Satellite-borne four-channel Ka-band data transmission channel system

Abstract

The invention discloses a satellite-borne four-channel Ka frequency band data transmission channel system, which relates to the technical field of satellite communication, and comprises the following steps of: acquiring original frequency spectrum data of frequency points in a Ka frequency band, synchronously receiving an interference hot area database, constructing an original frequency spectrum matrix, marking an interference area, and generating a frequency spectrum map with an interference mark; inputting the spectrum map with the interference mark into a pre-trained convolutional neural network, outputting the interference probability of each frequency band, identifying a safe frequency band, segmenting the safe frequency band into micro sub-bands, optimizing a distribution scheme through a Q-learning algorithm, and generating an optimized spectrum configuration table; and dynamically adjusting a modulation mode according to bandwidth parameters in the optimized spectrum configuration table, partitioning the high-speed AOS data according to frames, and generating a network coding check packet in combination with a channel allocation strategy in the optimized spectrum configuration table. According to the invention, the spectrum utilization rate, the communication reliability and the overall performance of the satellite-borne Ka-band data transmission system in a dynamic environment are improved.

Description

Technical Field

[0001] The present invention relates to the field of satellite communication technology, in particular to a satellite-borne four-channel Ka-band data transmission channel system. Background Art

[0002] With the continuous advancement of satellite communication technology, especially the increasing application of the Ka-band (26.5–40 GHz), intersatellite link communication systems have been developed. Currently, Ka-band-based data transmission channel technology mainly relies on key technologies such as multi-channel parallel transmission, adaptive modulation and coding (AMC), and dynamic resource scheduling. In order to improve data transmission rate and reliability, researchers have developed a variety of spectrum sensing methods to identify available spectrum resources and adopt advanced modulation and coding schemes (such as QPSK, 16APSK, etc.) to adapt to different channel conditions. In addition, network coding technology has also been introduced into space communication protocols to enhance the robustness and efficiency of data transmission. In particular, when dealing with interference problems in complex electromagnetic environments, traditional practices usually rely on pre-defined spectrum allocation strategies or simple energy detection algorithms to avoid known interference sources.

[0003] Although research has attempted to improve spectrum management and resource scheduling through machine learning and artificial intelligence technologies, most solutions have yet to fully implement end-to-end closed-loop control, from spectrum sensing, modeling, decision-making, to transmission. For example, while convolutional neural networks are used to predict spectrum states and identify interference areas, in practice, the effectiveness of such intelligent algorithms is significantly limited due to computational resource and real-time constraints, particularly in satellite-based environments. Furthermore, existing inter-satellite Ka-link communication systems generally use fixed or semi-fixed modulation schemes, failing to dynamically adjust based on channel quality. This not only affects system throughput but also reduces its anti-interference capabilities. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a satellite-borne four-channel Ka-band data transmission channel system that solves the problems of low spectrum utilization efficiency and weak anti-interference capability of existing satellite-borne communication systems.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] The present invention provides a satellite-borne four-channel Ka-band data transmission channel system, which includes a spectrum sensing module that collects original spectrum data of frequency points in the Ka-band, synchronously receives an interference hotspot database, constructs an original spectrum matrix, marks interference areas, and generates a spectrum map with interference marks;

[0008] The spectrum optimization module inputs the interference-marked spectrum map into a pre-trained convolutional neural network, outputs the interference probability, identifies safe frequency bands, divides the safe frequency bands into sub-bands, optimizes the allocation plan through the Q-learning algorithm, and generates an optimized spectrum configuration table.

[0009] The dynamic modulation module dynamically adjusts the modulation mode according to the bandwidth parameters in the optimized spectrum configuration table, divides the high-speed AOS data into blocks by frame, and generates a network coding check packet based on the channel allocation strategy in the optimized spectrum configuration table;

[0010] The link transmission module broadcasts the network coding check packet to the relay satellite through the inter-satellite Ka link. The relay satellite reconstructs the data according to the network coding rules, decodes the reconstructed data and verifies the CRC.

[0011] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, the original spectrum data includes signal amplitude and signal phase.

[0012] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, the steps of constructing the original spectrum matrix, marking the interference area, and generating a spectrum map with interference marks are as follows:

[0013] Construct the original spectrum matrix according to the signal amplitude and signal phase;

[0014] Based on the interference hotspot database, interference marking is performed in the original spectrum matrix through double judgment to generate an interference marking vector;

[0015] The original spectrum matrix is ​​fused with the interference marker vector to generate a spectrum map with interference markers.

