A method, program, device, and storage medium for satellite communication window optimization based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment.

Abstract

This invention belongs to the field of satellite navigation and communication technology, specifically relating to a method, program, device, and storage medium for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio (SNR) information in dynamic environments. For communication satellites with communication and navigation payloads on the same platform, this invention directly predicts the SNR time series. For communication satellites with communication and navigation payloads on different platforms, it uses known historical SNR and position information of the navigation satellite to calculate the motion state of the communication positioning terminal antenna, thereby indirectly predicting the future SNR sequence of communication satellites with unknown historical SNR. Based on the directly or indirectly predicted communication satellite SNR sequence, a suitable communication window for the communication satellite is found, reducing the bit error rate under adverse channel conditions. This invention achieves the prediction of SNR changes for satellites with unknown historical SNR.

Description

Technical Field

[0001] This invention belongs to the field of satellite navigation and communication technology, and specifically relates to a method, program, device and storage medium for satellite communication window optimization based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment. Background Technology

[0002] Satellite communication systems, with their unique advantages of wide coverage and lack of geographical limitations, have become an indispensable part of the global information infrastructure, playing a crucial role in emergency communications, ocean navigation, and Internet of Things (IoT) monitoring. A fundamental prerequisite for establishing a communication link is that the terminal antenna must be aligned with the target satellite to maximize the received signal strength. Even with a theoretically omnidirectional antenna covering 360 degrees, its radiation pattern is not an ideal sphere, exhibiting gain fluctuations; therefore, an optimal communication angle still exists in practical communication. Traditional satellite terminals are typically deployed at fixed ground stations or on stable platforms. However, with the continuous expansion of application scenarios, a large number of terminals need to be deployed on platforms that are constantly in motion, such as: ① Ocean monitoring: buoys bobbing with the waves, small lifeboats; ② Transportation: vehicles traveling on rough roads, violently swaying trains; ③ Emergency equipment: tethered drones or temporarily deployed portable devices. In these "communication on the move" scenarios, the pitch, roll, and yaw movements of the terminal can cause the antenna to deviate from the satellite, resulting in signal attenuation, a surge in bit error rate, and ultimately a decline in communication quality or even link interruption, which seriously restricts the reliability and continuity of the service.

[0003] To address the challenges posed by motion, the industry has proposed solutions primarily from two dimensions: the link layer and the physical layer. Link Layer Strategy: Retransmission and Adaptive Mechanisms. This involves employing Automatic Repeat Request (ARQ) or Hybrid Automatic Repeat Request (HARQ) protocols combined with Adaptive Coding and Modulation (ACM) technology. This approach monitors channel conditions in real time, triggering data retransmissions or switching to a more robust modulation and coding scheme when link quality is poor. While this method significantly improves link reliability, it is essentially a "remedial" measure. It does not utilize antenna motion information, and frequent retransmissions not only introduce additional latency but also significantly increase communication power consumption. For battery-powered remote IoT terminals, power consumption is a key constraint on their lifespan. Physical Layer Solution: Antenna Servo Tracking System. This uses mechanical or electronically scanned phased array antennas, equipped with inertial measurement units (IMUs), gyroscopes, and other sensors and servo control units. This system can calculate carrier attitude changes in real time and drive the antenna beam to always accurately point towards the satellite. This is a "proactive prevention" solution with excellent performance. However, this solution requires the introduction of high-precision sensors, complex servo mechanisms, and control algorithms, which significantly increases the size, weight, cost (SWaP-C), and power consumption of the terminal, making it difficult to integrate into miniaturized, low-cost terminals that are extremely sensitive to size, cost, and power consumption, and thus lacking universality.

[0004] In summary, current solutions exhibit a significant "capability gap": retransmission solutions sacrifice energy efficiency and real-time performance, making them suitable for applications that are not latency-sensitive but prioritize cost. Servo tracking solutions guarantee performance but sacrifice SWaP-C, while for the increasing number of miniaturized, low-power mobile communication terminals (such as IoT buoys and portable terminals), there is an urgent need for a new stabilization technology that balances high performance, low power consumption, low complexity, and low cost. Summary of the Invention

[0005] The purpose of this invention is to provide a method, program, device, and storage medium for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment. By analyzing the carrier-to-noise ratio sequence of received satellite navigation signals, the communication time window is optimized.

[0006] A method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment includes the following steps:

[0007] For communication satellites with communication and navigation payloads on the same platform, obtain the historical signal-to-noise ratio (SNR) time series of the communication satellite, obtain the fitting function of the SNR of the communication satellite with respect to time based on the fitting model, and predict the SNR time series of the communication satellite at the current moment and within the predicted time period thereafter.

