Beidou large-area sea millimeter-level high-precision positioning method

Abstract

The present application relates to the field of satellite navigation and precise positioning technology, in particular to a Beidou large-area sea millimeter-level high-precision positioning method. The specific implementation steps include: collecting and preprocessing marine environment state data and Beidou observation data, and extracting environmental influence parameters; based on the environment state data, satellite signal propagation data and antenna phase center data, a quantitative model is established to obtain propagation path delay data and antenna phase center change data; based on the Beidou cooperative satellite positioning algorithm, the observation data and the adaptive reference network information are processed to obtain the preliminary positioning result; a satellite signal error model is constructed to analyze the environmental influence parameters and data, obtain the positioning error value and adjust the preliminary positioning result; the positioning result is calculated and judged, divided into first and second types of working conditions, the second type of working condition is switched to the emergency mode, and the Beidou signal error is compensated. The present application improves the positioning accuracy, reliability and adaptability of the Beidou system in the marine environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of satellite navigation and precise positioning, in particular to a Beidou large-area sea millimeter-level high-precision positioning method. BACKGROUND

[0002] Global Navigation Satellite System (GNSS) and China's Beidou System (BDS) are the core of marine high-precision positioning and monitoring. In the scenario of real-time monitoring of large-scale cross-sea projects at the millimeter level, the accuracy and reliability of satellite positioning are very high. Although the existing GNSS technology has made important progress in improving positioning performance, there is still room for improvement in the wide and complex environment of large-area sea, especially in dynamically sensing and finely compensating the influence of marine environment on Beidou satellite signal propagation, and in the adaptive ability of positioning service guarantee under adverse marine working conditions.

[0003] Currently, Real-Time Kinematic (RTK), Precise Point Positioning (PPP) and its enhanced technology are the mainstream of GNSS high-precision positioning. However, when these technologies are used for millimeter-level high-precision positioning in dynamic marine environment, there are still some challenges. For the complex marine environment that affects millimeter-level Beidou positioning, including tidal pattern, strong surge impact and local atmospheric water vapor dramatic change, the real-time sensing accuracy and effectiveness of feature parameter extraction of existing methods are still insufficient, which limits the effectiveness of subsequent error compensation.

[0004] Although network RTK, PPP-RTK and other solutions provide support for wide-area precise satellite positioning, how to deeply and dynamically integrate the real-time sensing of fine marine environment influence features, including accurate antenna phase center displacement and high-precision path delay correction, into the core of collaborative satellite positioning algorithm, to ensure that users in large areas can obtain consistent millimeter-level positioning accuracy under sparse or dynamic marine reference network, is still a key technical bottleneck.

[0005] Although the navigation system has multi-modal and fault-tolerant design, when the marine environment changes dramatically and the Beidou satellite signal quality deteriorates, how to realize mode switching based on rapid intelligent identification and even prediction of adverse working conditions, and start more precise and enhanced Beidou signal propagation error compensation mechanism for this adverse working condition, to maximize the continuity of positioning service and ensure accuracy, the related technology still needs to be improved.

[0006] In view of the above situation, in order to overcome the deficiencies of the existing technology in realizing Beidou large-area sea millimeter-level high-precision satellite positioning, especially in improving the adaptive sensing and fine compensation ability of complex marine environment dynamic influence, optimizing the efficiency of collaborative satellite positioning solution based on environmental information, and enhancing the intelligent response and positioning performance guarantee level of specific adverse marine working conditions, the present application proposes a Beidou large-area sea millimeter-level high-precision positioning method to solve the above problems. SUMMARY

[0007] The purpose of the present application is to provide a Beidou large-area sea millimeter-level high-precision positioning method. The method realizes continuous and reliable millimeter-level high-precision positioning of a target in a large-area sea area by integrating ocean environment self-adaptive perception, fine compensation of cooperative positioning based on environmental influence, and adaptive adjustment and reinforcement processing of positioning strategy in extreme ocean working conditions.

