Intelligent transmission control method for burst meteor trail channel

Abstract

The application discloses a kind of intelligent transmission control methods for sudden burst meteor trail channel, comprising: receiving end receives the data frame transmitted by sending end, extracts embedded pilot symbol from data frame;According to the signal-to-noise ratio sequence of channel estimation pilot symbol;The signal-to-noise ratio sequence of channel is input into lightweight sparse channel prediction network, and the future signal-to-noise ratio sequence is obtained;The future signal-to-noise ratio sequence is input into transmission scheme decision network to determine the transmission scheme adapted, for the transmission of next data frame;Transmission scheme includes bandwidth configuration, modulation mode and channel coding rate;The transmission scheme adapted is packaged into reply frame, and the reply frame is returned to sending end, to realize transmission scheme intelligent control.The application can perceive channel state variation trend in advance, and realizes the dynamic optimization selection of transmission parameter using artificial intelligence technology, improves channel resource utilization and data transmission efficiency, enhances the adaptive capacity of system to fast time-varying meteor trail channel environment.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology, specifically relating to an intelligent transmission control method for burst meteor trail channels. Background Technology

[0002] When a meteor enters the Earth's atmosphere, it burns up due to high-speed friction, creating an ionized plasma trail. These meteor trails can temporarily reflect radio waves, enabling long-distance signal transmission over distances of up to 2000 kilometers—this is the principle of meteor trail communication. This communication method relies on meteor trails for intermittent signal transmission, making it naturally suitable for long-distance data exchange with tolerance for delays. Furthermore, its high security and stable link characteristics make it crucial for civilian and military emergency communications. However, compared to conventional communication methods, meteor trail communication suffers from low throughput and long information latency, severely limiting its widespread application. The problems stem primarily from two aspects: first, the highly random occurrence of meteors results in sudden, intermittent meteor trail channels with short availability and drastic fluctuations in status; second, existing communication systems have relatively fixed transmission modes, unable to flexibly adjust transmission rates based on real-time channel conditions, further exacerbating problems such as long latency and low system capacity.

[0003] In the field of wireless communication, adapting transmission schemes based on channel conditions, such as adaptive modulation and coding mechanisms, to improve the transmission efficiency of communication systems has become a widely researched topic in order to improve the transmission efficiency of communication systems. A common approach is to simulate various channel characteristics and generate a transmission scheme decision table through extensive simulations, allowing the system to dynamically select the optimal transmission scheme based on this table. However, this method has two major drawbacks when applied to meteor trail communication. First, the inherent characteristics and time-varying patterns of meteor trail channels still lack systematic research, resulting in simulation environments that cannot accurately depict channel characteristics and their dynamic changes, leading to significant deviations from real-world channel scenarios. This makes it quite difficult to construct transmission scheme decision tables or methods through simulation to determine the transmission scheme. Second, even if a transmission scheme decision table or method is obtained, the transmission scheme obtained based on the current channel conditions may fail due to the significant fast fading characteristics of meteor trail channels. Therefore, it is necessary to introduce artificial intelligence-based technologies to design intelligent transmission control methods with real-time transmission scheme prediction capabilities. Summary of the Invention

[0004] To address the aforementioned problems in the existing technology, this invention provides an intelligent transmission control method for burst meteor trail channels. In a first aspect, the present invention provides an intelligent transmission control method for burst meteor trail channels, comprising: The receiver receives data frames transmitted by the transmitter, extracts embedded pilot symbols from the data frames, and estimates the signal-to-noise ratio sequence of the channel based on the pilot symbols. The receiver inputs the signal-to-noise ratio (SNR) sequence of the channel into a trained lightweight sparse channel prediction network to predict the SNR changes of the received signal in the future, thus obtaining the future SNR sequence. The receiving end inputs the future signal-to-noise ratio sequence into the trained transmission scheme decision network to determine the appropriate transmission scheme for the transmission of the next data frame; The receiving end encapsulates the adapted transmission scheme into a reply frame and sends the reply frame back to the sending end to achieve intelligent transmission control of the transmission scheme.

[0005] In one embodiment of the present invention, estimating the signal-to-noise ratio sequence of the channel based on pilot symbols includes: Based on the pilot symbols, the estimated value of the channel matrix is ​​obtained using the least squares estimation algorithm; Estimate the signal-to-noise ratio sequence of the channel based on the estimated value of the channel matrix.

