Accelerated Method and Device for Precision Positioning Based on Satellite / LSTM Neural Network

Abstract

This application discloses a precise positioning acceleration method and apparatus based on low-Earth orbit satellites / LSTM neural networks. The method includes: a terminal acquiring wide-lane phase fractional deviation products and narrow-lane phase fractional deviation products generated and encoded by a server using an LSTM-based method, as well as a first GNSS combined observation value; determining the first GNSS wide-lane real ambiguity based on these; determining the initial GNSS de-ionization combined real ambiguity and other parameters based on a precise single-point positioning algorithm; performing inter-satellite single difference on the first GNSS wide-lane real ambiguity, and fixing its integer ambiguity using the wide-lane phase fractional deviation product; performing inter-satellite single difference on the initial GNSS de-ionization combined real ambiguity, and determining the first GNSS narrow-lane real ambiguity using the aforementioned wide-lane integer ambiguity, and then fixing its integer ambiguity using the narrow-lane phase fractional deviation product using a partial ambiguity fixing method; and calculating the user coordinate fixed solution using relevant parameters.

Description

Technical Field

[0001] This application relates to the field of satellite navigation and positioning technology, and relates to, but is not limited to, a precise positioning acceleration method and apparatus based on satellite / LSTM neural networks. Background Technology

[0002] The Global Navigation Satellite System (GNSS) uses radio ranging signals transmitted by navigation satellites covering the globe to determine the precise location and time status of ground users, providing positioning, navigation, and timing (PNT) services. Positioning is the core function and service of the GNSS system, and the algorithms and models involved in precise positioning technology have long been a hot topic of research and discussion.

[0003] Precise Point Positioning (PPP) is one of the most representative precise positioning technologies. It utilizes pre-estimated precise orbits and clock biases, comprehensively considering various error corrections for GNSS signals, to achieve high-precision absolute global positioning with a single receiver. Due to its high positioning accuracy, simple structure, and flexible operation, it has been widely used in many fields, such as precision agriculture, precision mapping, deformation monitoring, water vapor inversion, precision navigation, and positioning. However, because GNSS satellites orbit at high altitudes (typically above 20,000 kilometers), the satellite geometry changes slowly relative to the ground station, resulting in centimeter-level positioning accuracy often requiring tens of minutes. To shorten initialization time and improve positioning accuracy, Ambiguity Resolution (AR) has been proposed. However, reliable AR also requires long observation times, which limits the real-time application of PPP.

[0004] With the decreasing cost of commercial spaceflight, Low Earth Orbit (LEO) communication constellations such as SpaceX and OneWeb are booming. LEO satellites typically orbit at altitudes between 400 and 1500 kilometers, and their geometry changes approximately 40 times faster than that of Medium Earth Orbit (MEO) GNSS satellites, which facilitates faster PPP convergence. Some studies have shown that LEO constellations can significantly improve the performance of PPP and PPPAR, but few studies have investigated the contribution of LEO satellite floating-point solutions to PPP AR. Summary of the Invention

[0005] In view of this, embodiments of this application provide a precise positioning acceleration method and apparatus based on satellite / LSTM neural network to solve the problems of low positioning efficiency and slow convergence time.

[0006] The technical solution of this application embodiment is implemented as follows:

[0007] In a first aspect, embodiments of this application provide a precise positioning acceleration method based on satellite / LSTM neural networks, the method comprising:

[0008] The server generates wide-lane phase fractional deviation products and narrow-lane phase fractional deviation products based on the LSTM method, and sends them to the terminal after encoding them using a specific method.

[0009] The terminal acquires the encoded wide-lane phase fractional deviation product, narrow-lane phase fractional deviation product, and the first GNSS combined observation value; based on the first GNSS combined observation value, it determines the first GNSS wide-lane real ambiguity.

[0010] The terminal uses a precise single-point positioning algorithm based on the fusion of LEO and GNSS satellites to determine the floating-point solution of user coordinates, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, tropospheric delay, LEO receiver clock error, and GNSS receiver clock error.

[0011] The terminal performs inter-satellite single-difference processing on the first GNSS wide-lane real ambiguity, and uses the wide-lane phase fractional deviation product to fix the GNSS wide-lane real ambiguity after inter-satellite single-difference processing to obtain the first GNSS wide-lane integer ambiguity.

[0012] The terminal performs inter-satellite single-difference processing on the initial GNSS de-ionization combined real ambiguity. Based on the GNSS de-ionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity, the first GNSS narrow-lane real ambiguity is determined. The first GNSS narrow-lane real ambiguity is fixed using a partial ambiguity fixing method with the narrow-lane phase fractional deviation product to obtain the first GNSS narrow-lane integer ambiguity.

[0013] The terminal calculates the fixed solution for user coordinates based on the first GNSS combined observations, the first GNSS narrow lane integer ambiguity, the narrow lane phase fractional deviation product, the first GNSS wide lane integer ambiguity, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution for user coordinates.

[0014] Secondly, embodiments of this application provide a precision positioning acceleration device based on a satellite / LSTM neural network, comprising:

[0015] The server module and the terminal module, wherein:

[0016] The server module includes a generation and encoding module, and the terminal module includes an acquisition module, a determination module, a first fixing module, a second fixing module, and a calculation module.

[0017] The generation and encoding module is used to generate wide-lane phase fractional deviation products and narrow-lane phase fractional deviation products based on the LSTM method, and then encode them using a specific method before sending them to the terminal.

[0018] The acquisition module is used to acquire the encoded wide-lane phase fractional deviation product, narrow-lane phase fractional deviation product, and first GNSS combined observation value; and to determine the first GNSS wide-lane real ambiguity based on the first GNSS combined observation value.

[0019] The determining module is used to determine the floating-point solution of user coordinates, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock error, and the GNSS receiver clock error based on the precise single-point positioning algorithm fused by LEO satellites and GNSS satellites.

[0020] The first fixing module is used to perform inter-satellite single-difference processing on the first GNSS wide-lane real ambiguity, and use the wide-lane phase fractional deviation product to fix the GNSS wide-lane real ambiguity after inter-satellite single-difference processing to obtain the first GNSS wide-lane integer ambiguity.

[0021] The second fixing module is used to perform inter-satellite single-difference processing on the initial GNSS de-ionization combined real ambiguity, and based on the GNSS de-ionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity, determine the first GNSS narrow-lane real ambiguity, and use the narrow-lane phase fractional deviation product to fix the first GNSS narrow-lane real ambiguity using a partial ambiguity fixing method to obtain the first GNSS narrow-lane integer ambiguity;

[0022] The calculation module is used to calculate the fixed solution of user coordinates based on the first GNSS combined observations, the first GNSS narrow lane integer ambiguity, the narrow lane phase fractional deviation product, the first GNSS wide lane integer ambiguity, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution of user coordinates.

[0023] The beneficial effects of the technical solutions provided in this application include at least the following:

[0024] By integrating low-Earth orbit (LEO) satellites and GNSS systems, and combining wide-lane and narrow-lane phase fractional deviation products to fix ambiguity, high-precision calculation of user coordinate fixed solutions is achieved, improving the efficiency, accuracy, and reliability of precise positioning. In particular, through inter-satellite single-difference processing and ambiguity fixing strategies, the impact of errors is effectively reduced, and positioning results are optimized. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein:

[0026] Figure 1 A flowchart illustrating a precise positioning acceleration method based on a satellite / LSTM neural network, provided for an embodiment of this application;

[0027] Figure 2 A schematic diagram of the basic framework of GNSS data and product messages provided for embodiments of this application;

[0028] Figure 3 This is a schematic diagram of a specific data protocol for satellite phase deviation coding provided in an embodiment of this application;

[0029] Figure 4 A schematic diagram of a satellite phase deviation coding data protocol provided in an embodiment of this application;

[0030] Figure 5 A schematic diagram illustrating the initial fixed-time operation of a single-system PPP system with 1, 2, 3, 4, and all LEO satellites in different latitude regions, provided for embodiments of this application.

[0031] Figure 6 This is a schematic diagram of the composition structure of a precision positioning acceleration device based on a satellite / LSTM neural network, provided in an embodiment of this application. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0033] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0034] It should be noted that the terms "first, second, and third" used in the embodiments of this application are merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, and third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0035] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of this application pertain. It should also be understood that terms such as those defined in general dictionaries should be understood to have a meaning consistent with their meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0036] This application provides a method for accelerating precise positioning based on satellite / LSTM neural networks, applied to electronic devices. These electronic devices include, but are not limited to, mobile phones, laptops, tablets and handheld internet devices, multimedia devices, streaming media devices, mobile internet devices, wearable devices, or other types of electronic devices. The functions implemented by this method can be achieved by a processor in the electronic device calling program code. The program code can be stored in a computer storage medium; therefore, the electronic device includes at least a processor and a storage medium. The processor can be used to perform the precise positioning acceleration process based on satellite / LSTM neural networks, and the memory can be used to store the data required and generated during the precise positioning acceleration process based on satellite / LSTM neural networks.