[0008] To achieve the above purpose, the present application provides the following technical solutions:

[0009] A Beidou large-area sea millimeter-level high-precision positioning method, comprising:

[0010] Real-time collection and preprocessing of ocean environment state data and Beidou observation data, extraction of environmental influence parameters related to Beidou satellite signal propagation error and observation data quality from the ocean environment state data;

[0011] Establishing a quantitative model based on the environmental state data, Beidou satellite signal propagation data and antenna phase center data; the quantitative model analyzes the environmental influence parameters to obtain propagation path delay data and antenna phase center change data;

[0012] Processing the Beidou observation data and adaptive reference network information based on a Beidou cooperative satellite positioning algorithm to obtain a preliminary satellite positioning result;

[0013] Using historical quantitative environmental influence parameters, historical propagation path delay data and historical antenna phase center change data as input feature sets to construct a satellite signal error model, analyzing the current environmental influence parameters, propagation path delay data and antenna phase center change data to obtain a positioning error value; adjusting the preliminary satellite positioning result based on the positioning error value to obtain a positioning result;

[0014] Solving the positioning result to obtain a solution result; judging the solution result according to a solution result threshold to divide it into a first type of working condition and a second type of working condition; the second type of working condition switches the satellite positioning mode to an emergency mode, reinforces the compensation of Beidou signal error, and obtains a satellite positioning result.

[0015] Preferably, the ocean environment state data and Beidou observation data comprise:

[0016] The ocean environment state data includes tidal height and swell spectrum data collected in real time by an underwater pressure sensor network deployed in the positioning sea area, and temperature, pressure and water vapor content data collected in real time by an automatic weather station network deployed on the shore base and sea surface platform in the positioning area; the Beidou observation data includes raw pseudorange, raw carrier phase and raw Doppler observation values of multiple frequency points output in real time by a Beidou receiver carried by the positioning target.

[0017] Preferably, the environmental impact parameter acquisition process comprises:

[0018] Based on real-time tidal height and swell spectrum data, a three-dimensional instantaneous displacement of a GNSS receiving antenna phase center relative to a nominal position caused by water dynamic disturbance of the platform is calculated by a water dynamic response inversion algorithm based on linear wave theory; using multi-point temperature, air pressure and water vapor content data in the positioning area, a high-precision real-time correction value of the wet delay component of the tropospheric delay in the zenith direction above the positioning target is generated in real time by a bilinear interpolation method combined with regional three-dimensional meteorological grid information and application of an elevation difference correction factor.

[0019] Preferably, a quantitative model is established based on the environmental state data, Beidou satellite signal propagation data and antenna phase center data; the quantitative model analyzes the environmental impact parameters to obtain the propagation path delay data and the antenna phase center change data process, comprising:

[0020] A coupling quantitative model based on the physical response mechanism of swell disturbance affecting the receiving antenna phase center position and the real-time correction mechanism of meteorological gradient affecting the satellite signal propagation path delay is constructed by using historical marine environmental state data, actual delay characteristic data of Beidou satellite signal propagation and the calibration results of the receiving antenna phase center; the three-dimensional instantaneous displacement of the receiving antenna phase center and the real-time high-precision correction value of the zenith tropospheric wet delay are input as current dynamic parameters into the coupling quantitative model, and the current satellite signal propagation path delay data and the three-dimensional spatial accurate position change data of the receiving antenna phase center are output in real time by the calculation module driven by the physical mechanism built in the coupling quantitative model.

[0021] Preferably, a satellite signal error model is constructed to analyze the environmental impact parameters, the propagation path delay data and the antenna phase center change data to obtain the positioning error value process, comprising:

[0022] The historical quantitative environmental impact parameters, historical propagation path delay data and historical antenna phase center change data are taken as input feature sets, the deviation between the positioning results obtained by high-precision processing and the preliminary satellite positioning results is taken as training output target, a satellite signal error prediction model is trained by a machine learning algorithm to capture and utilize the time sequence dependence relationship between the input features; the quantitative environmental impact parameters, the propagation path delay data and the antenna phase center change data obtained in real time are input into the trained satellite signal error prediction model, and a three-dimensional positioning error correction vector representing the current comprehensive positioning error is output by the model.

[0023] Preferably, the positioning result is solved to obtain a solution result; the solution result is judged according to a solution result threshold to divide the first working condition and the second working condition, including:

[0024] A set of quality evaluation indexes capable of comprehensively reflecting the current positioning performance is calculated in real time for the positioning result sequence obtained after the error model adjustment, the indexes include a consistency representation index of the positioning result, a state representation index of the carrier phase ambiguity solution and a fluctuation representation index of the positioning residual sequence, and the indexes constitute a multi-dimensional positioning quality state vector; when a plurality of preset key performance indexes in the positioning quality state vector continuously deviate from the normal operation interval under the preset high-precision positioning condition within a predetermined evaluation time length, it is determined that the current marine working condition is the second working condition which adversely affects the Beidou satellite positioning solution; otherwise, it is determined that the first working condition which satisfies the preset high-precision positioning condition is met.