[0006] In one embodiment of the present invention, the expression for the estimated value of the channel matrix is ​​as follows: ; in, This represents the estimated value of the channel matrix. Indicates pilot symbol, Indicates the original pilot symbol, T This indicates the matrix transpose.

[0007] In one embodiment of the present invention, the expression for the signal-to-noise ratio sequence of the channel is as follows: ; in, The signal-to-noise ratio sequence representing the channel. This represents the estimated value of the channel matrix. Indicates pilot symbol, This represents the original pilot symbol.

[0008] In one embodiment of the present invention, the lightweight sparse channel prediction network operates as follows: The signal-to-noise ratio (SNR) sequence of the channel is obtained by sliding aggregation through one-dimensional convolution operation; The aggregated signal-to-noise ratio sequence is separated according to the period to obtain several sub-sequences; The subsequence is input into a neural network with shared parameters for prediction, and the predicted future subsequence is obtained. The future subsequences are recombined to generate the future signal-to-noise ratio sequence.

[0009] In one embodiment of the present invention, the forward propagation expression in the transmission scheme decision network is as follows: ; ; ; in, This represents the output of the first fully connected layer in the transmission scheme decision network. This represents the nonlinear activation function of a neuron. This represents the trainable weight matrix of the first fully connected layer. Indicates the future signal-to-noise ratio sequence. This represents the bias vector of the first fully connected layer. This represents the output of the second fully connected layer in the transmission scheme decision network. This represents the trainable weight matrix of the second fully connected layer. This represents the bias vector of the second fully connected layer. express Estimated performance scores for each transmission scheme This represents the trainable weight matrix of the third fully connected layer. This represents the bias vector of the third fully connected layer.

[0010] In one embodiment of the present invention, the training process of the transmission scheme decision network is as follows: The actual performance score of each transmission scheme is obtained based on the future signal-to-noise ratio sequence; Based on the actual performance scores of each transmission scheme, the transmission scheme decision network is trained through supervised learning, and mean squared error is used as the loss function of the transmission scheme decision network to complete the training of the transmission scheme decision network. In the training process of the transmission scheme decision network, the future signal-to-noise ratio sequence is used as the input of the transmission scheme decision network, the estimated performance score is used as the output of the transmission scheme decision network, and the actual performance score is used as the training sample label of the transmission scheme decision network.

[0011] In a second aspect, the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the steps of the intelligent transmission control method for burst meteor trail channels provided in the first aspect of the present invention.

[0012] Thirdly, the present invention provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, it implements the steps of the intelligent transmission control method for burst meteor trail channels provided in the first aspect of the present invention.

[0013] The beneficial effects of this invention are: The solution provided in this invention addresses the problem of transmission scheme decision-making failures caused by the strong burst and intermittent characteristics and unclear spatiotemporal variation patterns of meteor trail channels in existing methods. This invention employs a lightweight sparse channel prediction network to obtain the future signal-to-noise ratio (SNR) sequence and trains the neural network using collected training data to obtain the transmission scheme decision-making network. Leveraging the powerful feature learning capabilities of artificial intelligence, this invention can output transmission schemes that better reflect the actual channel conditions, effectively solving the pain point of decision failure in traditional methods and significantly improving the effectiveness of transmission decisions. To meet the stringent requirement of millisecond-level real-time response in meteor trail communication, the lightweight sparse channel prediction network used in this invention simplifies the cross-cycle prediction task by periodically separating the original SNR sequence. Compared to traditional prediction networks, this invention significantly reduces model complexity and the number of parameters, achieving a lightweight design, greatly shortening the running time of channel prediction and transmission scheme decision-making, significantly improving the timeliness of transmission scheme decisions, and ensuring efficient utilization of limited channel transmission time. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating an intelligent transmission control method for burst meteor trail channels provided in an embodiment of the present invention. Figure 2 The graph shows the change of the loss value of the lightweight sparse channel prediction network with the number of training iterations in an intelligent transmission control method for burst meteor trail channels provided in an embodiment of the present invention. Figure 3 The graph shows the performance score of the meteor trail communication system under different methods as a function of the number of training iterations. Figure 4 The graph shows the change in bit error rate of the meteor trail communication system with the number of training iterations under different methods. Figure 5 A graph showing the change in data throughput of a meteor trail communication system with the number of training iterations under different methods; Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0015] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0016] This invention addresses the problem that meteor trail channels are characterized by strong randomness and sudden intermittent nature, while existing transmission modes are relatively fixed and difficult to dynamically adjust the transmission scheme according to the real-time channel status. It provides an intelligent transmission control method, electronic device, and storage medium for burst meteor trail channels, thereby solving the technical problem that the transmission scheme in meteor trail communication systems is difficult to adapt to rapid channel changes and significantly improving the overall throughput of the system.

