Multi-link data transmission method, device and system in low-orbit satellite network

Abstract

The invention relates to the technical field of satellite communication, in particular to a multi-link data transmission method, device and system in a low-orbit satellite network. According to the invention, a multi-link data transmission mechanism is adopted, so that the satellite antenna is allowed to establish multiple communication connections with multiple satellites at the same time, and the limitation of single-link communication in the prior art is broken. Compared with an existing single-link transmission mode, bandwidth resources of multiple satellite links can be utilized to the maximum extent, and waste of link resources is avoided. By introducing the LSTM to predict the network performance, the problem of instability caused by network state fluctuation in the prior art is solved. The LSTM model can predict a future network state through historical data, so that when the network performance is about to deteriorate, a data distribution strategy can be adjusted in advance, and network instability caused by bandwidth fluctuation or satellite switching is avoided. Therefore, continuity and stability of data transmission can still be kept when the number of users is suddenly increased or satellites are switched.

Description

Technical Field

[0001] The present invention relates to the field of satellite communication technology, and in particular to a multi-link data transmission method, device and system in a low-orbit satellite network. Background Art

[0002] With the rapid development of low-orbit satellite networks, the network can provide Internet access worldwide, especially in areas lacking ground network coverage, and low-orbit satellite networks have shown their great potential. And because of its low latency, wide coverage and low infrastructure cost, it has gradually become an important part of the future B5G / 6G network. At present, thousands of low-orbit satellites successfully launched in relevant areas have provided high-speed Internet services to millions of users around the world. These satellite networks are particularly suitable for areas such as polar regions, oceans and aviation that are difficult to cover with traditional ground networks (such as cellular networks, WiFi, etc.). Users in these areas upload and download data through satellite networks, meeting users' needs for high-performance communications.

[0003] At present, the main application scenarios of low-orbit satellite networks include Internet access in remote areas, emergency communications in natural disasters, and network services at high altitudes and at sea. In the existing low-orbit satellite communication method, each terminal user can only establish a communication connection with one satellite through a satellite antenna at the same time to transmit data. Although this communication method meets basic network needs, with the increase in the number of satellites and the surge in user data, the existing technology has problems such as low communication efficiency, unstable network performance, and significant environmental impact. Summary of the invention

[0004] In view of this, the present invention provides a multi-link data transmission method and device in a low-orbit satellite network to solve the problem of low communication efficiency in the communication architecture between satellites and users in the prior art.

[0005] In a first aspect, the present invention provides a multi-link data transmission method in a low-orbit satellite network, which is applied to a satellite antenna. The satellite antenna is connected to multiple low-orbit satellites at the same time to form multiple links. The method includes: receiving a user's data transmission request and data to be transmitted; detecting the state of the low-orbit satellite and determining the number of links used for data transmission; based on the historical network performance data of each link, using a network performance prediction model to predict future network performance data, and the network performance prediction model is constructed using a long short-term memory network; based on the future network performance data and the data characteristics of the data to be transmitted, determining the data allocation strategy for each link, and allocating the data to be transmitted according to the data allocation strategy and transmitting it on each link.

[0006] In the present invention, by adopting a multi-link data transmission mechanism, a satellite antenna is allowed to establish multiple communication connections with multiple satellites at the same time, breaking the limitation of single-link communication in the prior art. Compared with the existing single-link transmission mode, the present invention can maximize the use of bandwidth resources of multiple satellite links and avoid the waste of link resources. At the same time, by introducing LSTM (Long Short-Term Memory, long short-term memory network) to predict network performance, the instability problem caused by network state fluctuation in the prior art is solved. Since the LSTM model can predict the future network state (such as link bandwidth and delay) through historical data, when the network performance is about to deteriorate, the system can adjust the data allocation strategy in advance, thereby avoiding network instability caused by bandwidth fluctuation or satellite switching. This technical means can ensure that the continuity and stability of data transmission can still be maintained when the number of users increases sharply or the satellite switches. In addition, when data transmission is performed, the data allocation strategy between multiple links can be dynamically adjusted according to the predicted link performance and data characteristics. This enables the present invention to provide a more stable and efficient data transmission solution in a dynamic and uncertain satellite network environment.

[0007] In an optional implementation, the method further includes: determining whether to trigger redundant transmission according to link quality and data characteristics of the data to be transmitted.

[0008] In the present invention, redundant transmission is used to solve the problem that links in traditional satellite communications are easily affected by the environment. The reliability of data transmission is effectively improved, especially under adverse conditions such as thunderstorms, which greatly reduces the risk of link interruption. The anti-interference ability and communication reliability are improved.

[0009] In an optional embodiment, triggering redundant transmission is determined based on link quality and data characteristics of the data to be transmitted, including: determining a link quality threshold for triggering redundant transmission based on the data characteristics of the data to be transmitted; obtaining link quality, judging the size of the link quality and the link quality threshold; triggering redundant transmission when the link quality is less than the link quality threshold.

[0010] In the present invention, the link quality threshold for triggering redundant transmission is determined according to data characteristics, thereby making the setting of the link quality threshold more flexible and convenient for coping with different transmission scenarios.

[0011] In an optional embodiment, after redundant transmission is triggered when the link quality is less than the link quality threshold, the method further includes: determining parameters of the erasure code based on data characteristics of the data to be transmitted; and based on the parameters of the erasure code, using the erasure code technology to construct redundant data for redundant transmission.

