Satellite communication channel allocation method and system based on navigation prediction model

Abstract

The application relates to the technical field of satellite communication, and discloses a satellite communication channel allocation method and system based on a navigation prediction model, which comprises the following steps: obtaining historical communication data and obstruction data, and respectively extracting characteristic data of the historical communication data and the obstruction data; obtaining a comprehensive influence number of the communication data according to the characteristic data of the communication data, and obtaining a comprehensive influence number of the obstruction data according to the characteristic data of the obstruction data; and evaluating a judgment coefficient of channel allocation by a third party according to the historical communication data and the obstruction data, wherein the judgment coefficient comprises sample data with the same comprehensive influence number of the communication data. The application constructs a channel allocation mechanism based on a navigation prediction model, significantly improves the reliability and efficiency of satellite communication, can dynamically select an optimal communication mode according to real-time environment and historical data, and effectively reduces communication interruption caused by meteorological obstruction or signal attenuation.

Description

Technical Field

[0001] This invention relates to the field of satellite communication technology, and in particular to a satellite communication channel allocation method and system based on a navigation prediction model. Background Technology

[0002] In the field of satellite communications, especially maritime satellite communications, communication quality is often unstable or even interrupted during ship navigation due to factors such as weather changes, geographical obstruction, and band attenuation.

[0003] Existing channel allocation methods are mostly based on static rules or simple thresholds, which are difficult to adapt dynamically to complex and ever-changing environmental conditions. They also lack the ability to make intelligent predictions by fusing historical data with real-time status, resulting in problems such as low efficiency in communication resource utilization and insufficient link reliability. Summary of the Invention

[0004] This invention provides a satellite communication channel allocation method and system based on a navigation prediction model to solve existing technical problems, thereby addressing the difficulty in dynamically adapting to complex and ever-changing environmental conditions.

[0005] To address the aforementioned technical problems, according to one aspect of the present invention, more specifically, a satellite communication channel allocation method based on a navigation prediction model, comprising the following steps:

[0006] S1. Obtain historical communication data and obstruction data, and extract characterization data from the historical communication data and obstruction data respectively.

[0007] S2. Obtain the comprehensive impact number of communication data based on the characterization data of communication data, and obtain the comprehensive impact number of obstruction data based on the characterization data of obstruction data.

[0008] S3. A third party evaluates the channel allocation judgment coefficient based on the historical communication and obstruction data. This judgment coefficient includes:

[0009] 1) Summarize sample data with the same comprehensive impact number of communication data, and obtain the judgment coefficient A from a third party for the summarized sample data;

[0010] 2) For sample data with the same overall impact on the hindering data, the judgment coefficient B is obtained from a third party for the summarized sample data;

[0011] 3) Then, for any sample data, a third party obtains the judgment coefficient C;

[0012] S4. Construct a prediction sub-model A based on the correlation between the judgment coefficient A and the comprehensive impact number of the obstruction data; construct a prediction sub-model B based on the correlation between the judgment coefficient B and the comprehensive impact number of the communication data.

[0013] S5. Determine the correlation between prediction sub-model A and prediction sub-model B based on the relationship between the judgment coefficient C and the comprehensive impact number of the obstruction data and the comprehensive impact number of the communication data.

[0014] S6. Construct a navigation prediction model based on the correlation between prediction sub-model A and prediction sub-model B, and determine the parameters of the navigation prediction model.

[0015] S7. Input the new communication data and obstruction data of the current state into the navigation prediction model, and determine whether to use mode A, mode B, mode C or mode D as the channel allocation state based on the judgment coefficient output by the navigation prediction model.

[0016] Furthermore, the characterization data of the communication data represents the overall signal quality ratio; the characterization data of the obstruction data represents the probability ratio of the signal being affected by obstruction.

[0017] Furthermore, in step S2, the overall impact number of the communication data is determined by the ratio of the overall satellite communication quality at the current location of the ship to the historical average overall satellite communication quality when the ship was at that location.