[0017] Below, we will first introduce an intelligent transmission control method for burst meteor trail channels provided by an embodiment of the present invention.

[0018] The implementation scenarios and parameter settings of the intelligent transmission control method for burst meteor trail channels provided in this embodiment of the invention can be as follows: Two meteor trail communication stations communicate point-to-point in half-duplex mode via a meteor trail channel. The master station (transmitter) transmits data frames to the slave station (receiver), and upon receiving the data frames, the slave station sends a reply frame back to the master station, containing the transmission scheme for the next data frame. Both the master and slave stations have a transmit power of 1000W and an antenna gain of 12dBi. The distance between the two stations is 1000km, and the meteor trail altitude is 80-120km. The data frame transmission scheme is a combination of bandwidth, modulation scheme, and coding rate. The bandwidth includes 20kHz, 40kHz, and 80kHz. The modulation schemes include QPSK (Quadrature Phase Shift Keying), 8QAM (Quadrature Amplitude Modulation), 16QAM, 32QAM, and 64QAM. The channel coding can use LDPC (Low Density Parity Check) coding. The coding rate can include 1 / 5, 3 / 10, 2 / 5, 1 / 2, 3 / 5, 5 / 6, and 8 / 9.

[0019] like Figure 1 As shown in the figure, the intelligent transmission control method for burst meteor trail channels provided by this embodiment of the invention may include the following steps: S1, the receiving end receives the data frame transmitted by the sending end, extracts the embedded pilot symbols from the data frame, and estimates the signal-to-noise ratio sequence of the channel based on the pilot symbols.

[0020] Specifically, embodiments of the present invention use a least squares estimation algorithm to estimate the channel matrix. The optimal estimate is achieved by minimizing the sum of squared errors between the predicted and actual observations. Assume the original pilot symbol transmitted by the transmitter is... Channel noise is nThen the pilot symbols received at the receiving end and estimation error It can be represented in the following form: ; .

[0021] By minimizing the sum of the squared errors between the received and transmitted signals, the expression for the estimated value of the channel matrix can be obtained as follows: ; in, This represents the estimated value of the channel matrix. Indicates pilot symbol, Indicates the original pilot symbol, T Indicates matrix transpose. This refers to the argmin function. Represents the channel matrix, This represents the square of the L2 norm.

[0022] In step S1, estimating the signal-to-noise ratio sequence of the channel based on the pilot symbols may include: Based on the pilot symbols, the estimated value of the channel matrix is ​​obtained using the least squares estimation algorithm; Estimate the signal-to-noise ratio sequence of the channel based on the estimated value of the channel matrix.

[0023] The expression for the signal-to-noise ratio sequence of the channel is as follows: ; in, The signal-to-noise ratio sequence representing the channel. This represents the estimated value of the channel matrix. Indicates pilot symbol, This represents the original pilot symbol.

[0024] S2, the receiver inputs the channel's signal-to-noise ratio (SNR) sequence into a trained lightweight sparse channel prediction network to predict the future SNR changes of the received signal and obtain the future SNR sequence.

[0025] The operation of a lightweight sparse channel prediction network can be described as follows: The signal-to-noise ratio (SNR) sequence of the channel is obtained by sliding aggregation through one-dimensional convolution operation; The aggregated signal-to-noise ratio sequence is separated according to the period to obtain several sub-sequences; The subsequence is input into a neural network with shared parameters for prediction, and the predicted future subsequence is obtained. The future subsequences are recombined to generate the future signal-to-noise ratio sequence.

[0026] Specifically, the signal-to-noise ratio sequence of the channel is defined. It can be a period of The time series is used. To enable the model to capture relevant information between data within a period, sliding aggregation is performed on the original sequence before separating the signal-to-noise ratio sequence into subsequences. Each aggregated data point contains information about its neighboring points within its period. Furthermore, since the aggregated value is a weighted average of surrounding points, the influence of outliers can be mitigated. This embodiment of the invention implements sliding aggregation through a one-dimensional convolution operation, using a kernel size that can be... The process can be described as follows: ; in, This represents the signal-to-noise ratio sequence after aggregation. This represents a one-dimensional convolution operation.