[0012] In this embodiment, the erasure code technology is used to construct redundant data, and the parameters of the erasure code are determined according to the data characteristics, so that the redundant data can be flexibly constructed according to the transmission situation.

[0013] In an optional implementation, the historical network performance data includes historical throughput data. Based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data, including: obtaining network performance data of multiple time steps at historical moments of each link, and normalizing the network performance data; processing the network performance data using a sliding window to construct training samples; training a long short-term memory network model including an attention mechanism based on the historical network performance data, using Adam as an optimizer during the training process, using mean square error as a loss function, adjusting the model parameters, and obtaining a network performance prediction model; and using the network performance prediction model to predict network performance data for a preset number of time steps in the future.

[0014] In this embodiment, a network performance prediction model based on a long short-term memory network is constructed. The model uses measured historical data to predict future network performance, and can accurately estimate the changes in link bandwidth and delay, thereby providing a basis for multi-link data allocation strategies. By introducing the LSTM model, the temporal dependencies in the network can be effectively captured, solving the problem of the difficulty in accurately predicting network performance in the prior art.

[0015] In an optional implementation, the data allocation strategy for each link is determined based on future network performance data and data characteristics of the data to be transmitted, including: using a deep reinforcement learning model to determine the data allocation strategy for each link based on future network performance data and data characteristics of the data to be transmitted.

[0016] In an optional embodiment, the deep reinforcement learning model includes state, behavior and reward, the state includes future network performance data and data characteristics of data to be transmitted, the data characteristics include the amount of data to be transmitted and the tolerated delay of transmission, the behavior includes the distribution method of the transmitted data on the link, and the reward is determined according to the efficiency and cost of data transmission.

[0017] In the present invention, by introducing a deep reinforcement learning algorithm, the data allocation strategy between multiple links can be dynamically adjusted according to the real-time link status and future network predictions. Compared with the existing static or experience-based allocation methods, the deep reinforcement learning algorithm has adaptive adjustment capabilities and can continuously optimize the delay and processing cost of data transmission. Therefore, the present application can maintain high transmission efficiency and low processing cost in different network environments, thereby greatly improving the flexibility and adaptability of the system.

[0018] In a second aspect, the present invention provides a multi-link data transmission device in a low-orbit satellite network, which is applied to a satellite antenna. The satellite antenna is connected to multiple low-orbit satellites at the same time to form multiple links. The device includes: a data receiving module, which is used to receive the user's data transmission request and data to be transmitted; a link quantity determination module, which is used to detect the state of the low-orbit satellite and determine the number of links used for data transmission; a network state prediction module, which is used to predict future network performance data based on the historical network performance data of each link and using a network performance prediction model; an intelligent data allocation module, which is used to determine the data allocation strategy for each link based on the future network performance data and the data characteristics of the data to be transmitted, and allocate the data to be transmitted according to the data allocation strategy and transmit it on each link.

[0019] In a third aspect, the present invention provides a multi-link data transmission system in a low-orbit satellite network, the system comprising: a user terminal, a satellite antenna and a low-orbit satellite network, the low-orbit satellite network comprising multiple low-orbit satellites, the satellite antenna using a preconfigured communication module to simultaneously establish connections with multiple low-orbit satellites to form multiple links; the user terminal is used to send data transmission requests and data to be transmitted to the satellite antenna; the satellite antenna is used to detect the status of the low-orbit satellite and determine the number of links used for data transmission; based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data; based on the future network performance data and the data characteristics of the data to be transmitted, a data allocation strategy for each link is determined, and the data to be transmitted is allocated according to the data allocation strategy and then transmitted on each link.

[0020] In a fourth aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the multi-link data transmission method in a low-orbit satellite network of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0021] In a fifth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the multi-link data transmission method in a low-orbit satellite network of the above-mentioned first aspect or any corresponding embodiment thereof.

[0022] In a sixth aspect, the present invention provides a computer program product, comprising computer instructions, wherein the computer instructions are used to enable a computer to execute the multi-link data transmission method in a low-orbit satellite network of the above-mentioned first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 is a schematic flow chart of a multi-link data transmission method in a low-orbit satellite network according to an embodiment of the present invention;

[0025] Figure 2 is a schematic diagram of the influence of historical data of different dimensions on the training results when training a network performance prediction model according to an embodiment of the present invention;

[0026] Figure 3 is a schematic flow chart of a multi-link data transmission method in a low-orbit satellite network according to another embodiment of the present invention;

[0027] Figure 4 is a schematic diagram of a network performance prediction model structure according to an embodiment of the present invention;

[0028] Figure 5 is a schematic diagram of the principle of an intelligent data allocation module based on deep reinforcement learning according to an embodiment of the present invention;

[0029] Figure 6 is a structural block diagram of a multi-link data transmission device in a low-orbit satellite network according to an embodiment of the present invention;

[0030] Figure 7 is a schematic diagram of a multi-link data transmission system in a low-orbit satellite network according to an embodiment of the present invention;

[0031] Figure 8 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] As described in the background technology, existing low-orbit satellite communication methods have problems such as low communication efficiency, unstable network performance, and significant environmental impact.

