Quality optimization method and system of satellite communication link

Abstract

The invention relates to the technical field of transmission, and provides a satellite communication link quality optimization method and system, and the method comprises the steps: obtaining the multi-dimensional quality data of a satellite communication link; inputting the multi-dimensional quality data into a pre-trained link quality prediction model to obtain a link quality grade output by the link quality prediction model; determining a link quality optimization strategy according to the link quality grade; the optimization strategy comprises optimization of satellite transmitting power, channel coding or an antenna pointing angle; wherein the link quality prediction model is obtained by training a link quality prediction training sample based on a long and short term memory network. According to the invention, comprehensive monitoring, accurate prediction and efficient optimization of the link quality can be realized, and the reliability and stability of the satellite communication link are improved.

Description

Technical Field

[0001] This invention relates to the field of transmission technology, and in particular to a method and system for optimizing the quality of satellite communication links. Background Technology

[0002] Satellite communication, as a globally covered communication method unrestricted by geographical conditions, plays an irreplaceable role in areas such as ocean communication, emergency rescue, and interconnection in remote areas. With the development of technologies such as 5G and the Internet of Things, satellite communication traffic is experiencing explosive growth, placing higher demands on the reliability, real-time performance, and stability of link transmission. However, satellite communication links are susceptible to complex environmental factors such as ionospheric scintillation, rainfall attenuation, multipath effects, and satellite orbital deviations, leading to fluctuations in the link signal-to-noise ratio (SNR), increased bit error rate (BER), and increased transmission delay (T), directly impacting the quality of communication services. Therefore, achieving real-time monitoring, accurate prediction, and dynamic optimization of satellite communication link quality has become a core technical requirement for ensuring the efficient operation of satellite communication systems, and is of significant practical importance for improving the reliability of satellite communication and user experience. Summary of the Invention

[0003] This invention provides a method and system for optimizing the quality of satellite communication links, which can achieve comprehensive monitoring, accurate prediction and efficient optimization of link quality, thereby improving the reliability and stability of satellite communication links.

[0004] This invention provides a method for optimizing the quality of a satellite communication link, comprising: acquiring multi-dimensional quality data of the satellite communication link; inputting the multi-dimensional quality data into a pre-trained link quality prediction model to obtain the link quality level output by the link quality prediction model; determining a link quality optimization strategy based on the link quality level; the optimization strategy includes optimizing satellite transmit power, channel coding, or antenna pointing angle; wherein the link quality prediction model is based on a long short-time memory network and trained using link quality prediction training samples.

[0005] According to the present invention, a method for optimizing the quality of a satellite communication link is provided, wherein the multi-dimensional quality data includes multiple data such as link signal-to-noise ratio, bit error rate, signal strength, transmission delay, and bandwidth utilization.

[0006] According to the present invention, a method for optimizing the quality of a satellite communication link is provided. The link quality prediction model includes an input layer, a hidden layer, and an output layer. The method further includes: introducing an attention mechanism through the input layer to dynamically assign weights to each dimension of the multi-dimensional quality data; using a long short-term memory network to capture long-term dependencies in the time series data through the hidden layer; and outputting the link quality level through the output layer, wherein the link quality level includes four levels: excellent, good, average, and poor.

[0007] According to the present invention, a method for optimizing the quality of a satellite communication link includes the following steps in the training process of the link quality prediction model: training the link quality prediction model using link quality prediction training samples, wherein the link quality prediction training samples include historical multi-dimensional quality data and their corresponding link quality level labels; adjusting the hyperparameters of the link quality prediction model, wherein the hyperparameters include the learning rate and the number of iterations; and employing adaptive regularization technology to dynamically adjust the regularization strength according to the training progress to prevent the model from overfitting.

[0008] According to a satellite communication link quality optimization method provided by the present invention, the step of determining a link quality optimization strategy based on the link quality level includes: maintaining the current link state when the link quality level is excellent; controlling the antenna pointing angle based on satellite orbit parameters and ground station location information when the link quality level is good; switching the channel coding mode and adaptively adjusting the Turbo / LDPC coding rate when the link quality level is medium; and adjusting the satellite transmit power when the link quality level is poor.