[0018] Furthermore, in step S2, the comprehensive impact number of the obstruction data is determined by the ratio of the cloud obstruction level of the ship in the current region and climate time to the historical average cloud obstruction level of the ship in that region and climate time.

[0019] Furthermore, adjust the parameters in the prediction sub-model A until the output value of the prediction sub-model A is close to the same as the judgment coefficient A;

[0020] Adjust the parameters in the prediction sub-model B until the output value of the prediction sub-model B is similar to the judgment coefficient B;

[0021] Adjust the parameters in the navigation prediction model until the output value of the navigation prediction model is similar to the judgment coefficient C.

[0022] Furthermore, in step S7, based on the enhanced, standard, and edge-level sample data of the satellite channel, a threshold for the division of the judgment coefficient is set for the automated classification of satellite communication channels.

[0023] Furthermore, based on the comparison results between the judgment coefficient and the division threshold, the satellite communication channel is automatically assigned to the corresponding mode A, mode B, mode C or mode D state.

[0024] Furthermore, the navigation prediction model is a fusion model constructed based on prediction sub-model A and prediction sub-model B through correlation analysis; the model receives the comprehensive influence number of obstruction data and the comprehensive influence number of communication data as input, and calculates and outputs judgment coefficients for use in deciding the channel allocation mode.

[0025] Furthermore, mode A indicates communication with satellites using only Ka-band channels;

[0026] Mode B represents communication with satellites using Ku-band and Ka-band channels respectively;

[0027] Mode C indicates communication with satellites using C-band and Ku-band channels respectively;

[0028] Mode D indicates communication with satellites using only C-band channels.

[0029] A satellite communication channel allocation method based on a navigation prediction model, wherein the satellite communication channel allocation system based on the navigation prediction model includes: a data acquisition module, a data representation extraction module, a comprehensive influence number calculation module, a judgment coefficient evaluation module, a prediction sub-model construction module, a correlation analysis module, a navigation prediction model construction and parameter determination module, a channel allocation decision module, and a threshold setting module.

[0030] This invention provides a satellite communication channel allocation method and system based on a navigation prediction model. Compared with existing technologies, the advantages of this method are:

[0031] 1. This invention significantly improves the reliability and efficiency of satellite communication by constructing a channel allocation mechanism based on a navigation prediction model. It can dynamically select the optimal communication mode according to real-time environment and historical data, effectively reducing communication interruptions caused by weather obstruction or signal attenuation.

[0032] 2. This invention enhances the objectivity and accuracy of channel allocation decisions by combining comprehensive influence data with third-party evaluation, enabling the system to adapt to complex communication conditions in different sea areas, climates, and ship types.

[0033] 3. This invention automates and automates the channel allocation process. By setting a judgment coefficient threshold, it automatically matches the optimal communication mode, reducing the need for manual intervention and improving system response speed and operation and maintenance efficiency.

[0034] 4. This invention optimizes the utilization efficiency of communication resources by flexibly switching between multi-band channel modes, extending equipment lifespan while ensuring communication quality and reducing overall communication costs.

[0035] 5. This invention has strong generalization and scalability. The model can be adapted to different ship sizes and historical data, making it suitable for various satellite communication scenarios and possessing good practicality and promotional value. Attached Figure Description

[0036] Figure 1 This is a flowchart of the present invention;

[0037] Figure 2 This is a schematic diagram of the judgment coefficient A and obstruction data for large ships in this invention;

[0038] Figure 3 This is a schematic diagram of the judgment coefficient B of large ships and communication data in this invention;

[0039] Figure 4 This is a schematic diagram of the judgment coefficient C for large ships and the obstruction data and communication data in this invention;

[0040] Figure 5 This is a schematic diagram of the judgment coefficient A and obstruction data for small vessels in this invention.

[0041] Figure 6 This is a schematic diagram of the judgment coefficient B and communication data for small ships in this invention.