[0033] Specifically, the existing single-link communication method limits the satellite antenna to only connect to one satellite at any time, and cannot fully utilize the multiple satellite resources in the satellite constellation. As the number of satellites and communication bandwidth increase, the single-link transmission method cannot meet the needs of efficient data transmission. This method shows obvious communication bottlenecks and link resource waste when users transmit large amounts of data or require low-latency communication. For example, when users upload large amounts of data, the data can only be transmitted through a single link, resulting in low communication efficiency and inability to utilize other available satellites, thus failing to achieve optimal resource allocation.

[0034] At the same time, the performance of existing satellite communication networks is limited by the volatility of link bandwidth and latency, especially when the number of users increases or satellites are switched. When the number of users in a certain area increases suddenly, the bandwidth of the satellite network can be quickly exhausted, resulting in a decrease in communication rate. In addition, due to the rapid changes in the position of low-orbit satellites relative to the earth, user terminals need to switch satellites frequently, but this switching process may cause short network interruptions or increased latency, which in turn affects the stability of data transmission.

[0035] In addition, the communication link between users and satellites is very susceptible to external environmental influences. For example, in severe weather conditions such as thunderstorms, the signal attenuation problem is significant, which can easily lead to the interruption of the communication link between users and satellites. When the link is interrupted, users will not be able to access Internet services normally. The fragility of this link makes the existing satellite communication framework less stable in certain environments, making it difficult to ensure continuous Internet access.

[0036] In addition, although multi-link transmission technology has been applied in the field of communication between satellites, the existing technology has not yet been deployed on the link between users and satellites. This makes it impossible for user terminals to establish connections with multiple satellites at the same time, limiting the further improvement of data transmission efficiency. Existing multipath transmission schemes focus more on communications within satellites, while multi-link transmission on the user side has not been widely studied and applied. Data upload on the user side still relies on a single-link communication architecture, further restricting the overall efficiency and performance of communications.

[0037] Therefore, how to introduce multi-link transmission technology into the existing communication architecture between satellites and users and how to maintain the stability of network performance and improve the network's anti-interference ability in an environment where the network status changes dynamically (including changes in the number of users or satellite switching, and changes that occur when satellite communication links are affected by the environment) are issues that need to be urgently addressed.

[0038] In view of this, the present invention allows the satellite antenna to establish multiple communication connections with multiple satellites at the same time by adopting a multi-link data transmission mechanism, breaking the limitation of single-link communication in the prior art. Compared with the existing single-link transmission mode, the present invention can maximize the use of bandwidth resources of multiple satellite links and avoid the waste of link resources. At the same time, by introducing LSTM (Long Short-Term Memory, long short-term memory network) to predict network performance, the instability problem caused by network state fluctuation in the prior art is solved. Since the LSTM model can predict the future network state (such as link bandwidth and delay) through historical data, when the network performance is about to deteriorate, the system can adjust the data allocation strategy in advance, thereby avoiding network instability caused by bandwidth fluctuation or satellite switching. This technical means can ensure that the continuity and stability of data transmission can still be maintained when the number of users increases sharply or the satellite switches. In addition, when data transmission is performed, the data allocation strategy between multiple links can be dynamically adjusted according to the predicted link performance and data characteristics. As a result, the present invention can provide a more stable and efficient data transmission solution in a dynamic and uncertain satellite network environment.

[0039] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0040] According to an embodiment of the present invention, an embodiment of a multi-link data transmission method in a low-orbit satellite network is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0041] In this embodiment, a multi-link data transmission method in a low-orbit satellite network is provided. Figure 1 is a flow chart of a multi-link data transmission method in a low-orbit satellite network according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0042] Step S101: receiving a user's data transmission request and data to be transmitted.

[0043] Specifically, the multi-link data transmission method can be used for data transmission between user terminals and low-orbit satellite networks. Among them, the low-orbit satellite network refers to a network composed of satellites operating in an orbit close to the earth's surface (usually at an altitude of 160 kilometers to 2000 kilometers), which can achieve large-scale global coverage. The low-orbit satellite network is composed of multiple low-orbit satellites. The satellites have high-bandwidth communication capabilities and can exchange data with satellite antennas to provide network access services for user terminals. The user terminal is composed of various types of users, such as mobile terminals, computers, smart vehicles, etc., which are used to send data transmission requests. In an embodiment of the present invention, the data generated by the user terminal may include data for services such as video on demand, file transfer, and video conferencing.

[0044] When data is transmitted between a user terminal and a low-orbit satellite network, the user terminal first sends a data transmission request to the satellite antenna through the wireless network between the user terminal and the satellite antenna. The satellite antenna is equipped with a communication module that supports multiple antennas or multiple links. For example, multiple-input multiple-output technology, multi-band support technology, multi-link aggregation technology, or adaptive modulation and coding technology can be integrated into the communication module. As a result, the satellite antenna can establish connections with multiple low-orbit satellites at the same time. At the same time, the satellite antenna has the ability to store data and perform intelligent computing, that is, the satellite antenna can select appropriate multiple links according to the connected low-orbit satellites to send the data transmitted by the user terminal to the low-orbit satellite.

[0045] Among them, multi-link refers to the technical means by which the user terminal can establish multiple communication links with the network to transmit data in parallel during the data transmission process. This multi-link transmission method breaks the limitations of traditional single-link transmission and allows data to be transmitted on multiple links simultaneously, thereby greatly improving data transmission efficiency and network performance stability.

[0046] When data transmission is required, the user terminal sends a data transmission request and the corresponding data to be transmitted to the satellite antenna, and the satellite antenna selects appropriate multiple links to send the data to be transmitted.