[0037] Figure 1 A flowchart illustrating a precise positioning acceleration method based on a satellite / LSTM neural network, provided for embodiments of this application, is shown below. Figure 1 As shown, the method includes at least the following steps:

[0038] Step S110: The server generates wide-lane phase fractional deviation products and narrow-lane phase fractional deviation products based on the LSTM method, and sends them to the terminal after encoding them using a specific method; the terminal obtains the encoded wide-lane phase fractional deviation products, narrow-lane phase fractional deviation products and the first GNSS combined observation value; based on the first GNSS combined observation value, the first GNSS wide-lane real ambiguity is determined;

[0039] Here, the method consists of two parts: a server and a user (i.e., a terminal). The server can calculate and generate a wide-lane (WL) fractional cycle bias (FCB) product (i.e., satellite-side wide-lane FCB) and a narrow-lane (NL) fractional cycle bias product (i.e., satellite-side narrow-lane FCB) for each satellite. Then, according to a fixed period, the coded wide-lane and narrow-lane fractional cycle bias products of all on-orbit satellites are sent to the terminal. When the terminal receives a signal (such as the first GNSS combined observation value) from a satellite (such as a satellite in a GNSS system), it can obtain the corresponding wide-lane and narrow-lane fractional cycle bias products based on the satellite's unique identifier and use them for ambiguity positioning (i.e., to achieve integer ambiguity resolution by correcting errors).

[0040] Step S120: The terminal determines the floating-point solution of user coordinates, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, and the GNSS receiver clock bias based on the precise single-point positioning algorithm fused by LEO satellites and GNSS satellites.

[0041] Here, performing PPP on both LEO and GNSS simultaneously can yield floating-point solutions for user coordinates, as well as parameters such as the combined real-valued ambiguity of LEO and GNSS ionosphere de-ambiguity.

[0042] Step S130: The terminal performs inter-satellite single-difference processing on the first GNSS wide-lane real ambiguity, and uses the wide-lane phase fractional deviation product to fix the GNSS wide-lane real ambiguity after inter-satellite single-difference processing to obtain the first GNSS wide-lane integer ambiguity.

[0043] Here, the real ambiguity of the first GNSS wide lane is used to make an inter-satellite single difference to eliminate the wide lane FCB at the receiver, and the wide lane ambiguity is fixed by rounding up the wide lane FCB product at the satellite end (i.e. the wide lane phase fractional deviation product).

[0044] Step S140: The terminal performs inter-satellite single-difference processing on the initial GNSS de-ionization combined real ambiguity. Based on the GNSS de-ionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity, the first GNSS narrow-lane real ambiguity is determined. The first GNSS narrow-lane real ambiguity is fixed using a partial ambiguity fixing method with the narrow-lane phase fractional deviation product to obtain the first GNSS narrow-lane integer ambiguity.

[0045] Here, the real ambiguity of the GNSS deionization combination is used to calculate the inter-satellite single difference and combined with the fixed wide-lane ambiguity to calculate the real narrow-lane ambiguity. The narrow-lane ambiguity is fixed using the LAMBDA (Least-square AMBiguity Decorrelation Adjustment) algorithm with the narrow-lane FCB product (i.e. the narrow-lane phase fractional deviation product).

[0046] In step S150, the terminal calculates the fixed solution for user coordinates based on the first GNSS combined observation, the first GNSS narrow lane integer ambiguity, the narrow lane phase fractional deviation product, the first GNSS wide lane integer ambiguity, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution for user coordinates.

[0047] In the above embodiments, by fusing low-Earth orbit (LEO) satellites and GNSS systems, and combining wide-lane and narrow-lane phase fractional deviation products to fix ambiguity, high-precision calculation of user coordinate fixed solutions is achieved, improving the efficiency, accuracy and reliability of precise positioning. In particular, through inter-satellite single-difference processing and ambiguity fixing strategies, the impact of errors is effectively reduced and the positioning results are optimized.

[0048] In some embodiments, the method further includes:

[0049] Step S101: The server obtains raw GNSS observations, which include the frequency, phase, pseudorange, wide lane carrier phase observations, and narrow lane pseudorange observations of the GNSS dual-frequency carrier.

[0050] Here, GNSS observations mainly include code pseudorange measurements. and carrier phase There are two types, and their observation equations can be expressed as shown in the following formulas (1) and (2):

[0051] Formula (1);

[0052] Formula (2);

[0053] Among them, subscript Station number, Indicates frequency point, superscript Number the satellite. This refers to GNSS systems, such as GPS, BDS-3, and other Code Division Multiple Access (CDMA) systems. This indicates the geometric distance between the receiver (i.e., the observation station) and the satellite. The speed of light in a vacuum For receiver clock bias, For satellite clock bias,

[0054] The total tropospheric delay in the direction of signal propagation can be divided into dry delay and wet delay. , and These represent the dry and wet components of the troposphere in the zenith direction. and Let dry delay and wet delay be the projection functions. for Ionospheric delay at frequency (GNSS frequency), and These represent the hardware delay deviations in pseudorange measurement at the receiver and satellite ends, respectively. and These are the uncalibrated phase delay (UPD) at the receiver and satellite ends, respectively. for Integer ambiguity at a frequency point, corresponding to a wavelength of , , These represent the observation noise errors for code pseudorange and carrier phase, respectively. There are also other error terms, such as Earth's rotation, satellite and receiver antenna phase center deviation and variation, antenna phase entanglement, solid tides, ocean tides, atmospheric tides, and relativistic effects, which can be corrected using existing models and are therefore not listed in the above equation.

[0055] In this embodiment of the application, the frequency of the GNSS dual-frequency carrier can be expressed as: f 1 and f 2, , express The dual-frequency carrier phase at different frequency points of the system, such as L1 and L2 of GPS, B1 and B3 of BDS-3, and E1 and E5a of Galileo, can be expressed as: and Dual-frequency pseudorange can be expressed as and The wide-lane carrier phase observation can be expressed as The pseudorange observation value of the narrow alley can be expressed as .

[0056] Step S102: The server performs a weighted calculation on the phase of the dual-frequency carrier based on the frequency of the GNSS dual-frequency carrier to obtain the first GNSS deionization carrier phase combination observation value.

[0057] Step S103: The server performs a weighted calculation on the dual-frequency pseudorange based on the frequency of the GNSS dual-frequency carrier to obtain the first GNSS ionospheric pseudorange combined observation value.

[0058] To eliminate first-order ionospheric delay, the original observations are usually combined to form ionospheric de-electrosphere combined observations (first GNSS ionospheric de-electrosphere pseudorange combined observations). Combined observations with the first GNSS ionospheric carrier phase The combination method can be expressed by the following formulas (3) and (4):

[0059] Formula (3);

[0060] Formula (4);

[0061] The combined observation equation can be expressed as shown in the following formula (5):

[0062] Formula (5);

[0063] in, Indicates the geometric distance between the station and the star. The speed of light in a vacuum For receiver clock bias, For satellite clock bias, The total tropospheric delay in the direction of signal propagation. and These are the hardware delay deviations in pseudorange measurement of the receiver and satellite ends without ionosphere. and These are the hardware delay deviations in the phase of the ionospherically-free combined carrier at the receiver and satellite ends, respectively. and For carrier wavelengths and ambiguity without ionosphere, , These represent the observation noise errors of pseudorange and carrier phase in ionospheric combination coding, respectively. The ambiguity of the IF combination, the pseudorange at the receiver end, and the hardware delay deviation of the phase at the satellite end are shown in formulas (6) to (10) below:

[0064] Formula (6);

[0065] Formula (7);

[0066] Formula (8);

[0067] Formula (9);

[0068] Formula (10);

[0069] in, and These are the hardware delay deviations in pseudorange measurement of the receiver and satellite ends without ionosphere. and For the receiver end at f 1. f Hardware delay deviation in code pseudorange measurement at 2 frequencies and For the receiver end at f 1. f Hardware delay deviation in code pseudorange measurement at 2 frequencies and These are the hardware delay deviations in the phase of the ionospherically-free combined carrier at the receiver and satellite ends, respectively. and For the receiver end at f 1. f Carrier phase hardware delay deviation at 2 frequencies and For satellite end in f 1. f Carrier phase hardware delay deviation at 2 frequencies and For carrier wavelengths and ambiguity without ionosphere, and Receiver end at f 1. f Carrier phase ambiguity when observing satellite s at frequency 2.