[0009] The satellite communication link quality optimization method provided by the present invention further includes: activating a backup link when the optimized link quality data does not meet the preset improvement requirements.

[0010] According to a satellite communication link quality optimization method provided by the present invention, after obtaining multi-dimensional quality data of the satellite communication link, the method further includes: preprocessing the multi-dimensional quality data; the preprocessing method includes removing random noise by using a sliding window mean filtering method and removing outliers by using the Laida criterion.

[0011] According to the satellite communication link quality optimization method provided by the present invention, the loss function of the link quality prediction model is: , in, For the first a Actual link quality value for each sample. For the first a Predicted values ​​for each sample, N This represents the number of samples.

[0012] This invention also provides a satellite communication link quality optimization system, comprising: an acquisition module for acquiring multi-dimensional quality data of a satellite communication link; a level determination module for inputting the multi-dimensional quality data into a pre-trained link quality prediction model to obtain a link quality level output by the link quality prediction model; and a strategy determination module for determining a link quality optimization strategy based on the link quality level; wherein the optimization strategy includes optimizing satellite transmit power, channel coding, or antenna pointing angle; and wherein the link quality prediction model is based on a long short-time memory network and trained using link quality prediction training samples.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the satellite communication link quality optimization method as described above.

[0014] The present invention provides a method and system for optimizing the quality of satellite communication links, which can realize comprehensive monitoring, accurate prediction and efficient optimization of link quality, thereby improving the reliability and stability of satellite communication links. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a satellite communication link quality optimization method provided by the present invention.

[0017] Figure 2 This is a schematic diagram illustrating the specific process of a satellite communication link quality optimization method provided by the present invention.

[0018] Figure 3 This is a schematic diagram of the structure of a satellite communication link quality optimization system provided by the present invention.

[0019] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0021] Satellite communication, as a globally covered communication method unrestricted by geographical conditions, plays an irreplaceable role in areas such as ocean communication, emergency rescue, and interconnection in remote areas. With the development of technologies such as 5G and the Internet of Things, satellite communication traffic is experiencing explosive growth, placing higher demands on the reliability, real-time performance, and stability of link transmission. However, satellite communication links are susceptible to complex environmental factors such as ionospheric scintillation, rainfall attenuation, multipath effects, and satellite orbital deviations, leading to fluctuations in the link's signal-to-noise ratio (SNR), increased bit error rate (BER), and increased transmission latency (T), directly impacting the quality of communication services. Therefore, achieving real-time monitoring, accurate prediction, and dynamic optimization of satellite communication link quality has become a core technical requirement for ensuring the efficient operation of satellite communication systems, and is of significant practical importance for improving the reliability of satellite communication and user experience.

[0022] This invention provides a method for real-time monitoring and optimization of satellite communication link quality. By constructing a multi-dimensional monitoring system, a dynamic sampling mechanism, an improved LSTM (Bit Error Rate) prediction model with attention mechanism, and a hierarchical optimization and emergency switching strategy, it achieves comprehensive monitoring, accurate prediction, and efficient optimization of link quality, thereby improving the reliability and stability of satellite communication links.

[0023] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a method for optimizing the quality of a satellite communication link provided by the present invention.

[0024] Please refer to Figure 2 , Figure 2 A schematic diagram illustrating the specific process of a satellite communication link quality optimization method provided by the present invention.

[0025] This invention provides a method for optimizing the quality of a satellite communication link, comprising: 101: Obtain multi-dimensional quality data of satellite communication links.

[0026] As a preferred embodiment, the multi-dimensional quality data includes link signal-to-noise ratio, bit error rate, signal strength, transmission delay, and bandwidth utilization.

[0027] Considering that using a single or a few indicators to assess link quality cannot fully reflect the link quality status, and that data collection uses a fixed sampling frequency, it cannot adapt to the monitoring needs of links with different bandwidths, resulting in insufficient sampling in high-bandwidth scenarios and waste of resources in low-bandwidth scenarios; at the same time, single-channel asynchronous transmission is prone to poor data synchronization, affecting the accuracy of subsequent analysis.

