A high-precision marine navigation and positioning system based on Beidou

Abstract

The present invention belongs to the technical field of navigation and positioning systems. The present invention discloses a high-precision marine navigation and positioning system based on Beidou, including: a star selection module for selecting suitable satellites; a verification data module for receiving and parsing satellite signals and obtaining verification data; a navigation terminal data module for obtaining initial navigation data; an error matching module for matching verification data with initial navigation data to obtain matching errors; a weather data module for regularly collecting and storing weather information; a weather correction module for extracting weather difference features and correcting errors using a pre-trained model; and a positioning determination module for calculating the current position of the navigation terminal using the corrected errors. The system takes into account the impact of differences in weather conditions on positioning accuracy, uses a weather data module to regularly collect weather data from base stations and navigation terminals, and adjusts the errors through a weather correction module, further improving positioning accuracy.

Description

Technical Field

[0001] The present invention relates to the field of navigation and positioning systems, and more specifically, to a Beidou-based marine high-precision navigation and positioning system. Background Art

[0002] Beidou satellite navigation technology has experienced rapid development in the past few decades, providing a variety of application solutions for many industries, including driving navigation, ship navigation, exploration activities, etc. In the marine field, high-precision navigation and positioning systems are particularly important. They provide key support for scientific research activities such as offshore oil exploration, resource development, and environmental monitoring, and have greatly promoted scientific research results in these fields.

[0003] Satellite navigation and positioning systems obtain location information by receiving and analyzing signals transmitted by satellites. However, in practical applications, various factors, such as the influence of the atmospheric ionosphere and clock errors, will cause errors in signal propagation, thereby affecting positioning accuracy. In order to improve positioning accuracy, existing navigation systems generally use differential correction technology. This technology relies on the base station to receive satellite signals and calculate propagation errors, and then send the error data to the navigation terminal to reduce system errors.

[0004] However, in the ocean navigation positioning scenario, since the distance between the base station and the navigation terminal is usually far, there will be a significant delay in the transmission of error data. This delay not only increases the waiting time of the navigation terminal, but may also introduce more errors during this period. In addition, a major limitation of this technology is that it ignores the impact of error diffusion due to differences in weather conditions between the base station and the navigation terminal. Specifically, even if the base station and the navigation terminal receive signals from the same satellite, due to the different weather environments of both parties, different errors may occur when the signal passes through the atmosphere. If the corresponding correction processing is not performed, these errors will affect the accuracy of the differential correction and ultimately reduce the accuracy of navigation positioning. Summary of the invention

[0005] In order to overcome the problem that the prior art does not take into account the communication delay between the base station and the navigation terminal, which may introduce more errors during the delay period and lead to a decrease in navigation and positioning accuracy, the present invention proposes a Beidou-based ocean high-precision navigation and positioning system to solve the above problem.

[0006] The present invention provides the following technical solutions:

[0007] A Beidou-based marine high-precision navigation and positioning system, comprising:

[0008] A satellite selection module is used to select at least four Beidou satellites from all visible Beidou satellites as selected satellites;

[0009] A verification data module, used to receive and analyze the signals sent by all selected satellites through the base station to obtain corresponding verification data, and store them in a verification database, wherein the verification data includes a transmission error, a first azimuth angle, a first altitude angle and a base station time;

[0010] A navigation terminal data module, used for receiving and parsing signals sent by all selected satellites through a navigation terminal to obtain corresponding initial navigation data, wherein the initial navigation data includes an initial pseudorange, a second azimuth angle, a second altitude angle and a navigation time;

[0011] An error matching module is used to match the initial navigation data with the verification data in the verification database to obtain a corresponding matching error;

[0012] A weather data module is used to collect weather data at the base station and the navigation terminal at preset time intervals, the weather data including temperature, air pressure and air humidity, record the weather data at the base station as first weather data, record the weather data at the navigation terminal as second weather data, and store the first weather data and the second weather data in a weather database after adding a time stamp;

[0013] The weather correction module is used to extract two sets of weather data from the weather database according to the navigation time, and use the difference extraction method to extract the weather difference features between the two sets of weather data, and input the weather difference features and matching errors into the pre-trained difference correction model to obtain the correction error;

[0014] The positioning determination module is used to obtain the initial pseudorange and correction error corresponding to all selected satellites, correct the initial pseudorange, obtain the corrected distance corresponding to all selected satellites, and calculate the current position of the navigation terminal based on the corrected distance.