[0070] Step S104: Based on the first GNSS deionization carrier phase combination observation value and the first GNSS deionization pseudorange combination observation value, the server analyzes the coupling relationship between the original GNSS deionization combination real ambiguity and the hardware delay, and reparameterizes the original GNSS deionization combination real ambiguity and the hardware delay to obtain the GNSS deionization combination real ambiguity containing hardware delay bias.

[0071] To reduce satellite orbit and clock errors, PPP typically employs precise ephemeris and precise satellite clock bias products. Since precise satellite clock bias is generated from ionospherically-free pseudorange observation data from ground-based tracking networks, the data provided by IGS (International GNSS Service) is... Absorbed the pseudorange hardware delay bias of the satellite-end ionospheric combination The pseudorange hardware delay bias at the receiver varies depending on the satellite system (this difference is called inter-system bias (ISB)), and it is absorbed by the receiver clock bias. Therefore, the receiver clock biases of different satellite systems are different. The ambiguity parameter is closely related to the phase hardware delay bias and is difficult to separate. At the same time, due to the introduction of the satellite clock based on pseudorange observation, the pseudorange hardware delay bias at the satellite end is introduced into the phase observation equation and absorbed by the ambiguity parameter; since the pseudorange observation equation and the phase observation equation jointly estimate the receiver clock bias, the pseudorange hardware delay bias at the receiver end is also introduced into the phase observation equation and absorbed by the ambiguity parameter. Therefore, after the satellite clock bias is corrected by precision products, formula (5) can be rewritten as formula (11) as follows:

[0072] Formula (11);

[0073] In the formulas, the reparameterized receiver clock bias and ambiguity are shown in formulas (12) to (15). In addition, the parameters to be estimated also include the receiver's three-dimensional position and the tropospheric wet delay in the zenith direction, i.e., the parameter vector to be estimated is... .

[0074] Formula (12);

[0075] Formula (13);

[0076] Formula (14);

[0077] Formula (15);

[0078] in, It is the reparameterized receiver-side UPD. It is a reparameterized satellite-end UPD. It is a reparameterized receiver clock bias that incorporates receiver hardware delay. It is a reparameterized GNSS deionization combined real-valued ambiguity with hardware delay bias. As can be seen from the above equation, the IF combined ambiguity parameter... It contains hardware delay biases at the receiver and satellite ends, and loses integer characteristics. Since PPP uses single-station positioning, it cannot directly separate these factors. Therefore, real-number solutions are usually used in parameter estimation to combine the two for estimation.

[0079] In formula (13), the reparameterized IF combined real ambiguity (i.e., the GNSS deionization combined real ambiguity with hardware delay bias) The ambiguity can be expressed as a combination of the real ambiguity of the second GNSS wide lane and the real ambiguity of the second GNSS narrow lane in the following formula (16):

[0080] Formula (16);

[0081] in, For the second GNSS wide lane real ambiguity, The real ambiguity of the second GNSS narrow alleyway.

[0082] Step S105: The server determines the second GNSS wide-lane real ambiguity based on the wide-lane carrier phase observation and the narrow-lane pseudorange observation.

[0083] In some embodiments, step S105, "the server determines the second GNSS wide-lane real ambiguity based on the wide-lane carrier phase observation and the narrow-lane pseudorange observation," includes the following steps:

[0084] Step S1051: The server constructs a first HMW combined carrier phase observation value based on the difference between the wide-lane carrier phase observation value and the narrow-lane pseudorange observation value.

[0085] Step S1052: The server uses a multi-epoch LSTM-assisted neural network method to smooth the first HMW combined carrier phase observation value. Based on the smoothed first HMW combined carrier phase observation value and the wide-lane combined wavelength, the second GNSS wide-lane real ambiguity is determined.

[0086] The second GNSS wide-lane real ambiguity can be calculated from the first HMW combined carrier phase observation value, which is defined as the difference between the wide-lane carrier phase observation value and the narrow-lane pseudorange observation value, as shown in formulas (17) to (19) below:

[0087] Formula (17);

[0088] Formula (18);

[0089] Formula (19);

[0090] in, Indicates the real ambiguity of the second GNSS width lane. This represents the first HMW combined carrier phase observation after smoothing. This represents the wide-lane carrier phase observation value. This represents the pseudorange observation value of the narrow alley. Indicates the combined wavelength of the wide-lane circuit. This represents the wide-lane integer ambiguity (i.e., the second GNSS wide-lane integer ambiguity) that includes the influence of wide-lane hardware delay integers. and The fractional part FCB of the receiver-side and satellite-side UPD is represented. and express f 1 and f For wavelength 2, the integer part of UPD is absorbed by the wide-lane integer ambiguity without affecting its integer characteristics. Since the HMW combined observation noise is large, multi-epoch smoothing is required to reduce the influence of observation noise and multipath effect. The smoothing method is shown in the following formula (20):

[0091] Formula (20);

[0092] In the formula, Represents the epoch number. This represents the smoothed first HMW combined carrier phase observation. To increase the accuracy of the smoothing process, this application introduces a neural network-assisted algorithm for multi-epoch smoothing. An improved model of the Long Short-Term Memory (LSTM) neural network is used, which introduces memory units and uses gating units to solve gradient explosion and vanishing problems. It can capture long-term effects and has the ability to perform highly nonlinear dynamic mapping and store past information. The steps of LSTM-assisted smoothing are as follows:

[0093] Determine the model input and output: Use the smoothed observations up to the i-th epoch and the observations at the i-th epoch as the parameters of the LSTM model;

[0094] Perform standardization: Calculate the mean and standard deviation of the training dataset, and use normalization methods to unify the dimensions of the training dataset and train it under the same standard, which can effectively reduce prediction error.

[0095] Setting LSTM network parameters: The accuracy of LSTM training is closely related to the network parameters, such as the number of hidden layers and the number of neurons in the hidden layer. Training speed and training error are related to the number of hidden layers. In this application, one hidden layer is set during training. Regarding the number of neurons in the hidden layer, this application adopts... The empirical formula sets the number of neurons, where This represents the number of features in the input layer. This represents the number of features in the output layer. The number of neurons is then adjusted based on the training results. The Adam algorithm uses moment estimation, which has advantages such as adaptability, fast computation speed, and small memory footprint. Therefore, Adam is chosen to accelerate training and meets the real-time requirements of practical applications.

[0096] When training the network, an error threshold is set, and RMSE is chosen as the loss function, representing...

[0097] In the formula, The output features are obtained by standardizing the input features. For the corresponding output features, the weights are iteratively adjusted, and training stops when the RMSE is less than the threshold.

[0098] LSTM prediction: Input the smoothed observations up to the i-th epoch and the observations at the i-th epoch into the trained network to output a standardized prediction sequence.

[0099] Destandardization: Restore the original dimensions of the sequence obtained in the previous step, and destandardize to obtain the smoothed observation value of the i-th epoch. .

[0100] HMW combined observations are constructed based on wide-lane carrier phase and narrow-lane pseudorange observations, and smoothed by a multi-epoch LSTM-assisted neural network to reduce the impact of observation noise and multipath effects, thereby improving the calculation accuracy of the second GNSS wide-lane real ambiguity.

[0101] Step S106: The server generates the wide-lane phase decimal deviation product based on the linear relationship between the fractional part of the second GNSS wide-lane real ambiguity and the phase decimal deviation; based on the wide-lane phase decimal deviation product, the second GNSS wide-lane real ambiguity is fixed to obtain the second GNSS wide-lane integer ambiguity.

[0102] Step S107: The server determines the second GNSS narrow lane real ambiguity based on the GNSS de-ionization combined real ambiguity with hardware delay bias and the second GNSS wide lane integer ambiguity;

[0103] Once the wide-lane phase fractional deviation product is obtained, the fixed second GNSS wide-lane integer ambiguity can be substituted into formula (16) to obtain the second GNSS narrow-lane real ambiguity, as shown in formulas (21) to (23) below:

[0104] Formula (21);

[0105] Formula (22);

[0106] Formula (23);

[0107] In the formula, Indicates the real ambiguity of the second GNSS narrow lane. It is a reparameterized GNSS de-ionospheric composite real ambiguity with hardware delay bias. Indicates the integer ambiguity of the second GNSS wide lane. This indicates the presence of narrow alleyway integer ambiguity (i.e., second GNSS narrow alleyway integer ambiguity). and This indicates the narrow-lane FCB at the receiver and satellite ends.