[0028] The method of this invention constructs a multi-dimensional monitoring index system, including link signal-to-noise ratio (SNR), bit error rate (BER), signal strength (RSSI), transmission delay (T), and bandwidth utilization (U). Real-time data for each index are synchronously collected by a distributed sensing module on the satellite payload. The sampling frequency is dynamically adjusted to 0.1 to 0.3 times the bandwidth value based on the link bandwidth. The distributed sensing module includes a signal-to-noise ratio sensor, a bit error rate detector, a signal strength acquisition unit, a delay tester, and a bandwidth monitoring unit. Each module transmits data via an on-board bus in a time-division multiplexing manner, with a bus transmission rate of no less than 1 Gbps.

[0029] As a preferred embodiment, after acquiring the multi-dimensional quality data of the satellite communication link, the method further includes: preprocessing the multi-dimensional quality data; the preprocessing method includes removing random noise by using the sliding window mean filtering method and removing outliers by using the Laida criterion.

[0030] In this embodiment, the collected raw monitoring data (multi-dimensional quality data) is preprocessed. A sliding window mean filter is used to remove random noise, with the window length set to 5-15 sampling points. Then, the Laida criterion is used to remove outliers; that is, when a data point deviates from the mean by more than three times the standard deviation, it is considered an outlier and linear interpolation is used to fill in the gaps. The calculation formula for the sliding window mean filter is: in, For the filtered first Data points, For the original data points, This is the half-length of the window, with a value ranging from 2 to 7.

[0031] 102: Input multi-dimensional quality data into a pre-trained link quality prediction model to obtain the link quality level output by the link quality prediction model.

[0032] As a preferred embodiment, the link quality prediction model includes an input layer, a hidden layer, and an output layer; the method further includes: introducing an attention mechanism through the input layer to dynamically assign weights to each dimension in the multi-dimensional quality data; using a long short-term memory network in the hidden layer to capture long-term dependencies in the time series data; and outputting the link quality level through the output layer, which includes four levels: excellent, good, medium, and poor.

[0033] As a preferred embodiment, the training process of the link quality prediction model includes: training the link quality prediction model using link quality prediction training samples, which include historical multi-dimensional quality data and their corresponding link quality level labels; adjusting the hyperparameters of the link quality prediction model, including the learning rate and the number of iterations; and using adaptive regularization technology to dynamically adjust the regularization strength according to the training progress to prevent the model from overfitting.

[0034] In this embodiment, a link quality prediction model is constructed based on an improved Long Short-Term Memory (LSTM) network. An attention mechanism is introduced into the input layer of the LSTM network to dynamically assign different weights to each monitoring indicator, highlighting the influence of core indicators such as SNR and BER. The weight values ​​are calculated using the Softmax function. The improved LSTM network has 64-128 hidden layer neurons and undergoes 100-200 training iterations. The model's prediction error is controlled within 5%, providing accurate support for optimization strategies.

[0035] The calculation process for dynamic weights is as follows: Let the original feature vector of each monitoring indicator be... The corresponding initial weights are Dynamic weights are calculated through the attention mechanism layer. ,in, ,in, , For the first i The attention scores for each metric are obtained through neural network training. e j For the first j Attention score.

[0036] The criteria for determining the link quality level are: when When, the grade is "Excellent"; when When, the grade is "Good"; when When, the level is "medium"; when or At that time, the grade was "poor".

[0037] The training process of the improved LSTM network uses link quality prediction training samples and the Adam optimizer. The initial learning rate is set to 0.001, and the learning rate decays by 50% every 50 iterations. The loss function is the mean squared error (MSE).

[0038] An adaptive regularization technique is employed, dynamically adjusting the regularization strength based on training progress. Regularization is used to prevent overfitting, especially when dealing with complex link quality data. Initially, the regularization strength is low, allowing the model to quickly learn the main features of the data. As training progresses, the regularization strength is gradually increased to limit model complexity and prevent overfitting. The adjustment of the regularization strength is based on the model's performance on the validation set. When the error on the validation set begins to increase, the regularization strength is further increased.

[0039] As a preferred embodiment, the loss function of the link quality prediction model is: , in, For the first a Actual link quality value for each sample. For the first a Predicted values ​​for each sample, N This represents the number of samples.