[0016] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, the spectrum map with interference marks is input into a pre-trained convolutional neural network, the interference probability of each frequency band is output, and the safe frequency band is identified. The specific steps are as follows:

[0017] Extracting spectrum feature data from the spectrum map with interference markers, normalizing the spectrum feature data, and generating a standardized feature matrix;

[0018] Collect historical interference-marked spectrum maps for convolutional neural network training;

[0019] Input the normalized feature matrix into the trained convolutional neural network to generate interference probability;

[0020] A safety threshold is set based on the historical interference probability. When the interference probability is less than or equal to the safety threshold, the frequency band is determined to be safe.

[0021] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, wherein: the safe frequency band is divided into neutrino bands, the specific steps are as follows:

[0022] Collect the highest modulation order, calculate the minimum bandwidth, and obtain the neutrino band width;

[0023] The safety band is divided into sub-bands according to the total width of the safety band and the width of the sub-band.

[0024] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, wherein: the allocation scheme is optimized by the Q-learning algorithm to generate an optimized spectrum configuration table, the specific steps are as follows:

[0025] Measure the signal-to-noise ratio, interference power, and channel coherence bandwidth of the neutrino band center frequency point and generate a neutrino band state parameter table;

[0026] Based on the neutrino band state parameter table, an initial Q-learning table is constructed, and the strategy optimization iteration is performed through the Q-learning algorithm to generate an optimized spectrum configuration table.

[0027] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, the modulation mode is dynamically adjusted according to the bandwidth parameters in the optimized spectrum configuration table. The specific steps are as follows:

[0028] Extract the bandwidth parameters of the neutrino band from the optimized spectrum configuration table and construct a mapping rule between the bandwidth parameters and the modulation order;

[0029] Modulation modes are allocated to the micro sub-bands according to the mapping rules, and a modulation parameter configuration table is generated.

[0030] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, wherein: the high-speed AOS data is divided into blocks by frame, and the network coding verification packet is generated in combination with the channel allocation strategy in the optimized spectrum configuration table. The specific steps are as follows:

[0031] Collect high-speed AOS data streams and divide them into codable units to generate raw data block sequences;

[0032] The channel priorities are extracted from the optimized spectrum configuration table, the original data blocks in the original data block sequence are allocated according to the channel priorities, and a network coding check packet is generated.

[0033] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, wherein: the network coding verification packet is broadcast to the relay satellite via the inter-satellite Ka link, and the specific steps are as follows:

[0034] Query neighboring satellites, obtain link margins, and select relay satellites based on the link margins;

[0035] According to the optimized spectrum configuration table, combined with the modulation parameter configuration table, configure the broadcast frequency band and obtain the transmission frequency;

[0036] Based on the transmission frequency, the network coding verification packet is sent to the relay satellite using the inter-satellite Ka link.

[0037] As a preferred solution of the satellite-borne four-channel Ka-band data transmission channel system of the present invention, wherein: the relay satellite reconstructs data according to the network coding rule, decodes the reconstructed data and verifies the CRC, the specific steps are as follows:

[0038] The relay satellite receives the network coding check packet, reconstructs the data block sequence through Viterbi soft decision demodulation, and recovers the lost data blocks to generate a complete data block group;

[0039] Based on the complete data block group, the data block CRC-16 value is calculated and the check status code is generated.

[0040] The present invention achieves the following beneficial effects: by collecting the amplitude and phase of Ka-band signals to construct an interference-marked spectrum map, and using a convolutional neural network to intelligently identify spectrum interference states, it accurately identifies safe frequency bands in complex electromagnetic environments. Furthermore, by combining a Q-learning algorithm to dynamically optimize the allocation of neutrino-band resources, it improves the resource scheduling capability and transmission efficiency of intersatellite communication systems under multi-interference, time-varying channel conditions. These two core steps respectively address the existing problems of spectral response lag and resource allocation rigidity, thereby improving the spectrum utilization, communication reliability, and overall performance of satellite-borne Ka-band data transmission systems in dynamic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 Schematic diagram of the satellite's four-channel Ka-band data transmission channel system.

[0043] Figure 2 Flowchart for generating the neutrino belt state parameter table.

[0044] Figure 3 Flowchart for constructing an optimized spectrum configuration table.

[0045] Figure 4 Flowchart for generating network coding verification packet. DETAILED DESCRIPTION

[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0047] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0048] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0049] Reference Figures 1 to 4 , is an embodiment of the present invention, which provides a satellite-borne four-channel Ka-band data transmission channel system, including the following steps:

[0050] The spectrum sensing module collects the original spectrum data of each frequency point in the Ka band, simultaneously receives the interference hotspot database, constructs the original spectrum matrix, marks the interference area, and generates a spectrum map with interference marks;

[0051] Collect the original spectrum data of each frequency point in the Ka band, including signal amplitude and signal phase, and simultaneously receive the interference hotspot database;

[0052] Furthermore, the starting frequency of the onboard receiver is set to 26.5 GHz, the ending frequency is set to 40 GHz, and the scanning step is set to 50 MHz to generate a scanning frequency list; each frequency point is traversed according to the scanning frequency list, the radio frequency signal is down-converted to a 1.5 GHz intermediate frequency through a superheterodyne mixer, and the signal amplitude and phase are measured through an I / Q demodulator; the data file uploaded by the ground station is received through the S-band telemetry link, and the interference area definition fields in the data file are parsed, including the starting frequency, ending frequency, and interference level, to generate an interference hot zone database.