[0108] Step S108: The server generates the narrow lane phase decimal deviation product based on the linear relationship between the fractional part of the second GNSS narrow lane real ambiguity and the phase decimal deviation.

[0109] From formulas (21) to (23), it can be seen that the narrow lane FCB absorbs the wide lane FCB. By adopting an estimation process similar to that of the wide lane FCB, the narrow lane FCB can be obtained.

[0110] In the above embodiments, by acquiring the original GNSS observations and performing combined calculations, the ionospheric combined ambiguity and hardware delay are re-parameterized, the hardware delay bias is separated and processed, and wide-lane and narrow-lane phase fractional bias products are generated, providing reliable data support for subsequent ambiguity fixing and enhancing the robustness of the positioning model.

[0111] In some embodiments, step S106, "the server generates the wide-lane phase decimal deviation product based on the linear relationship between the fractional part of the second GNSS wide-lane real ambiguity and the phase decimal deviation," includes the following steps:

[0112] Step S1061: The server constructs a first virtual observation equation based on the linear relationship between the fractional part of the second GNSS wide-lane real ambiguity and the fractional phase deviation.

[0113] Assuming the ground observation network consists of It consists of several stations, each capable of observing... If there are 10 satellites, then according to formula (17), the first virtual observation equation shown in formula (24) can be obtained:

[0114] Formula (24);

[0115] Step S1062: The server sets the phase fractional deviation of a certain receiver or satellite to 0 as a reference, and calculates the wide-lane phase fractional deviation product based on the reference and the first virtual observation equation.

[0116] in, For virtual observations, i.e. the fractional part of the wide-lane ambiguity, the normal equation above is rank deficient. In order for the equation to be solved, it is necessary to take the FCB=0 of a certain receiver or satellite as a reference and treat it as a virtual observation, and then perform wide-lane FCB product estimation. In the embodiments of this application, the receiver FCB of the first station can be selected as the reference.

[0117] Step S108, "the server generates the narrow-lane phase decimal deviation product based on the linear relationship between the fractional part of the second GNSS narrow-lane real ambiguity and the phase decimal deviation," includes the following steps:

[0118] Step S1081: The server constructs a second virtual observation equation based on the linear relationship between the fractional part of the second GNSS narrow lane real ambiguity and the fractional phase deviation.

[0119] In step S1082, the server sets the phase fractional deviation of a certain receiver or satellite to 0 as a benchmark, and calculates the narrow lane phase fractional deviation product based on the benchmark and the second virtual observation equation.

[0120] By employing a similar process to constructing the first virtual observation equation, a second virtual observation equation can be constructed, and then narrow-lane FCB product estimation can be performed.

[0121] In the above embodiments, by constructing virtual observation equations and setting a benchmark (the phase fractional deviation of a certain receiver or satellite is 0), the rank deficiency problem of the normal equations is solved, and stable calculation of phase fractional deviation products for wide lane and narrow lane is achieved, ensuring product accuracy and laying the foundation for ambiguity fixation.

[0122] In some embodiments, the method further includes:

[0123] Step S1091: The server compresses the wide-lane phase fractional deviation product and the narrow-lane phase fractional deviation product in binary format; based on the basic message framework, the compressed wide-lane phase fractional deviation product and the compressed narrow-lane phase fractional deviation product are encapsulated in a structured manner. The basic message framework includes a synchronization header, a reserved flag bit, a message length, and a data message. The data message includes information type, information reference time, data synchronization flag, information body, and data verification information.

[0124] In step S1092, the server determines the transmission rules based on a preset encoding protocol, and transmits the encapsulated wide-lane phase fractional deviation product and the encapsulated narrow-lane phase fractional deviation product based on the transmission rules. The transmission rules include broadcasting by satellite system and satellite splitting, and transmission by frequency point and signal type. The server deducts the integer part of a preset week or more from the wide-lane phase fractional deviation product and the narrow-lane phase fractional deviation product, and assigns a unique message number to the wide-lane phase fractional deviation product and the narrow-lane phase fractional deviation product of different systems.

[0125] The message design for the aforementioned corrections to the ambiguity of wide and narrow lanes is a crucial step in the practical application of this technology. In real-time engineering applications, it is essential to minimize the amount of data transmitted while maintaining a certain level of accuracy. Therefore, a specific message format needs to be designed based on theory and practical application to compress the network-end product encoding into binary format. GNSS real-time data is transmitted or stored in a machine-readable binary format. Binary format effectively compresses data size compared to ASCII format, which is beneficial for the network transmission of large amounts of data.

[0126] To ensure simplicity in program implementation and maximize data compression, the stability of data transmission is increased. This application's embodiments reference the PPP-B2b data format, the QZSS centimeter-level enhanced service (CLAS) information format, and the IGS state domain real-time precision correction product format, employing a binary format with bits as the smallest storage unit. An RTCM3-based binary phase fractional deviation message data format is designed.

[0127] Figure 2 The basic framework for GNSS data and product messages mainly consists of a synchronization header, reserved flags, message length, and the data message itself. The synchronization header identifies the message type; the machine program determines the message encoding format by recognizing the synchronization header. Different types of information on the same information channel must have different synchronization headers. The message length records the length of the data message; this part can be omitted for fixed-length messages. Figure 2 The first line of the data message includes the information type and the information reference time (corresponding to...). Figure 2 'Time' in the data, and data synchronization markers (corresponding to...) Figure 2 The 'synchronization flag' in the data, and the information subject (i.e., the data subject) Figure 2 The second line contains the 'data message' and data verification information (corresponding to...). Figure 2 (CRC in the text). Information type is used to distinguish different data or product types; information reference time records information time information, which usually includes the second within a week or the second within a day to save broadcast traffic; data synchronization mark is used to connect segmented messages; data check verifies whether the message is complete and usable.

[0128] To avoid excessively large data packets, correction values ​​are broadcast system by system, with a maximum of 20 satellites per system. Synchronization markers are also set to ensure the integrity of products across systems. Different observation types at the same frequency have the same phase accuracy, differing from each other by 1 / 4 or an integer multiple of 1 / 4 wavelength. This information is identical for all satellites and can be provided by the receiver. This deviation has been pre-corrected for MSM signals in RTCM. Therefore, wide-lane FCB and narrow-lane FCB are broadcast for each satellite at each frequency, supporting a maximum of 5 frequencies. These deviations are coded in the message for user use. The message primarily involves wide-lane FCB and narrow-lane FCB.

[0129] Figure 3 The specific data protocol for satellite phase deviation coding is given. Figure 4 For the satellite phase deviation encoding data protocol, it should be noted that: signals of different types at the same frequency differ from each other by 1 / 4 or an integer multiple of 1 / 4 wavelength. This information will be recorded in the observation values. Therefore, when broadcasting satellite phase deviation products, two sets of reference type products need to be broadcast for each frequency point; satellite phase deviation is reduced by integers greater than 100 cycles.

[0130] In this application embodiment, the phase deviation encoding message number for each system is defined as follows: GPS is 1601, GALILEO is 1602, BDS is 1603, GLONASS is 1604, and other systems can continue to use the message numbering.

[0131] In the above embodiments, binary compression and structured encapsulation are used to process products with fractional phase deviations. Combined with preset transmission rules (splitting by system / satellite, distinguishing by frequency point, etc.), the amount of data transmission is reduced while ensuring accuracy, thereby improving transmission efficiency and stability and meeting the needs of real-time engineering applications.

[0132] In some embodiments, the first GNSS combined observation value includes wide-lane carrier phase combined observation value and narrow-lane pseudorange combined observation value. Step S110, "the terminal determines the first GNSS wide-lane real ambiguity based on the first GNSS combined observation value", includes the following steps:

[0133] Step S1101: The terminal constructs a second HMW combined carrier phase observation value by the difference between the wide-lane carrier phase combined observation value and the narrow-lane pseudorange combined observation value;

[0134] In step S1102, the terminal smooths the second HMW combined carrier phase observation value using a multi-epoch recursive averaging method, and determines the first GNSS wide-lane real ambiguity based on the smoothed second HMW combined carrier phase observation value and the wide-lane combined wavelength.