[0042] Figure 7 This is a schematic diagram of the judgment coefficient C for small ships and the obstruction data and communication data in this invention;

[0043] Figure 8 This is a schematic diagram of the band channel for Mode A, Mode B, Mode C, or Mode D in this invention. Detailed Implementation

[0044] To make the technical solution of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0045] Example 1

[0046] like Figure 1 As shown, the satellite communication channel allocation method based on the navigation prediction model first obtains historical communication data and obstruction data, and then extracts characterization data from the historical communication data and obstruction data respectively.

[0047] The specific calculation formula for the normalization sub-formula of a single indicator in the representation data of communication data is as follows:

[0048] ;

[0049] In the formula, This represents the normalized value of the i-th communication quality parameter. It is a dimensionless value ranging from [0, 1]. =1 indicates that the parameter is in its optimal state. =0 indicates that the parameter is in the worst-case state;

[0050] This represents the raw measurement value of the i-th communication quality parameter. For example, this could be the real-time measured signal-to-noise ratio (SNR) value (in dB), bit error rate (BER) value, or received signal strength indication (RSSI) value (in dBm).

[0051] This represents the theoretical or practical optimal or upper limit value of the i-th parameter. This value is typically set based on the physical limits or experience of the satellite communication system. For example, for signal-to-noise ratio (SNR), it might be the theoretical maximum value that the modem can decode without loss.

[0052] This represents the worst-case or lower limit of the i-th parameter within the system's acceptable range. When the parameter falls below this value, the communication link is considered completely interrupted.

[0053] Therefore, the specific formula for calculating the overall signal quality ratio in the characterization data of communication data is as follows:

[0054] ;

[0055] In the formula, This represents the current overall satellite communication quality. It is an overall score that integrates all key performance indicators, ranging from [0, 1]. The closer it is to 1, the better the overall communication quality;

[0056] This represents the weight of the i-th parameter, a coefficient greater than 0, indicating its importance in the overall quality assessment. For example, in the Ka-band where rain attenuation is severe, the instantaneous signal-to-noise ratio (SNR) may have a higher weight than the signal strength index (RSSI). All weights are summed to 1 for ease of calculation and understanding.

[0057] n represents the total number of key performance indicators selected. For example, if signal-to-noise ratio, bit error rate, and signal strength are selected, then n=3.

[0058] Among them, the probability ratio of the signal being affected by occlusion, which represents the obstruction of data, is specifically calculated using the following formula:

[0059] ;

[0060] ;

[0061] In the formula, This represents the current probability of cloud cover, which is a probability value between (0, 1). =0.85 means that the model predicts there is an 85% chance that signal obstruction will occur due to clouds;

[0062] e represents the natural constant, approximately equal to 2.71828, and is the base of the natural logarithm;

[0063] z represents a linear predictor, which is a linear combination of all input feature variables. The value of z can be any real number (-∞, +∞), and it is mapped to a probability space of (0, 1) through a logistic function;

[0064] This represents the intercept (or bias) term of the logistic regression model. It is a constant term, representing the bias when all feature values... , ,..., Basic logic output when both are 0;

[0065] , ,..., These represent the coefficients of the logistic regression model. Each coefficient... Its corresponding features were quantified. For the final occlusion probability The direction and magnitude of the influence.

[0066] Positive coefficient ( >0): Feature Increasing the value will lead to an increased occlusion probability. Increase;

[0067] , ,..., The magnitude of the absolute value of the coefficient will affect The degree of intensity.

[0068] , ,..., These represent the meteorological characteristic variables input to the model. These are raw data acquired in real time and related to cloud cover. For example:

[0069] This indicates the optical thickness of the cloud layer as determined by satellite cloud imagery.

[0070] This indicates the rainfall intensity (mm / h) provided by weather radar data.

[0071] This indicates the relative humidity (%) provided by the atmospheric profile data.

[0072] Indicates the height of the cloud top (in meters).

[0073] The comprehensive impact number of communication data is obtained from the characterization data of communication data, and the comprehensive impact number of obstruction data is obtained from the characterization data of obstruction data.

