the second hidden layer has 64 neurons; the third hidden layer has 32 neurons; the ReLU activation function is used for the hidden layers; the softmax activation function is used for the output layer.

[0107] Specifically, the scoring model includes:

[0108] S303-1: Calculate the out-of-distribution likelihood value through scoring according to the in-distribution likelihood ratio;

[0109] The scoring calculation expression is:

[0110]

[0111] where Score(x out ) is the out-of-distribution likelihood value, c represents the c-th complete mixture component parameter, C represents the number of complete mixture component parameters, is the weight, |Σ c | represents the determinant of the covariance matrix of the c-th complete mixture component, x in represents the latent space representation, μ c is the mean vector, T represents the transpose calculation, and the out-of-distribution likelihood value is obtained by performing negative logarithm calculation on the in-distribution likelihood ratio;

[0112] S303-2: Preset a scoring threshold, and obtain the inter-satellite communication anomaly mode through scoring judgment according to the out-of-distribution likelihood value and the scoring threshold;

[0113] The scoring judgment includes judging whether the out-of-distribution likelihood value is greater than the scoring threshold. If so, the inter-satellite communication anomaly mode is of unknown type; if not, the inter-satellite communication anomaly mode is of known type.

[0114] In this embodiment, the system includes a sensor, an inferencer, an actuator, an HMI interface, a communication controller, and a ground receiver. The sensor monitors the satellite environment information and status data in real time and stores them in a database. The inferencer analyzes satellite faults based on the constellation behavior perception model and the abnormal pattern reconstruction model, identifies the fault type and location, and transmits the results to the actuator. After receiving the fault information, the actuator feeds it back to the communication controller. The communication controller sends the message to the ground receiver and transmits the repair result to the actuator to guide subsequent repair operations. The HMI interface displays the satellite status, fault information, and repair progress through a human-machine interaction interface for operators to monitor and make decisions.

[0115] A satellite communication fault diagnosis and repair system for an aircraft based on deep learning, comprising a data acquisition module, a fault detection module, and a type diagnosis module:

[0116] The data acquisition module is used to obtain the communication satellite topology network, collect the observation parameters of the communication satellite topology network through a dual-mode heterogeneous receiving terminal, and perform data archiving to generate a constellation measurement data set;

[0117] The fault detection module is used to obtain abnormal node location information through the constellation behavior perception model according to the constellation measurement data set and the communication satellite topology network;

[0118] The type diagnosis module is used to obtain the inter-satellite communication abnormal pattern through the abnormal pattern reconstruction model according to the abnormal node location information; perform fault repair according to the inter-satellite communication abnormal pattern.

[0119] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the computer-readable storage medium include: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0120] A computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0121] The program code contained on a computer-readable medium can be transmitted using any appropriate medium, including but not limited to wireless, wire, optical fiber cable, RF, etc., or any suitable combination of the above. The computer program code for performing the operations of the present invention can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., by using an Internet service provider to connect through the Internet).

[0122] As described above, the above are only the preferred embodiments of the present invention and do not impose any formal limitations on the present invention. Although the present invention has been disclosed above in preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or refinements to the above-disclosed technical content to obtain equivalent embodiments with equivalent changes within the scope of the technical solution of the present invention. However, any simple modification, equivalent change, and refinement made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A method for diagnosing and repairing satellite communication faults of aircraft based on deep learning, characterized in that: include: Acquire a communication satellite topology network, collect observation parameters of the communication satellite topology network through a dual-mode heterogeneous receiving terminal, and archive the data to generate a constellation measurement data set; Obtaining abnormal node positioning information through a constellation behavior perception model according to the constellation measurement data set and the communication satellite topology network; According to the abnormal node positioning information, an abnormal mode of inter-satellite communication is obtained through an abnormal mode reconstruction model; and fault repair is performed according to the abnormal mode of inter-satellite communication.

2. The method for diagnosing and repairing satellite communication faults of aircraft based on deep learning according to claim 1, characterized in that: The constellation behavior perception model includes: Obtaining an initial network topology, positioning parameters, and candidate abnormal nodes through preliminary fault detection according to the constellation measurement data set and the communication satellite topology network; The abnormal node location information is obtained through back-generation detection according to the initial network topology, the candidate abnormal nodes, and the location parameters.

3. The method for diagnosing and repairing aviation aircraft satellite communication faults based on deep learning according to claim 2, characterized in that: The preliminary fault detection specifically includes: Acquire measurement data and predicted pseudorange in the measurement data set, and obtain pseudorange deviation through residual calculation according to the measurement data and the predicted pseudorange; A normalized pseudorange deviation is obtained by normalizing the pseudorange deviation; Obtaining a maximum normalized pseudorange deviation by sorting the normalized pseudorange deviations, and marking a satellite corresponding to the maximum normalized pseudorange deviation as a candidate abnormal node; According to the candidate abnormal nodes, a maximum correlation coefficient is obtained by correlation calculation and sorting; a correlation threshold is preset, and if the maximum correlation coefficient is greater than the correlation threshold, the satellite corresponding to the maximum correlation is a candidate abnormal node; if the maximum correlation is not greater than the correlation threshold, the satellite corresponding to the maximum correlation is a normal node; The initial network topology is obtained by excluding the candidate abnormal nodes in the communication satellite topology network; and the positioning parameters are obtained according to the initial network topology.