[0015] Preferably, the step of obtaining the matching error includes:

[0016] Acquire initial navigation data, and record the selected satellite corresponding to the initial navigation data as the target satellite;

[0017] Obtain any verification data from the verification database corresponding to the target satellite;

[0018] Calculate the value index of the verification data. The calculation formula of the value index is as follows:

[0019] ,

[0020] In the formula, represents the value index, represents the first azimuth, represents the second azimuth, represents the first altitude angle, represents the second altitude angle, represents a natural constant, Indicates the preset time coefficient, Indicates the base station time, Indicates navigation time;

[0021] The value index of each verification data corresponding to the target satellite in the verification database is calculated, and the transmission error in the verification data with the largest value index is obtained as the matching error.

[0022] Preferably, extracting two sets of weather data from the weather database according to the navigation time includes: obtaining the weather data in the weather database closest to the navigation time. The first weather data is extracted in chronological order to obtain the temperature, air pressure and air humidity as a set of weather data;

[0023] Get the weather database closest to the navigation time The second weather data is extracted in chronological order, and the temperature, air pressure and air humidity are used as another set of weather data;

[0024] Said A preset integer greater than 2.

[0025] Preferably, the difference extraction method comprises:

[0026] Obtain two sets of weather data, select one set to be marked as a first data set, and the other set to be marked as a second data set;

[0027] Performing characterization processing on the first data group and the second data group to obtain a characteristic sequence;

[0028] Mark the weather data originally marked as the first data group as the second data group, and mark the weather data originally marked as the second data group as the first data group, thereby completing the label exchange of the two groups of weather data;

[0029] After completing the tag exchange of the two sets of weather data, the first data set and the second data set are again subjected to characterization processing to obtain a feature sequence;

[0030] The two obtained feature sequences are used to form weather difference features.

[0031] Preferably, the characterization processing method includes:

[0032] The temperature, air pressure and air humidity in the first data group are obtained to obtain the corresponding first temperature array, first air pressure array and first air humidity array; the temperature, air pressure and air humidity in the second data group are obtained to obtain the corresponding second temperature array, second air pressure array and second air humidity array;

[0033] Subtract the data at the same position in the second temperature array from each data in the first temperature array to obtain a temperature difference array, and divide each data in the temperature difference array by the data at the same position in the first temperature array to obtain a temperature change array;

[0034] Use each data in the first air pressure array to subtract the data at the same position in the second air pressure array to obtain an air pressure difference array, and use each data in the air pressure difference array to divide the data at the same position in the first air pressure array to obtain an air pressure change array;

[0035] Subtract the data at the same position in the second air humidity array from each data in the first air humidity array to obtain an air humidity difference array, and divide each data in the air humidity difference array by the data at the same position in the first air humidity array to obtain an air humidity change array;

[0036] Calculate and arrange in order the statistical features of the temperature difference array, the pressure difference array, the air humidity difference array, the temperature change array, the pressure change array, and the air humidity change array to obtain a feature sequence, wherein the statistical features include the mean, variance, standard deviation, maximum value, and minimum value of the array.

[0037] Preferably, the training process of the differential correction model includes:

[0038] Obtain historical data, obtain a learning data set based on the historical data, and divide the learning data set into a training set and a validation set in a ratio of 8:2;

[0039] Using a fully connected neural network as the basic framework, setting the activation function, the hierarchical structure of the fully connected neural network and the number of neurons in each layer, a machine learning model is constructed as a differential correction model;

[0040] Set the hyperparameters of the differential correction model, randomly set the weights and bias values ​​of the differential correction model, input the first error and weather difference characteristics in the training set into the differential correction model, and calculate the gradient of the loss function to the differential correction model parameters through the back propagation algorithm;

[0041] Use the optimizer to update the weights and biases in the differential correction model according to the gradient of the differential correction model parameters;

[0042] When the accuracy of the differential correction model reaches the preset index, the training of the differential correction model is completed.

[0043] Preferably, the acquiring of historical data and obtaining a learning data set based on the historical data includes:

[0044] Get any moment in history The differential errors obtained by the next two base stations through parsing the signals sent by the Beidou satellite are recorded as the first error and the second error respectively;

[0045] Get the distance between the two base stations The most recent weather data, thus obtaining two sets of weather data, and obtaining weather difference features through difference extraction method;

[0046] Use the first error, second error and weather difference features to form a learning data.