[0008] For communication satellites with different communication and navigation payload platforms, navigation satellites are selected to form a prediction combination. The historical signal-to-noise ratio (SNR) time series of each navigation satellite in the prediction combination are obtained. The average SNR at each sampling moment is calculated to obtain a historical SNR average sequence. The gradient mean and gradient standard deviation of the historical SNR average sequence are taken as time-domain features. Fourier transform is used to calculate the spectral amplitude of the historical SNR average sequence. A certain number of frequency indices with the largest spectral amplitude are taken as frequency-domain features. The time-domain features and frequency-domain features are combined to construct a feature vector. A prediction model is constructed and trained. The feature vector is input into the prediction model, and the prediction model outputs the predicted motion parameter vector of the communication positioning terminal. Based on the historical motion trajectory of the communication satellite, the antenna direction vector of the communication positioning terminal, the vector pointing from the antenna of the communication positioning terminal to the communication satellite, and the distance between the communication positioning terminal and the communication satellite are predicted for the current moment and the subsequent prediction time period. This allows for the prediction of the SNR time series of the communication satellite for the current moment and the subsequent prediction time period.

[0009] Select one of the middle positions of the signal-to-noise ratio time series of the communication satellite at the current time and within the predicted time period thereafter as the dynamic threshold; if the predicted signal-to-noise ratio of the communication satellite at the current time is greater than the dynamic threshold, the communication positioning terminal communicates with the communication satellite.

[0010] Furthermore, for communication satellites with different platforms for communication and navigation payloads, selection... The navigation satellites constitute the predicted combination Get the current time and before Within a time interval Signal-to-noise ratio time series of each navigation satellite ;

[0011] in, For the first Navigation satellites at sampling time The signal-to-noise ratio, the time step between each sampling time is , , , .

[0012] Furthermore, the average signal-to-noise ratio at each sampling time is calculated. The historical average signal-to-noise ratio sequence is obtained, and the gradient mean of the historical average signal-to-noise ratio sequence is taken. and gradient standard deviation As a temporal domain feature, specifically:

[0013]

[0014] ,

[0015] in, .

[0016] Calculate the historical signal-to-noise ratio average sequence using Fourier transform The spectral amplitude is selected based on the largest spectral amplitude. Frequency index As a frequency domain feature, the time domain feature and the frequency domain feature are combined to construct a feature vector. .

[0017] Furthermore, the feature vector Input the prediction model, and the prediction model outputs the predicted motion parameter vector of the communication and positioning terminal. ;

[0018] The current moment and afterwards The time points are used as the prediction interval. ;

[0019] Motion parameter vector prediction based on communication and positioning terminals Based on the historical motion trajectory of communication satellites, the prediction interval is obtained. Each time point Antenna direction vector prediction for local communication positioning terminal Vector prediction of the antenna pointing to the communication satellite of the communication positioning terminal. Distance prediction between communication positioning terminal and communication satellite This allows for the prediction of communication satellites within the prediction range. Each time point Signal-to-noise ratio of communication satellites The signal-to-noise ratio time series was obtained. ;

[0020]

[0021] in, Indicates the communication positioning terminal at a certain time point. The angle of deviation between the antenna main lobe and the direction of the communication satellite. ; The half-power beamwidth of the antenna for the communication positioning terminal; The antenna transmission power of the communication positioning terminal; This refers to the antenna noise power of the communication positioning terminal.

[0022] Furthermore, the training prediction model specifically includes:

[0023] Get Predicted Combinations Historical signal-to-noise ratio time series of various navigation satellites in China , The total number of historical sampling time points obtained; the historical motion trajectories of the communication positioning terminal, communication satellites, and each navigation satellite are obtained;

[0024] Calculate the average signal-to-noise ratio at each sampling time. The historical signal-to-noise ratio time-averaged sequence was obtained. ;

[0025] Set time steps Historical signal-to-noise ratio time-averaged series The data is split into segments, and each segment after splitting has a historical signal-to-noise ratio time-mean subsequence as follows: , Rebuild The index of each element in the text. , , ;

[0026] For each historical signal-to-noise ratio time mean subsequence Calculate the gradient mean of the average signal-to-noise ratio. and gradient standard deviation Fourier transform was used to calculate the time-mean subsequence of each historical signal-to-noise ratio. The spectral amplitude is selected, and the one with the largest spectral amplitude is chosen. Frequency index As a frequency domain feature, the gradient mean of the average signal-to-noise ratio is used. and gradient standard deviation As a time-domain feature, construct the time-mean subsequence of each historical signal-to-noise ratio. eigenvectors ;

[0027] Set the prediction time steps The time-mean subsequences of each historical signal-to-noise ratio Corresponding end time point After The time points are used as the prediction interval. Based on the historical motion trajectories of the communication positioning terminal, communication satellites, and navigation satellites, the position of the communication positioning terminal within the predicted range is obtained. Motion parameter vector in Obtain the prediction interval Each time point Antenna direction vector of the communication positioning terminal The antenna of the communication positioning terminal points to the vector of the communication satellite. Distance between the communication positioning terminal and the communication satellite The antenna of the communication positioning terminal is pointed towards the vector of the navigation satellite with the highest signal-to-noise ratio. ; ;

[0028] Calculate the prediction interval Each time point Signal-to-noise ratio of communication satellites This constitutes a signal-to-noise ratio time series. ;

[0029] eigenvectors Input the prediction model, and the prediction model outputs the predicted motion parameter vector of the communication and positioning terminal. Prediction of motion parameter vectors based on communication positioning terminals Calculate the prediction interval Each time point Prediction of antenna direction vector of communication positioning terminal The antenna of the communication positioning terminal points to the vector of the communication satellite. Distance between the communication positioning terminal and the communication satellite Then calculate the prediction interval. Each time point Prediction of signal-to-noise ratio of communication satellites This constitutes the signal-to-noise ratio prediction time series. ;

[0030] use The prediction model is trained using a set of samples.