[0053] Construct the original spectrum matrix, mark the interference area, and generate a spectrum map with interference marks;

[0054] The original spectrum matrix expression is:

[0055]

[0056] Among them, S is the original spectrum matrix, f i is the frequency, A iis the signal amplitude, φ i is the signal phase, t i is the timestamp, i is the row dimension index;

[0057] The row dimension expression of the original spectrum matrix is:

[0058]

[0059] Among them, n is the row dimension of the original spectrum matrix, f s is the starting frequency, f e is the end frequency, Δf is the scanning step;

[0060] Based on the interference hotspot database, interference marking is performed in the original spectrum matrix through double judgment to generate an interference marking vector;

[0061] For example, when the amplitude value is greater than -90 dBm, it is marked as interference; when the frequency falls into any area of ​​the interference hot zone database, it is marked as interference and an interference mark vector is generated.

[0062] The original spectrum matrix is ​​fused with the interference marker vector to generate a spectrum map with interference markers;

[0063] Furthermore, the interference mark vector is added as the fifth column to the original spectrum matrix, and the rows with the fifth column value of 1 are defined to correspond to the interference area, thereby generating a spectrum map with interference marks.

[0064] The spectrum optimization module inputs the interference-marked spectrum map into a pre-trained convolutional neural network, outputs the interference probability of each frequency band, identifies safe frequency bands, divides them into sub-bands, optimizes the allocation plan using the Q-learning algorithm, and generates an optimized spectrum configuration table.

[0065] Extracting spectrum feature data from the spectrum map with interference markers, normalizing the spectrum feature data, and generating a standardized feature matrix;

[0066] Furthermore, the first four columns of the spectrum map, including frequency, amplitude, phase, and timestamp, are extracted as spectrum feature data, and the interference mark in the fifth column is used as a verification label. The spectrum feature data is standardized to generate a standardized feature matrix.

[0067] Collect historical interference-marked spectrum maps for convolutional neural network training;

[0068] Furthermore, the first four columns of the historical interference-marked spectrum map are used as spectrum feature data, and the interference mark in the fifth column is used as a verification label; the spectrum feature data is standardized to make the data obey the normal distribution; a convolutional neural network architecture with three convolutional layers and two fully connected layers is constructed, and each convolutional layer uses a 3×3 convolution kernel size and a ReLU activation function; the convolutional neural network weight parameters are updated using the backpropagation algorithm, and the binary cross entropy loss function is minimized by gradient descent; the early stopping strategy is used during training to prevent overfitting, and the training is terminated when the accuracy of the verification set no longer improves after three consecutive training rounds; finally, the convolutional neural network weight parameters are obtained that can output the interference probability of each frequency point based on the input standardized feature matrix.

[0069] Input the normalized feature matrix into the trained convolutional neural network to generate interference probability;

[0070] Furthermore, the expression of the three-layer convolution operation is:

[0071] Y (1) =ReLU(W (1) *X+b (1) );

[0072] Y (2) =ReLU(W (2) *Y (1) +b (2) );

[0073] Y (3) =ReLU(W (3) *Y (2) +b (3) );

[0074] Among them, Y (1) Y is the output feature map of the first convolution layer, (2) Y is the output feature map of the first convolution layer, (3) is the first layer convolution output feature map, W (1) is the first layer convolution kernel weight matrix, W (2) is the second layer convolution kernel weight matrix, W (3) is the third layer convolution kernel weight matrix, b (1) is the bias vector of the first layer, b (2) is the bias vector of the second layer, b (3) is the bias vector of the third layer, X is the normalized feature moment, and ReLU is the rectified linear unit activation function;

[0075] The output interference probability expression of the fully connected layer is:

[0076] P=σ(W (4) vec(Y (3) )+b (4) );

[0077] Among them, P is the interference probability, σ is the Sigmoid activation function, W (4) is the weight matrix of the fully connected layer, b (4) is the fully connected layer bias vector, and vec is the flattening operation.

[0078] A safety threshold is set based on the historical interference probability. When the interference probability is less than or equal to the safety threshold, the frequency band is determined to be safe.