[0027] The aggregated signal-to-noise ratio sequence is periodically separated into There are subsequences, each with a length of 0. .

[0028] The process of separating subsequences can be represented as: ; in, Indicates the first The first subsequence One value, The first element in the aggregated signal-to-noise ratio sequence represents the... One value, This indicates the length of each subsequence.

[0029] Since time series sequences often exhibit periodicity, the resulting subsequences are input into a neural network with shared parameters for prediction. The actual prediction process is equivalent to using the separated subsequences as network input to predict future subsequences: ; in, Indicates the first i A future subsequence, Indicates the first i Subsequences This represents a neural network with shared parameters used for prediction.

[0030] Will A length of The future subsequences are recombined to generate a sequence of total length. K The future signal-to-noise ratio sequence.

[0031] The process of recombination can be represented as: ; in, The first term in the future signal-to-noise ratio sequence One value, Indicates the first i The nth future subsequence Values.

[0032] In embodiments of the present invention, a supervised learning method can be selected to train a lightweight sparse channel prediction network, which uses the channel's signal-to-noise ratio sequence. As network input, future signal-to-noise ratio sequence As the network output, the estimated future true signal-to-noise ratio sequence is obtained. The training sample labels are used as the basis for the network. Simultaneously, the classic mean squared error is employed as the loss function for the lightweight sparse channel prediction network. This function measures the predicted values. Compared with the true value The difference between them is expressed by the following formula: ; in, This represents the loss function of a lightweight sparse channel prediction network. This represents the L2 norm.

[0033] After the lightweight sparse channel prediction network training converges, the receiver inputs the channel's signal-to-noise ratio sequence into the lightweight sparse channel prediction network to predict the future signal-to-noise ratio changes of the received signal.

[0034] For example, the neural network sharing parameters in the lightweight sparse channel prediction network is a two-layer multilayer perceptron. Here, the period of the original sequence... It can be set to 2, which is the length of the input signal-to-noise ratio sequence. L It can be set to 6, which is the length of the future signal-to-noise ratio sequence output by the network. K The learning rate can be set to 10. For lightweight sparse channel prediction networks, the learning rate can be set to 1×10. -4 The number of training iterations can be set to 2×10. 4 .

[0035] S3, the receiver inputs the future signal-to-noise ratio sequence into the trained transmission scheme decision network to determine the appropriate transmission scheme for the transmission of the next data frame.

[0036] The transmission scheme decision network is a three-layer MLP (Multi-Layer Perceptron), and its input is the future signal-to-noise ratio sequence predicted in step S2. The output is an estimated performance score for all transmission schemes (bandwidth, modulation scheme, and coding rate). .

[0037] The expression for forward propagation in the transmission scheme decision network is as follows: ; ; ; in, This represents the output of the first fully connected layer in the transmission scheme decision network. This represents the nonlinear activation function of a neuron. This represents the trainable weight matrix of the first fully connected layer. Indicates the future signal-to-noise ratio sequence. This represents the bias vector of the first fully connected layer. This represents the output of the second fully connected layer in the transmission scheme decision network. This represents the trainable weight matrix of the second fully connected layer. This represents the bias vector of the second fully connected layer. express Estimated performance scores for each transmission scheme This represents the trainable weight matrix of the third fully connected layer. This represents the bias vector of the third fully connected layer.

[0038] During the training of the transmission scheme decision network, this embodiment of the invention considers different transmission schemes to represent different data carrying capacities and bit error rates, thereby obtaining different future signal-to-noise ratio sequences. Input M The actual performance score of a transmission scheme is expressed as follows: ; in, Indicates the first m The actual performance score of each transmission scheme Indicates the number of symbols contained in a data segment of a data frame. Indicates the first m Channel coding rate of each transmission scheme, Indicates the first m The modulation order of a transmission scheme, Indicates the first m Bit error rate under each transmission scheme.

[0039] The training process of the transmission scheme decision network can be as follows: The actual performance score of each transmission scheme is obtained based on the future signal-to-noise ratio sequence; Based on the actual performance scores of each transmission scheme, the transmission scheme decision network is trained through supervised learning, and mean squared error is used as the loss function of the transmission scheme decision network to complete the training of the transmission scheme decision network. In the training process of the transmission scheme decision network, the future signal-to-noise ratio sequence is used as the input of the transmission scheme decision network, the estimated performance score is used as the output of the transmission scheme decision network, and the actual performance score is used as the training sample label of the transmission scheme decision network.