[0047] Step S102, detecting the status of the low-orbit satellite and determining the number of links for data transmission. Specifically, after receiving the data transmission request, the satellite antenna needs to first screen the available low-orbit satellites according to the status of the connected low-orbit satellites. When screening the available satellites, the satellite antenna can send satellite status query requests to multiple connected low-orbit satellites to obtain detailed status information of the satellites, such as working mode, load condition, and remaining life, so as to determine the available low-orbit satellites. After determining the available low-orbit satellites, a link for data transmission can be formed based on the connection with the available low-orbit satellites.

[0048] Step S103, based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data, and the network performance prediction model is constructed using a long short-term memory network. Among them, the network performance data includes data such as bandwidth, delay, and throughput. The network performance data is related to the time series to a certain extent, and the long short-term memory network has a significant advantage in time series prediction. Therefore, this embodiment constructs a network performance prediction model by training a long short-term memory network, and uses the network performance prediction model to predict future network performance data. Specifically, during training, the long and short-term network can be trained based on the network performance data of each link at the historical moment, so as to construct a network performance prediction model corresponding to each link. In addition, the predicted future network performance data provides future network status change information for subsequent data allocation, reducing the impact of network fluctuations on the data allocation strategy.

[0049] Step S104, based on the future network performance data and the data characteristics of the data to be transmitted, the data allocation strategy of each link is determined, and the data to be transmitted is allocated according to the data allocation strategy and then transmitted on each link. Among them, the data characteristics of the data to be transmitted include the data volume of the data to be transmitted, the expected transmission tolerance delay and other characteristics. When determining the data allocation strategy, it can be determined based on the future network performance data and data characteristics, and the data transmission efficiency and transmission cost as the target. Therefore, a suitable allocation strategy can be determined by constructing a current function or using relevant algorithms such as machine learning algorithms. Based on this, when the network state, that is, the network performance data, changes, the determined data allocation strategy will also change, that is, the data allocation strategy can be adaptively adjusted according to the dynamic network environment, so as to provide users with a better data transmission experience, such as less delay or data processing cost, higher reliability, etc.

[0050] Specifically, the data allocation strategy can be understood as the ratio of allocating the data to be transmitted to each link. For example, if the number of links determined for data transmission is 3, the data to be transmitted can be divided into multiple data blocks according to the ratio of 40%, 10% and 50%, and transmitted according to the corresponding links.

[0051] In this embodiment, a multi-link data transmission method in a low-orbit satellite network is provided, and the process includes the following steps:

[0052] Step S201: Receive a user's data transmission request and data to be transmitted. Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0053] Step S202: Detect the status of the low-orbit satellite and determine the number of links used for data transmission. Figure 1Step S102 of the illustrated embodiment will not be described in detail here.

[0054] Step S203: Based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data; see Figure 1 Step S103 of the illustrated embodiment will not be described in detail here.

[0055] Step S204, based on the link quality and the data characteristics of the data to be transmitted, it is determined whether to trigger redundant transmission. Specifically, before data allocation and transmission, it is necessary to determine whether to use a redundant strategy for data transmission. Among them, when the performance of some links is interfered by the environment (specifically determined by the link quality) or the user has high requirements for data transmission reliability (specifically determined by the data characteristics of the data to be transmitted), a redundant strategy can be used to determine the continuity and reliability of the overall data transmission, thereby improving the anti-interference ability of the network without significantly increasing bandwidth consumption.

[0056] Specifically, the above step S204 includes:

[0057] Step S2041, determining the link quality threshold for triggering redundant transmission according to the data characteristics of the data to be transmitted. When determining to trigger redundant transmission, the link quality threshold for triggering redundant transmission can be determined first by the data characteristics of the data to be transmitted, the data characteristics include data type and service reliability, and the link quality threshold includes an average signal-to-noise ratio threshold or a packet loss rate threshold, etc. When determining the link quality threshold according to the data characteristics, a high-reliability service such as online video conferencing or important file transmission can be set to have a lower link quality threshold; while a low-sensitivity service such as ordinary file downloading may not trigger redundant transmission or have a higher threshold.

[0058] Step S2042, obtain link quality, and determine the link quality and the link quality threshold. Specifically, after determining the link quality threshold, the link quality of each link within a period of time can be obtained, for example, the average signal-to-noise ratio or packet loss rate of each link within a period of time is calculated as the link quality of the link. After determining the link quality of each link, it is compared with the link quality threshold to determine whether to trigger redundant transmission.

[0059] Step S2043, when the link quality is less than the link quality threshold, redundant transmission is triggered. Specifically, when the link quality is less than the link quality threshold, redundant transmission is triggered. For example, the link quality threshold is an average signal-to-noise ratio threshold of 20 dB, and the packet loss rate is 5%. When the average signal-to-noise ratio is lower than 20 dB, or the packet loss rate is lower than 5%, the redundant transmission strategy is started.

[0060] Step S2044, determining the parameters of the erasure code according to the data characteristics of the data to be transmitted; according to the parameters of the erasure code, using the erasure code technology to construct redundant data for redundant transmission. In this embodiment, when it is determined that redundant transmission is triggered, the erasure code technology is used to construct redundant data. Among them, the parameters of the erasure code can be expressed as (n, k), where n represents the total number of generated data blocks (including original data and redundant data), k represents the number of original data blocks (i.e., the number of links), and nk is the number of redundant data blocks. When executing the redundant data transmission strategy, the satellite antenna needs to dynamically configure n and k according to business requirements and link conditions. For services that require higher reliability in the redundant transmission strategy, a higher redundancy ratio can be set (such as n = 10, k = 7, that is, 3 redundant blocks are added, and the redundancy ratio is 30%). For services that require lower reliability in the redundant transmission strategy, a lower redundancy ratio can be used (such as n = 10, k = 9, and the redundancy ratio is 10%).