[0079] Furthermore, a safety threshold is set according to the historical interference probability. For example, the safety threshold is 0.1. All frequency points are traversed. When the interference probability is less than or equal to the safety threshold, it is marked as safe and the value is 1. When the interference probability is greater than the safety threshold, it is marked as unsafe and the value is 0. Find all continuous index intervals that satisfy the frequency safety value equal to 1 and the next frequency safety value is also equal to 1 as the safe frequency band, and extract the physical frequency range to generate the safe frequency band.

[0080] Collect the highest modulation order, calculate the minimum bandwidth, and obtain the neutrino band width;

[0081] Furthermore, the highest modulation order supported by the onboard modem is read and the minimum bandwidth that meets the highest modulation order requirement is calculated. The expression is:

[0082]

[0083] Where Δf min is the minimum bandwidth, R is the system nominal bit rate, α is the roll-off coefficient, and M is the modulation order;

[0084] According to the minimum bandwidth that meets the highest-order modulation requirement, the sub-band width is set to an integer multiple of the minimum bandwidth. For example, the sub-band width is set to 2 times the minimum bandwidth.

[0085] Furthermore, the highest modulation order stored in the onboard modem is queried, and the system nominal code rate and roll-off factor are obtained at the same time. Based on the highest modulation order, the system nominal code rate and the roll-off factor, the minimum bandwidth calculation operation is performed, and an integer multiple greater than or equal to the minimum bandwidth is selected as the neutrino band width.

[0086] The safety band is divided into sub-bands according to the total width of the safety band and the width of the sub-band;

[0087] Furthermore, the total width of the safe frequency band is obtained, and the number of divisible sub-bands is calculated according to the width of the sub-band to generate continuous sub-bands.

[0088] Measure the signal-to-noise ratio, interference power, and channel coherence bandwidth of the neutrino band center frequency point and generate a neutrino band state parameter table;

[0089] Furthermore, the spectrum analyzer is first activated and configured in sweep mode. The spectrum analyzer's receive port is connected to the onboard RF front-end output interface, and the sweep range is set to cover all sub-bands generated by the safe band splitting operation. For each sub-band center frequency, the spectrum analyzer automatically performs three independent sweeps. The first sweep captures the signal power spectral density distribution and calculates the signal-to-noise ratio (SNR) by the difference between the power spectrum peak and the noise floor. The second sweep activates the interference detection algorithm and calculates the residual spectral energy integral as the interference power after eliminating the main signal power. The third sweep uses a wideband linear frequency modulation signal as excitation, and the channel coherence bandwidth is determined by the zero-point interval of the received signal autocorrelation function. The results of each sweep are recorded in real time in the measurement log buffer, forming a temporary data set containing a frequency index field. After traversing all sub-bands, the frequency index field, signal-to-noise ratio field, interference power field, and channel coherence bandwidth field are extracted from the temporary data set and sorted in ascending frequency order to generate a structured table. The structured table contains four columns of data entities. The first column is the physical coordinates of the center frequency point of the neutrino band, the second column is the signal-to-noise ratio, the third column is the interference power, and the fourth column is the channel coherence bandwidth. The output table is named the neutrino band state parameter table.

[0090] Based on the neutrino band state parameter table, an initial Q-learning table is constructed, and the strategy optimization iteration is performed through the Q-learning algorithm to generate an optimized spectrum configuration table;

[0091] Furthermore, during initialization, all data entries in the neutrinoband state parameter table are read, extracting four parameters: the neutrinoband center frequency coordinates, signal-to-noise ratio, interference power, and channel coherence bandwidth. Based on these four parameters, a three-dimensional matrix data structure is constructed, with row indices strictly corresponding to the neutrinoband number sequence, column indices corresponding to the state type classification, and depth indices corresponding to the sampling point sequence in the time dimension. Each matrix cell stores the quantized result of the neutrinoband state at a specific sampling moment, forming the underlying data framework of the initial Q-learning table.

[0092] The Q-learning algorithm strategy optimization iterative process is initiated. The state space is defined as the output value of the channel quality assessment function, which is generated by a weighted algorithm using the signal-to-noise ratio and interference power in the neutrino-band state parameter table. The action space is defined as the complete set of channel allocation strategies, including the four channel combination modes supported by the onboard four-channel Ka-band data transmission channel system. During each iteration of action selection, the standard ε-greedy strategy balances exploration and exploitation. A random number is generated and compared with a fixed coefficient ε. If the random number is less than or equal to ε, a channel allocation strategy is randomly selected from the action space. Otherwise, the channel allocation strategy action number corresponding to the maximum Q value in the current three-dimensional matrix is ​​selected. After executing the selected action, the transmission success rate indicator of the intersatellite link is monitored, and the Q value storage unit in the three-dimensional matrix is ​​updated using the Bellman equation.