[0040] Specifically, upon obtaining M After determining the actual performance scores of each transmission scheme, a supervised learning approach is chosen to train the transmission scheme decision network, which will then use the future signal-to-noise ratio sequence. As network input, Actual performance score as network output As the training sample labels, the classic mean squared error is used as the loss function for the transmission scheme decision network, and its expression is as follows: ; in, This represents the loss function of the transmission scheme decision network.

[0041] The receiver inputs the future signal-to-noise ratio sequence into the transmission scheme decision network. This is to determine the appropriate transmission scheme (bandwidth, modulation method, and coding rate) for the transmission of the next data frame.

[0042] Due to the output of the transmission scheme decision network Each element The corresponding one is to adopt the first m Given a given transmission scheme, the estimated performance score of the next transmitted data frame. Therefore, the final determined suitable transmission scheme can be expressed as: ; in, Indicates the adapted transmission scheme. This refers to the argmax function.

[0043] For example, the learning rate for training the transmission scheme decision network can be set to 5 × 10⁻⁶. -5 The number of training iterations can be set to 1×10. 5 Furthermore, a performance evaluation is performed every 50 training steps. To evaluate the performance of the transmission scheme decision network, the master station can transmit data to the slave station over a 6-minute meteor trail channel and record its performance score, bit error rate, and data throughput.

[0044] S4, the receiving end encapsulates the adapted transmission scheme into a reply frame and sends the reply frame back to the sending end to realize intelligent transmission control of the transmission scheme.

[0045] Specifically, the receiving end encapsulates the adapted transmission scheme into a reply frame and sends the reply frame back to the sending end. The sending end extracts the bandwidth, modulation scheme, and coding rate indicated by the transmission scheme in the reply frame and sends a data frame to the receiving end in the next transmission time slot.

[0046] The intelligent transmission control method for burst meteor trail channels provided in this invention aims to improve the transmission efficiency of meteor trail communication systems in bursty, intermittent, and fast-fading channels. This intelligent transmission control method achieves the following two design objectives: First, it has channel prediction capabilities. Due to the fast time-varying fading characteristics of the channel, the receiver needs to predict the channel state at future moments and select an appropriate transmission scheme for the transmitter based on this prediction.

[0047] Second, ensure low decision-making latency. Channel prediction and transmission scheme selection must be completed in a very short time to ensure their effectiveness, adapt to the rapid dynamic changes of the channel, and make full use of the limited available channel time.

[0048] To address the problem that existing transmission scheme decision-making methods fail due to the strong burst and intermittent characteristics of meteor trail channels and unclear spatiotemporal variation patterns, this invention employs a lightweight sparse channel prediction network to obtain the future signal-to-noise ratio sequence. The neural network is then trained using collected training data to obtain the transmission scheme decision-making network. Leveraging the powerful feature learning capabilities of artificial intelligence, this intelligent transmission control method can output transmission schemes that better reflect the actual channel conditions, effectively solving the pain point of decision failure in traditional methods and significantly improving the effectiveness of transmission decisions.

[0049] To address the stringent requirement of millisecond-level real-time response in meteor trail communication for transmission scheme decisions, the lightweight sparse channel prediction network employed in this invention simplifies the cross-cycle prediction task by periodically separating the original signal-to-noise ratio sequence. Compared to prediction networks such as RNNs (Recurrent Neural Networks), LSTMs (Long Short-Term Memory), and Transformers, this network significantly reduces model complexity and the number of parameters, achieving a lightweight design. This greatly shortens the runtime of channel prediction and transmission scheme decision-making, significantly improves the timeliness of transmission scheme decisions, and ensures efficient utilization of limited channel transmission time.

[0050] The following verification of the beneficial effects of the intelligent transmission control method for burst meteor trail channels provided in this embodiment of the invention is conducted through specific implementation results and effect verification.

[0051] The embodiments of this invention use the following two comparison algorithms as a reference: Lookup table method: The transmission scheme is selected based on a predefined mapping between the estimated signal-to-noise ratio and the corresponding transmission scheme; Fixed transmission scheme: The sending end uses a fixed transmission scheme to transmit data frames, with a bandwidth of 20kHz, a modulation method of 8QAM, and an LDPC coding rate of 1 / 5.