[0074] The composite impact number of communication data is determined by the ratio of the composite satellite communication quality at the current location of the ship (which can be expressed as the signal-to-quality ratio; a higher signal-to-quality ratio indicates better composite satellite communication quality) to the historical average composite satellite communication quality when the ship is at that location. Therefore, the specific formula for calculating the composite impact number of communication data is as follows:

[0075] ;

[0076] In the formula, The overall signal quality ratio is a dimensionless ratio used to measure the current level of communication quality relative to historical norms.

[0077] >1 indicates that the current quality is better than the historical average;

[0078] =1 indicates that the current quality is equal to the historical average level;

[0079] <1 indicates that the current quality is worse than the historical average.

[0080] This represents the historical average overall satellite communication quality. This is the historical average quality over all periods / times and locations under the same geographical location and climate season / time period. The average value (arithmetic mean or exponentially weighted average) represents the "normal" or "expected" level of communication quality at that location during that period.

[0081] The composite impact number of obstruction data is determined by comparing the cloud obstruction level of the ship in the current region and weather time (this cloud obstruction level can be represented by the probability ratio of signal obstruction; the higher the probability ratio of signal obstruction, the greater the cloud obstruction level), with the historical average cloud obstruction level of the ship in that region and weather time. Therefore, the specific formula for calculating the composite impact number of obstruction data is as follows:

[0082] ;

[0083] In the formula, This represents the probability ratio of a signal being affected by obstruction. It is a dimensionless ratio used to measure the current risk of obstruction relative to historical norms.

[0084] >1 indicates that the current wind protection risk is higher than the historical average risk.

[0085] =1 indicates the current wind protection level, and the risk level is equal to the historical average risk.

[0086] <1 indicates that the current occlusion risk is lower than the historical average risk.

[0087] This represents the historical average probability of cloud cover. This is the probability of cloud cover over all historical data within the same region and at the same climatic time (e.g., the same month, the same time period). The average value represents the "normal" level of occlusion risk in that area during that time period.

[0088] By systematically extracting data representations and calculating comprehensive impact numbers, a scientific and reliable data preprocessing foundation was established. Its advantage lies in transforming complex and multi-source communication and environmental data into standardized and quantifiable evaluation indicators, providing high-quality and highly comparable input data for subsequent intelligent prediction models, and significantly improving the theoretical foundation and data processing consistency of the entire method.

[0089] Example 2

[0090] like Figure 1 As shown, a third party evaluates the channel allocation judgment coefficient based on historical communication and obstruction data. This judgment coefficient includes:

[0091] 1) Summarize sample data with the same comprehensive impact number of communication data, and obtain the judgment coefficient A from a third party for the summarized sample data;

[0092] The specific steps for a third party to obtain the judgment coefficient A include, for example:

[0093] First, third-party technical personnel obtained sample data with the same overall impact from 100 historical communication data points (these 100 historical communication sample data points are evenly distributed).

[0094] 2. If the communication quality of one of the sample data exceeds that of the other 50 sample data, then the third-party acquisition judgment coefficient A of that sample data is 50%.

[0095] 2) For sample data with the same overall impact on the hindering data, the judgment coefficient B is obtained from a third party for the summarized sample data;

[0096] The specific steps for a third party to obtain the judgment coefficient B include, for example:

[0097] First, third-party technical personnel obtained sample data with the same comprehensive impact of 100 historical obstruction data points (these 100 historical communication sample data points are evenly distributed).

[0098] 2. If the communication quality of one of the sample data exceeds that of the other 50 sample data, then the third-party acquisition judgment coefficient B of that sample data is 50%.

[0099] 3) Then, for any sample data, a third party obtains the judgment coefficient C;

[0100] The specific steps for a third party to obtain the judgment coefficient C include, for example:

[0101] First, third-party technical personnel obtained a sample of data with a uniform distribution of the combined impact count and the combined impact count of 100 historical communication data points.

[0102] 2. If the overall communication quality of one sample data exceeds that of the other 50 sample data, then the third-party acquisition judgment coefficient C of that sample data is 50%.