[0093] After completing a fixed number of iterations, the optimization process is terminated. The storage cells associated with each neutrino band number in the three-dimensional matrix are scanned row by row to locate the position of the maximum Q value and parse the corresponding channel allocation strategy action number. The three elements of the neutrino band number, channel allocation strategy action number, and center frequency point physical coordinates are mapped into structured data entries and arranged in ascending order according to the center frequency point physical coordinates to generate an optimized spectrum configuration table.

[0094] The dynamic modulation module dynamically adjusts the modulation mode according to the bandwidth parameters in the optimized spectrum configuration table, divides the high-speed AOS data into blocks by frame, and generates a network coding check packet based on the channel allocation strategy in the optimized spectrum configuration table;

[0095] Extract the bandwidth parameters of the neutrino band from the optimized spectrum configuration table and construct a mapping rule between the bandwidth parameters and the modulation order;

[0096] The optimized spectrum configuration table is then traversed, reading the combinations of the sub-band number and bandwidth parameter fields recorded in the table row by row. The modulation configuration reference table, which contains three columns: the lower and upper bandwidth limits, and the corresponding modulation order, is then loaded from the onboard modem's non-volatile memory. For each sub-band bandwidth parameter, a full table scan of the modulation configuration reference table is performed. If the bandwidth parameter falls within the lower and upper bandwidth limits of a row in the modulation configuration reference table, the modulation order is extracted.

[0097] Allocate modulation modes to micro sub-bands according to mapping rules and generate a modulation parameter configuration table;

[0098] Specifically, an intermediate mapping data set is established, which contains three columns of entities. The first column is the neutrino band number of the optimized spectrum configuration table, the second column is the bandwidth parameter of the optimized spectrum configuration table, and the last column is the modulation order matched by the modulation configuration reference table. After traversing all neutrino bands, the intermediate mapping data set is arranged in ascending order according to the neutrino band number to form a modulation parameter configuration table.

[0099] Collect high-speed AOS data streams and divide them into codable units to generate raw data block sequences;

[0100] Furthermore, the 1553B bus interface of the onboard data management unit is activated to receive the high-speed AOS data stream bit sequence from the satellite platform payload in real time. The bus interface's buffer management strategy is configured to a cyclic overwrite mode, and the receive buffer threshold is set to twice the maximum AOS transmission frame length. When the buffer data volume reaches the receive buffer threshold, the protocol parsing engine is triggered to scan the high-speed AOS data bit sequence, identify the fixed-length synchronization code characteristics of the AOS transmission frame, and locate the starting position of the main frame header. The synchronization code field, frame count field, and frame length indicator field in the main frame header are extracted, and the physical boundary of the current AOS transmission frame is determined based on the frame length indicator field. The payload area after the main frame header is stripped, and the operation control field and frame error control field in the main frame header are retained as protocol metadata. The payload area is divided into data segment entities of fixed size, and the corresponding protocol metadata is attached to each data segment entity to form a high-speed AOS data stream. The modulation order field recorded in the modulation parameter configuration table is loaded, and the coding block size mapping table in the onboard memory is queried. The coding block size mapping table stores the correspondence between the modulation order and the maximum codable unit length. The target length of the codable unit is determined by matching the current modulation order with the coding block size mapping table. When the high-speed AOS data stream exceeds the target length, a secondary segmentation operation is performed to generate a sequence of sub-data units. When the high-speed AOS data stream is less than the target length, idle bytes are padded to the target length. Each codable unit is assigned a globally unique index number. The index number field, protocol metadata field, and data unit field are combined to form a three-column structured record. All records are sorted in ascending order by index number to generate the original data block sequence.

[0101] extracting channel priorities from the optimized spectrum configuration table, allocating original data blocks in the original data block sequence according to the channel priorities, and generating a network coding check packet;

[0102] Furthermore, the data storage structure of the spectrum configuration table is located and optimized, and the channel allocation strategy coding field of all row records is read. The channel allocation strategy coding field consists of four binary digits, each of which corresponds to the priority status of a physical transmission channel in the satellite's four-channel Ka-band data transmission channel system. According to the binary weight parsing rules, the highest weight bit represents the priority of physical transmission channel one, the second highest weight bit represents the priority of physical transmission channel two, and so on, the lowest weight bit represents the priority of physical transmission channel four. A higher priority indicates a higher priority, and a physical transmission channel priority sorting list is generated.

[0103] The raw data block allocation operation is initiated, loading all records of the raw data block sequence. A round-robin allocation mechanism is implemented, following the descending order of the physical transmission channel priority list. Starting with the highest-priority physical transmission channel, raw data blocks are selected from the raw data block sequence, one at a time, for each physical transmission channel. After allocation to the lowest-priority physical transmission channel, allocation continues by returning to the highest-priority physical transmission channel, forming a closed-loop polling process. After all raw data blocks are allocated, each physical transmission channel receives a set of raw data blocks proportional to its priority.