[0052] The graph shows how the loss value of a lightweight sparse channel prediction network changes with the number of training iterations. Figure 2 As shown, the trend of the loss function of the lightweight sparse channel prediction network demonstrates its good convergence characteristics during training. In the initial training phase (around 0-3000 steps), the loss value rapidly decreases from approximately 6.6 to below 3, indicating that the lightweight sparse channel prediction network can quickly learn the main temporal features of the input channel data. Subsequently, in the 3000-8000 step range, the loss value continues to decrease, but the rate of decrease gradually slows, decreasing from approximately 3 to approximately 1. During this stage, the network begins to learn more detailed channel variation patterns and finely optimizes the model parameters. After 8000 steps, the loss function decreases more gradually and stabilizes around 0.8, with only minor fluctuations, indicating that the model has basically converged, the training process is stable, and there are no obvious oscillations or divergences. Finally, after approximately 20,000 training steps, the network reaches a stable state, demonstrating that the lightweight sparse channel prediction network can effectively fit the channel temporal features and possesses good training stability and convergence performance.

[0053] The performance score of the meteor trail communication system varies with the number of training iterations under different methods, as shown in the figure. Figure 3 As shown, the performance score obtained by the embodiment of the present invention is 33.8% and 1723.42% higher than that of the lookup table method and the fixed transmission scheme, respectively. The lookup table method, due to its difficulty in adapting to fast-fading meteor trail channels by selecting the transmission scheme for the next transmission slot based on the current signal-to-noise ratio, often becomes ineffective. Meanwhile, the fixed transmission scheme performs significantly worse than any adaptive transmission method. In contrast, the embodiment of the present invention predicts the future channel signal-to-noise ratio, obtains the channel state change trend in advance, and combines artificial intelligence transmission control technology to achieve dynamic optimization selection of the transmission scheme. This allows for more effective matching of channel conditions, improves channel resource utilization and transmission efficiency, and ultimately achieves significantly better performance than the comparative methods.

[0054] The bit error rate of the meteor trail communication system varies with the number of training iterations under different methods, as shown in the figure. Figure 4 As shown, after training convergence, the final bit error rate of the embodiment of the present invention is approximately 0.0011, while the bit error rate of the lookup table method is approximately 0.0025. This demonstrates the superior ability of the embodiment of the present invention in reducing bit error rate and improving transmission efficiency. Furthermore, although the fixed transmission scheme achieves a lower average bit error rate due to its focus on high-reliability transmission, the cost is a significant decrease in data throughput.

[0055] The graph shows the data throughput of the meteor trail communication system as a function of the number of training iterations under different methods, as shown in the figure. Figure 5 As shown, the data throughput corresponding to this embodiment of the invention is approximately 920kb. In comparison, the lookup table method and the fixed transmission scheme achieve data throughputs of 708kb and 93kb, respectively. It is worth noting that compared to the lookup table method and the fixed transmission scheme, the intelligent transmission control method for burst meteor trail channels provided by this embodiment of the invention achieves performance improvements of 29.9% and 889.2%, respectively. This demonstrates that selecting the transmission scheme based on the predicted signal-to-noise ratio of the next transmission slot achieves better performance in fast-fading meteor trail channels compared to the lookup table method, which relies on the current signal-to-noise ratio. Furthermore, compared to the fixed transmission scheme, this intelligent transmission control method significantly improves data throughput.

[0056] Secondly, embodiments of the present invention also provide an electronic device, such as... Figure 6 As shown, it includes a processor 001, a communication interface 002, a memory 003, and a communication bus 004, wherein the processor 001, the communication interface 002, and the memory 003 communicate with each other through the communication bus 004. The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the steps of any of the intelligent transmission control methods for burst meteor trail channels provided in the first aspect of the present invention.

[0057] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0058] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0059] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0060] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0061] The method provided in this invention can be applied to electronic devices. Specifically, the electronic device can be a desktop computer, a portable computer, a smart mobile terminal, a server, etc. No limitation is made herein; any electronic device that can implement this invention falls within the protection scope of this invention.

[0062] Thirdly, corresponding to the intelligent transmission control method for burst meteor trail channels provided in the first aspect, this embodiment of the invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps of any of the intelligent transmission control methods for burst meteor trail channels provided in the first aspect of this invention.