[0040] 103: Determine the link quality optimization strategy based on the link quality level; the optimization strategy includes optimizing satellite transmit power, channel coding or antenna pointing angle; among which, the link quality prediction model is based on a long short time memory network and is trained using link quality prediction training samples.

[0041] As a preferred embodiment, a link quality optimization strategy is determined based on the link quality level, including: maintaining the current link state when the link quality level is excellent; controlling the antenna pointing angle based on satellite orbit parameters and ground station location information when the link quality level is good; switching the channel coding mode and adaptively adjusting the Turbo / LDPC coding rate when the link quality level is medium; and adjusting the satellite transmit power when the link quality level is poor.

[0042] In this embodiment, based on the link quality level output by the link quality prediction model, which includes four levels: Excellent, Good, Medium, and Poor, the corresponding optimization strategy is triggered: When the rating is "poor", priority is given to adjusting the satellite's transmission power, with the power adjustment amount being... According to the formula Calculate, where k is the power adjustment coefficient, with a value ranging from 0.8 to 1.2.

[0043] When the level is "Medium", optimize the channel coding method and switch to an adaptive coding mode of Turbo code or LDPC code. When switching the channel coding method, adjust the coding rate synchronously. The coding rate of Turbo code can adaptively switch between 1 / 2, 2 / 3, and 3 / 4, and the coding rate of LDPC code can adaptively switch between 1 / 2, 3 / 5, and 4 / 5. The coding rate adjustment is determined based on the bandwidth utilization U. When U > 80%, a high coding rate is selected, and when U < 50%, a low coding rate is selected.

[0044] When the rating is "Good", fine-tune the antenna pointing angle, adjusting the precision to [insert precision here]. The antenna pointing angle adjustment employs dual-axis linkage control, calculating the optimal pointing angle based on satellite orbit parameters and ground station location information. During the adjustment process, signal strength feedback values ​​are collected in real time. RSSI Increased to -60 dBm Stop adjusting when the value is 10 or above.

[0045] When the rating is "Excellent", maintain the current link status.

[0046] The system provides real-time feedback on the optimized link quality data and compares it with the indicators before optimization. If the link quality level does not improve after optimization, the above optimization steps are repeated until the level reaches "good" or above, or an alarm signal is issued when there is no improvement after three consecutive optimizations.

[0047] As a preferred embodiment, the method further includes: activating a backup link if the optimized link quality data does not meet the preset improvement requirements.

[0048] If three consecutive optimizations fail to improve the situation, a link switching mechanism will be activated to switch communication services to a backup satellite link. Seamless switching technology will be used during the switching process, with a switching latency of no more than 100ms. At the same time, an alarm message containing an analysis of the cause of the link failure will be sent to the ground control center to ensure the continuity of critical services such as emergency communication and high-speed data transmission, thereby improving system reliability.

[0049] The present invention also includes a link quality data storage and traceability step: storing the preprocessed raw data (multi-dimensional quality data), prediction results (link quality level) and optimization strategy execution records to the on-board solid-state storage module, with a storage period of 7 to 30 days, and supporting the ground station to retrieve historical data for link quality analysis via remote control commands.

[0050] Example 1 This embodiment is applied to a broadband communication link of a low-Earth orbit satellite constellation. This link mainly serves high-speed data transmission services in remote areas on the ground. The link is designed with a bandwidth of 10Gbps and a communication distance of approximately 1500km. It faces link quality fluctuation problems caused by ionospheric scintillation and multipath effects. The specific implementation process is as follows: Step S1: The multi-dimensional data acquisition module constructs a monitoring index system including link signal-to-noise ratio (SNR), bit error rate (BER), signal strength (RSSI), transmission delay (T), and bandwidth utilization (U). A distributed sensing module is deployed on the satellite payload. This module consists of a high-precision SNR sensor with a measurement range of -10 to 30 dB and a high-speed bit error rate detector with a detection range of 10 dB. -9 ~10 -3 The system consists of a signal strength acquisition unit with a sampling accuracy of 0.1 dBm, a nanosecond-level delay tester with a measurement error ≤ 1 ms, and a bandwidth monitoring unit. Each module transmits data via an onboard PCIe 4.0 bus in a time-division multiplexing manner, with a bus transmission rate set to 1.2 Gbps to meet the requirements for synchronous transmission of multiple indicators. The sampling frequency is dynamically adjusted according to the link bandwidth. In this embodiment, the bandwidth is 10 Gbps, and the sampling frequency is set to 0.2 times the bandwidth value, i.e., 2 Gbps, to ensure the timeliness and completeness of data acquisition.