[0031] Furthermore, the loss function for training the prediction model is:

[0032]

[0033] in, , , , These are the weighting coefficients;

[0034] For motion smoothing constraint loss:

[0035]

[0036] For direction error loss:

[0037]

[0038] To perform feature regularization loss, a feature regularization term is constructed for the feature vector. Prediction of motion parameter vectors of communication and positioning terminals Let the vector with more elements be denoted as . The vector with fewer elements is Vectors with more elements Perform clipping so that the clipped vector The number of elements in a vector with fewer elements Consistent;

[0039]

[0040] To compensate for the loss of spectral consistency, a spectral domain penalty mechanism based on Fourier transform is introduced:

[0041]

[0042] in, The Fourier transform is used to calculate the spectral amplitude of the sequence.

[0043] Furthermore, the signal-to-noise ratio time series of the communication satellite at the current moment and within the predicted time period thereafter... Choose one from the middle positions as the dynamic threshold. If the predicted signal-to-noise ratio of the communication satellite at the current moment is greater than the dynamic threshold, i.e. Then the communication positioning terminal communicates with the communication satellite.

[0044] A computer device / equipment / system includes a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above-described method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment.

[0045] A computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implements the steps of the above-described method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment.

[0046] A computer program product includes a computer program / instructions that, when executed by a processor, implement the steps of the satellite communication window optimization method based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment described above.

[0047] The beneficial effects of this invention are as follows:

[0048] This invention directly predicts the signal-to-noise ratio (SNR) time series for communication satellites with communication and navigation payloads on the same platform. For communication satellites with communication and navigation payloads on different platforms, it uses the known historical SNR and position information of the navigation satellites to calculate the motion state of the communication positioning terminal antenna, thereby indirectly predicting the future SNR sequence of communication satellites with unknown historical SNR. Based on the directly or indirectly predicted communication satellite SNR sequence, a suitable communication window for the communication satellite is found, reducing the bit error rate under adverse channel conditions. This invention achieves the prediction of SNR changes for satellites with unknown historical SNR. Attached Figure Description

[0049] Figure 1 This is an overall flowchart of an embodiment of the present invention.

[0050] Figure 2 This is an example of the prediction effect of direct prediction on satellite signal-to-noise ratio and peak communication point determination in an embodiment of the present invention.

[0051] Figure 3 This is a comparison chart of direct prediction of communication window selection, bit error rate, and throughput in an embodiment of the present invention.

[0052] Figure 4 This is a diagram illustrating the prediction effect of indirect prediction on satellite signal-to-noise ratio and the determination of peak communication points in an embodiment of the present invention.

[0053] Figure 5 This is a comparison chart of indirect prediction communication window selection, bit error rate, and throughput in an embodiment of the present invention. Detailed Implementation

[0054] The present invention will now be further described with reference to the accompanying drawings.

[0055] This invention provides a method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment, specifically including the following:

[0056] For communication satellites with communication and navigation payloads on the same platform, the historical signal-to-noise ratio (SNR) time series of the communication satellite is obtained. Based on the fitting model, the fitting function of the SNR of the communication satellite with respect to time is obtained, and the SNR time series of the communication satellite at the current moment and within the predicted time period thereafter is predicted. An item is selected as a dynamic threshold from the middle position of the SNR time series of the communication satellite at the current moment and within the predicted time period thereafter. If the predicted SNR value of the communication satellite at the current moment is greater than the dynamic threshold, the communication positioning terminal communicates with the communication satellite.

[0057] For communication satellites with different communication and navigation payload platforms, perform the following steps:

[0058] Step 1: Select The navigation satellites constitute the predicted combination Obtain the historical signal-to-noise ratio time series of each navigation satellite. , For the first Navigation satellites at sampling time The signal-to-noise ratio, the time step between each sampling time is , , , ; Obtain the historical motion trajectories of communication positioning terminals, communication satellites, and various navigation satellites;

[0059] Step 2: Calculate the average signal-to-noise ratio at each sampling time. The historical signal-to-noise ratio time-averaged sequence was obtained. ;

[0060]

[0061] Step 3: Set the number of time steps Historical signal-to-noise ratio time-averaged series The data is split into segments, and each segment after splitting has a historical signal-to-noise ratio time-mean subsequence as follows: , Rebuild The index of each element in the text. , , ;

[0062] Step 4: For each historical signal-to-noise ratio time-mean subsequence Calculate the gradient mean of the average signal-to-noise ratio. and gradient standard deviation ;

[0063] ,

[0064] in, ;