[0061] In this embodiment, the problem that the link in traditional satellite communication is susceptible to environmental influences is solved through a redundant data transmission strategy. In the event of severe weather or link interruption, the redundant data transmission mechanism allows redundant data to be transmitted through multiple links, ensuring that even if some links have problems, the receiving end can still recover the data. The mechanism uses erasure code technology to allow satellite antennas of different satellite network providers to flexibly set the increased redundant data ratio according to data transmission and processing costs and environmental requirements. For example, for high-reliability services (such as video conferencing) in scenarios that require redundant transmission, the satellite antenna can increase a higher proportion of redundant data (such as 30%). For services with relatively low reliability requirements (such as ordinary file downloads), a lower proportion of redundancy (such as 10%) is added, and even no redundancy is added when the network environment is good. By adding a small amount of redundant data blocks, the reliability of data transmission is effectively improved, especially under adverse conditions such as thunderstorms, the risk of link interruption is greatly reduced. Therefore, the present invention can ensure the continuity and anti-interference ability of data transmission when facing complex environments.

[0062] Step S205: determine the data allocation strategy for each link based on the future network performance data and the data characteristics of the data to be transmitted, and distribute the data to be transmitted on each link according to the data allocation strategy. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0063] In this embodiment, a multi-link data transmission method in a low-orbit satellite network is provided, and the method comprises the following steps:

[0064] Step S301: Receive the user's data transmission request and the data to be transmitted; see Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0065] Step S302, detect the status of the low-orbit satellite and determine the number of links used for data transmission; see Figure 1 Step S102 of the illustrated embodiment will not be described in detail here.

[0066] Step S303: Based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data. In this embodiment, throughput data is mainly used as network performance data for model training and prediction.

[0067] Specifically, the above step S303 includes:

[0068] Step S3031, obtain the network performance data of multiple time steps of each link at the historical moment, and normalize the network performance data; specifically, the network performance data of each time step can be obtained by real-time monitoring of the link status of the low-orbit satellite network. In order to improve the speed of model training, the obtained network performance data can be normalized, for example, the normalization formula can be used to scale the data to between 0 and 1.

[0069] Step S3032, use a sliding window to process the network performance data and construct a training sample; specifically, before using the acquired network performance data for model training, the data can be processed in batches according to the time series, and each batch of data can include data of m consecutive time steps (for example, the throughput of the past 10 time steps). For example, the acquired data can be processed in a sliding window manner to form a training sample.

[0070] Step S3033, based on the historical network performance data, the long short-term memory network model including the attention mechanism is trained. During the training process, Adam is used as the optimizer, and the mean square error is used as the loss function. The model parameters are adjusted to obtain a network performance prediction model.

[0071] In this embodiment, the long short-term memory network model includes an input layer, an LSTM hidden layer, a fully connected layer and an output layer. Among them, the input layer is used to receive network performance data of past time steps, such as the network throughput of the past 10 time steps. The LSTM hidden layer contains one or more LSTM units for capturing the temporal dependency of data. Each LSTM unit can retain historical information for a long time and is suitable for processing time series data. The fully connected layer contains 32 neurons, which are responsible for mapping the output of the LSTM hidden layer to a scalar, that is, the predicted value of the network throughput of the future time step. Finally, the output layer outputs the predicted throughput value. The LSTM hidden layer of the present invention uses a two-layer LSTM structure to enhance the prediction ability of the model, the first layer contains 50 neurons, and the second layer contains 100 neurons to capture complex temporal dependencies. In the LSTM model, the weights of different features are usually automatically learned by the model. At the same time, an attention mechanism is added to the long short-term memory network model to further optimize the model so that the model automatically adjusts the importance of the input features. For example, the network throughput at a certain moment may have a greater impact on the predicted throughput, while the impact at other moments may be relatively small. The attention mechanism can help the model flexibly adjust the weights of each feature at each time step.

[0072] Specifically, when training the LSTM network model, the input of the model is the historical throughput sequence of the network, which usually needs to include throughput data of multiple past time steps. In the present invention, it is assumed that the throughput of the past m time steps is used to predict the future throughput. Among them, the value of m can be flexibly changed according to different scenarios. The output of the model is the throughput prediction value of the future n time points. In the present invention, the values ​​of m and n can be set as needed. Figure 2 As shown, the present invention provides a result diagram of the influence of different m values ​​on the training results. In this scenario, the value of n is set to 10, and it can be observed that relying solely on the throughput information at a past moment will result in a higher prediction error. This limitation stems from insufficient historical data and the inability to provide global information on throughput changes. In addition, increasing the dimension of historical throughput information does not always reduce training losses. For example, the training effect is poor when m=25. This phenomenon has an adverse effect on the accuracy of the prediction due to outdated and redundant information. When m=15 and m=20, the training effect is the same. Therefore, in order to reduce computational complexity, this scenario can set m=15 when training the network performance prediction model. It should be noted that in other scenarios, this value can be flexibly adjusted according to the requirements of the characteristics.