[0025] Preferably, the second working condition switches the satellite positioning mode to an emergency mode, intensifies the compensation of the Beidou signal error, and obtains a satellite positioning result, including:

[0026] The emergency mode is: according to the specific cause of the identified second working condition, activating the corresponding compensation algorithm module, including enabling a multipath effect suppression algorithm, a carrier phase cycle slip detection and repair algorithm, and a dynamic compensation algorithm based on the close coupling of inertial measurement unit data and Beidou observation data; based on the online evaluation results of the real-time quality of each Beidou satellite signal, the satellite signals affected by the current adverse working condition are lowered, and signals from specific frequency bands of the Beidou system with stronger anti-interference capability are preferentially used to maintain the overall stability and continuity of the positioning solution; by executing the above targeted error compensation and data fusion weight optimization adjustment, the satellite positioning result which remains continuous and stable under the second working condition is output by the positioning algorithm under the emergency mode.

[0027] Compared with the prior art, the present application has the following beneficial effects:

[0028] 1、The present application establishes a refined quantitative model for the influence of marine environmental factors on the Beidou satellite signal propagation and the phase center of the receiving antenna, and real-time extracts quantitative environmental influence parameters related to satellite signal propagation error and observation data quality to dynamically integrate into the core of the cooperative satellite positioning algorithm, which realizes accurate compensation of complex marine unique error sources, thereby significantly improving the absolute accuracy and result reliability of millimeter-level positioning.

[0029] 2、The application enhances the applicability and overall performance of the Beidou system in achieving consistent millimeter-level high-precision collaborative positioning in large sea areas by adopting a Beidou collaborative satellite positioning algorithm capable of processing mixed adaptive reference network information composed of shore-based fixed stations and sea surface mobile platforms, and combining a mechanism for dynamically adjusting algorithm internal parameters and applying environmental impact compensation according to a preset performance objective function.

[0030] 3、The application improves the adaptive response capability of the Beidou positioning system in adverse sea conditions, the continuity of positioning services and the reliability of the final results by solving the positioning results and intelligently dividing the sea working conditions, switching to an emergency mode when adverse sea conditions are identified, starting the Beidou signal propagation error intensive compensation for the working condition, and dynamically optimizing the comprehensive response strategy of data fusion weight. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 A flowchart of a Beidou millimeter-level high-precision positioning method for the application embodiment is provided.

[0032] Figure 2 A flowchart of ocean environment and Beidou data collection for the application embodiment is provided.

[0033] Figure 3 A schematic diagram of the core mechanism of the coupling quantization model for the application embodiment is provided.

[0034] Figure 4 A schematic diagram of the satellite signal error prediction model structure for the application embodiment is provided.

[0035] Figure 5 A schematic diagram of the key algorithm module in the emergency mode for the application embodiment is provided. DETAILED DESCRIPTION

[0036] Next, the technical solutions of the application will be explained in detail and clearly in combination with specific examples of the application and related drawings. It needs to be made clear that the examples described below only represent a part of the application, not the whole. Based on these examples, those skilled in the art can deduce other possible implementation manners of the application without creative effort, and these manners are also protected by the application.

[0037] Example 1: Beidou millimeter-level high-precision positioning method in large sea areas under strong tidal current simulation conditions of A cross-sea bridge

[0038] The present embodiment provides a Beidou millimeter-level high-precision positioning method in large sea areas, which is suitable for high-precision positioning requirements in the strong tidal current environment of A cross-sea bridge. Referring to Figure 1The specific implementation steps of the method proposed in this invention include: S1. Real-time acquisition and preprocessing of marine environmental state data and BeiDou observation data that affect the propagation of BeiDou satellite signals, and extraction of environmental impact parameters related to BeiDou satellite signal propagation error and observation data quality from the marine environmental state data; S2. Establishment of a quantization model based on environmental state data, BeiDou satellite signal propagation data, and antenna phase center data; analysis of environmental impact parameters by the quantization model to obtain propagation path delay data and antenna phase center change data; S3. Processing BeiDou observation data and adaptive reference network information based on the BeiDou collaborative satellite positioning algorithm to obtain preliminary satellite positioning results; S4. Construction of a satellite signal error model to analyze environmental impact parameters, propagation path delay data, and antenna phase center change data to obtain positioning error values; adjustment of the preliminary satellite positioning results based on the positioning error values ​​to obtain positioning results; S5. Solving the positioning results to obtain solution results; judging the solution results based on the threshold of the solution results to divide them into a first type of working condition and a second type of working condition; S6. In the second type of working condition, switching the satellite positioning mode to emergency mode to enhance compensation for BeiDou signal error and obtain satellite positioning results.