[0051] Step S2: Data Preprocessing. The collected raw data undergoes two-stage preprocessing: The first stage uses a sliding window mean filter to remove random noise. Considering the noise characteristics of low-Earth orbit satellite links, the window length is set to half the window length m=4 for 10 sampling points. The filtering calculation formula is as follows: , where y n Let x be the nth data point after filtering. n The first stage uses the original data points; the second stage removes outliers using the Laida criterion and calculates the mean of each indicator data. and standard deviation When a data point satisfies If an outlier is identified, a linear interpolation method is used to fill in the missing data. This involves constructing a linear function based on three valid data points before and after the outlier and calculating the fill value at the outlier location to ensure the continuity of the data sequence.

[0052] Step S3: Link quality prediction is based on an improved Long Short-Term Memory (LSTM) network to build a prediction model. The model input is preprocessed 5-dimensional monitoring index data, and an attention mechanism is introduced into the input layer to highlight the impact of key indicators. The original feature vector is set. , corresponding to SNR, BER, RSSI, T and U respectively, initial weights Attention scores are obtained through training at the attention mechanism layer. Then, the dynamic weights are calculated using the Softmax function. Finally obtained It is evident that SNR and BER are the core indicators affecting link quality. The improved LSTM network has 128 hidden layer neurons. The activation functions for the forget gate, input gate, and output gate are all Sigmoid functions, while the output layer uses a Linear function. The training process uses the Adam optimizer with an initial learning rate of 0.001, which decays by 50% every 50 iterations. The training iterations are 150, and the loss function is the mean squared error. The training sample size was N=10000 groups, and the final model prediction error was controlled within 5%. According to the preset judgment criteria, the model's output link quality level was "Medium", with SNR=8.2dB and BER=4.5×10⁻⁶. -5 T=950ms, U=60%.

[0053] Step S4: Execution of the hierarchical optimization strategy. Since the link quality level is "medium," the channel coding optimization strategy is triggered, switching to the adaptive coding mode of LDPC codes. Considering the current bandwidth utilization U=60%, which is within the 50%~80% range, a coding rate of 3 / 5 is selected. This coding rate ensures both transmission reliability and bandwidth efficiency. The coding switching process is implemented through the onboard FPGA chip, with the switching latency controlled within 20ms to ensure uninterrupted service transmission.

[0054] Step S5: After optimizing feedback and iterative optimization, real-time link quality data is collected, resulting in SNR=12.5dB and BER=7.8×10. -7 With T=720ms and U=62%, the rating is improved to "Good" according to the judgment criteria. The preprocessed raw data, prediction results, and optimization strategy execution records are stored in the on-board solid-state storage module with a storage period of 15 days, allowing the ground station to retrieve historical data for link quality trend analysis via remote control commands.

[0055] Example 2 This embodiment is applied to the emergency communication link of a high-orbit geostationary satellite. This link is used for emergency command and communication in natural disaster scenarios. It requires high reliability and rapid recovery capability. The link bandwidth is 5Gbps, the communication distance is about 35,800km, and it is significantly affected by rainfall attenuation. The specific implementation process is as follows: Step S1: Multi-dimensional data acquisition adopts the same multi-dimensional monitoring index system as in Example 1. The parameters of each unit of the distributed sensing module are adapted to the characteristics of the high-orbit link. Among them, the measurement range of the latency tester is extended to 0~2000ms, and the bandwidth monitoring unit supports full-load monitoring of 5Gbps. The on-board bus transmission rate is set to 1.5Gbps, and the sampling frequency is dynamically adjusted to 1Gbps according to the bandwidth, which is 0.2 times the bandwidth value, to ensure that data changes can be quickly captured when the signal changes due to rainfall attenuation.

