An AI-based aerospace electromagnetic environment big data processing system

Abstract

The present invention relates to the technical field of electromagnetic spectrum management, and discloses an AI-based aerospace electromagnetic environment big data processing system, comprising an electromagnetic information collection module, which is used to collect electromagnetic information; a No. 1 ordered information construction module, which constructs No. 1 ordered information of two structural dimensions based on the electromagnetic information; a link quality determination module, which inputs the No. 1 ordered information into a quality model, and the quality model outputs the results of each link quality; and a spectrum allocation module, which inputs the No. 1 ordered information into an allocation model, and the allocation model outputs the results of spectrum allocation of each communication node. The present invention realizes intelligent dynamic allocation of spectrum resources in an aerospace-ground communication network by analyzing electromagnetic information, and further considers atmospheric environment information. Compared with only allocating spectrum resources, the system provides a spectrum allocation solution with high allocation quality, low cost and high utilization rate while considering the influence of the atmosphere.

Description

Technical Field

[0001] The present invention relates to the field of electromagnetic spectrum management technology, and more specifically, to an AI-based aerospace electromagnetic environment big data processing system. Background Art

[0002] Low-orbit satellite constellations, ground communication stations, and aerial platforms constitute a complex three-dimensional communication system. However, the increasing complexity of the aerospace electromagnetic environment and the increasing scarcity of electromagnetic spectrum resources have made the intelligent and dynamic allocation of electromagnetic spectrum resources a key and difficult issue in current research. By constructing a hierarchical decision-making model and reward mechanism, efficient allocation and dynamic adjustment of spectrum resources are achieved. However, atmospheric refraction and scattering cause drastic fluctuations in link quality, affecting the accuracy of spectrum resource allocation decisions, resulting in frequent adjustments to allocation plans and degraded system performance. Summary of the Invention

[0003] The purpose of the present invention is to provide an AI-based aerospace electromagnetic environment big data processing system in order to solve the above problems.

[0004] The present invention provides an AI-based aerospace electromagnetic environment big data processing system, comprising:

[0005] An electromagnetic information collection module is used to collect electromagnetic information, including communication node information and atmospheric environment information; communication nodes include satellites, ground stations, and aerial platforms;

[0006] The No. 1 ordered information construction module constructs No. 1 ordered information in two structural dimensions based on electromagnetic information. In the first structural dimension, objects are used as basic structural elements, including satellites, ground stations, air platforms, and atmospheric environments; in the second structural dimension, No. 1 time sequence units are used as basic structural elements.

[0007] A link quality determination module inputs the first order information into a quality model, and the quality model outputs the results of the quality of each link;

[0008] The spectrum allocation module inputs the first sequential information and the second timing unit into the allocation model, and the allocation model outputs the spectrum allocation result of each communication node.

[0009] Furthermore, the method for obtaining the first order information includes:

[0010] Extract the object of number one order information;

[0011] Extract specific features associated with the object and divide the extracted specific features into multiple parts according to the data collection cycle. Each part is mapped to a first time sequence unit. The order of the first time sequence unit is consistent with the time order of the collection cycle of the specific features.

[0012] Establish a mapping between the extracted objects and their associated specific features;

[0013] For any two objects, if there is a correlation between the associated specific features related to the electromagnetic spectrum management task or there is a correlation between the objects related to the electromagnetic spectrum management task, an object relationship is established between the two objects.

[0014] Furthermore, specific characteristics associated with the satellite: satellite location, current allocated frequency, allocated bandwidth, and transmit power;

[0015] Specific characteristics associated with the ground site: location of the ground site, current allocated frequency, allocated bandwidth, and transmit power;

[0016] Specific characteristics associated with the airborne platform: the location of the airborne platform, the current allocated frequency, the allocated bandwidth, and the transmit power;

[0017] Specific characteristics associated with the atmospheric environment: atmospheric extent, refractive index, rainfall rate, atmospheric pressure, temperature distribution, and water vapor pressure.

[0018] Furthermore, the relevance of risk detection tasks includes:

[0019] There is frequency overlap between the two communication nodes and they are within their respective communication coverage areas;

[0020] If there is a correlation between two communication nodes and their signal propagation paths pass through the corresponding atmospheric environments, then the two communication nodes are respectively associated with the atmospheric environments;

[0021] The signal propagation path is the spatial trajectory of electromagnetic wave propagation between two communication nodes.

[0022] Furthermore, the quality model includes a No. 1 backbone network, which includes a first timing layer, a second graph network layer, and a first output layer, wherein the first timing layer inputs No. 1 sequential information and outputs a first hidden feature to the second graph network layer, the second graph network layer outputs a second hidden feature to the first output layer, and the first output layer outputs the results of the quality of each link.