[0039] Furthermore, real-time acquisition and preprocessing of marine environmental status data and BeiDou observation data affecting BeiDou satellite signal propagation are performed, and environmental impact parameters related to BeiDou satellite signal propagation errors and observation data quality are extracted from the marine environmental status data; (referencing...) Figure 2 Corresponding to step S1 above, the specific implementation process includes:

[0040] A sensor array consisting of multiple underwater pressure sensors was deployed on the A-bridge to collect tidal height data in real time at a sampling frequency of 1 Hz. These sensors are also equipped with wave monitoring capabilities and, in conjunction with wave sensors, collect swell spectrum data. Multiple meteorological stations were deployed at both ends of the A-bridge's shoreline and on the bridge itself. These stations collect temperature, air pressure, and water vapor content data in real time at 30-second sampling intervals.

[0041] A BeiDou multi-frequency receiver is installed on the positioning target to simultaneously receive signals from five BeiDou frequency points: B1I, B1C, B2a, B2b, and B3I. It outputs the raw pseudorange, raw carrier phase, and raw Doppler observations in real time. The receiver's sampling rate is set to 1Hz to meet the requirements of real-time high-precision positioning.

[0042] This application embodiment acquires comprehensive and timely basic data, including tidal field dynamics, fine swell characteristics, regional multi-point temperature, air pressure, water vapor partial pressure, and original observations from multiple BeiDou frequencies, through multi-modal real-time data acquisition over a large sea area. This provides rich and reliable raw information input for refined environmental impact modeling and high-precision parameter extraction of the BeiDou millimeter-level high-precision positioning method over a large sea area.

[0043] Based on real-time acquired tidal height and surge spectrum data, a platform hydrodynamic response inversion algorithm based on linear wave theory is used to calculate the three-dimensional instantaneous offset of the GNSS receiving antenna phase center relative to its nominal position caused by hydrodynamic disturbances to the platform. The algorithm first decomposes the surge spectrum into multiple frequency components, then calculates the force and torque exerted on the platform for each frequency component, and finally calculates the platform's displacement and attitude changes using a dynamic model, thereby obtaining the three-dimensional instantaneous offset of the antenna phase center.

[0044] By utilizing multi-point temperature, air pressure, and water vapor content data of the positioning area, and employing a bilinear interpolation method that combines regional three-dimensional meteorological grid information with an elevation difference correction factor, a high-precision real-time correction value for the wet delay component of the tropospheric propagation delay in the zenith direction above the positioning target is generated. This method first constructs a regional three-dimensional meteorological grid using meteorological station data, with a grid resolution of 100m × 100m × 10m. Then, meteorological parameters at the positioning target location are calculated using bilinear interpolation. Finally, the tropospheric wet delay is calculated using the Saastamoinen model and corrected using an elevation difference correction factor. In region A, the tropospheric wet delay typically ranges from 10 cm to 30 cm; after correction using this method, the residual error of the wet delay can be controlled within 5 mm.

[0045] This application's embodiments successfully quantify complex marine environmental impacts into key parameters that can be directly used for high-precision positioning error compensation through targeted, physical model-based inversion and interpolation algorithms. This provides high-quality, highly relevant input for the implementation of the BeiDou millimeter-level high-precision positioning method over large areas of the sea, improving the ability to capture and quantify the dynamic impacts of the marine environment.

[0046] Furthermore, a quantitative model is established based on environmental state data, BeiDou satellite signal propagation data, and antenna phase center data; the quantitative model analyzes environmental impact parameters to obtain propagation path delay data and antenna phase center change data; referencing Figure 3 The core mechanism of the coupled quantization model is illustrated in the diagram, corresponding to step S2 above. The specific implementation process includes:

[0047] Using marine environmental status data collected over the past three months in the A-cross-sea bridge area and actual delay characteristics data of BeiDou satellite signal propagation, combined with the calibration results of the receiving antenna phase center, a coupled quantization model was constructed. This model includes two core mechanisms: first, a physical response mechanism based on the impact of wave disturbances on the position of the receiving antenna phase center; and second, a real-time meteorological gradient correction mechanism based on the delay effect of atmospheric parameters on the satellite signal propagation path.