[0047] Obtain multiple learning data to form a learning data set.

[0048] Preferably, the activation function is set to a ReLU function, the hierarchical structure of the fully connected neural network has 9 layers, including 1 input layer, 1 output layer and 7 hidden layers, the input layer and the output layer are each provided with 64 neurons, and the hidden layer is provided with 128 neurons; the loss function is a mean square error function, and the optimizer is an Adam optimizer.

[0049] Preferably, the step of determining whether the accuracy of the differential correction model reaches a preset index comprises:

[0050] Obtain any data from the validation set as validation data, input the first error and weather difference characteristics in the validation data into the differential correction model, and obtain an output value;

[0051] Using the formula Calculate the proximity corresponding to the verification data, where Indicates proximity, Represents the output value, represents the second error in the validation data;

[0052] Obtain at least ten verification data and calculate the corresponding proximity, and calculate the average of at least ten proximity as the accuracy;

[0053] Compare the accuracy rate with the preset index. If the accuracy rate is greater than or equal to the preset index, the judgment result is that the accuracy rate reaches the preset index.

[0054] Preferably, calculating the current position of the navigation terminal according to the corrected distance includes:

[0055] Get the corrected distances corresponding to at least four selected satellites;

[0056] Using the known position and corrected distance of each selected satellite, construct the equation of a sphere with the four selected satellite positions as the center and the corresponding corrected distance as the radius;

[0057] The intersection of these spherical equations is solved using the least squares method, and this intersection is the current position of the navigation terminal.

[0058] The present invention provides a Beidou-based marine high-precision navigation and positioning system, which has the following beneficial effects:

[0059] 1. The verification data module will receive and analyze the signals sent by all selected satellites, and store the corresponding verification data in the database. When these data need to be used for error correction, the error matching module directly obtains the corresponding data from the database for matching to obtain the error, without waiting for real-time data transmission. This greatly reduces the waiting time caused by data transmission delays, thereby reducing the errors introduced during the delay period and improving the operating efficiency and positioning accuracy of the navigation and positioning system.

[0060] 2. By considering the difference in weather conditions between the base station and the navigation terminal, the accuracy of satellite navigation positioning is significantly improved. Specifically, when the navigation terminal and the base station receive signals from the same satellite, different errors may occur due to differences in weather conditions. If such errors are not effectively considered and processed, they will affect the final positioning accuracy. The weather data module will collect weather data from the base station and the navigation terminal, and then the weather correction module will adjust the errors based on these data to reduce the errors caused by differences in weather conditions, thereby improving positioning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 A schematic diagram of a module of a Beidou-based marine high-precision navigation and positioning system of the present invention;

[0062] Figure 2 The present invention is a schematic diagram of the training process of a differential correction model of a Beidou-based marine high-precision navigation and positioning system. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] Embodiment 1:

[0065] See also Figure 1 In this embodiment, a Beidou-based marine high-precision navigation and positioning system includes:

[0066] A satellite selection module is used to select at least four Beidou satellites from all visible Beidou satellites as selected satellites;

[0067] A verification data module, used to receive and analyze the signals sent by all selected satellites through the base station to obtain corresponding verification data, and store them in a verification database, wherein the verification data includes a transmission error, a first azimuth angle, a first altitude angle and a base station time;

[0068] A navigation terminal data module, used for receiving and parsing signals sent by all selected satellites through a navigation terminal to obtain corresponding initial navigation data, wherein the initial navigation data includes an initial pseudorange, a second azimuth angle, a second altitude angle and a navigation time;

[0069] In this embodiment, it should be noted that a Beidou satellite is selected from all visible Beidou satellites. It is a technology widely used in existing satellite navigation systems; the transmission error is the differential error calculated by the base station; it should be noted that when the base station and the navigation terminal calculate the distance between the satellite and the satellite by parsing the satellite signal, it is usually based on the time difference between signal sending and receiving multiplied by the theoretical speed of the signal. However, since signal propagation may be affected by factors such as tropospheric errors, the actual propagation speed is not always the theoretical value, so there is an error in the calculated distance. Since the position of the base station is known, its true distance from the satellite can be accurately calculated. By comparing the true distance with the distance obtained based on signal analysis, the differential error is obtained.