[0135] Here, the first GNSS wide-lane real ambiguity can be expressed by the following formula (25):

[0136] Formula (25);

[0137] in, Indicates the real ambiguity of the first GNSS width lane. This represents the smoothed second HMW combined carrier phase observation. Indicates the combined wavelength of the wide-lane circuit. This represents the observed value of the wide-lane carrier phase combination. This represents the combined observation value of the narrow alley pseudorange. Indicates carrier phase, Indicates the pseudorange of the code. Indicates frequency point, subscript Station number, superscript Number the satellite. It represents GNSS systems, such as GPS, BDS-3, and other Code Division Multiple Access (CDMA) systems.

[0138] The second HMW combined carrier phase observation can be smoothed using the following formula (26) to reduce the impact of observation noise and multipath effects:

[0139] Formula (26);

[0140] in, Represents the epoch number. This represents the smoothed second HMW combined carrier phase observation.

[0141] In the above embodiments, HMW combined observations are constructed by combining wide-lane carrier phase and narrow-lane pseudorange observations. Multi-epoch recursive averaging smoothing is used to effectively reduce observation noise and improve the accuracy of the first GNSS wide-lane real ambiguity, providing a reliable input for subsequent wide-lane ambiguity fixing.

[0142] In some embodiments, step S120, "the terminal determines the user coordinate floating-point solution, initial LEO de-ionization combined real ambiguity, initial GNSS de-ionization combined real ambiguity, tropospheric delay, LEO receiver clock bias, and GNSS receiver clock bias based on a precise single-point positioning algorithm fused from LEO and GNSS satellites," includes the following steps:

[0143] Step S1201: The terminal determines the ionosphere-depleted pseudorange observation values ​​of the user receiver for the LEO satellite and GNSS satellite based on the geometric distance from the user receiver to the LEO satellite and the GNSS satellite, the speed of light, the receiver clock error, the tropospheric delay, and the pseudorange observation error term.

[0144] In step S1202, the terminal determines the first LEO deionized carrier phase combination observation value, the first LEO deionized pseudorange combination observation value, and the second GNSS deionized carrier phase combination observation value for the LEO satellite and the GNSS satellite, respectively, based on the geometric distance from the user receiver to the LEO satellite and the GNSS satellite, the speed of light, the receiver clock error, the tropospheric delay, the wavelength corresponding to the deionization combination, the initial LEO deionization combination real ambiguity, the initial GNSS deionization combination real ambiguity, and the carrier phase observation error value.

[0145] In step S1203, the terminal constructs a function model based on the first LEO de-ionization carrier phase combination observation, the first LEO de-ionization pseudorange combination observation, the second GNSS de-ionization carrier phase combination observation, and the second GNSS de-ionization pseudorange combination observation. The function model is solved using the Kalman filter parameter estimation method to obtain the user coordinate floating-point solution, the initial LEO de-ionization combination real ambiguity, the initial GNSS de-ionization combination real ambiguity, the tropospheric delay, the LEO receiver clock bias, and the GNSS receiver clock bias.

[0146] Among them, the user coordinate floating-point solution and the LEO and GNSS de-ionization combined real ambiguity (i.e., the initial LEO de-ionization combined real ambiguity) can be obtained by using the Kalman filter parameter estimation method based on the LEO / GNSS PPP function model shown in the following formula (27). Real ambiguity combined with initial GNSS deionization ) and other parameters.

[0147] Formula (27);

[0148] Among them, superscript K Indicates GNSS system, Indicates the LEO system. and This indicates the geometric distance between the station (receiver, i.e., the observation station) and the satellite. This represents the combined observation value of the second GNSS ionospheric pseudorange. This represents the second GNSS deionization carrier phase combination observation value. This represents the combined observation value of the first LEO ionospheric pseudorange. This represents the first LEO deionization carrier phase combination observation value; The speed of light in a vacuum and It is a reparameterized receiver clock bias that incorporates receiver hardware delay. and For the tropospheric delay in the direction of signal propagation, The wavelength corresponding to the deionization combination, and To eliminate real-valued ambiguity in ionospheric combinations, and This is for pseudorange measurement without ionosphere (i.e., pseudorange observation error term). and This represents the observation noise error of the carrier phase without ionosphere (i.e., the carrier phase observation error value).

[0149] In the above embodiments, a precise single-point positioning algorithm based on the fusion of LEO and GNSS is constructed to build a function model containing multiple types of observations. The floating-point solution of user coordinates and various error parameters are obtained through Kalman filtering. This fully utilizes the advantage of the rapid geometric changes of LEO satellites to accelerate positioning convergence and improve parameter estimation accuracy.

[0150] In some embodiments, step S140, "the terminal determines the first GNSS narrow lane real ambiguity based on the GNSS de-ionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity," includes the following steps:

[0151] Step S1401: The terminal performs a weighted calculation on the GNSS deionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity based on the frequency of the GNSS dual-frequency carrier to obtain the first GNSS narrow-lane real ambiguity.

[0152] The first GNSS narrow lane real ambiguity is calculated by performing inter-satellite single difference on the combined GNSS deionization real ambiguity and combining it with the fixed first GNSS wide lane integer ambiguity, as shown in the following formula (28):

[0153] Formula (28);

[0154] in, Indicates the real ambiguity of the first GNSS narrow alleyway. Indicates the integer ambiguity of the first GNSS width lane. This represents the real-valued ambiguity of the GNSS deionization combination after inter-satellite single-difference. Wherein, Indicates a reference star.

[0155] Formula (29);

[0156] As shown in formula (29) above, Indicates the integer ambiguity of the first GNSS narrow lane. For the satellite-side narrow-lane FCB (i.e., narrow-lane phase fractional deviation product), the above formula uses the narrow-lane FCB product and employs the LAMBDA algorithm partial ambiguity strategy to fix narrow-lane ambiguities, obtaining the first GNSS narrow-lane integer ambiguity. The main idea is to arrange the elevation angles of the satellite sequence whose narrow-lane ambiguities need to be fixed. The elevation angle sorting method is based on the idea that the lower the elevation angle, the worse the ambiguity accuracy. First, the ambiguities of the top 3 satellites with the highest elevation angles are fixed. If the ambiguity is successfully fixed, the 4th satellite with the highest elevation angle is added for ambiguity fixing. If the ambiguity is successfully fixed, the 5th satellite is added, and so on, until the ambiguity cannot be fixed after adding more satellites. At this point, ambiguity fixing stops, the last added satellite is removed, and the satellite sequence after the last successful ambiguity fixing is used. This multi-loop traversal partial ambiguity strategy based on elevation angles improves the success rate of ambiguity fixing and saves fixing time.

[0157] In the above embodiments, by performing a weighted calculation on the real ambiguity of the GNSS deionization combination after inter-satellite single difference and the integer ambiguity of the wide lane, the real ambiguity of the first GNSS narrow lane is accurately derived, providing an accurate initial value for fixing the narrow lane ambiguity and further improving the positioning accuracy.

[0158] In some embodiments, the first GNSS combined observation value includes the frequency of the GNSS dual-frequency carrier. Step S150, "the terminal calculates the fixed solution of user coordinates based on the first GNSS combined observation value, the first GNSS narrow-lane integer ambiguity, the first GNSS narrow-lane real ambiguity, the first GNSS wide-lane integer ambiguity, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution of user coordinates," includes the following steps:

[0159] Step S1501: The terminal reshapes the initial GNSS de-ionization combined real ambiguity based on the frequency of the GNSS dual-frequency carrier, the first GNSS narrow lane integer ambiguity, the narrow lane phase fractional deviation product, and the first GNSS wide lane integer ambiguity, to obtain the reshaped GNSS de-ionization combined real ambiguity.

[0160] The real ambiguity of the initial GNSS deionization combination is reshaped after verification, as shown in the following formula (30):

[0161] Formula (30);

[0162] in, This represents the real-valued ambiguity of the GNSS de-ionization combination after reshaping. Indicates the integer ambiguity of the first GNSS narrow lane. This indicates the narrow-lane phase fractional deviation product (i.e., satellite-end narrow-lane FCB). This represents the integer ambiguity of the first GNSS width lane.

[0163] In step S1502, the terminal determines the fixed solution of user coordinates based on the initial LEO de-ionization combination real ambiguity, the initial GNSS de-ionization combination real ambiguity, the reshaped GNSS de-ionization combination real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution of user coordinates.