[0065] Step 5: Calculate the time-mean subsequence of each historical signal-to-noise ratio using Fourier transform. The spectral amplitude is selected, and the one with the largest spectral amplitude is chosen. Frequency index As a frequency domain feature, the gradient mean of the average signal-to-noise ratio is used. and gradient standard deviation As a time-domain feature, construct the time-mean subsequence of each historical signal-to-noise ratio. eigenvectors ;

[0066] Step 6: Set the prediction time steps The time-mean subsequences of each historical signal-to-noise ratio Corresponding end time point After The time points are used as the prediction interval. Based on the historical motion trajectories of the communication positioning terminal, communication satellites, and navigation satellites, the position of the communication positioning terminal within the predicted range is obtained. Motion parameter vector in Obtain the prediction interval Each time point Antenna direction vector of the communication positioning terminal The antenna of the communication positioning terminal points to the vector of the communication satellite. Distance between the communication positioning terminal and the communication satellite The antenna of the communication positioning terminal is pointed towards the vector of the navigation satellite with the highest signal-to-noise ratio. ; ;

[0067] Calculate the prediction interval Each time point Signal-to-noise ratio of communication satellites This constitutes a signal-to-noise ratio time series. ;

[0068]

[0069] in, This indicates the angle of deviation between the main lobe of the communication positioning terminal's antenna and the direction of the communication satellite. ; The half-power beamwidth of the antenna for the communication positioning terminal; The antenna transmission power of the communication positioning terminal; Antenna noise power for the communication positioning terminal;

[0070] Step 7: Construct a prediction model. The input to the prediction model is the feature vector. The output is a predicted motion parameter vector of the communication positioning terminal. Prediction of motion parameter vectors based on communication positioning terminals Calculate the prediction interval Each time point Prediction of antenna direction vector of communication positioning terminal The antenna of the communication positioning terminal points to the vector of the communication satellite. Distance between the communication positioning terminal and the communication satellite Then calculate the prediction interval. Each time point Prediction of signal-to-noise ratio of communication satellites This constitutes the signal-to-noise ratio prediction time series. ;

[0071] use The prediction model is trained using a set of samples, and the loss function is:

[0072]

[0073] in, , , , These are the weighting coefficients;

[0074] For motion smoothing constraint loss:

[0075]

[0076] For direction error loss:

[0077]

[0078] To perform feature regularization loss, a feature regularization term is constructed for the feature vector. Prediction of motion parameter vectors of communication and positioning terminals Let the vector with more elements be denoted as . The vector with fewer elements is Vectors with more elements Perform clipping so that the clipped vector The number of elements in a vector with fewer elements Consistent;

[0079]

[0080] To compensate for the loss of spectral consistency, a spectral domain penalty mechanism based on Fourier transform is introduced:

[0081]

[0082] in, To calculate the spectral amplitude of the sequence using Fourier transform;

[0083] Step 8: Get the current time and before Within a time interval Predicting Combinations Signal-to-noise ratio time series of various navigation satellites in China Extracting feature vectors The data is input into the trained prediction model to obtain the current time and subsequent time of the communication positioning terminal. Within a time interval Motion parameter vector prediction ;

[0084] Prediction of motion parameter vectors based on communication positioning terminals Calculate the prediction interval Each time point Prediction of antenna direction vector of communication positioning terminal The antenna of the communication positioning terminal points to the vector of the communication satellite. Distance between the communication positioning terminal and the communication satellite Then calculate the prediction interval. Each time point Prediction of signal-to-noise ratio of communication satellites This constitutes the signal-to-noise ratio prediction time series. ;

[0085] from Choose one from the middle positions as the dynamic threshold. ;like Then it communicates with communication satellites.

[0086] Example 1:

[0087] The following is in conjunction with the appendix Figure 1 The method proposed in this invention is further described in detail through simulation of the motion changes of the floating base station antenna under the impact of ocean waves. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0088] 1. Satellite position acquisition and data preprocessing

[0089] 1.1 Simulation Data Generation and Initialization

[0090] In practical applications, the carrier-to-noise ratio (SNR) sequence of the navigation signals from the receivable navigation satellites is first obtained using a satellite navigation module, along with the spatial location of the navigation satellite numbers. Then, a Savitzky-Golay filter (15-point window, 3rd-order polynomial) is used to denoise the original SNR data while preserving effective signal characteristics.

[0091] 1.2 Extreme Point Detection

[0092] Peak / trough identification: A height threshold (median), minimum interval (10 seconds), and significance level (2dB) are set to detect true SNR extreme points, which serve as the benchmark for subsequent evaluation. The timestamps of these extreme points and their corresponding SNR values ​​are recorded for dynamic tolerance matching.

[0093] 2. Algorithm Branch Selection

[0094] Based on whether the communication payload and the navigation signal transmission payload are on the same satellite, the satellite communication method is analyzed in two cases: communication and navigation payloads on the same platform and communication and navigation payloads on different satellite platforms. The former corresponds to the direct prediction algorithm for SNR, and the latter corresponds to the indirect prediction algorithm for SNR.