[0070] The first azimuth angle is the azimuth angle between the base station and the satellite, the first altitude angle is the altitude angle between the base station and the satellite, and the base station time is the time when the base station receives the satellite signal; the initial pseudorange is the pseudorange from the satellite to the navigation terminal obtained by the navigation terminal decoding the satellite signal, the second azimuth angle is the azimuth angle between the navigation terminal and the satellite, the second altitude angle is the altitude angle between the navigation terminal and the satellite, and the navigation time is the time when the navigation terminal receives the satellite signal.

[0071] An error matching module is used to match the initial navigation data with the verification data in the verification database to obtain a corresponding matching error;

[0072] The step of obtaining the matching error comprises:

[0073] Acquire initial navigation data, and record the selected satellite corresponding to the initial navigation data as the target satellite;

[0074] Obtain any verification data from the verification database corresponding to the target satellite;

[0075] Calculate the value index of the verification data. The calculation formula of the value index is as follows:

[0076] ,

[0077] In the formula, represents the value index, represents the first azimuth, represents the second azimuth, represents the first altitude angle, represents the second altitude angle, represents a natural constant, Indicates the preset time coefficient, Indicates the base station time, Indicates navigation time;

[0078] The value index of each verification data corresponding to the target satellite in the verification database is calculated, and the transmission error in the verification data with the largest value index is obtained as the matching error.

[0079] In this embodiment, first, the initial navigation data is obtained; then, the verification data corresponding to the target satellite is extracted, and the value index of each data is calculated. Finally, the transmission error in the verification data with the largest value index is obtained as the matching error. Through this technical solution, we can effectively find the most valuable verification data from the verification database, and use the error of the data as the matching error. It should be noted that the value is mainly reflected in the similarity between the time and space relationship between the navigation terminal and the satellite, and the time and space relationship between the base station and the satellite. The higher the similarity of these two sets of relationships, the higher the value of the data, that is, the closer the transmission error is to the error when the navigation terminal parses the satellite signal to calculate the distance, which can be used to correct the distance calculated by the navigation terminal. In this way, the error can be corrected quickly without waiting for the base station to send differential data, thereby ensuring the real-time nature of navigation.

[0080] A weather data module is used to collect weather data at the base station and the navigation terminal at preset time intervals, the weather data including temperature, air pressure and air humidity, record the weather data at the base station as first weather data, record the weather data at the navigation terminal as second weather data, and store the first weather data and the second weather data in a weather database after adding a time stamp;

[0081] The weather correction module is used to extract two sets of weather data from the weather database according to the navigation time, and use the difference extraction method to extract the weather difference features between the two sets of weather data, and input the weather difference features and matching errors into the pre-trained difference correction model to obtain the correction error;

[0082] The step of extracting two sets of weather data from the weather database according to the navigation time includes: obtaining the weather data in the weather database closest to the navigation time. The first weather data is extracted in chronological order to obtain the temperature, air pressure and air humidity as a set of weather data;

[0083] Get the weather database closest to the navigation time The second weather data is extracted in chronological order, and the temperature, air pressure and air humidity are used as another set of weather data;

[0084] Said A preset integer greater than 2.

[0085] In this embodiment, it should be noted that these weather data are known factors that cause tropospheric errors. Weather data can be collected through corresponding sensors or obtained by connecting to a meteorological system. After collection or acquisition, they are stored in a weather database and retrieved. A typical value for is 10.

[0086] The difference extraction method comprises:

[0087] Obtain two sets of weather data, select one set to be marked as a first data set, and the other set to be marked as a second data set;

[0088] Performing characterization processing on the first data group and the second data group to obtain a characteristic sequence;

[0089] Mark the weather data originally marked as the first data group as the second data group, and mark the weather data originally marked as the second data group as the first data group, thereby completing the label exchange of the two groups of weather data;

[0090] After completing the tag exchange of the two sets of weather data, the first data set and the second data set are again subjected to characterization processing to obtain a feature sequence;

[0091] The two obtained feature sequences are used to form weather difference features.