[0073] In addition, during the training process, the number of iterations is set to 1000 and 16 data are sampled as the input of the model in each iteration. It should be noted that the hyperparameters in the model structure, such as the data sampling size, can be tuned according to the actual scenario, and can be automatically adjusted through grid search or random search to find the optimal combination. In addition, this embodiment uses Adam as an optimizer, which has good convergence performance during the training process of LSTM. At the same time, this embodiment sets the learning rate to 0.001 and uses the mean squared error (MSE) as the loss function. It should be noted that these parameter settings can be changed with different scales of networks and data requirements.

[0074] Step S3034, using the network performance prediction model to predict the network performance data of the preset time steps in the future. Specifically, after the network performance prediction model is trained and determined, it can be used to predict the network performance data, that is, the throughput data of multiple time steps before the current moment can be input into the network performance prediction model, and the throughput data of multiple time steps in the future can be output. In addition, the network performance prediction model can also be updated based on the predicted throughput data and the throughput data obtained in real time, so that the network performance prediction model can keep up with the changes in the network environment.

[0075] In this embodiment, by testing the network performance at different time periods, the characteristics of network performance fluctuations over time within a day are accurately captured, revealing the unique characteristics of low-orbit satellite networks. The collection of these data provides an in-depth understanding of the dynamic changes of the link and provides a basis for the subsequent design of the data allocation algorithm.

[0076] In this embodiment, by introducing a network performance prediction model based on LSTM, the instability problem caused by network state fluctuations in the prior art is solved. Since the LSTM model can predict future network states (such as link bandwidth and delay) through historical data, when the network performance is about to deteriorate, the system can adjust the data allocation strategy in advance to avoid network instability caused by bandwidth fluctuations or satellite switching. This technical means can ensure that the continuity and stability of data transmission can still be maintained when the number of users increases sharply or the satellite switches. In addition, through the above steps, the network performance prediction model can effectively predict the future network throughput. Combined with the multi-link data allocation mechanism in the present invention, the data allocation strategy can be adjusted in advance, the utilization of network resources can be optimized, and the efficiency of data transmission can be improved.

[0077] Step S304, determining the data allocation strategy for each link based on the future network performance data and the data characteristics of the data to be transmitted, and allocating the data to be transmitted according to the data allocation strategy and transmitting it on each link.

[0078] Specifically, the above step S304 includes:

[0079] Step S3041, using a deep reinforcement learning model, determine the data allocation strategy for each link based on future network performance data and data characteristics of the data to be transmitted. The deep reinforcement learning model can optimize the data allocation strategy of the satellite antenna between multiple links. Specifically, through dynamic perception of the environment (user data and network status), the data allocation between different links can be intelligently adjusted to ensure the efficiency and stability of data transmission.

[0080] Specifically, before using the deep reinforcement learning model to determine the data allocation strategy, it is necessary to model the environment of deep reinforcement learning, that is, state, action and reward. In this embodiment, the state is defined as the data characteristics of the data to be transmitted and the network state information. Among them, the data characteristics of the data to be transmitted include the amount of data to be transmitted and the tolerant delay of the transmission; the network state information includes the future network throughput of the available links predicted by the network performance prediction model and the current and historical network throughput. Behavior is defined as the distribution method of the data to be transmitted on the available links. Among them, the distribution method is given in the form of a ratio, and its constraint is that the distribution ratio of the data on the available links is equal to 1. For example, if the number of available satellite links is 3, then the behavior (0.4, 0.1, 0.5) is an effective strategy, where 0.4, 0.1, and 0.5 respectively represent that 40%, 10%, and 50% of the transmitted data are distributed to three links for transmission. In the present invention, the satellite antenna needs to select a suitable behavior in the effective behavior space to maximize the reward. Reward is a measure of the quality of a certain behavior (data allocation strategy). The goal is to optimize the efficiency of data transmission and reduce the transmission cost, which is determined by two factors: the lower the delay of data transmission, the higher the reward, and the lower the cost of transmission, the higher the reward. In the deep reinforcement learning training process of the present invention, the algorithm converges to the optimal strategy through continuous iteration, achieving low-latency and low-cost data transmission.

[0081] In this embodiment, by introducing a deep reinforcement learning algorithm, the data allocation strategy between multiple links can be dynamically adjusted according to the real-time link status and future network predictions. Compared with the existing static or experience-based allocation methods, the deep reinforcement learning algorithm has adaptive adjustment capabilities and can continuously optimize the delay and processing cost of data transmission. Therefore, this embodiment can maintain high transmission efficiency and low processing cost in different network environments, thereby greatly improving the flexibility and adaptability of the system.

[0082] In this embodiment, the user's data is divided into multiple data blocks, which are transmitted in parallel through multiple satellite links, thereby maximizing the utilization of link resources. Specifically, the data allocation strategy of this embodiment combines the designed network performance prediction model with deep reinforcement learning technology, and adaptively adjusts the data allocation strategy in a dynamic network environment according to the data characteristics of each user and the real-time status of the available links (such as bandwidth, delay, etc.) to cope with the fluctuations of different link conditions, and ultimately optimizes the efficiency and delay of data transmission while minimizing the data processing cost. Therefore, by predicting the future network status, the method can consider the upcoming link bandwidth and delay changes in advance in the data allocation strategy, thereby avoiding the impact of bandwidth fluctuations or network instability caused by satellite switching on data transmission, and ensuring the optimality of the data transmission strategy in the future time period. This innovative design combining the prediction model and the adaptive algorithm enables the present invention to provide a more stable and efficient data transmission solution in a dynamic and uncertain satellite network environment.