[0023] Furthermore, a link indicates that there is an association between two communicating nodes.

[0024] Furthermore, the allocation model includes a No. 2 backbone network, which includes a third timing layer, a fourth timing layer, a fifth graph network layer, a feature fusion layer and a second output layer, wherein the third timing layer inputs the No. 1 sequential information and outputs the third hidden feature to the feature fusion layer, wherein the fourth timing layer inputs the No. 2 timing unit and outputs the fourth hidden feature to the feature fusion layer, the feature fusion layer outputs the fused feature to the fifth graph network layer, the fifth graph network layer outputs the fifth hidden feature to the second output layer, and the second output layer outputs the result of the spectrum allocation of each communication node.

[0025] Furthermore, based on the link quality data output by the quality model, second structured data including a structural dimension is constructed. The basic structural element included in this structural dimension is a second timing unit. The structuring method includes:

[0026] Extract the associated information of the link quality determination task, divide the extracted specific features into multiple parts according to the data collection cycle, map one part to a No. 2 timing unit, and the order of the No. 2 timing units is consistent with the time sequence of the collection cycle of the associated information of the link quality determination task.

[0027] Furthermore, the link quality determination module is trained through supervised learning, and the loss function calculation method of the link quality determination module is:

[0028] ;

[0029] in, Represents the loss value of the link quality judgment module, represents the total number of link samples, Indicates the The actual quality value of the link samples, Indicates the output of the link quality determination module. The quality value of each link sample;

[0030] The spectrum allocation module is trained through Q-learning, with instant rewards The calculation method is:

[0031] ;

[0032] in, represents the total number of assigned samples, and Indicates the Second and The output of the secondary link quality determination module is The quality value of the assigned samples, Indicates execution The switching cost when Indicates execution The spectrum utilization rate when 、 、 are the first, second, and third hyperparameters;

[0033] Spectrum utilization The calculation method is:

[0034] ;

[0035] in, Indicates that in the strategy The number of communication nodes allocated under Indicates the The bandwidth allocated to each communication node, Indicates the The transmission power allocated to each communication node is Indicates the total available bandwidth, Indicates the maximum available transmit power;

[0036] Switching cost The calculation method is:

[0037] ;

[0038] in, Indicates that in the strategy Next The communication nodes are assigned frequencies, Indicates that at the last moment Time The communication nodes are assigned frequencies, Indicates the total available frequency range.

[0039] The present invention provides a computer storage medium for storing computer-readable instructions, which, when read, can execute the aforementioned AI-based aerospace electromagnetic environment big data processing system.

[0040] The beneficial effects of the present invention are: by analyzing electromagnetic information, the present invention realizes the intelligent dynamic allocation of spectrum resources in the air-space-ground communication network, and further considers atmospheric environmental information. Compared with only allocating spectrum resources, the system provides a spectrum allocation solution with high allocation quality, low cost and high utilization rate while taking into account the influence of the atmosphere. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0042] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. In addition, features described with respect to some examples may also be combined in other examples.

[0043] At least one embodiment of the present invention discloses an AI-based aerospace electromagnetic environment big data processing system, such as Figure 1 Shown, including:

[0044] An electromagnetic information collection module 101 is used to collect electromagnetic information, including communication node information and atmospheric environment information;

[0045] In one embodiment of the present invention, the communication nodes include satellites, ground stations, and aerial platforms;

[0046] The first order information construction module 102 constructs the first order information in two structural dimensions based on the electromagnetic information. In the first structural dimension, the basic structural elements are objects, including satellites, ground stations, aerial platforms, and atmospheric environments. In the second structural dimension, the basic structural elements are the first time sequence units.

[0047] In some embodiments of the present invention, the method for obtaining the first order information includes:

[0048] Extract the object of number one order information;

[0049] Extract specific features associated with the object and divide the extracted specific features into multiple parts according to the data collection cycle. Each part is mapped to a first time sequence unit. The order of the first time sequence unit is consistent with the time order of the collection cycle of the specific features.

[0050] Establish a mapping between the extracted objects and their associated specific features;

[0051] For any two objects, if there is a correlation between the associated specific features related to the electromagnetic spectrum management task or there is a correlation between the objects related to the electromagnetic spectrum management task, an object relationship is established between the two objects.