[0092] The characterization processing method comprises:

[0093] The temperature, air pressure and air humidity in the first data group are obtained to obtain the corresponding first temperature array, first air pressure array and first air humidity array; the temperature, air pressure and air humidity in the second data group are obtained to obtain the corresponding second temperature array, second air pressure array and second air humidity array;

[0094] Subtract the data at the same position in the second temperature array from each data in the first temperature array to obtain a temperature difference array, and divide each data in the temperature difference array by the data at the same position in the first temperature array to obtain a temperature change array;

[0095] Use each data in the first air pressure array to subtract the data at the same position in the second air pressure array to obtain an air pressure difference array, and use each data in the air pressure difference array to divide the data at the same position in the first air pressure array to obtain an air pressure change array;

[0096] Subtract the data at the same position in the second air humidity array from each data in the first air humidity array to obtain an air humidity difference array, and divide each data in the air humidity difference array by the data at the same position in the first air humidity array to obtain an air humidity change array;

[0097] Calculate and arrange in order the statistical features of the temperature difference array, the pressure difference array, the air humidity difference array, the temperature change array, the pressure change array, and the air humidity change array to obtain a feature sequence, wherein the statistical features include the mean, variance, standard deviation, maximum value, and minimum value of the array.

[0098] In this embodiment, the weather difference feature reflects the change characteristics of the weather data between the base station and the navigation terminal. The difference between the matching error caused by different weather and the real error of the navigation terminal is captured through machine learning methods, and then the matching error is further corrected to obtain a corrected error closer to the real error, thereby improving the positioning accuracy.

[0099] The positioning determination module is used to obtain the initial pseudorange and correction error corresponding to all selected satellites, correct the initial pseudorange, obtain the corrected distance corresponding to all selected satellites, and calculate the current position of the navigation terminal based on the corrected distance.

[0100] Calculating the current position of the navigation terminal according to the corrected distance includes:

[0101] Get the corrected distances corresponding to at least four selected satellites;

[0102] Using the known position and corrected distance of each selected satellite, construct the equation of a sphere with the four selected satellite positions as the center and the corresponding corrected distance as the radius;

[0103] The intersection of these spherical equations is solved using the least squares method, and this intersection is the current position of the navigation terminal.

[0104] In this embodiment, the initial pseudorange is corrected, specifically, the correction error is subtracted from the initial pseudorange, and the spherical equation may be established on the basis of the earth coordinate system.

[0105] Embodiment 2:

[0106] See also Figure 2 Based on Example 1, the training process of the differential correction model includes:

[0107] S1. Obtain historical data, obtain a learning data set based on the historical data, and divide the learning data set into a training set and a validation set in a ratio of 8:2;

[0108] S2. Use a fully connected neural network as the basic framework, set the activation function, the hierarchical structure of the fully connected neural network and the number of neurons in each layer, and build a machine learning model as a differential correction model;

[0109] S3, setting the hyperparameters of the differential correction model, randomly setting the weights and bias values ​​of the differential correction model, inputting the first error and weather difference characteristics in the training set into the differential correction model, and calculating the gradient of the loss function to the differential correction model parameters through the back propagation algorithm;

[0110] S4, using the optimizer to update the weights and biases in the differential correction model according to the gradient of the differential correction model parameters;

[0111] S5. When the accuracy of the differential correction model reaches a preset indicator, the training of the differential correction model is completed.

[0112] The acquiring of historical data and obtaining a learning data set according to the historical data includes:

[0113] Get history The differential errors obtained by any two base stations at the time by analyzing the signals sent by the Beidou satellite are recorded as the first error and the second error respectively;

[0114] Get the distance between the two base stations The most recent weather data, thus obtaining two sets of weather data, and obtaining weather difference features through difference extraction method;

[0115] Use the first error, second error and weather difference features to form a learning data.

[0116] Obtain multiple learning data to form a learning data set.

[0117] The step of judging whether the accuracy of the differential correction model reaches a preset index comprises:

[0118] Obtain any data from the validation set as validation data, input the first error and weather difference characteristics in the validation data into the differential correction model, and obtain an output value;

[0119] Using the formula Calculate the proximity corresponding to the verification data, where Indicates proximity, Represents the output value, represents the second error in the validation data;

[0120] Obtain at least ten verification data and calculate the corresponding proximity, and calculate the average of at least ten proximity as the accuracy;

[0121] Compare the accuracy rate with the preset index. If the accuracy rate is greater than or equal to the preset index, the judgment result is that the accuracy rate reaches the preset index.