[0095] 3. For direct prediction of the SNR of communication satellites with navigation payloads, among the received multi-satellite navigation signals, the SNR time series of navigation signals located on the same platform as the short message communication payload is first extracted based on satellite identifiers and ephemeris data. After outlier removal and smoothing of this series, its statistical characteristics and short-term variation trend model are established, and this model is mapped to the evolution law of the SNR of the short message communication link. Based on this, combined with the system packet error rate requirements and available power budget, the period with higher predicted SNR and smaller fluctuations is selected as the transmission window, and the optimal transmission duration required to meet reliable transmission is calculated.

[0096] 3.1 Parameter Set Selection

[0097] The stationarity of the filtered sequence was verified using the ADF test, in conjunction with the initial... Determining the order of difference between non-seasonal and seasonal values ​​based on the estimated value and Based on the ACF peak detection function, the average adjacent peak positions are 39, therefore the final selected seasonal cycle parameter is... Then, iterate through the candidate parameter set. A set of valid parameters was selected for model fitting; preliminary optimization was performed on the valid parameter sets that met the above conditions, and four sets of optimal parameters were selected based on the AIC criterion for the main model selection, combined with seasonal parameters. The feedback callback is dynamically selected.

[0098] 3.2 SARIMA Model Construction

[0099] Based on the above parameter set and initial... A SARIMA model is constructed and trained. ACF peak detection is performed based on the model residuals, and seasonal parameters are dynamically retrieved. Used for the next round of model fitting. Continuous feedback and parameter adjustment allow the model to dynamically adapt to changes in the SNR of the communication satellite. Within a sliding window size... Predicting steps Under the conditions, the model is initially ,maximum The final optimal parameter set is .

[0100] 3.3 Sliding Window Dynamic Prediction and Peak Detection

[0101] Based on the parameters obtained from the main training loop, the rolling window prediction continues on the test set, and peak detection is performed on the final prediction sequence. The prediction performance of the model is judged by the accuracy of the peak detection time point and amplitude, as well as their joint accuracy.

[0102] 4. Indirect prediction of SNR for communication satellites without navigation payloads: When communication and navigation payloads are on different platforms (such as Tiantong satellite mobile communication), select the navigation satellite closest to the zenith. Using this as a benchmark, find three more satellites to form a satellite group, maximizing the coverage area of ​​the satellite group in the sky. When calculating the coverage area, the volume of the convex polygon formed by connecting the four satellites can be used as a standard. Extract the signal-to-noise ratio (SNR) sequence of the navigation signals of the navigation satellites in this satellite group. Estimate the antenna trajectory based on the selected SNR sequence, and then predict the direction of maximum antenna gain. Select the transmission time and duration based on the predicted change in the direction of maximum antenna gain.

[0103] 4.1 Obtaining the signal-to-noise ratio sequence and feature vectors Extraction

[0104] Multiple satellites are generated, and the satellite combination with the widest dispersion and strongest overall signal-to-noise ratio is selected. In this simulation, four satellite combinations are selected, meaning each group contains four satellites. The dispersion can be calculated using the volume of the convex polygon formed by the multiple satellites. The average signal-to-noise ratio (SNR) time series of the known satellite combinations is obtained. The signal-to-noise ratio intensity at time is denoted as The interval between two samplings extract Its time-domain and frequency-domain characteristics:

[0105] Temporal characteristics: Calculate the mean and standard deviation of the first derivative (gradient) of the attitude sequence within each window;

[0106] Frequency domain features: Perform FFT on the SNR sequence, extract the indices of the three frequency components with the largest amplitude as feature vectors for lightweight regularization, and save the entire spectrum sequence under the current window for the calculation of spectrum consistency regularization terms.

[0107]

[0108]

[0109] Let k be the complex frequency domain representation of the component with frequency k. This represents the time series of signal-to-noise ratio (SNR) within the current window, where the range of the current window is... to At that time, Recorded as , This indicates the amplitude of the frequency component, selecting the three main frequency indices with the largest amplitudes. As a feature, N in the formula refers to the number of sampling points in the current window.

[0110] The final 5-dimensional feature vector is formed. .

[0111] 4.2 Platform Motion Modeling and Parameter Vectors The establishment

[0112] This implementation case uses Phillips wave theory as a model to construct a wave spectrum model: defining the frequency component range. Generate frequency sequences representing short-period perturbations and their corresponding angular frequency values. The energy density of each frequency component was estimated using the Phillips spectral model based on the wind speed parameters.

[0113]

[0114] in, Phillips constant, It is the acceleration due to gravity; Main frequency angular frequency, This refers to wind speed.

[0115] Based on the above energy density, calculate the amplitude of each frequency component:

[0116]

[0117] Randomly initialize the phase of each frequency component:

[0118]

[0119] Antenna attitude change curve generated by superimposed harmonic components:

[0120]

[0121] in, , These represent the pitch and roll angles of the ocean waves in the initial state of the system, respectively. , These represent the initial phases of the pitch and roll angles, respectively.

[0122] Introducing long-period surge components to supplement low-frequency disturbances:

[0123]

[0124] in, For the amplitude of the surge wave; For the swell cycle; This is the initial phase (a fixed constant). The pitch angle after incorporating low-frequency swell disturbances.