[0104] Execute the network coding verification packet generation operation. For each set of original data blocks assigned to the physical transmission channel, apply the Reed-Solomon coding algorithm to generate the verification data block. The original data blocks and verification data blocks are encapsulated in a fixed ratio into a network coding verification packet entity. The physical transmission channel number field and data block index field are appended, and the network coding verification packet is output.

[0105] The link transmission module broadcasts the network coding check packet to the relay satellite via the inter-satellite Ka link. The relay satellite reconstructs the data according to the network coding rules, decodes the reconstructed data and verifies the CRC.

[0106] Query neighboring satellites, obtain link margins, and select relay satellites based on the link margins;

[0107] Furthermore, the onboard routing protocol stack's neighbor discovery function is activated, periodically broadcasting link state request signaling. Link state notification messages from neighboring satellites are received, and the orbital parameter and device identification fields in the messages are parsed to generate a dynamic neighbor satellite list. For each node in the dynamic neighbor satellite list, a two-way ranging operation is initiated, transmitting a ranging request pulse to the target satellite and recording the precise transmission timestamp; receiving the ranging response pulse returned by the target satellite and recording the precise reception timestamp. The timestamp data is processed according to the propagation delay calculation formula stored in the onboard device, outputting the spatial propagation loss value for the current link. The received signal strength indicator is synchronously measured and combined with the spatial propagation loss value to calculate the link margin index (LMI). This generates a link quality assessment matrix, whose row index corresponds to the neighbor satellite node identifier, and whose column index contains the orbital altitude difference angle, link margin index, and carrier frequency offset value. Comparison and selection logic is executed, scanning all rows of the LQA matrix to locate the row with the largest LMI value. The corresponding neighbor satellite node identifier is then extracted as the relay satellite selection result.

[0108] According to the optimized spectrum configuration, the broadcast frequency band is configured in combination with the modulation parameter configuration table to obtain the transmission frequency;

[0109] Specifically, read the neutrino band number field, channel allocation strategy coding field, and center frequency physical coordinate field of each row record; synchronously load the corresponding row record of the modulation parameter configuration table, and extract the modulation order field associated with the neutrino band number field. According to the binary weight parsing result of the channel allocation strategy coding field, bind the neutrino band to the physical transmission channel. When the highest weight bit of the coding field is 1, it is bound to physical transmission channel one. When the second highest weight bit is 1, it is bound to physical transmission channel two. Subsequent bits are bound to physical transmission channel three and physical transmission channel four in turn, forming a mapping relationship set between physical transmission channels and neutrino bands. Access the frequency configuration template fixed in the non-volatile memory of the onboard transmitter. The frequency configuration template stores the correspondence between the modulation order field, the center frequency offset field, and the bandwidth occupancy coefficient field; match the frequency configuration template by the modulation order field to extract the center frequency offset field and the bandwidth occupancy coefficient field. For each set of micro-bands bound to a physical transmission channel, the frequency band boundary determination operation is performed: the minimum physical coordinate value of the micro-band center frequency point is used as the frequency band start boundary, the maximum physical coordinate value of the micro-band center frequency point superimposed on the bandwidth occupancy coefficient field is used as the frequency band end boundary, and the center frequency offset field is superimposed to generate the final frequency band range. The center transmit frequency is calculated based on the frequency band start boundary and the frequency band end boundary, and four sets of structured data entities are output: physical transmission channel one, frequency band start boundary, frequency band end boundary, and center transmit frequency; physical transmission channel two, frequency band start boundary, frequency band end boundary, and center transmit frequency; physical transmission channel three, frequency band start boundary, frequency band end boundary, and center transmit frequency; physical transmission channel four, frequency band start boundary, frequency band end boundary, and center transmit frequency.

[0110] Based on the transmission frequency, the network coding verification packet is sent to the relay satellite using the inter-satellite Ka link;