[0083] As a specific application example of the embodiment of the present invention, Figure 3 , Figure 4 and Figure 5 As shown, the multi-link data transmission method in the low-orbit satellite network can be implemented by the following process:

[0084] Step S01, the user sends data and a transmission request to a satellite antenna.

[0085] Step S02: The satellite antenna detects the status of available satellites in low orbit and determines the number of links for data transmission.

[0086] Step S03: the network status prediction module of the satellite antenna uses the measured historical data to perform link future throughput evaluation, that is, predicts the future network throughput based on the historical network throughput.

[0087] Step S04: The satellite antenna determines whether a redundant strategy is needed for data transmission. If necessary, the satellite antenna will analyze and construct redundant data. This process is completed by an intelligent data distribution module deployed on the satellite antenna.

[0088] Step S05, the intelligent data distribution module of the satellite antenna determines the distribution ratio of data on each link according to the future state of each link and the characteristics of the data to be sent; this process is completed by the intelligent data distribution module deployed on the satellite antenna.

[0089] Specifically, the intelligent data allocation module is used to obtain the future network throughput predicted by the network performance prediction module and to perform data segmentation and allocation for different types of user data. When performing data allocation, the module also determines the data transmission strategy. That is, whether redundant transmission is required. When redundant transmission is required, erasure coding technology is used to construct redundant data. At the same time, the intelligent data allocation module uses deep intensity learning to simultaneously consider user data characteristics such as data volume and future network status (determined by historical network status and future network status) to make the optimal data allocation strategy.

[0090] Step S06: The satellite antenna distributes data to each link for data transmission.

[0091] In this embodiment, a multi-link data transmission device in a low-orbit satellite network is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made are not repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0092] This embodiment provides a multi-link data transmission device in a low-orbit satellite network, which is applied to a satellite antenna. The satellite antenna is connected to multiple low-orbit satellites at the same time to form multiple links, such as Figure 6 As shown, the device comprises:

[0093] The data receiving module 61 is used to receive a user's data transmission request and data to be transmitted.

[0094] The link quantity determination module 62 is used to detect the status of the low-orbit satellite and determine the number of links used for data transmission.

[0095] The network status prediction module 63 is used to predict future network performance data based on the historical network performance data of each link using a network performance prediction model.

[0096] The intelligent data allocation module 64 is used to determine the data allocation strategy for each link based on the future network performance data and the data characteristics of the data to be transmitted, and allocate the data to be transmitted on each link according to the data allocation strategy.

[0097] In an optional implementation, the device further includes: a redundant transmission triggering module, configured to determine whether to trigger redundant transmission according to link quality and data characteristics of the data to be transmitted.

[0098] In an optional implementation, the redundant transmission trigger module is specifically used to: determine the link quality threshold for triggering redundant transmission based on the data characteristics of the data to be transmitted; obtain the link quality, judge the size of the link quality and the link quality threshold; trigger redundant transmission when the link quality is less than the link quality threshold.

[0099] In an optional implementation, the redundant transmission trigger module is further used to: determine parameters of the erasure code according to data characteristics of the data to be transmitted; and construct redundant data for redundant transmission using the erasure code technology according to the parameters of the erasure code.

[0100] In an optional implementation, the historical network performance data includes historical throughput data, and the network status prediction module is specifically used to: obtain network performance data of multiple time steps at each link's historical moments, and normalize the network performance data; process the network performance data using a sliding window to construct training samples; train a long short-term memory network model that includes an attention mechanism based on the historical network performance data, use Adam as an optimizer during the training process, use mean square error as a loss function, adjust the model parameters, and obtain a network performance prediction model; use the network performance prediction model to predict network performance data for a preset number of time steps in the future.

[0101] In an optional implementation, the intelligent data allocation module is specifically used to: use a deep reinforcement learning model to determine the data allocation strategy for each link based on future network performance data and data characteristics of the data to be transmitted.

[0102] In an optional embodiment, the deep reinforcement learning model includes state, behavior and reward, the state includes future network performance data and data characteristics of data to be transmitted, the data characteristics include the amount of data to be transmitted and the tolerated delay of transmission, the behavior includes the distribution method of the transmitted data on the link, and the reward is determined according to the efficiency and cost of data transmission.

[0103] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0104] The present embodiment provides a multi-link data transmission system in a low-orbit satellite network, the system comprising: a user terminal, a satellite antenna and a low-orbit satellite network, the low-orbit satellite network comprising multiple low-orbit satellites, the satellite antenna using a preconfigured communication module to simultaneously establish connections with multiple low-orbit satellites to form multiple links; the user terminal is used to send data transmission requests and data to be transmitted to the satellite antenna; the satellite antenna is used to detect the status of the low-orbit satellite and determine the number of links used for data transmission; based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data; based on the future network performance data and the data characteristics of the data to be transmitted, a data allocation strategy for each link is determined, and the data to be transmitted is allocated according to the data allocation strategy and then transmitted on each link.