[0125] In addition, a scaling factor is introduced. Proportional adjustment factor and offset adjustment factor Make fine adjustments:

[0126]

[0127] Finally, Gaussian perturbations are added to simulate unstructured fluctuations:

[0128]

[0129] and These represent the pitch and roll angles after adding random perturbations, respectively. Ultimately, the pitch and roll angles of the antenna's motion direction can be converted into three-dimensional unit direction vectors from the current state of the ocean waves to characterize the pointing of the antenna's main lobe.

[0130]

[0131] The control variables involved in the above modeling process are uniformly represented as a nine-dimensional parameter vector. ,parameter Specifically, it includes:

[0132]

[0133] in, Indicates wind speed, which determines the location of the dominant frequency; This is the angle scaling factor, used to adjust the scaling relationship between the wave angle and the actual angle, that is, to stretch the dimensionless small angle calculated by the wave model into the amplitude range that the antenna may actually swing. These are the amplitude and period of the surge, respectively. These are used to adjust the pitch and roll ratios and offset, respectively. The strength of the random perturbation is defined. By setting boundary conditions for the parameter vector, efficient constraints and searches can be performed in subsequent optimization stages. Through fitting... By obtaining the values ​​of the intermediate parameters, the motion state of the antenna platform can be approximately simulated.

[0134] 4.3 Construct a constructor to build a prediction model

[0135] 4.3.1 Loss Function Design

[0136] To accurately fit the antenna's true pointing trajectory, an inverse optimization algorithm based on the signal-to-noise ratio (SNR) sequence of the received navigation signal is established to estimate the parameter vector of the enhanced antenna motion model. The model parameters are updated using the received satellite navigation signal-to-noise ratio sequence. To estimate the current surging state of the ocean waves, the statistical error between the signal-to-noise ratio obtained by the antenna pointing in the model output and the signal-to-noise ratio caused by the actual ocean waves is minimized in the observable event of when the antenna is pointed at any satellite, thereby achieving antenna trajectory estimation.

[0137] The phrase "updating model parameters using the received satellite navigation signal-to-noise ratio sequence" To achieve antenna trajectory estimation, the specific steps include: employing a dynamic window mechanism to update motion model parameters in real time. The algorithm predicts the extreme points of antenna pointing. First, a sliding window mechanism is used to define the window size, ensuring real-time capture of dynamic changes in antenna pointing. Within each prediction period, the algorithm extracts time-domain and frequency-domain features from the SNR sequence within the current window and uses these features to update the motion model parameters. The motion model parameters are updated in real-time using a hybrid optimizer that combines differential genetic algorithm (DE) and quasi-Newton method (L-BFGS-B) to achieve a synergistic effect of global search and local optimization. During the parameter update process, the unknown parameter vector is incorporated. The model parameters are optimized by minimizing a comprehensive loss function, which includes a motion smoothing term, a direction error term, a feature regularization term, and a spectral consistency term.

[0138] (1) Motion smoothing term: In order to avoid the attitude prediction being discontinuous due to excessive parameter updates, the smoothness constraint of parameter vector changes is defined in the form of first-order difference.

[0139] (2) Direction error term: To ensure that the estimated antenna pointing is consistent with the actual navigation satellite direction in physical space, a direction error term is constructed to measure the direction vector of the predicted antenna main lobe. With the direction vector of currently active satellites The Euclidean distance between them;

[0140] (3) Feature regularization term: To enhance the fusion between the model and the dynamic environment features, a feature regularization term is constructed to regularize the time steps. Dynamic feature vector at the location With the corresponding optimization parameter vector Apply dot product constraints, at this time Pick The first five items and Alignment is performed to encourage the ability of features to interpret pose changes;

[0141] (4) Spectrum consistency term: based on the antenna direction vector Combined with the antenna gain model, the predicted target satellite received signal-to-noise ratio under this parameter vector is calculated;

[0142] (5) Based on this, predict the signal-to-noise ratio sequence. Compared with the true signal-to-noise ratio sequence A spectral domain penalty mechanism based on Fast Fourier Transform (FFT) is introduced to ensure that the predicted signal-to-noise ratio sequence is consistent with the real signal in the frequency domain characteristics.

[0143] In summary, by applying weight constraints to the above terms, the final loss function is as follows:

[0144]

[0145] In the loss function These are, respectively, motion smoothing constraint, orientation matching error, feature regularization constraint, and spectral consistency penalty. These are the corresponding coefficients. In short-term prediction, the satellite's own motion is approximately considered negligible. In this example application, each weight is set to... The model update cycle is The initial parameter vector to be optimized is The predicted antenna direction vector is The target satellite's direction vector is .

[0146] 4.3.2 Fitting and Two-Stage Optimization Process

[0147] Before optimization begins, initialize the hybrid optimizer and motion model, including initial parameter vectors. and extracted feature vectors The model was trained using the loss function described above. A two-stage optimization process followed: first, global optimization was performed using Differential Evolution (DE) with a population size of 20, a mutation factor range of (0.5, 1.5), and a crossover probability of 0.8, for 105 iterations (70% of the total iterations); then, local optimization was performed using the DE results as initial values ​​and the L-BFGS-B algorithm for fine-tuning to ensure convergence to a local optimum. During each optimization process, features and optimization parameters were dynamically adjusted based on the data within the current time window, enabling the model to adapt to data changes. A sliding window method was used for prediction, and the training data was periodically updated and the model was retrained to adapt to dynamic data changes.