[0052] In some embodiments of the present invention, the specific characteristics associated with the satellite include: the satellite's location, current allocated frequency, allocated bandwidth, and transmit power;

[0053] Specific characteristics associated with the ground site: location of the ground site, current allocated frequency, allocated bandwidth, and transmit power;

[0054] Specific characteristics associated with the airborne platform: the location of the airborne platform, the current allocated frequency, the allocated bandwidth, and the transmit power;

[0055] Specific characteristics associated with the atmospheric environment: atmospheric extent, refractive index, rainfall rate, atmospheric pressure, temperature distribution, water vapor pressure;

[0056] In some embodiments of the present invention, the relevance related to the risk detection task includes:

[0057] There is frequency overlap between the two communication nodes and they are within their respective communication coverage areas;

[0058] If there is a correlation between two communication nodes and their signal propagation paths pass through the corresponding atmospheric environments, then the two communication nodes are respectively associated with the atmospheric environments;

[0059] In one embodiment of the present invention, the signal propagation path is a spatial trajectory of electromagnetic wave propagation between two communication nodes.

[0060] In one embodiment of the present invention, the refractive index is calculated using a refractive index model:

[0061]

[0062] in, In spatial position and time The refractive index at 、 represents the spatial coordinates on the horizontal plane, is the height, It's time, is the rainfall rate (mm / h), is the rainfall influence coefficient, is the height attenuation coefficient, is the surface air pressure, is the pressure elevation, is the ground temperature, is the temperature lapse rate, is the surface water vapor pressure, is the water vapor pressure height, The middle shows the variation of rainfall rate with altitude.

[0063] The link quality determination module 103 inputs the first orderliness information into a quality model. The quality model includes a first backbone network. The first backbone network includes a first timing layer, a second graph network layer, and a first output layer. The first timing layer inputs the first orderliness information and outputs a first hidden feature to the second graph network layer. The second graph network layer outputs a second hidden feature to the first output layer. The first output layer outputs the quality results of each link.

[0064] In one embodiment of the present invention, a link indicates that there is an association between two communication nodes;

[0065] The spectrum allocation module 104 inputs the first order information and the second timing unit into an allocation model. The allocation model includes a second backbone network. The second backbone network includes a third timing layer, a fourth timing layer, a fifth graph network layer, a feature fusion layer, and a second output layer. The third timing layer inputs the first order information and outputs the third hidden feature to the feature fusion layer. The fourth timing layer inputs the second timing unit and outputs the fourth hidden feature to the feature fusion layer. The feature fusion layer outputs the fused feature to the fifth graph network layer. The fifth graph network layer outputs the fifth hidden feature to the second output layer. The second output layer outputs the spectrum allocation result for each communication node.

[0066] In one embodiment of the present invention, second structured data including a structural dimension is constructed based on link quality data output by a quality model. The basic structural element included in this structural dimension is a second timing unit. The structuring method includes:

[0067] Extract the associated information of the link quality determination task, divide the extracted specific features into multiple parts according to the data collection cycle, map one part to a No. 2 timing unit, and the order of the No. 2 timing units is consistent with the time sequence of the collection cycle of the associated information of the link quality determination task.

[0068] In one embodiment of the present invention, the first timing layer and the third timing layer both adopt a timing neural network, such as RNN, LSTM or GRU.

[0069] An expression of the first, third, and fourth timing layers is as follows:

[0070]

[0071]

[0072]

[0073]

[0074] in:

[0075] : No. The reset gate vector of time steps is used to control the degree of reset of historical information;

[0076] : No. The update gate vector of time steps is used to control the update degree of new information;

[0077] : No. Candidate hidden state vectors for time steps;

[0078] : No. The hidden state vector of each time step is the first, third, or fourth hidden feature (the first temporal layer outputs the first hidden feature, the third temporal layer outputs the third hidden feature, and the fourth temporal layer outputs the fourth hidden feature);

[0079] : No. The hidden state vector of time steps

[0080] : No. The input vector of time steps corresponds to the first or second sequential unit (the first and third sequential layers are the first sequential unit, and the fourth sequential layer is the second sequential unit);

[0081] 、 、 : Reset the gate, update the weight matrix of the gate and candidate state, which is a trainable parameter;

[0082] 、 、 : Reset the gate, update the input weight matrix of the gate and candidate state, which is a trainable parameter;

[0083] 、 、 : Reset gate, update gate and bias vector of candidate state, which are trainable parameters;

[0084] : Hadamard product of vectors (element-wise multiplication);

[0085] : Sigmoid activation function;

[0086] : Hyperbolic tangent activation function;

[0087] : Time step index, the value range is , is the sequence length, when hour, As the initial state

[0088] The network layer in the second figure and the network layer in the fifth figure both adopt a multi-layer structure. The calculation formula of the layer is as follows:

[0089]

[0090] in:

[0091] Indicates the Objects in Layer The hidden state vector of

[0092] Indicates the Objects in Layer The hidden state vector of

[0093] Representation and Objects The set of directly connected neighbor nodes;

[0094] Representation and Objects The set of directly connected neighbor nodes;

[0095] Representation and Objects A collection of objects with object connections;

[0096] Representation and Objects A collection of objects with object connections;

[0097] Indicates the The learnable weight matrix of the layer;

[0098] Indicates the network layer index, the value range is ,in is the total number of layers, when hour, , where, for the second graph network layer, , Indicates the When the specific features of all the first sequential units associated with the object are input into the first sequential layer, the last first hidden feature is output. is the total number of sequential units No. 1, for the network layer in Figure 5, , representing an object The fusion features output by the corresponding feature fusion layer are hour, This is the final output node feature (the second graph network layer outputs the second hidden feature, and the fifth graph network layer outputs the fifth hidden feature, but the trainable parameters of the second and fifth graph network layers are different);

[0099] The expression of the feature fusion layer is as follows:

[0100]

[0101] in, Indicates the The fusion features associated with each object, represents the feature combination function (splicing function or summation function), Indicates the When the specific features of all the first sequential units associated with the object are input into the first sequential layer, the last first hidden feature is output. is the total number of sequential units No. 1, It means that the second sequential unit outputs the last fourth hidden feature when inputting the fourth sequential layer. is the total number of sequential units No. 2.

[0102] The expression of the first output layer is as follows:

[0103]

[0104] in Represents a link quality vector, the cth component value of the link quality vector represents the quality value of the cth link, Indicates the Layer The second hidden feature of the object, Represents the set of all objects, and FC represents fully connected.

[0105] The expression of the second output layer is as follows:

[0106]

[0107] in Represents an allocation vector. The dth component value of the allocation vector represents the probability value of the dth strategy. The strategy with the largest probability value is selected as the output. The strategy group contains all executable strategies. A strategy represents the spectrum allocation scheme of a communication node, including frequency allocation, bandwidth allocation and power allocation. Indicates the Layer The fifth hidden feature of the object, Represents the collection of all objects, Indicates the total number of layers, and FC means fully connected.

[0108] In one embodiment of the present invention, the link quality determination module is trained through supervised learning, and the loss function calculation method of the link quality determination module is:

[0109]

[0110] in, Represents the loss value of the link quality judgment module, represents the total number of link samples, Indicates the The actual quality value of the link samples, Indicates the output of the link quality determination module. The quality value of each link sample.

[0111] In one embodiment of the present invention, the spectrum allocation module is trained by Q-learning, where the immediate reward The calculation method is as follows:

[0112]

[0113] in, represents the total number of assigned samples, and Indicates the Second and The output of the secondary link quality determination module is The quality value of the assigned samples, Indicates execution The switching cost when Indicates execution The spectrum utilization rate when 、 、 are the first, second, and third hyperparameters;

[0114] In one embodiment of the present invention, the spectrum utilization rate The calculation method is:

[0115]

[0116] in, Indicates that in the strategy The number of communication nodes allocated under Indicates the The bandwidth allocated to each communication node, Indicates the The transmission power allocated to each communication node is Indicates the total available bandwidth, Indicates the maximum available transmit power;

[0117] Switching cost The calculation method is:

[0118]

[0119] in, Indicates that in the strategy Next The communication nodes are assigned frequencies, Indicates that at the last moment Time The communication nodes are assigned frequencies, Indicates the total available frequency range.

[0120] In at least one embodiment of the present invention, a computer storage medium is provided for storing computer-readable instructions, which, when read, can execute the aforementioned AI-based aerospace electromagnetic environment big data processing system.

[0121] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. An AI-based aerospace electromagnetic environment big data processing system, characterized by: include: An electromagnetic information collection module is used to collect electromagnetic information, including communication node information and atmospheric environment information; Communications nodes include satellites, ground stations, and aerial platforms; The No. 1 ordered information construction module constructs No. 1 ordered information in two structural dimensions based on electromagnetic information. In the first structural dimension, objects are used as basic structural elements, including satellites, ground stations, air platforms, and atmospheric environments; in the second structural dimension, No. 1 time sequence units are used as basic structural elements. A link quality determination module inputs the first order information into a quality model, and the quality model outputs the results of the quality of each link; The second sequential unit is constructed based on the output of the quality model. The construction method includes: Extracting relevant information for the link quality determination task, dividing the extracted specific features into multiple parts according to the data collection period, mapping each part to a second sequential unit, and the order of the second sequential units is consistent with the time order of the collection period of the relevant information for the link quality determination task; The spectrum allocation module inputs the first sequential information and the second timing unit into the allocation model, and the allocation model outputs the spectrum allocation result of each communication node.