[0103] Specific examples include:

[0104] A. Forecasting of large ships

[0105] like Figure 2 As shown, a prediction sub-model A is constructed based on the correlation between the judgment coefficient A and the comprehensive influence number of the hindering data; and the parameters in the prediction sub-model A are adjusted until the output value of the prediction sub-model A is similar to the judgment coefficient A.

[0106] The red dots in the figure represent the distribution of the combined influence of the judgment coefficient A and the hindering data in the 100 collected sample data. Based on this 100 sample data, a model is fitted, and then:

[0107] (Formula 1);

[0108] In Formula 1 above, , Used for control The trend of image change and the trend of data change of judgment coefficient A tend to be the same constant. And by... Figure 2 The data in the middle can be determined , When, in Formula 1 The data tend to overlap with the judgment coefficient A.

[0109] In Formula 1, k(x) represents the mathematical expression of the constructed predictive sub-model A, where x represents the combined influence number of the input hindering data.

[0110] like Figure 3As shown, a prediction sub-model B is constructed based on the correlation between the judgment coefficient B and the comprehensive influence number of communication data; and the parameters in the prediction sub-model B are adjusted until the output value of the prediction sub-model B is similar to the judgment coefficient B.

[0111] The blue dots in the diagram represent the distribution of the judgment coefficient B and the combined influence number of the communication data among the 100 collected sample data. Based on this 100 sample data, a model is fitted, and then:

[0112] (Formula 2);

[0113] In formula 2 above, by Figure 3 The data in the middle can be determined , When, in Formula 2 The trend of image change tends to overlap with the trend of data change of judgment coefficient B.

[0114] In Formula 2, k(y) represents the mathematical expression of the constructed prediction sub-model B, where y represents the comprehensive impact number of the input communication data.

[0115] like Figure 4 As shown, the correlation between prediction sub-model A and prediction sub-model B is determined based on the relationship between the judgment coefficient C and the comprehensive influence number of obstruction data and the comprehensive influence number of communication data, respectively; and, the parameters in the navigation prediction model are adjusted until the output value of the navigation prediction model is similar to the judgment coefficient C.

[0116] And from the above Figure 4 Being able to know, when , hour, and The trend of change coincides with the trend of change of the corresponding judgment coefficient, and at this time, the output terms of Formula 1 and Formula 2 are correlated. Therefore:

[0117] The predictive sub-model A and predictive sub-model B are positively correlated. Combining this with the characteristic relationship from Formula 1 and Formula 2 above, we have:

[0118] ;

[0119] In the formula, k represents the mathematical expression of the navigation prediction model for large ships, x represents the comprehensive impact number of the input obstruction data of large ships, and y represents the comprehensive impact number of the input communication data of large ships.

[0120] Comparing multiple sets of data yields the following:

[0121] Table 1. Judgment coefficients output by the large ship navigation prediction model

[0122]

[0123] When the amount of data becomes infinite, a judgment coefficient k can be used as a threshold to determine the communication signal mode used. According to the data in Table 1 above, mode A is optimal when k ≥ 77.3%. Mode B is optimal when k ≥ 68.6%. Mode C is optimal when k ≤ 56.4%. Mode C is optimal when k ≤ 48.4%.

[0124] B. Prediction of small vessels

[0125] like Figure 5 As shown, a prediction sub-model A is constructed based on the correlation between the judgment coefficient A and the comprehensive influence number of the hindering data; and the parameters in the prediction sub-model A are adjusted until the output value of the prediction sub-model A is similar to the judgment coefficient A.

[0126] The red dots in the figure represent the distribution of the combined influence of the judgment coefficient A and the hindering data in the 100 collected sample data. Based on this 100 sample data, a model is fitted, and then:

[0127] (Formula 1);

[0128] In Formula 1 above, , Used for control The trend of image change and the trend of data change of judgment coefficient A tend to be the same constant. And by... Figure 5 The data in the middle can be determined , When, in Formula 1 The data tend to overlap with the judgment coefficient A.

[0129] In Formula 1, k(x) represents the mathematical expression of the constructed predictive sub-model A, where x represents the combined influence number of the input hindering data.