[0148] 4.4 Real-time prediction and dynamic assessment

[0149] 4.4.1 Communication Signal-to-Noise Ratio Prediction

[0150] In this simulation, the initial window size is 300 seconds, sliding every 30 seconds to update the model parameters. The aforementioned time-frequency domain features are extracted, and the parameters of these features are updated each time the window slides. Tolerance calculation: The matching threshold is dynamically adjusted based on the SNR range within the window. In this simulation, it is set to 0.2 times the range within the window, with 5.0 set as the minimum tolerance value to ensure that the tolerance is not too small when SNR changes are small. This logic can adapt to different signal strength variations: in windows with large SNR changes, the tolerance increases accordingly, allowing for larger time deviations to match the predicted and true points; while in windows with small SNR changes, the tolerance remains at a small, fixed value to improve matching accuracy.

[0151] The system then detects and matches extreme values. For each predicted extreme value, it searches for the true extreme value within its tolerance range. If a true extreme value exists, it is marked as a correct match; otherwise, it is recorded as a false alarm. The model assumes that if the signal strength does not change abruptly, the signal strength in the vicinity of the predicted extreme value should be at a relatively good level. The model is updated periodically, and forced retraining is performed after every 10 prediction windows to prevent long-term drift. Regarding the accuracy of the prediction, the simulation directly compares the actual signal-to-noise ratio (SNR) of the unloaded satellite with the predicted SNR. However, in practice, when the actual SNR of the unloaded satellite is unavailable, the SNR of communication with the loaded satellite can be calculated by simulating the antenna's motion trajectory and compared with the actual collected SNR. When the prediction performance is poor or the environment changes abruptly, the system will retrain the model completely.

[0152] 4.5 Communication Window Optimization and Communication Strategy Settings

[0153] Based on the above prediction results, a communication window is selected, and communication occurs when the signal-to-noise ratio (SNR) is higher than the dynamic threshold. The dynamic threshold can be set according to the actual application situation, based on the fluctuation range of the current SNR. The higher the threshold is set, the higher the reliability, but the effectiveness will be reduced to some extent because the time period for meeting the communication threshold is shorter.

[0154] In this simulation, a threshold of -3dB of the maximum signal-to-noise ratio within the sliding window was used as the preferred communication window. The scheme of this application was compared with direct communication, traditional ARQ communication, HARQ communication, etc., and a comparison curve of bit error rate and throughput was plotted (as attached). Figure 3 , 5 ).

[0155] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment, characterized in that: For communication satellites with communication and navigation payloads on the same platform, obtain the historical signal-to-noise ratio (SNR) time series of the communication satellite, obtain the fitting function of the SNR of the communication satellite with respect to time based on the fitting model, and predict the SNR time series of the communication satellite at the current moment and within the predicted time period thereafter. For communication satellites with different communication and navigation payload platforms, navigation satellites are selected to form a prediction combination. The historical signal-to-noise ratio (SNR) time series of each navigation satellite in the prediction combination is obtained. The average SNR at each sampling moment is calculated to obtain a historical SNR average sequence. The gradient mean and gradient standard deviation of the historical SNR average sequence are taken as time-domain features. Fourier transform is used to calculate the spectral amplitude of the historical SNR average sequence. A certain number of frequency indices with the largest spectral amplitude are taken as frequency-domain features. The time-domain features and frequency-domain features are combined to construct a feature vector. A prediction model is constructed and trained. The feature vector is input into the prediction model, and the prediction model outputs the predicted motion parameter vector of the communication positioning terminal. Based on the historical motion trajectory of the communication satellite, the antenna direction vector of the communication positioning terminal, the vector pointing from the antenna of the communication positioning terminal to the communication satellite, and the distance between the communication positioning terminal and the communication satellite are predicted for the current moment and the subsequent prediction time period. This allows for the prediction of the SNR time series of the communication satellite for the current moment and the subsequent prediction time period. Select one position from the middle position of the signal-to-noise ratio time series of the communication satellite at the current time and within the predicted time period thereafter as the dynamic threshold; if the predicted signal-to-noise ratio of the communication satellite at the current time is greater than the dynamic threshold, the communication positioning terminal communicates with the communication satellite.

2. The method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment according to claim 1, characterized in that: For communication satellites with different platforms for communication and navigation payloads, select The navigation satellites constitute the predicted combination Get the current time and before Within a time interval Signal-to-noise ratio time series of each navigation satellite ; in, For the first Navigation satellites at sampling time The signal-to-noise ratio, the time step between each sampling time is , , , .