[0164] Among them, the fixed solution of user coordinates can be calculated based on the conditional adjustment, as shown in the following formula (31):

[0165] Formula (31);

[0166] In the formula, This represents the initial GNSS de-ionization combination real ambiguity parameters obtained through parameter estimation methods. Other parameters include the user coordinate floating-point solution, receiver type difference, zenith tropospheric wet delay, and initial LEO de-ionization combination real ambiguity parameter. This represents the variance-covariance matrix of the initial GNSS ionospheric de-ambiguity combination real ambiguity parameters obtained through parameter estimation methods. This represents the variance and covariance matrix of the initial GNSS ionospheric de-ambiguity combination real ambiguity parameters and other parameters obtained through parameter estimation methods. To reconstruct GNSS ionospheric composite ambiguity, For other parameters, the solution is fixed.

[0167] In the above embodiments, the GNSS deionization combined real ambiguity is reshaped based on the dual-frequency carrier frequency and the fixed wide-lane and narrow-lane integer ambiguities. Multiple types of parameters are integrated to calculate the fixed solution of user coordinates. The fixed ambiguity information is fully utilized to significantly improve the accuracy and stability of the positioning results.

[0168] The following are four experiments using the GPS system as an example to enhance the number of low-Earth orbit (LEO) satellites. Taking two LEO satellites as an example, the method for controlling the number of LEO satellites is explained as follows: When there are more than two LEO satellites, the first two satellites are retained based on their elevation angles. If one of the retained LEO satellites experiences a drop in elevation angle (i.e., its elevation angle falls below the cutoff elevation angle), the satellite with the larger elevation angle is selected to replace it.

[0169] Figure 5 This demonstrates the time to first fix (TTFF) for single-system PPP (Precise Point Positioning with Ambiguity Resolution) augmentation in different latitude regions using 1, 2, 3, 4, and all LEO satellites. 0 indicates no LEO satellite augmentation; 1, 2, 3, and 4 satellites represent 1, 2, 3, and 4 LEO satellites respectively; and - indicates no control over the number of LEO satellites (no limit on the number of LEO satellites, all available satellites participate in augmentation), i.e., augmentation by all observed LEO satellites. One LEO satellite has almost no effect on accelerating the first fix of ambiguity; 2, 3, 4, and all LEO satellites can accelerate the time to first fix (TTFF) of approximately 15 minutes for single-system PPP to 9.0, 7.5, 6.8, and 6.4 minutes, respectively.

[0170] To address the limitation of real-time applications of PPP (Peripherally Argumentative Positioning) due to the long observation time required for reliable ambiguity fixation solutions, this application provides a precise single-point positioning method and apparatus based on low-Earth orbit satellite-assisted ambiguity fixation. By employing long short-term memory (LSTM) to smooth HMW (High-Magnitude Multi-Way) combined observations over multiple epochs, the impact of observation noise and multipath effects is reduced, and the accuracy of the smoothed observations is improved. To ensure simple program implementation and maximize data compression, thereby increasing data transmission stability, the application references my country's PPP-B2b data format, Japan's QZSS (Centimeter-Level Enhancement Service) information format (CLAS), and the IGS (In-State Domain Real-Time Precise Correction) product format. A binary format with bits as the smallest storage unit is adopted, and an RTCM3-based binary phase fractional deviation message data format is designed and implemented.

[0171] Furthermore, after solving the parameters based on the LEO / GNSS PPP function model, partial ambiguity fixation is performed on the GNSS ambiguities. Since the LEO ambiguity observation arc is relatively short, it is not conducive to ambiguity fixation. Partial ambiguity fixation of the GNSS ambiguities can improve the success rate of ambiguity fixation and reduce algorithm complexity and solution time. In addition, this method does not require a large number of low-Earth orbit satellites; rapid PPP AR can be achieved with only 3-4 LEO satellites.

[0172] Based on the foregoing embodiments, this application further provides a precision positioning acceleration device based on satellite / LSTM neural network. The device includes various modules and units included in each module, which can be implemented by a processor in an electronic device; of course, it can also be implemented by specific logic circuits. In the implementation process, the processor can be a central processing unit (CPU), a microprocessor (MPU), a digital signal processor (DSP), or a field-programmable gate array (FPGA), etc.

[0173] Figure 6 A schematic diagram of the composition structure of a precision positioning acceleration device based on a satellite / LSTM neural network provided in this application embodiment is shown below. Figure 6 As shown, the device 600 includes: a server module 61 and a terminal module 62, wherein:

[0174] The server module 61 includes a generation and encoding module 611, and the terminal module 62 includes an acquisition module 621, a determination module 622, a first fixing module 623, a second fixing module 624, and a calculation module 625.

[0175] The generation and encoding module 611 is used to generate wide-lane phase fractional deviation products and narrow-lane phase fractional deviation products based on the LSTM method, and then encode them using a specific method before sending them to the terminal.

[0176] The acquisition module 621 is used to acquire the encoded wide-lane phase fractional deviation product, narrow-lane phase fractional deviation product, and first GNSS combined observation value; and to determine the first GNSS wide-lane real ambiguity based on the first GNSS combined observation value.

[0177] The determining module 622 is used to determine the floating-point solution of user coordinates, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock error, and the GNSS receiver clock error based on the precise single-point positioning algorithm fused by LEO satellites and GNSS satellites.

[0178] The first fixing module 623 is used to perform inter-satellite single difference processing on the first GNSS wide lane real ambiguity, and use the wide lane phase fractional deviation product to fix the GNSS wide lane real ambiguity after inter-satellite single difference processing to obtain the first GNSS wide lane integer ambiguity.

[0179] The second fixing module 624 is used to perform inter-satellite single-difference processing on the initial GNSS de-ionization combined real ambiguity, and based on the GNSS de-ionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity, determine the first GNSS narrow-lane real ambiguity, and use the narrow-lane phase fractional deviation product to fix the first GNSS narrow-lane real ambiguity using a partial ambiguity fixing method to obtain the first GNSS narrow-lane integer ambiguity;

[0180] The calculation module 625 is used to calculate the fixed solution of user coordinates based on the first GNSS combined observations, the first GNSS narrow lane integer ambiguity, the narrow lane phase fractional deviation product, the first GNSS wide lane integer ambiguity, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution of user coordinates.

[0181] In some possible embodiments, the generation and encoding module 611 includes: an acquisition submodule, configured to acquire raw GNSS observations, the raw GNSS observations including the frequency, phase, pseudorange, wide-lane carrier phase observations, and narrow-lane pseudorange observations of the GNSS dual-frequency carrier; a first calculation submodule, configured to perform a weighted calculation on the phase of the dual-frequency carrier based on the frequency of the GNSS dual-frequency carrier to obtain a first GNSS de-ionization carrier phase combination observation; and to perform a weighted calculation on the pseudorange based on the frequency of the GNSS dual-frequency carrier to obtain a first GNSS de-ionization pseudorange combination observation; and a parameterization submodule, configured to perform a parameterization calculation on the first GNSS de-ionization carrier phase combination observation and the first GNSS dual-frequency carrier phase combination observation. The S-shaped ionospheric pseudorange combined observations are analyzed. By examining the coupling relationship between the original GNSS ionospheric combined real ambiguity and the hardware delay, the original GNSS ionospheric combined real ambiguity and the hardware delay are reparameterized to obtain GNSS ionospheric combined real ambiguity with hardware delay bias. A first generation submodule is used to determine the second GNSS wide-lane real ambiguity based on the wide-lane carrier phase observations and the narrow-lane pseudorange observations. Based on the linear relationship between the fractional part of the second GNSS wide-lane real ambiguity and the fractional phase bias, the wide-lane phase bias product is generated. Based on the wide-lane phase bias product, the second GNSS wide-lane real ambiguity is fixed to obtain the second GNSS wide-lane integer ambiguity.

[0182] The second generation submodule is used to determine the second GNSS narrow lane real ambiguity based on the GNSS de-ionization combined real ambiguity with hardware delay bias and the second GNSS wide lane integer ambiguity; and to generate the narrow lane phase fractional deviation product based on the linear relationship between the fractional part of the second GNSS narrow lane real ambiguity and the phase fractional deviation.

[0183] In some possible embodiments, the first generation submodule includes: a first construction unit, configured to construct a first virtual observation equation based on the linear relationship between the fractional part of the second GNSS wide-lane real ambiguity and the phase fractional deviation; and a first solution unit, configured to set the phase fractional deviation of a receiver or satellite to 0 as a reference, and to calculate the wide-lane phase fractional deviation product based on the reference and the first virtual observation equation; the second generation submodule includes: a second construction unit, configured to construct a second virtual observation equation based on the linear relationship between the fractional part of the second GNSS narrow-lane real ambiguity and the phase fractional deviation; and a second solution unit, configured to set the phase fractional deviation of a receiver or satellite to 0 as a reference, and to calculate the narrow-lane phase fractional deviation product based on the reference and the second virtual observation equation.