[0111] Specifically, the onboard processor accesses the transmit frequency data entity, which contains four structured records: physical transmission channel one, its frequency band start boundary, frequency band end boundary, and center transmit frequency; physical transmission channel two, its frequency band start boundary, frequency band end boundary, and center transmit frequency; physical transmission channel three, its frequency band start boundary, frequency band end boundary, and center transmit frequency; and physical transmission channel four, its frequency band start boundary, frequency band end boundary, and center transmit frequency. The transmit frequency data entity is sequentially traversed based on the physical transmission channel number field. For each physical transmission channel, the upconverter's carrier generation parameters are configured: the center transmit frequency field is written to the frequency synthesizer's control register, the frequency band start boundary field is set to the lower sideband cutoff point, and the frequency band end boundary field is set to the upper sideband cutoff point, completing RF front-end initialization. Network coding verification packet modulation is initiated, and four entities in the network coding verification packet set are loaded: the verification packet sequence associated with physical transmission channel one, the verification packet sequence associated with physical transmission channel two, the verification packet sequence associated with physical transmission channel three, and the verification packet sequence associated with physical transmission channel four. For each verification packet sequence, the network coding verification packet entity is extracted in index order and input into the baseband signal port of the quadrature amplitude modulator. The modulator generates a carrier signal based on the center transmit frequency field of the current physical transmission channel and performs IQ quadrature mixing to up-convert the baseband signal to a Ka-band RF signal. Inter-satellite Ka link broadcast transmission is executed, activating the beamforming circuitry of the Ka-band phased array antenna to align it with the relay satellite's orbital coordinates. The RF signal generated by each physical transmission channel is boosted to its rated radiated power by a power amplifier and fed into the antenna array of radiating elements. The antenna control unit periodically transmits a RF signal pulse sequence according to the Space Data System protocol standard. The network coding verification packet entity is encoded into a phase-modulated waveform of the pulse sequence and propagates through free space to the relay satellite receiver. Bit error rate metrics are monitored in real time throughout the transmission process. If no acknowledgment signaling is received for three consecutive pulse sequences, an automatic retransmission mechanism is triggered to resend the current verification packet sequence.

[0112] The relay satellite receives the network coding check packet, reconstructs the data block sequence through Viterbi soft decision demodulation, and recovers the lost data blocks to generate a complete data block group;

[0113] The relay satellite then activates its Ka-band receiver array to capture the RF signal pulse sequence from the transmitting satellite. After down-converting the received RF signal pulse sequence to baseband, it is input into the Viterbi soft-decision demodulator processing unit. The Viterbi soft-decision demodulator processing unit executes a continuous phase detection algorithm, reconstructing the modulation symbol sequence using trellis path metric calculations to generate a bit stream of the original data block sequence and the check data block. Data recovery is initiated, and the Reed-Solomon decoder's error correction matrix is ​​loaded. The index number field and the physical transmission channel number field in the bitstream are parsed, and the data block sequence is reassembled in ascending index order. If a discontinuous index number is detected, the check data block activation mechanism is triggered: the associated check data block is extracted, and finite field arithmetic operations are applied to reconstruct the contents of the missing data block. Once all index numbers are arranged consecutively, a complete data block group consisting of the original data block and the recovered data block is output.

[0114] Based on the complete data block group, calculate the data block CRC-16 value and generate the verification status code;

[0115] Furthermore, the three elements of the complete data block group record, namely the index number field, the protocol metadata field, and the data unit field, are extracted. The data unit fields are arranged in ascending order according to the index number, and a byte array concatenation operation is performed to generate a continuous data stream bit sequence. The initialization function of the cyclic redundancy check calculation unit is activated, the initial value of the sixteen-bit register is set to the all-one state, and the generator polynomial defined in the ISO 3309 standard is selected. The shift register bit operation process is started, and the data stream bit sequence is shifted into the register one by one starting from the most significant bit: when the most significant bit of the register is in the logic one state, the generator polynomial exclusive OR operation is performed and shifted left by one bit; when the most significant bit of the register is in the logic zero state, the left shift operation is directly performed by one bit. After all bits are input, the sixteen-bit state value of the register is taken as the CRC-16 calculation result of the current complete data block group.

[0116] In summary, the present invention achieves accurate identification of safe frequency bands in complex electromagnetic environments by: collecting Ka-band signal amplitude and phase to construct an interference-marked spectrum map; and using a convolutional neural network to intelligently identify spectrum interference states. Furthermore, it combines a Q-learning algorithm to dynamically optimize the allocation of neutrino-band resources, improving the resource scheduling capability and transmission efficiency of intersatellite communication systems under multi-interference, time-varying channel conditions. These two core steps address the existing issues of spectral response lag and resource allocation rigidity, respectively, and enhance the spectrum utilization, communication reliability, and overall performance of satellite-borne Ka-band data transmission systems in dynamic environments.