[0130] like Figure 6 As shown, a prediction sub-model B is constructed based on the correlation between the judgment coefficient B and the comprehensive influence number of communication data; and the parameters in the prediction sub-model B are adjusted until the output value of the prediction sub-model B is similar to the judgment coefficient B.

[0131] The blue dots in the diagram represent the distribution of the judgment coefficient B and the combined influence number of the communication data among the 100 collected sample data. Based on this 100 sample data, a model is fitted, and then:

[0132] (Formula 2);

[0133] In formula 2 above, by Figure 6 The data in the middle can be determined , When, in Formula 2 The trend of image change tends to overlap with the trend of data change of judgment coefficient B.

[0134] In Formula 2, k(y) represents the mathematical expression of the constructed prediction sub-model B, where y represents the comprehensive impact number of the input communication data.

[0135] like Figure 7 As shown, the correlation between prediction sub-model A and prediction sub-model B is determined based on the relationship between the judgment coefficient C and the comprehensive influence number of obstruction data and the comprehensive influence number of communication data, respectively; and, the parameters in the navigation prediction model are adjusted until the output value of the navigation prediction model is similar to the judgment coefficient C.

[0136] And from the above Figure 7 Being able to know, when , hour, and The trend of change coincides with the trend of change of the corresponding judgment coefficient, and at this time, the output terms of Formula 1 and Formula 2 are correlated. Therefore:

[0137] The predictive sub-model A and predictive sub-model B are positively correlated. Combining this with the characteristic relationship from Formula 1 and Formula 2 above, we have:

[0138] ;

[0139] In the formula, k represents the mathematical expression of the navigation prediction model for small vessels, x represents the comprehensive impact number of the input obstruction data for small vessels, and y represents the comprehensive impact number of the input communication data for small vessels.

[0140] Comparing multiple sets of data yields the following:

[0141] Table 2. Judgment coefficients output by the small vessel navigation prediction model

[0142]

[0143] When the amount of data becomes infinite, a judgment coefficient k can be used as a threshold to determine the communication signal mode used. According to the data in Table 2 above, mode A is optimal when k ≥ 77.3%. Mode B is optimal when k ≥ 68.6%. Mode C is optimal when k ≤ 56.4%. Mode C is optimal when k ≤ 48.4%.

[0144] Introducing a third-party evaluation mechanism to obtain judgment coefficients has the advantage of enhancing the credibility and authority of model training data through objective and neutral external evaluation, effectively avoiding subjective bias, making the constructed prediction sub-model more generalizable and practically valuable, and providing a solid basis for model fusion and decision-making.

[0145] Example 3

[0146] like Figure 1 As shown, a navigation prediction model is constructed based on the correlation between prediction sub-model A and prediction sub-model B, and the parameters of the navigation prediction model are determined.

[0147] Furthermore, the current state's new communication data and obstruction data are input into the navigation prediction model, and the magnitude of the judgment coefficients output by the navigation prediction model determines whether to adopt mode A, mode B, mode C, or mode D as the channel allocation state. Based on the enhancement level, standard level, and edge level sample data of satellite channels, the classification threshold of the judgment coefficients is set for the automatic classification of satellite communication channels.

[0148] like Figure 8 As shown, based on the comparison results between the judgment coefficient and the division threshold, the satellite communication channel is automatically assigned to the corresponding mode A, mode B, mode C or mode D state.

[0149] Mode A indicates communication with satellites using only Ka-band channels;

[0150] Mode B indicates communication with satellites using Ku-band and Ka-band channels respectively;

[0151] Mode C indicates communication with satellites using both C-band and Ku-band channels;

[0152] Mode D indicates communication with satellites using only C-band channels.

[0153] Ka band refers to the satellite communication band with a frequency range of 26.5-40 GHz. Its advantages are large available bandwidth and high transmission rate, supporting high-speed data communication and multimedia services. However, the signal is easily attenuated by meteorological factors such as rain and clouds, and adaptive modulation and coding technology is required to ensure the reliability of the link under severe weather conditions.