2. The AI-based aerospace electromagnetic environment big data processing system according to claim 1 is characterized in that: The method for obtaining the number one sequential information includes: Extract the object of number one order information; Extract specific features associated with the object and divide the extracted specific features into multiple parts according to the data collection cycle. Each part is mapped to a first time sequence unit. The order of the first time sequence unit is consistent with the time order of the collection cycle of the specific features. Establish a mapping between the extracted objects and their associated specific features; For any two objects, if there is a correlation between the associated specific features related to the electromagnetic spectrum management task or there is a correlation between the objects related to the electromagnetic spectrum management task, an object relationship is established between the two objects.

3. The AI-based aerospace electromagnetic environment big data processing system according to claim 1, characterized in that: Specific characteristics associated with the satellite: satellite location, current allocated frequency, allocated bandwidth, and transmit power; Specific characteristics associated with the ground site: location of the ground site, current allocated frequency, allocated bandwidth, and transmit power; Specific characteristics associated with the airborne platform: the location of the airborne platform, the current allocated frequency, the allocated bandwidth, and the transmit power; Specific characteristics associated with the atmospheric environment: atmospheric extent, refractive index, rainfall rate, atmospheric pressure, temperature distribution, and water vapor pressure.

4. The AI-based aerospace electromagnetic environment big data processing system according to claim 1, characterized in that: Relevance to the risk detection task includes: There is frequency overlap between the two communication nodes and they are within their respective communication coverage areas; If there is a correlation between two communication nodes and their signal propagation paths pass through the corresponding atmospheric environments, then the two communication nodes are respectively associated with the atmospheric environments; The signal propagation path is the spatial trajectory of electromagnetic wave propagation between two communication nodes.

5. The AI-based aerospace electromagnetic environment big data processing system according to claim 1, characterized in that: The quality model includes a No. 1 backbone network, which includes a first timing layer, a second graph network layer, and a first output layer. The first timing layer inputs No. 1 sequential information and outputs a first hidden feature to the second graph network layer. The second graph network layer outputs a second hidden feature to the first output layer. The first output layer outputs the results of the quality of each link.

6. The AI-based aerospace electromagnetic environment big data processing system according to claim 5, characterized in that: A link indicates the existence of an association between two communicating nodes.

7. The AI-based aerospace electromagnetic environment big data processing system according to claim 1, characterized in that: The allocation model includes a No. 2 backbone network, which includes a third timing layer, a fourth timing layer, a fifth graph network layer, a feature fusion layer and a second output layer, wherein the third timing layer inputs the No. 1 sequential information and outputs the third hidden feature to the feature fusion layer, wherein the fourth timing layer inputs the No. 2 timing unit and outputs the fourth hidden feature to the feature fusion layer, the feature fusion layer outputs the fused feature to the fifth graph network layer, the fifth graph network layer outputs the fifth hidden feature to the second output layer, and the second output layer outputs the results of the spectrum allocation of each communication node.

8. The AI-based aerospace electromagnetic environment big data processing system according to claim 1, characterized in that: The link quality determination module is trained through supervised learning. The loss function calculation method of the link quality determination module is: ； in, Represents the loss value of the link quality judgment module, represents the total number of link samples, Indicates the The actual quality value of the link samples, Indicates the output of the link quality determination module. The quality value of each link sample; The spectrum allocation module is trained through Q-learning, with instant rewards The calculation method is: ； in, represents the total number of assigned samples, and Indicates the Second and The output of the secondary link quality determination module is The quality value of the assigned samples, Indicates execution The switching cost when Indicates execution The spectrum utilization rate when 、 、 are the first, second, and third hyperparameters; Spectrum utilization The calculation method is: ； in, Indicates that in the strategy The number of communication nodes allocated under Indicates the The bandwidth allocated to each communication node, Indicates the The transmission power allocated to each communication node is Indicates the total available bandwidth, Indicates the maximum available transmit power; Switching cost The calculation method is: ； in, Indicates that in the strategy Next The communication nodes are assigned frequencies, Indicates that at the last moment Time The communication nodes are assigned frequencies, Indicates the total available frequency range.

9. A computer storage medium, characterized in that It is used to store computer-readable instructions, which, when read, can execute an AI-based aerospace electromagnetic environment big data processing system as described in any one of claims 1-8.

Images