[0056] Step S2: The data preprocessing sliding window length is set to 8 sampling points, the half-length of the window is m=3, and the filtering calculation formula is as follows: This effectively balances noise removal with data real-time performance. Outliers caused by sudden rainfall events, such as a 10dB drop in SNR, are removed using the Raida criterion, and linear interpolation is used to complete the data, preventing outliers from affecting prediction accuracy.

[0057] Step S3: Improve the LSTM prediction model for link quality prediction. Set the number of hidden layer neurons to 64, train for 100 iterations, and obtain dynamic weights after training with the attention mechanism. Transmission delay (T) is weighted at 18% due to the characteristics of high-orbit links. The model inputs preprocessed data and outputs a link quality level of "poor," corresponding to SNR = 4.3 dB and BER = 2.1 × 10⁻⁶. -4 T=1520ms, U=45%.

[0058] Step S4: Execute the initial optimization using the tiered optimization strategy: The tier is "poor," so prioritize adjusting the satellite's transmit power. (Settings...) The power adjustment coefficient k = 1.0, according to the formula Calculated The transmit power was increased by 5.7dB using an onboard power amplifier. The adjustment process lasted 30ms, and the power amplifier's operating status was monitored in real time to prevent overload. Secondary optimization: After optimization, the collected data showed an SNR of 7.1dB and a BER of 8.3×10⁻⁶. -5 With T=1100ms and U=48%, the channel level remains "medium," triggering the channel coding optimization strategy and switching to Turbo coding mode. Since U=48%<50%, a low coding rate of 1 / 2 is selected to improve anti-interference capability, and the coding switching delay is controlled within 15ms.

[0059] Step S5: Optimize Feedback and Emergency Handling. After secondary optimization, the link metrics are updated to SNR=11.3dB and BER=2.9×10. -6 With T=680ms and U=50%, the rating is improved to "Good," and the optimization process is stopped. Relevant data is stored in the onboard solid-state storage module with a storage period of 30 days. If the rating does not improve after the first optimization, the optimization strategy in step S4 will be repeated. If there is no improvement after three consecutive optimizations, the backup satellite link switching mechanism will be activated, employing seamless switching technology to ensure a switching latency of ≤100ms. Simultaneously, an alarm message will be sent to the ground control center, including the cause of the fault (heavy rainfall attenuation), current link indicators, and optimization execution records.

[0060] The satellite communication link quality optimization system provided by the present invention is described below. The satellite communication link quality optimization system described below and the satellite communication link quality optimization method described above can be referred to each other.

[0061] Please refer to Figure 3 , Figure 3 This is a schematic diagram of the structure of a satellite communication link quality optimization system provided by the present invention.

[0062] This invention also provides a satellite communication link quality optimization system, comprising: an acquisition module 301 for acquiring multi-dimensional quality data of the satellite communication link; a level determination module 302 for inputting the multi-dimensional quality data into a pre-trained link quality prediction model to obtain the link quality level output by the link quality prediction model; and a strategy determination module 303 for determining a link quality optimization strategy based on the link quality level; the optimization strategy includes optimizing satellite transmit power, channel coding, or antenna pointing angle; wherein the link quality prediction model is based on a long short-term memory network and trained using link quality prediction training samples.

[0063] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4 As shown, the electronic device may include a processor 401, a communications interface 402, a memory 403, and a communication bus 404. The processor 401, communications interface 402, and memory 403 communicate with each other via the communication bus 404. The processor 401 can call logical instructions in the memory 403 to execute a satellite communication link quality optimization method. This method includes: acquiring multi-dimensional quality data of the satellite communication link; inputting the multi-dimensional quality data into a pre-trained link quality prediction model to obtain the link quality level output by the link quality prediction model; determining a link quality optimization strategy based on the link quality level; the optimization strategy includes optimizing satellite transmit power, channel coding, or antenna pointing angle; wherein the link quality prediction model is based on a long short-term memory network and trained using link quality prediction training samples.