[0184] In some possible embodiments, the generation and encoding module 611 further includes: a compression submodule, used to compress the wide-lane phase fractional deviation product and the narrow-lane phase fractional deviation product in binary format; a structured encapsulation processing submodule, used to perform structured encapsulation processing on the compressed wide-lane phase fractional deviation product and the compressed narrow-lane phase fractional deviation product based on the basic message framework, wherein the basic message framework includes a synchronization header, a reserved flag bit, a message length, and a data message, and the data message includes information type, information reference time, data synchronization flag, information body, and data verification information; and a transmission rule determination submodule, used to determine transmission rules based on a preset encoding protocol, and transmit the encapsulated wide-lane phase fractional deviation product and the encapsulated narrow-lane phase fractional deviation product based on the transmission rules; the transmission rules include broadcasting by satellite system and satellite splitting, transmission by frequency point and signal type, deducting integer parts of preset weeks or more from the wide-lane phase fractional deviation product and the narrow-lane phase fractional deviation product, and assigning a unique message number to the wide-lane phase fractional deviation product and the narrow-lane phase fractional deviation product of different systems.

[0185] In some possible embodiments, the first generation submodule further includes: a third construction unit, configured to construct a first HMW combined carrier phase observation value based on the difference between the wide-lane carrier phase observation value and the narrow-lane pseudorange observation value; and a smoothing processing unit, configured to smooth the first HMW combined carrier phase observation value using a multi-epoch LSTM-assisted neural network method, and determine the second GNSS wide-lane real ambiguity based on the smoothed first HMW combined carrier phase observation value and the wide-lane combined wavelength.

[0186] In some possible embodiments, the first GNSS combined observation value includes a wide-lane carrier phase combined observation value and a narrow-lane pseudorange combined observation value. The acquisition module 621 includes: a first construction submodule, used to construct a second HMW combined carrier phase observation value by the difference between the wide-lane carrier phase combined observation value and the narrow-lane pseudorange combined observation value; and a smoothing processing submodule, used to smooth the second HMW combined carrier phase observation value using a multi-epoch recursive averaging method, and determine the first GNSS wide-lane real ambiguity based on the smoothed second HMW combined carrier phase observation value and the wide-lane combined wavelength.

[0187] In some possible embodiments, the determining module 622 includes: a first determining submodule, used to determine the pseudorange observation values ​​of the de-ionization combination of the LEO satellite and the GNSS satellite based on the geometric distance from the user receiver to the LEO satellite and the GNSS satellite, the speed of light, the receiver clock error, the tropospheric delay, and the pseudorange observation error term; and a second determining submodule, used to determine the first LEO de-ionization carrier phase combination observation value of the LEO satellite based on the geometric distance from the user receiver to the LEO satellite and the GNSS satellite, the speed of light, the receiver clock error, the tropospheric delay, the wavelength corresponding to the de-ionization combination, the initial LEO de-ionization combination real ambiguity, the initial GNSS de-ionization combination real ambiguity, and the carrier phase observation error value. The system includes: measured values, first LEO de-ionization pseudorange combined observation values, and second GNSS de-ionization carrier phase combined observation values ​​and second GNSS de-ionization pseudorange combined observation values; a solution submodule, used to construct a function model based on the first LEO de-ionization carrier phase combined observation values, the first LEO de-ionization pseudorange combined observation values, the second GNSS de-ionization carrier phase combined observation values, and the first GNSS de-ionization pseudorange combined observation values, and solve the function model using the Kalman filter parameter estimation method to obtain the user coordinate floating-point solution, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, and the GNSS receiver clock bias.

[0188] In some possible embodiments, the second fixed module 624 includes: a second calculation submodule, used to perform a weighted calculation on the GNSS deionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity based on the frequency of the GNSS dual-frequency carrier, to obtain the first GNSS narrow-lane real ambiguity.

[0189] In some possible embodiments, the first GNSS combined observation includes the frequency of the GNSS dual-frequency carrier. The calculation module 625 includes: a reshaping submodule, used to reshape the initial GNSS de-ionization combined real ambiguity based on the frequency of the GNSS dual-frequency carrier, the first GNSS narrow-lane integer ambiguity, the narrow-lane phase fractional deviation product, and the first GNSS wide-lane integer ambiguity, to obtain the reshaped GNSS de-ionization combined real ambiguity; and a third determining submodule, used to determine the user coordinate fixed solution based on the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the reshaped GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the user coordinate floating-point solution.

[0190] It should be noted that the descriptions of the above device embodiments are similar to those of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0191] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0192] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0193] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0194] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units; some or all of the units may be selected to achieve the purpose of the embodiments of this application according to actual needs. In addition, each functional unit in the embodiments of this application may be fully integrated into one processing unit, or each unit may be a separate unit, or two or more units may be integrated into one unit; the integrated unit may be implemented in hardware or in the form of hardware plus software functional units.

[0195] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause the device automatic test line to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0196] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined to obtain new method embodiments without conflict. The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined to obtain new method embodiments or device embodiments without conflict.

[0197] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A precise positioning acceleration method based on satellite / LSTM neural network, characterized in that, include: The server generates wide-lane phase fractional deviation products and narrow-lane phase fractional deviation products based on the LSTM method, and sends them to the terminal after encoding them using a specific method. The terminal acquires the encoded wide-lane phase fractional deviation product, narrow-lane phase fractional deviation product, and the first GNSS combined observation value; based on the first GNSS combined observation value, it determines the first GNSS wide-lane real ambiguity. The terminal uses a precise single-point positioning algorithm based on the fusion of LEO and GNSS satellites to determine the floating-point solution of user coordinates, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, tropospheric delay, LEO receiver clock error, and GNSS receiver clock error. The terminal performs inter-satellite single-difference processing on the first GNSS wide-lane real ambiguity, and uses the wide-lane phase fractional deviation product to fix the GNSS wide-lane real ambiguity after inter-satellite single-difference processing to obtain the first GNSS wide-lane integer ambiguity. The terminal performs inter-satellite single-difference processing on the initial GNSS de-ionization combined real ambiguity. Based on the GNSS de-ionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity, the first GNSS narrow-lane real ambiguity is determined. The first GNSS narrow-lane real ambiguity is fixed using a partial ambiguity fixing method with the narrow-lane phase fractional deviation product to obtain the first GNSS narrow-lane integer ambiguity. The terminal calculates the fixed solution for user coordinates based on the first GNSS combined observations, the first GNSS narrow lane integer ambiguity, the narrow lane phase fractional deviation product, the first GNSS wide lane integer ambiguity, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution for user coordinates.

2. The precision positioning acceleration method according to claim 1, characterized in that, Also includes: The server acquires raw GNSS observations, which include the frequency, phase, pseudorange, wide-lane carrier phase observations, and narrow-lane pseudorange observations of the GNSS dual-frequency carrier. The server performs a weighted calculation on the phase of the dual-frequency carrier based on the frequency of the GNSS dual-frequency carrier to obtain the first GNSS deionization carrier phase combination observation value; The server performs a weighted calculation on the dual-frequency pseudorange based on the frequency of the GNSS dual-frequency carrier to obtain the first GNSS ionospheric pseudorange combined observation value. The server, based on the first GNSS deionization carrier phase combination observation value and the first GNSS deionization pseudorange combination observation value, analyzes the coupling relationship between the original GNSS deionization combination real ambiguity and the hardware delay, and re-parameterizes the original GNSS deionization combination real ambiguity and the hardware delay to obtain the GNSS deionization combination real ambiguity containing hardware delay bias. The server determines the second GNSS wide-lane real ambiguity based on the wide-lane carrier phase observation and the narrow-lane pseudorange observation; The server generates the wide-lane phase decimal deviation product based on the linear relationship between the fractional part of the second GNSS wide-lane real ambiguity and the phase decimal deviation; based on the wide-lane phase decimal deviation product, the second GNSS wide-lane real ambiguity is fixed to obtain the second GNSS wide-lane integer ambiguity. The server determines the second GNSS narrow lane real ambiguity based on the GNSS de-ionization combined real ambiguity with hardware delay bias and the second GNSS wide lane integer ambiguity; The server generates the narrow-lane phase decimal deviation product based on the linear relationship between the fractional part of the second GNSS narrow-lane real ambiguity and the phase decimal deviation.