[0122] In this embodiment, for the special case of navigation and positioning in the marine environment, due to the sparse distribution of base stations in the ocean, the navigation terminal and the base station are far apart, resulting in significant differences in the weather conditions between the two. This weather difference will significantly affect the propagation of satellite signals in the troposphere, resulting in large inconsistencies in the tropospheric errors of satellite signals received between the navigation terminal and the base station. This inconsistency will further affect the accuracy of the positioning correction value calculated by the base station based on the differential technology, thereby reducing the accuracy of navigation positioning.

[0123] Using the weather differences between the two base stations and their corresponding differential errors as training data, a differential error correction model was constructed based on a fully connected neural network. Through a large amount of historical data training, combined with a fully connected neural network and back propagation algorithm for model optimization, the model can learn and identify the potential relationship between weather differences and differential errors. After the model training is completed, the navigation positioning error can be corrected in real time according to the weather conditions of the current base station and navigation terminal through the trained differential correction model. This method improves the accuracy and applicability of differential error correction by effectively capturing and utilizing weather difference characteristics, thereby significantly improving the accuracy of navigation positioning in marine environments.

[0124] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0125] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0126] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0127] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A Beidou-based marine high-precision navigation and positioning system, characterized in that: include: A satellite selection module is used to select at least four Beidou satellites from all visible Beidou satellites as selected satellites; A verification data module, used to receive and analyze the signals sent by all selected satellites through the base station to obtain corresponding verification data, and store them in a verification database, wherein the verification data includes a transmission error, a first azimuth angle, a first altitude angle and a base station time; A navigation terminal data module, used for receiving and parsing signals sent by all selected satellites through a navigation terminal to obtain corresponding initial navigation data, wherein the initial navigation data includes an initial pseudorange, a second azimuth angle, a second altitude angle and a navigation time; An error matching module is used to match the initial navigation data with the verification data in the verification database to obtain a corresponding matching error; A weather data module is used to collect weather data at the base station and the navigation terminal at preset time intervals, the weather data including temperature, air pressure and air humidity, record the weather data at the base station as first weather data, record the weather data at the navigation terminal as second weather data, and store the first weather data and the second weather data in a weather database after adding a time stamp; The weather correction module is used to extract two sets of weather data from the weather database according to the navigation time, and use the difference extraction method to extract the weather difference features between the two sets of weather data, and input the weather difference features and matching errors into the pre-trained difference correction model to obtain the correction error; The positioning determination module is used to obtain the initial pseudorange and correction error corresponding to all selected satellites, correct the initial pseudorange, obtain the corrected distance corresponding to all selected satellites, and calculate the current position of the navigation terminal based on the corrected distance.

2. The Beidou-based marine high-precision navigation and positioning system according to claim 1, characterized in that: The step of obtaining the matching error comprises: Acquire initial navigation data, and record the selected satellite corresponding to the initial navigation data as the target satellite; Obtain any verification data from the verification database corresponding to the target satellite; Calculate the value index of the verification data. The calculation formula of the value index is as follows: ， In the formula, represents the value index, represents the first azimuth, represents the second azimuth, represents the first altitude angle, represents the second altitude angle, represents a natural constant, Indicates the preset time coefficient, Indicates the base station time, Indicates navigation time; The value index of each verification data corresponding to the target satellite in the verification database is calculated, and the transmission error in the verification data with the largest value index is obtained as the matching error.

3. The Beidou-based marine high-precision navigation and positioning system according to claim 2, characterized in that: The step of extracting two sets of weather data from the weather database according to the navigation time includes: obtaining the weather data in the weather database closest to the navigation time. The first weather data is extracted in chronological order to obtain the temperature, air pressure and air humidity as a set of weather data; Get the weather database closest to the navigation time The second weather data is extracted in chronological order, and the temperature, air pressure and air humidity are used as another set of weather data; Said It is a preset integer greater than 2.

4. The Beidou-based marine high-precision navigation and positioning system according to claim 3 is characterized in that: The difference extraction method comprises: Obtain two sets of weather data, select one set to be marked as a first data set, and the other set to be marked as a second data set; Performing characterization processing on the first data group and the second data group to obtain a characteristic sequence; Mark the weather data originally marked as the first data group as the second data group, and mark the weather data originally marked as the second data group as the first data group, thereby completing the label exchange of the two groups of weather data; After completing the tag exchange of the two sets of weather data, the first data set and the second data set are again subjected to characterization processing to obtain a feature sequence; The two obtained feature sequences are used to form weather difference features.