[0063] For the electronic device / storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and relevant details can be found in the description of the method embodiments.

[0064] It should be noted that the electronic device and storage medium in the embodiments of the present invention are respectively electronic devices and storage media that apply the above-mentioned intelligent transmission control method for burst meteor trail channels. Therefore, all embodiments of the above-mentioned intelligent transmission control method for burst meteor trail channels are applicable to the electronic device and storage medium, and can achieve the same or similar beneficial effects.

[0065] It should be noted that, in the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0066] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A smart transmission control method for burst meteor trail channels, characterized in that, include: The receiving end receives the data frames transmitted by the sending end and extracts the embedded pilot symbols from the data frames; The signal-to-noise ratio sequence of the channel is estimated based on the pilot symbols; The receiver inputs the signal-to-noise ratio (SNR) sequence of the channel into a trained lightweight sparse channel prediction network to predict the SNR changes of the received signal in the future, thus obtaining the future SNR sequence. The receiving end inputs the future signal-to-noise ratio sequence into the trained transmission scheme decision network to determine the appropriate transmission scheme for the transmission of the next data frame; The receiving end encapsulates the adapted transmission scheme into a reply frame and sends the reply frame back to the sending end to achieve intelligent transmission control of the transmission scheme.

2. The intelligent transmission control method for burst meteor trail channels according to claim 1, characterized in that, The signal-to-noise ratio (SNR) sequence for estimating the channel based on pilot symbols includes: Based on the pilot symbols, the estimated value of the channel matrix is ​​obtained using the least squares estimation algorithm; Estimate the signal-to-noise ratio sequence of the channel based on the estimated value of the channel matrix.

3. The intelligent transmission control method for burst meteor trail channels according to claim 2, characterized in that, The expression for the estimated value of the channel matrix is ​​as follows: ； in, This represents the estimated value of the channel matrix. Indicates pilot symbol, Indicates the original pilot symbol, T This indicates the matrix transpose.

4. The intelligent transmission control method for burst meteor trail channels according to claim 2, characterized in that, The expression for the signal-to-noise ratio sequence of the channel is as follows: ； in, The signal-to-noise ratio sequence representing the channel. This represents the estimated value of the channel matrix. Indicates pilot symbol, This represents the original pilot symbol.

5. The intelligent transmission control method for burst meteor trail channels according to claim 2, characterized in that, The lightweight sparse channel prediction network operates as follows: The signal-to-noise ratio (SNR) sequence of the channel is obtained by sliding aggregation through one-dimensional convolution operation; The aggregated signal-to-noise ratio sequence is separated according to the period to obtain several sub-sequences; The subsequence is input into a neural network with shared parameters for prediction, and the predicted future subsequence is obtained. The future subsequences are recombined to generate the future signal-to-noise ratio sequence.

6. The intelligent transmission control method for burst meteor trail channels according to claim 1, characterized in that, The forward propagation expression in the transmission scheme decision network is as follows: ； ； ； in, This represents the output of the first fully connected layer in the transmission scheme decision network. This represents the nonlinear activation function of a neuron. This represents the trainable weight matrix of the first fully connected layer. Indicates the future signal-to-noise ratio sequence. This represents the bias vector of the first fully connected layer. This represents the output of the second fully connected layer in the transmission scheme decision network. This represents the trainable weight matrix of the second fully connected layer. This represents the bias vector of the second fully connected layer. express Estimated performance scores for each transmission scheme This represents the trainable weight matrix of the third fully connected layer. This represents the bias vector of the third fully connected layer.

7. The intelligent transmission control method for burst meteor trail channels according to claim 6, characterized in that, The training process of the transmission scheme decision network is as follows: The actual performance score of each transmission scheme is obtained based on the future signal-to-noise ratio sequence; Based on the actual performance scores of each transmission scheme, the transmission scheme decision network is trained through supervised learning, and mean squared error is used as the loss function of the transmission scheme decision network to complete the training of the transmission scheme decision network. In the training process of the transmission scheme decision network, the future signal-to-noise ratio sequence is used as the input of the transmission scheme decision network, the estimated performance score is used as the output of the transmission scheme decision network, and the actual performance score is used as the training sample label of the transmission scheme decision network.

8. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the steps of the intelligent transmission control method for burst meteor trail channels as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the intelligent transmission control method for burst meteor trail channels as described in any one of claims 1-7.

Images