[0105] Specifically, Figure 7 As shown, the multi-link data transmission system can work according to the following process:

[0106] First, the user terminal (i.e., the transmitter) sends a data transmission request to the satellite antenna through the wireless network between the user and the satellite antenna. The satellite antenna then detects the link between the user and the satellite that can be used for data transmission and analyzes the user's data request (including data size and service type). Next, the satellite antenna needs to predict and evaluate the network status of the currently available link. After performing the network status evaluation, the satellite antenna chooses whether to adopt a redundant data transmission strategy for different service types and network conditions. If a redundant data transmission strategy is required, the satellite antenna will add redundant data to the user data. After that, the satellite antenna determines the proportion of data allocated to each link based on the user's data characteristics and network status. This process uses deep reinforcement learning methods to determine how the data is divided on the available links. Next, the satellite antenna divides the data according to the data allocation decision. After the division is completed, the satellite antenna allocates the data blocks to the available links for transmission. When the data blocks are transmitted on different available links, they are transmitted to the low-orbit satellite (receiving end) through the ground station. After the low-orbit satellite network receives the data blocks transmitted by different links, it synchronizes and merges the data to obtain the complete data to be transmitted.

[0107] The embodiment of the present invention also provides a computer device having the above Figure 6 A multi-link data transmission device in a low-orbit satellite network is shown.

[0108] See also Figure 8 , Figure 8 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 8As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 8 A processor 10 is taken as an example.

[0109] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0110] The memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiment.

[0111] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created by the use of a computer device based on the presentation of a small program landing page, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0112] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0113] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0114] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0115] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0116] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A multi-link data transmission method in a low-orbit satellite network, characterized in that: Applied to a satellite antenna, the satellite antenna is connected to multiple low-orbit satellites at the same time to form multiple links, and the method includes: Receive user's data transmission request and data to be transmitted; Detect the status of low-orbit satellites and determine the number of links for data transmission; Based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data, wherein the network performance prediction model is constructed using a long short-term memory network; The data allocation strategy for each link is determined based on the future network performance data and the data characteristics of the data to be transmitted, and the data to be transmitted is allocated on each link according to the data allocation strategy.

2. The method according to claim 1, characterized in that The method further comprises: Whether to trigger redundant transmission is determined based on link quality and data characteristics of the data to be transmitted.

3. The method according to claim 2, characterized in that Triggering redundant transmission is determined based on link quality and data characteristics of the data to be transmitted, including: Determining a link quality threshold for triggering redundant transmission according to data characteristics of the data to be transmitted; Obtain link quality, and determine the link quality and link quality threshold; Redundant transmission is triggered when the link quality is less than a link quality threshold.

4. The method according to claim 3, characterized in that After triggering redundant transmission when the link quality is less than a link quality threshold, the method further includes: Determining the parameters of the erasure code according to the data characteristics of the data to be transmitted; According to the parameters of the erasure code, the erasure code technology is used to construct redundant data for redundant transmission.

5. The method according to claim 1, characterized in that Historical network performance data includes historical throughput data. Based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data, including: Obtain network performance data of multiple time steps at each link's historical moment, and normalize the network performance data; Use sliding windows to process network performance data and construct training samples; Based on historical network performance data, the long short-term memory network model with attention mechanism is trained. During the training process, Adam is used as the optimizer, mean square error is used as the loss function, and the model parameters are adjusted to obtain the network performance prediction model. The network performance prediction model is used to predict the network performance data for a preset time step in the future.

6. The method according to claim 1, characterized in that Determining a data allocation strategy for each link based on the future network performance data and data characteristics of the data to be transmitted includes: A deep reinforcement learning model is used to determine the data allocation strategy for each link based on the future network performance data and the data characteristics of the data to be transmitted.

7. The method according to claim 6, characterized in that The deep reinforcement learning model includes state, behavior and reward. The state includes future network performance data and data characteristics of data to be transmitted. The data characteristics include the amount of data to be transmitted and the tolerated delay of transmission. The behavior includes the distribution method of the transmitted data on the link. The reward is determined according to the efficiency and cost of data transmission.

8. A multi-link data transmission device in a low-orbit satellite network, characterized in that: Applied to a satellite antenna, the satellite antenna is connected to multiple low-orbit satellites at the same time to form multiple links, and the device includes: A data receiving module, used to receive a user's data transmission request and data to be transmitted; A link quantity determination module, used for detecting the state of the low-orbit satellite and determining the number of links used for data transmission; A network status prediction module is used to predict future network performance data based on the historical network performance data of each link using a network performance prediction model; The intelligent data allocation module is used to determine the data allocation strategy for each link based on the future network performance data and the data characteristics of the data to be transmitted, and allocate the data to be transmitted in each link according to the data allocation strategy.

9. A multi-link data transmission system in a low-orbit satellite network, characterized in that: The system comprises: a user terminal, a satellite antenna and a low-orbit satellite network, wherein the low-orbit satellite network comprises a plurality of low-orbit satellites, and the satellite antenna uses a pre-configured communication module to simultaneously establish connections with the plurality of low-orbit satellites to form a plurality of links; The user terminal is used to send a data transmission request and data to be transmitted to the satellite antenna; The satellite antenna is used to detect the status of low-orbit satellites and determine the number of links used for data transmission; based on the historical network performance data of each link, a network performance prediction model is used to predict future network performance data; based on the future network performance data and the data characteristics of the data to be transmitted, the data allocation strategy for each link is determined, and the data to be transmitted is allocated according to the data allocation strategy and then transmitted on each link.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the multi-link data transmission method in a low-orbit satellite network as described in any one of claims 1 to 7.

Images