4. The method for diagnosing and repairing satellite communication faults of aircraft based on deep learning according to claim 2, characterized in that: The back generation detection comprises: Obtaining a global detection statistic by global statistic calculation according to the candidate abnormal node and the positioning parameter; A global threshold is preset, and a final candidate node seat without abnormality is obtained through recovery judgment according to the global detection statistic and the global threshold; The recovery judgment includes judging whether the global detection statistic is less than the global threshold. If yes, the candidate abnormal node is restored to a normal node; if no, it is still a candidate abnormal node; The abnormal node location information is obtained by least square estimation according to the final abnormal node seat without candidate.

5. The method for diagnosing and repairing aviation aircraft satellite communication faults based on deep learning according to claim 1, characterized in that: The abnormal pattern reconstruction model includes: Acquire telemetry data of a faulty satellite, and obtain a latent space representation through a compression model according to the telemetry data of the faulty satellite; Probabilistic models; The intersatellite communication abnormality pattern is obtained through a scoring model according to the likelihood within the distribution.

6. The method for diagnosing and repairing aviation aircraft satellite communication faults based on deep learning according to claim 5, characterized in that: The compression model includes: Obtaining faulty satellite telemetry reduced dimension data by performing dimension reduction on the faulty satellite telemetry data; The latent space representation is obtained according to the dimensionality reduction data of the faulty satellite telemetry through a low-dimensional representation optimization model.

7. The method for diagnosing and repairing satellite communication faults of aircraft based on deep learning according to claim 6, characterized in that: The low-dimensional representation optimization model includes: According to the dimensionality reduction data of the faulty satellite telemetry, an initial low-dimensional representation of the sample is obtained by minimizing the reconstruction loss function; The reconstruction loss function is expressed as: Among them, L rec represents the reconstruction loss function, N represents the number of faulty satellite telemetry dimensionality reduction data, i represents the i-th faulty satellite telemetry dimensionality reduction data, x i represents the dimensionality reduction of the faulty satellite telemetry, Attention(x′ i ,ai) represents the attention mechanism, x′ i represents the data after reconstruction loss function, a i represents the attention weight, λ is the regularization parameter, j represents the jth weight, M represents the number of weights, and w j represents weight; The latent space representation is obtained by calculating the contrast loss function according to the initial low-dimensional representation of the sample.

8. The method for diagnosing and repairing aviation aircraft satellite communication faults based on deep learning according to claim 6, characterized in that: The probability model includes: Obtaining component type probabilities through an artificial neural network according to the latent space representation; Obtaining complete mixed component parameters by maximizing the probability of the component types, wherein the complete mixed component parameters include weight parameters, mean vector parameters, and covariance matrix parameters; Obtaining the in-distribution likelihood by probability calculation based on the complete mixture component parameters and the latent space representation; The probability calculation expression is: Where P(x) is the likelihood within the distribution, c represents the cth complete mixture component parameter, C represents the number of complete mixture component parameters, is the weight, |∑ c | represents the determinant of the covariance matrix of the cth complete mixture component, tr represents the trace operation, x represents the latent space representation, μ c is the mean vector, and T represents the transposed calculation.

9. The method for diagnosing and repairing aviation aircraft satellite communication faults based on deep learning according to claim 5, characterized in that: Scoring models include: Obtaining an out-of-distribution likelihood value by scoring according to the in-distribution likelihood; The score calculation expression is: Among them, Score(x out ) is the out-of-distribution likelihood value, c represents the cth complete mixture component parameter, C represents the number of complete mixture component parameters, is the weight, |∑ c | represents the determinant of the covariance matrix of the cth complete mixture component, x in represents the latent space representation, μ c is the mean vector, T represents the transposition calculation, and the out-of-distribution likelihood value is obtained by calculating the negative logarithm of the in-distribution likelihood; Preset a scoring threshold, and obtain the intersatellite communication abnormality mode through scoring judgment according to the out-of-distribution likelihood value and the scoring threshold; The scoring judgment includes judging whether the out-of-distribution likelihood value is greater than the scoring threshold. If yes, the intersatellite communication abnormality pattern is of unknown type; if no, the intersatellite communication abnormality pattern is of known type.

10. A deep learning-based aviation aircraft satellite communication fault diagnosis and repair system, characterized in that: Including data acquisition module, fault detection module, type diagnosis module: The data acquisition module is used to acquire a communication satellite topology network, collect observation parameters of the communication satellite topology network through a dual-mode heterogeneous receiving terminal, and perform data archiving to generate a constellation measurement data set; The fault detection module is used to obtain abnormal node positioning information through a constellation behavior perception model according to the constellation measurement data set and the communication satellite topology network; The type diagnosis module is used to obtain an abnormal inter-satellite communication mode through an abnormal mode reconstruction model according to the abnormal node positioning information; and perform fault repair according to the abnormal inter-satellite communication mode.

Images