[0154] Ku band refers to the satellite communication band with a frequency range of 12-18 GHz. It is characterized by a good balance between capacity and anti-attenuation performance and is widely used in fixed and mobile satellite communications in maritime, aviation, broadcasting and television and other fields. Although it also suffers from rain attenuation, it is less severe than that of Ka band and is one of the mainstream satellite communication frequency bands.

[0155] The C-band refers to the satellite communication band with a frequency range of 4-8 GHz. It has strong diffraction capability and rain attenuation resistance, is less affected by weather, and has stable and reliable signal propagation. It is especially suitable for rainy equatorial regions and critical communication services with high continuity requirements, but the available bandwidth is relatively narrow.

[0156] By constructing an integrated navigation prediction model and achieving automatic channel mode matching, its advantage lies in realizing end-to-end intelligence from data to decision-making. The system can dynamically and adaptively select the optimal communication strategy based on real-time input, which not only greatly improves channel utilization efficiency and communication reliability, but also reduces the system's dependence on human experience.

[0157] Example 4

[0158] A satellite communication channel allocation system based on a navigation prediction model includes: a data acquisition module, a data representation extraction module, a comprehensive influence number calculation module, a judgment coefficient evaluation module, a prediction sub-model construction module, a correlation analysis module, a navigation prediction model construction and parameter determination module, a channel allocation decision module, and a threshold setting module.

[0159] The data acquisition module is used to acquire historical communication data and obstruction data.

[0160] The data representation extraction module is used to extract representation data for communication data and obstruction data respectively. The representation data for communication data is the overall signal quality ratio, and the representation data for obstruction data is the probability ratio of signal being affected by obstruction.

[0161] The comprehensive impact number calculation module is used to calculate the comprehensive impact number of communication data based on the characterization data of communication data, and to calculate the comprehensive impact number of obstruction data based on the characterization data of obstruction data.

[0162] The judgment coefficient evaluation module is used to analyze historical communication and obstruction data through third parties to obtain judgment coefficients A, B, and C.

[0163] The prediction sub-model construction module is used to construct prediction sub-model A based on the correlation between judgment coefficient A and the comprehensive impact number of obstruction data, and to construct prediction sub-model B based on the correlation between judgment coefficient B and the comprehensive impact number of communication data.

[0164] The correlation analysis module is used to determine the correlation between prediction sub-model A and prediction sub-model B based on the relationship between the judgment coefficient C and the combined influence number of communication data and obstruction data.

[0165] The navigation prediction model construction and parameter determination module is used to construct a navigation prediction model based on the correlation between prediction sub-model A and prediction sub-model B, and to determine its parameters.

[0166] The channel allocation decision module is used to input the latest communication data and obstruction data into the navigation prediction model, and automatically select the channel allocation mode (mode A, B, C or D) based on the output judgment coefficient.

[0167] The threshold setting module is used to set the division threshold of the judgment coefficient based on the enhanced level, standard level, and edge level sample data to realize the automatic classification of channel state.

[0168] The systematization of the entire method into a multi-module collaborative hardware and software system has the advantage of providing a complete and implementable engineering solution. Each module has a clear responsibility and well-defined interfaces, supporting the system's scalability, maintainability, and large-scale deployment, and has good engineering applicability and industrialization prospects.