3. The precision positioning acceleration method according to claim 2, characterized in that, The server generates the wide-lane phase fractional deviation product based on the linear relationship between the fractional part of the second GNSS wide-lane real ambiguity and the phase fractional deviation, including: The server constructs a first virtual observation equation based on the linear relationship between the fractional part of the second GNSS wide-lane real ambiguity and the fractional phase deviation. The server sets the phase fractional deviation of a certain receiver or satellite to 0 as a benchmark, and calculates the wide-lane phase fractional deviation product based on the benchmark and the first virtual observation equation. The server generates the narrow-lane phase fractional deviation product based on the linear relationship between the fractional part of the second GNSS narrow-lane real ambiguity and the phase fractional deviation, including: The server constructs a second virtual observation equation based on the linear relationship between the fractional part of the second GNSS narrow lane real ambiguity and the fractional phase deviation. The server sets the phase fractional deviation of a certain receiver or satellite to 0 as a benchmark, and calculates the narrow lane phase fractional deviation product based on the benchmark and the second virtual observation equation.

4. The precision positioning acceleration method according to claim 2, characterized in that, Also includes: The server compresses the wide-lane phase decimal deviation product and the narrow-lane phase decimal deviation product in binary format; Based on the basic message framework, the compressed wide-lane phase fractional deviation product and the compressed narrow-lane phase fractional deviation product are subjected to structured encapsulation processing. The basic message framework includes a synchronization header, reserved flag bits, message length and data message. The data message includes information type, information reference time, data synchronization flag, information body and data verification information. The server determines the transmission rules based on a preset encoding protocol, and transmits the encapsulated wide-lane phase fractional deviation product and the encapsulated narrow-lane phase fractional deviation product based on the transmission rules. The transmission rules include broadcasting by satellite system and satellite splitting, and transmission by frequency point and signal type. The server deducts the integer part of the wide-lane phase fractional deviation product and the narrow-lane phase fractional deviation product from the integer part of the preset week, and assigns a unique message number to the wide-lane phase fractional deviation product and the narrow-lane phase fractional deviation product of different systems.

5. The precision positioning acceleration method according to claim 2, characterized in that, The server determines the second GNSS wide-lane real ambiguity based on the wide-lane carrier phase observations and the narrow-lane pseudorange observations, including: The server constructs a first HMW combined carrier phase observation based on the difference between the wide-lane carrier phase observation and the narrow-lane pseudorange observation. The server uses a multi-epoch LSTM-assisted neural network method to smooth the first HMW combined carrier phase observation value. Based on the smoothed first HMW combined carrier phase observation value and the wide-lane combined wavelength, the second GNSS wide-lane real ambiguity is determined.

6. The precision positioning acceleration method according to claim 1, characterized in that, The first GNSS combined observations include wide-lane carrier phase combined observations and narrow-lane pseudorange combined observations. Determining the first GNSS wide-lane real ambiguity based on the first GNSS combined observations includes: The terminal constructs a second HMW combined carrier phase observation value by the difference between the wide-lane carrier phase combined observation value and the narrow-lane pseudorange combined observation value; The terminal uses a multi-epoch recursive averaging method to smooth the second HMW combined carrier phase observation value. Based on the smoothed second HMW combined carrier phase observation value and the wide-lane combined wavelength, the first GNSS wide-lane real ambiguity is determined.

7. The precision positioning acceleration method according to claim 1, characterized in that, The terminal uses a precise point positioning algorithm based on the fusion of LEO and GNSS satellites to determine the user's floating-point coordinates, initial LEO de-ionization combined real ambiguity, initial GNSS de-ionization combined real ambiguity, tropospheric delay, LEO receiver clock bias, and GNSS receiver clock bias, including: The terminal determines the ionosphere-depleted pseudorange observation values ​​of the user receiver for LEO and GNSS satellites based on the geometric distance from the user receiver to the LEO and GNSS satellites, the speed of light, the receiver clock error, the tropospheric delay, and the pseudorange observation error term. The terminal determines the first LEO deionized carrier phase combination observation value and the first LEO deionized pseudorange combination observation value of the LEO satellite, as well as the second GNSS deionized carrier phase combination observation value and the second GNSS deionized pseudorange combination observation value of the GNSS satellite, based on the geometric distance from the user receiver to the LEO satellite and the GNSS satellite, the speed of light, the receiver clock error, the tropospheric delay, the wavelength corresponding to the deionization combination, the initial LEO deionization combination real ambiguity, the initial GNSS deionization combination real ambiguity, and the carrier phase observation error value. The terminal constructs a function model based on the first LEO de-ionization carrier phase combination observation, the first LEO de-ionization pseudorange combination observation, the second GNSS de-ionization carrier phase combination observation, and the second GNSS de-ionization pseudorange combination observation. The function model is solved using the Kalman filter parameter estimation method to obtain the user coordinate floating-point solution, the initial LEO de-ionization combination real ambiguity, the initial GNSS de-ionization combination real ambiguity, the tropospheric delay, the LEO receiver clock bias, and the GNSS receiver clock bias.

8. The precision positioning acceleration method according to claim 1, characterized in that, The determination of the first GNSS narrow-lane real ambiguity based on the GNSS de-ionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity includes: The terminal performs a weighted calculation on the real ambiguity of the GNSS deionization combination after inter-satellite single-difference processing and the integer ambiguity of the first GNSS wide lane, based on the frequency of the GNSS dual-frequency carrier, to obtain the real ambiguity of the first GNSS narrow lane.

9. The precision positioning acceleration method according to claim 1, characterized in that, The first GNSS combined observations include the frequencies of the GNSS dual-frequency carriers. The terminal, based on the first GNSS combined observations, the first GNSS narrow-lane integer ambiguity, the narrow-lane phase fractional deviation product, the first GNSS wide-lane integer ambiguity, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the user coordinate floating-point solution, calculates the user coordinate fixed solution, including: The terminal reshapes the initial GNSS de-ionization combined real ambiguity based on the frequency of the GNSS dual-frequency carrier, the first GNSS narrow lane integer ambiguity, the narrow lane phase fractional deviation product, and the first GNSS wide lane integer ambiguity, to obtain the reshaped GNSS de-ionization combined real ambiguity. The terminal determines the fixed solution for user coordinates based on the initial LEO de-ionization combination real ambiguity, the initial GNSS de-ionization combination real ambiguity, the reshaped GNSS de-ionization combination real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution for user coordinates.

10. A precision positioning acceleration device based on satellite / LSTM neural network, characterized in that, include: The server module and the terminal module, wherein: The server module includes a generation and encoding module, and the terminal module includes an acquisition module, a determination module, a first fixing module, a second fixing module, and a calculation module. The generation and encoding module is used to generate wide-lane phase fractional deviation products and narrow-lane phase fractional deviation products based on the LSTM method, and then encode them using a specific method before sending them to the terminal. The acquisition module is used to acquire the encoded wide-lane phase fractional deviation product, narrow-lane phase fractional deviation product, and first GNSS combined observation value; and to determine the first GNSS wide-lane real ambiguity based on the first GNSS combined observation value. The determining module is used to determine the floating-point solution of user coordinates, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock error, and the GNSS receiver clock error based on the precise single-point positioning algorithm fused by LEO satellites and GNSS satellites. The first fixing module is used to perform inter-satellite single-difference processing on the first GNSS wide-lane real ambiguity, and use the wide-lane phase fractional deviation product to fix the GNSS wide-lane real ambiguity after inter-satellite single-difference processing to obtain the first GNSS wide-lane integer ambiguity. The second fixing module is used to perform inter-satellite single-difference processing on the initial GNSS de-ionization combined real ambiguity, and based on the GNSS de-ionization combined real ambiguity after inter-satellite single-difference processing and the first GNSS wide-lane integer ambiguity, determine the first GNSS narrow-lane real ambiguity, and use the narrow-lane phase fractional deviation product to fix the first GNSS narrow-lane real ambiguity using a partial ambiguity fixing method to obtain the first GNSS narrow-lane integer ambiguity; The calculation module is used to calculate the fixed solution of user coordinates based on the first GNSS combined observations, the first GNSS narrow lane integer ambiguity, the narrow lane phase fractional deviation product, the first GNSS wide lane integer ambiguity, the initial LEO de-ionization combined real ambiguity, the initial GNSS de-ionization combined real ambiguity, the tropospheric delay, the LEO receiver clock bias, the GNSS receiver clock bias, and the floating-point solution of user coordinates.

Images