[0117] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A satellite-borne four-channel Ka-band data transmission channel system, characterized by: include, The spectrum sensing module collects raw spectrum data of frequency points within the Ka band, simultaneously receives the interference hotspot database, constructs the original spectrum matrix, marks the interference areas, and generates a spectrum map with interference marks; The spectrum optimization module inputs the interference-marked spectrum map into a pre-trained convolutional neural network, outputs the interference probability, identifies safe frequency bands, divides the safe frequency bands into sub-bands, optimizes the allocation plan through the Q-learning algorithm, and generates an optimized spectrum configuration table. The dynamic modulation module dynamically adjusts the modulation mode according to the bandwidth parameters in the optimized spectrum configuration table, divides the high-speed AOS data into blocks by frame, and generates a network coding check packet based on the channel allocation strategy in the optimized spectrum configuration table; The link transmission module broadcasts the network coding check packet to the relay satellite through the inter-satellite Ka link. The relay satellite reconstructs the data according to the network coding rules, decodes the reconstructed data and verifies the CRC.

2. The satellite-borne four-channel Ka-band data transmission channel system according to claim 1, characterized in that: The original spectrum data includes signal amplitude and signal phase.

3. The satellite-borne four-channel Ka-band data transmission channel system according to claim 2, characterized in that: The original spectrum matrix is ​​constructed, and the interference area is marked to generate a spectrum map with interference marks. The specific steps are as follows: Construct the original spectrum matrix according to the signal amplitude and signal phase; Based on the interference hotspot database, interference marking is performed in the original spectrum matrix through double judgment to generate an interference marking vector; The original spectrum matrix is ​​fused with the interference marker vector to generate a spectrum map with interference markers.

4. The satellite-borne four-channel Ka-band data transmission channel system according to claim 3, characterized in that: The spectrum map with interference marks is input into the pre-trained convolutional neural network to output the interference probability of each frequency band and identify the safe frequency band. The specific steps are as follows: Extracting spectrum feature data from the spectrum map with interference markers, normalizing the spectrum feature data, and generating a standardized feature matrix; Collect historical interference-marked spectrum maps for convolutional neural network training; Input the normalized feature matrix into the trained convolutional neural network to generate interference probability; A safety threshold is set based on the historical interference probability. When the interference probability is less than or equal to the safety threshold, the frequency band is determined to be safe.

5. The satellite-borne four-channel Ka-band data transmission channel system according to claim 4, characterized in that: The specific steps of dividing the safe frequency band into micro sub-bands are as follows: Collect the highest modulation order, calculate the minimum bandwidth, and obtain the neutrino band width; The safety band is divided into sub-bands according to the total width of the safety band and the width of the sub-band.

6. The satellite-borne four-channel Ka-band data transmission channel system according to claim 5, characterized in that: The Q-learning algorithm is used to optimize the allocation scheme and generate an optimized spectrum configuration table. The specific steps are as follows: Measure the signal-to-noise ratio, interference power, and channel coherence bandwidth of the neutrino band center frequency point and generate a neutrino band state parameter table; Based on the neutrino band state parameter table, an initial Q-learning table is constructed, and the strategy optimization iteration is performed through the Q-learning algorithm to generate an optimized spectrum configuration table.

7. The satellite-borne four-channel Ka-band data transmission channel system according to claim 6, characterized in that: The specific steps of dynamically adjusting the modulation mode according to the bandwidth parameters in the optimized spectrum configuration table are as follows: Extract the bandwidth parameters of the neutrino band from the optimized spectrum configuration table and construct a mapping rule between the bandwidth parameters and the modulation order; Modulation modes are allocated to the micro sub-bands according to the mapping rules, and a modulation parameter configuration table is generated.

8. The satellite-borne four-channel Ka-band data transmission channel system according to claim 7, characterized in that: The high-speed AOS data is divided into blocks by frame, and the network coding verification packet is generated by combining the channel allocation strategy in the optimized spectrum configuration table. The specific steps are as follows: Collect high-speed AOS data streams and cut them into codable units to generate raw data block sequences; The channel priorities are extracted from the optimized spectrum configuration table, the original data blocks in the original data block sequence are allocated according to the channel priorities, and a network coding check packet is generated.

9. The satellite-borne four-channel Ka-band data transmission channel system according to claim 8, characterized in that: The specific steps of broadcasting the network coding verification packet to the relay satellite via the inter-satellite Ka link are as follows: Query neighboring satellites, obtain link margins, and select relay satellites based on the link margins; According to the optimized spectrum configuration table and the modulation parameter configuration table, configure the broadcast frequency band and obtain the transmission frequency; Based on the transmission frequency, the network coding verification packet is sent to the relay satellite using the inter-satellite Ka link.

10. The satellite-borne four-channel Ka-band data transmission channel system according to claim 9, characterized in that: The relay satellite reconstructs the data according to the network coding rule, decodes the reconstructed data and verifies the CRC. The specific steps are as follows: The relay satellite receives the network coding check packet, reconstructs the data block sequence through Viterbi soft decision demodulation, and recovers the lost data blocks to generate a complete data block group; Based on the complete data block group, the data block CRC-16 value is calculated and the check status code is generated.

Images