[0169] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A satellite communication channel allocation method based on a navigation prediction model, characterized in that, Includes the following steps: S1. Obtain historical communication data and obstruction data, and extract characterization data from the historical communication data and obstruction data respectively. S2. Obtain the comprehensive impact number of communication data based on the characterization data of communication data, and obtain the comprehensive impact number of obstruction data based on the characterization data of obstruction data. The overall impact factor of communication data is determined by the ratio of the overall satellite communication quality at the current ship location to the historical average overall satellite communication quality when the ship was at that location; and the overall satellite communication quality is expressed as the signal-to-weighted average (SSA) ratio. The formula for calculating the SSA ratio is: ； In the above formula, Indicates the overall signal quality ratio; This represents the normalized value of the i-th communication quality parameter; To represent the corresponding weights; The overall impact of cloud obstruction data is determined by the ratio of the cloud obstruction level of the ship in the current region and weather time to the historical average cloud obstruction level of the ship in that region and weather time. Furthermore, the cloud obstruction level is expressed as the probability ratio of signal obstruction, calculated using the following formula: ； ； In the above formula, This indicates the probability that the signal is affected by obstruction. , ... Representing meteorological characteristic variables; , ... denoted by , and z represents the regression coefficient; z represents the linear predictor, which is a linear combination of all input feature variables. S3. A third party evaluates the channel allocation judgment coefficient based on the historical communication and obstruction data. This judgment coefficient includes: 1) Summarize sample data with the same comprehensive impact number of communication data, and obtain the judgment coefficient A from a third party for the summarized sample data; 2) For sample data with the same overall impact on the hindering data, the judgment coefficient B is obtained from a third party for the summarized sample data; 3) Then, for any sample data, a third party obtains the judgment coefficient C; S4. Construct a prediction sub-model A based on the correlation between the judgment coefficient A and the comprehensive impact number of the obstruction data; construct a prediction sub-model B based on the correlation between the judgment coefficient B and the comprehensive impact number of the communication data. S5. Determine the correlation between prediction sub-model A and prediction sub-model B based on the relationship between the judgment coefficient C and the comprehensive impact number of the obstruction data and the comprehensive impact number of the communication data. S6. Construct a navigation prediction model based on the correlation between prediction sub-model A and prediction sub-model B, and determine the parameters of the navigation prediction model. S7. Input the new communication data and obstruction data of the current state into the navigation prediction model, and determine whether to use mode A, mode B, mode C or mode D as the channel allocation state based on the judgment coefficient output by the navigation prediction model. The communication data characterization data represents the overall signal quality ratio; the obstruction data characterization data represents the probability ratio of signal being affected by obstruction.

2. The satellite communication channel allocation method based on a navigation prediction model according to claim 1, characterized in that: Adjust the parameters in the prediction sub-model A until the output value of the prediction sub-model A is close to the same as the judgment coefficient A; Adjust the parameters in the prediction sub-model B until the output value of the prediction sub-model B is similar to the judgment coefficient B; Adjust the parameters in the navigation prediction model until the output value of the navigation prediction model is similar to the judgment coefficient C.

3. The satellite communication channel allocation method based on a navigation prediction model according to claim 1, characterized in that: In step S7, based on the enhanced, standard, and edge-level sample data of the satellite channel, the threshold for the division of the judgment coefficient is set for the automatic classification of satellite communication channels.

4. The satellite communication channel allocation method based on a navigation prediction model according to claim 3, characterized in that: Based on the comparison between the judgment coefficient and the division threshold, the satellite communication channel is automatically assigned to the corresponding mode A, mode B, mode C or mode D state.

5. The satellite communication channel allocation method based on a navigation prediction model according to claim 1, characterized in that: The navigation prediction model is a fusion model constructed based on prediction sub-model A and prediction sub-model B through correlation analysis. The model receives the combined impact number of obstruction data and the combined impact number of communication data as input, and calculates and outputs judgment coefficients for use in deciding the channel allocation mode.

6. The satellite communication channel allocation method based on a navigation prediction model according to claim 1, characterized in that: Mode A represents communication with satellites using only Ka-band channels; Mode B represents communication with satellites using Ku-band and Ka-band channels respectively; Mode C indicates communication with satellites using C-band and Ku-band channels respectively; Mode D indicates communication with satellites using only C-band channels.

7. A satellite communication channel allocation system based on a navigation prediction model, characterized in that, The satellite communication channel allocation method based on the navigation prediction model as described in any one of claims 1-6, the satellite communication channel allocation system based on the navigation prediction model, includes: a data acquisition module, a data characterization extraction module, a comprehensive influence number calculation module, a judgment coefficient evaluation module, a prediction sub-model construction module, a correlation analysis module, a navigation prediction model construction and parameter determination module, a channel allocation decision module, and a threshold setting module.

Images