3. The method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment according to claim 2, characterized in that: Calculate the average signal-to-noise ratio at each sampling time. The historical average signal-to-noise ratio sequence is obtained, and the gradient mean of the historical average signal-to-noise ratio sequence is taken. and gradient standard deviation As a temporal domain feature, specifically: ， in, ; Calculate the historical signal-to-noise ratio average sequence using Fourier transform The spectral amplitude is selected based on the largest spectral amplitude. Frequency index As a frequency domain feature, the time domain feature and the frequency domain feature are combined to construct a feature vector. .

4. The method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment according to claim 3, characterized in that: eigenvectors Input the prediction model, and the prediction model outputs the predicted motion parameter vector of the communication and positioning terminal. ; Current moment and afterwards The time points are used as the prediction interval. ; Motion parameter vector prediction based on communication and positioning terminals Based on the historical motion trajectory of communication satellites, the prediction interval is obtained. Each time point Antenna direction vector prediction for local communication positioning terminal Vector prediction of the antenna pointing to the communication satellite of the communication positioning terminal. Distance prediction between communication positioning terminal and communication satellite This allows for the prediction of communication satellites within the prediction range. Each time point Signal-to-noise ratio of communication satellites The signal-to-noise ratio time series was obtained. ; in, Indicates the communication positioning terminal at a certain time point. The angle of deviation between the antenna main lobe and the direction of the communication satellite. ; The half-power beamwidth of the antenna for the communication positioning terminal; The antenna transmission power of the communication positioning terminal; This refers to the antenna noise power of the communication positioning terminal.

5. The method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment according to claim 4, characterized in that: The training prediction model is specifically as follows: Get Predicted Combinations Historical signal-to-noise ratio time series of various navigation satellites in China , The total number of historical sampling time points obtained; Acquire the historical motion trajectories of the communication positioning terminal, communication satellites, and various navigation satellites; Calculate the average signal-to-noise ratio at each sampling time. The historical signal-to-noise ratio time-averaged sequence was obtained. ; Set time steps Historical signal-to-noise ratio time-averaged series The data is split into segments, and each segment after splitting has a historical signal-to-noise ratio time-mean subsequence as follows: , Rebuild The index of each element in the text. , , ; For each historical signal-to-noise ratio time mean subsequence Calculate the gradient mean of the average signal-to-noise ratio. and gradient standard deviation Fourier transform was used to calculate the time-mean subsequence of each historical signal-to-noise ratio. The spectral amplitude is selected, and the one with the largest spectral amplitude is chosen. Frequency index As a frequency domain feature, the gradient mean of the average signal-to-noise ratio is used. and gradient standard deviation As a time-domain feature, construct the time-mean subsequence of each historical signal-to-noise ratio. eigenvectors ; Set the prediction time steps The time-mean subsequences of each historical signal-to-noise ratio Corresponding end time point After The time points are used as the prediction interval. Based on the historical motion trajectories of the communication positioning terminal, communication satellites, and navigation satellites, the position of the communication positioning terminal within the predicted range is obtained. Motion parameter vector in Obtain the prediction interval Each time point Antenna direction vector of the communication positioning terminal The antenna of the communication positioning terminal points to the vector of the communication satellite. Distance between the communication positioning terminal and the communication satellite The antenna of the communication positioning terminal is pointed towards the vector of the navigation satellite with the highest signal-to-noise ratio. ; ; Calculate the prediction interval Each time point Signal-to-noise ratio of communication satellites This constitutes a signal-to-noise ratio time series. ; eigenvectors Input the prediction model, and the prediction model outputs the predicted motion parameter vector of the communication and positioning terminal. Prediction of motion parameter vectors based on communication positioning terminals Calculate the prediction interval Each time point Prediction of antenna direction vector of communication positioning terminal The antenna of the communication positioning terminal points to the vector of the communication satellite. Distance between the communication positioning terminal and the communication satellite Then calculate the prediction interval. Each time point Prediction of signal-to-noise ratio of communication satellites This constitutes the signal-to-noise ratio prediction time series. ; use The prediction model is trained using a set of samples.

6. The method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment according to claim 5, characterized in that: The loss function for training the prediction model is: in, , , , These are the weighting coefficients; For motion smoothing constraint loss: For direction error loss: To perform feature regularization loss, a feature regularization term is constructed for the feature vector. Prediction of motion parameter vectors of communication and positioning terminals Let the vector with more elements be denoted as . The vector with fewer elements is Vectors with more elements Perform clipping so that the clipped vector The number of elements in the vector with fewer elements Consistent; To compensate for the loss of spectral consistency, a spectral domain penalty mechanism based on Fourier transform is introduced: in, The spectral amplitude of the sequence is calculated using Fourier transform.

7. The method for optimizing satellite communication windows based on satellite navigation signal carrier-to-noise ratio information in a dynamic environment according to claim 1, characterized in that: Signal-to-noise ratio time series of communication satellites at the current moment and within the predicted time period thereafter. Choose one from the middle positions as the dynamic threshold. If the predicted signal-to-noise ratio of the communication satellite at the current moment is greater than the dynamic threshold, i.e. Then the communication positioning terminal communicates with the communication satellite.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 7.

10. A computer program product comprising computer instructions, characterized in that: When executed by a processor, the computer instructions implement the steps of the method according to any one of claims 1 to 7.

Images