5. The Beidou-based marine high-precision navigation and positioning system according to claim 4, characterized in that: The characterization processing method comprises: The temperature, air pressure and air humidity in the first data group are obtained to obtain the corresponding first temperature array, first air pressure array and first air humidity array; the temperature, air pressure and air humidity in the second data group are obtained to obtain the corresponding second temperature array, second air pressure array and second air humidity array; Subtract the data at the same position in the second temperature array from each data in the first temperature array to obtain a temperature difference array, and divide each data in the temperature difference array by the data at the same position in the first temperature array to obtain a temperature change array; Use each data in the first air pressure array to subtract the data at the same position in the second air pressure array to obtain an air pressure difference array, and use each data in the air pressure difference array to divide the data at the same position in the first air pressure array to obtain an air pressure change array; Subtract the data at the same position in the second air humidity array from each data in the first air humidity array to obtain an air humidity difference array, and divide each data in the air humidity difference array by the data at the same position in the first air humidity array to obtain an air humidity change array; Calculate and arrange in order the statistical features of the temperature difference array, the pressure difference array, the air humidity difference array, the temperature change array, the pressure change array, and the air humidity change array to obtain a feature sequence, wherein the statistical features include the mean, variance, standard deviation, maximum value, and minimum value of the array.

6. The Beidou-based marine high-precision navigation and positioning system according to claim 5, characterized in that: The training process of the differential correction model includes: Obtain historical data, obtain a learning data set based on the historical data, and divide the learning data set into a training set and a validation set in a ratio of 8:2; Using a fully connected neural network as the basic framework, setting the activation function, the hierarchical structure of the fully connected neural network and the number of neurons in each layer, a machine learning model is constructed as a differential correction model; Set the hyperparameters of the differential correction model, randomly set the weights and bias values ​​of the differential correction model, input the first error and weather difference characteristics in the training set into the differential correction model, and calculate the gradient of the loss function to the differential correction model parameters through the back propagation algorithm; Use the optimizer to update the weights and biases in the differential correction model according to the gradient of the differential correction model parameters; When the accuracy of the differential correction model reaches the preset index, the training of the differential correction model is completed.

7. The Beidou-based marine high-precision navigation and positioning system according to claim 6, characterized in that: The acquiring of historical data and obtaining a learning data set according to the historical data includes: Get any moment in history The differential errors obtained by the next two base stations through parsing the signals sent by the Beidou satellite are recorded as the first error and the second error respectively; Get the distance between the two base stations The most recent weather data, thus obtaining two sets of weather data, and obtaining weather difference features through difference extraction method; Use the first error, second error and weather difference features to form a learning data. Obtain multiple learning data to form a learning data set.

8. The Beidou-based marine high-precision navigation and positioning system according to claim 7, characterized in that: The activation function is set to the ReLU function, the hierarchical structure of the fully connected neural network has 9 layers, including 1 input layer, 1 output layer and 7 hidden layers, the input layer and the output layer are each provided with 64 neurons, and the hidden layer is provided with 128 neurons; the loss function is the mean square error function, and the optimizer is the Adam optimizer.

9. The Beidou-based marine high-precision navigation and positioning system according to claim 8, characterized in that: The step of judging whether the accuracy of the differential correction model reaches a preset index comprises: Obtain any data from the validation set as validation data, input the first error and weather difference characteristics in the validation data into the differential correction model, and obtain an output value; Using the formula Calculate the proximity corresponding to the verification data, where Indicates proximity, Represents the output value, represents the second error in the validation data; Obtain at least ten verification data and calculate the corresponding proximity, and calculate the average of at least ten proximity as the accuracy; Compare the accuracy rate with the preset index. If the accuracy rate is greater than or equal to the preset index, the judgment result is that the accuracy rate reaches the preset index.

10. The Beidou-based marine high-precision navigation and positioning system according to claim 5, characterized in that: Calculating the current position of the navigation terminal according to the corrected distance includes: Get the corrected distances corresponding to at least four selected satellites; Using the known position and corrected distance of each selected satellite, construct the equation of a sphere with the four selected satellite positions as the center and the corresponding corrected distance as the radius; The intersection of these spherical equations is solved using the least squares method, and this intersection is the current position of the navigation terminal.

Images