[0064] Furthermore, the logical instructions in the aforementioned memory 403 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0065] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the satellite communication link quality optimization method provided by the above methods. The method includes: acquiring multi-dimensional quality data of the satellite communication link; inputting the multi-dimensional quality data into a pre-trained link quality prediction model to obtain the link quality level output by the link quality prediction model; determining a link quality optimization strategy based on the link quality level; the optimization strategy includes optimizing satellite transmission power, channel coding, or antenna pointing angle; wherein the link quality prediction model is based on a long short-time memory network and trained using link quality prediction training samples.

[0066] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements a satellite communication link quality optimization method provided by the methods described above. The method includes: acquiring multi-dimensional quality data of the satellite communication link; inputting the multi-dimensional quality data into a pre-trained link quality prediction model to obtain the link quality level output by the link quality prediction model; determining a link quality optimization strategy based on the link quality level; the optimization strategy includes optimizing satellite transmit power, channel coding, or antenna pointing angle; wherein the link quality prediction model is based on a long short-term memory network and trained using link quality prediction training samples.

[0067] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0068] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the quality of a satellite communication link, characterized in that, include: Acquire multi-dimensional quality data of satellite communication links; The multi-dimensional quality data is input into a pre-trained link quality prediction model to obtain the link quality level output by the link quality prediction model. Based on the link quality level, a link quality optimization strategy is determined; the optimization strategy includes optimizing satellite transmit power, channel coding, or antenna pointing angle. The link quality prediction model is based on a long short-term memory network and is trained using link quality prediction training samples.

2. The satellite communication link quality optimization method according to claim 1, characterized in that, The multi-dimensional quality data includes multiple metrics such as link signal-to-noise ratio, bit error rate, signal strength, transmission delay, and bandwidth utilization.

3. The satellite communication link quality optimization method according to claim 1, characterized in that, The link quality prediction model includes an input layer, a hidden layer, and an output layer; The method further includes: An attention mechanism is introduced through the input layer to dynamically assign weights to each dimension of the multi-dimensional quality data. The hidden layer is used to capture long-term dependencies in time-series data using a long short-term memory network. The output layer outputs the link quality level, which includes four levels: excellent, good, medium, and poor.

4. The satellite communication link quality optimization method according to claim 1, characterized in that, The training process of the link quality prediction model includes: The link quality prediction model is trained using link quality prediction training samples, which include historical multi-dimensional quality data and their corresponding link quality level labels. Adjust the hyperparameters of the link quality prediction model, including the learning rate and the number of iterations; An adaptive regularization technique is used to dynamically adjust the regularization strength according to the training progress in order to prevent the model from overfitting.

5. The satellite communication link quality optimization method according to claim 1, characterized in that, The step of determining the link quality optimization strategy based on the link quality level includes: If the link quality level is excellent, maintain the current link status; When the link quality level is good, the antenna pointing angle is controlled based on satellite orbit parameters and ground station location information; When the link quality level is medium, switch the channel coding mode and adaptively adjust the Turbo / LDPC coding rate. If the link quality level is poor, adjust the satellite transmission power.

6. The satellite communication link quality optimization method according to claim 1, characterized in that, Also includes: If the optimized link quality data does not meet the preset improvement requirements, the backup link will be activated.

7. The satellite communication link quality optimization method according to claim 1, characterized in that, After acquiring the multi-dimensional quality data of the satellite communication link, the process also includes: The multi-dimensional quality data is preprocessed; The preprocessing methods include using a sliding window mean filtering method to remove random noise and using the Laida criterion to remove outliers.

8. The satellite communication link quality optimization method according to any one of claims 1 to 7, characterized in that, The loss function of the link quality prediction model is: ， in, For the first a Actual link quality value for each sample. For the first a Predicted values ​​for each sample, N This represents the number of samples.

9. A satellite communication link quality optimization system, characterized in that, include: The acquisition module is used to acquire multi-dimensional quality data of the satellite communication link; The grade determination module is used to input the multi-dimensional quality data into a pre-trained link quality prediction model to obtain the link quality grade output by the link quality prediction model. The strategy determination module is used to determine a link quality optimization strategy based on the link quality level; the optimization strategy includes optimizing satellite transmit power, channel coding, or antenna pointing angle. The link quality prediction model is based on a long short-term memory network and is trained using link quality prediction training samples.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the satellite communication link quality optimization method as described in any one of claims 1 to 8.

Images