[0048] The physical response mechanism employs a fluid-structure coupling analysis method, combining the hydrodynamic model under surge action with the platform structural dynamic model to establish the dynamic response relationship between surge, platform, and antenna. This mechanism considers the effects of surges of different frequencies and directions on the platform, as well as the inherent frequency characteristics of the platform structure, and can accurately predict the dynamic changes of the antenna phase center under specific surge conditions.

[0049] The real-time meteorological gradient correction mechanism employs a three-dimensional variational assimilation technique, fusing multi-point meteorological observation data with numerical weather prediction model results to construct a high-resolution three-dimensional atmospheric parameter field. This mechanism can capture spatial gradient changes in atmospheric parameters, particularly the abrupt changes in atmospheric parameters at the land-sea interface, thereby more accurately estimating the delay effects along the satellite signal propagation path.

[0050] The three-dimensional instantaneous offset of the receiving antenna phase center and the real-time high-precision correction value of the zenith tropospheric wet delay are used as current dynamic parameters and input into the coupled quantization model. Through the calculation module driven by the physical mechanism built into the coupled quantization model, the current propagation path delay data of each Beidou satellite signal and the three-dimensional spatial precise position change data of the receiving antenna phase center are output in real time.

[0051] Under strong tidal surge conditions, model-calculated satellite signal propagation path delay data show significant differences in signal delay at different elevation angles. The signal delay for low-elevation satellites (below 15°) can reach a maximum of 3.5 cm, while the signal delay for high-elevation satellites (above 60°) is typically within 0.8 cm. Three-dimensional spatial position change data of the antenna phase center show a clear periodicity under strong tidal surge conditions. The period is similar to the main period of the tidal surge, and the amplitude is proportional to the height of the tidal surge.

[0052] This application's embodiments, by constructing and applying this refined quantification model, accurately calculate the complex three-dimensional positional changes of the antenna phase center under the simulated strong tidal surge conditions of the A-cross-sea bridge. Through a precise atmospheric delay correction model, the high-precision correction of zenith wet delay is combined with the precise projection of dry delay onto the satellite's line-of-sight direction, obtaining significant differences in satellite signal propagation path delay at different elevation angles. These precise, component-based error model outputs provide core physical quantities for subsequent positioning algorithms to directly and accurately compensate for major systematic errors in the marine environment, serving as crucial model support for achieving the ultimate millimeter-level positioning target.

[0053] Furthermore, based on the BeiDou collaborative satellite positioning algorithm, BeiDou observation data and adaptive reference network information are processed to obtain preliminary satellite positioning results; corresponding to step S3 above, the specific implementation process includes:

[0054] An adaptive reference network consisting of multiple fixed reference stations was established on the land on both sides of the A-bridge. Each reference station is equipped with a high-precision BeiDou multi-frequency receiver and a stable communication link to provide real-time observation data and network correction data.

[0055] Then, the BeiDou observation data and adaptive reference network information are processed using the BeiDou collaborative satellite positioning algorithm. This algorithm first performs quality checks on the raw observation data, removing gross errors and outliers. Then, it uses network correction data provided by the adaptive reference network to correct the observation data, including satellite orbit errors, satellite clock errors, ionospheric delay, and tropospheric delay. Next, the algorithm constructs observation equations using double-difference phase observations, estimates parameters using the Kalman filtering method, and calculates the preliminary position of the positioning target.

[0056] Under strong tidal conditions, the algorithm also employs an adaptive weighting strategy, dynamically adjusting the weights of observations based on the signal-to-noise ratio, elevation angle, and severity of multipath effects of satellite signals to improve the robustness of positioning solutions. Simultaneously, the algorithm leverages the multi-frequency observation advantages of the BeiDou system, reducing the impact of ionospheric delay through inter-frequency combination to further improve positioning accuracy.

[0057] The preliminary satellite positioning results obtained through the BeiDou collaborative satellite positioning algorithm achieved centimeter-level accuracy in both horizontal and vertical directions under strong tidal conditions. However, to meet the millimeter-level positioning requirements, further optimization of the positioning results is needed using an error model.

[0058] Furthermore, a satellite signal error model is constructed to analyze environmental impact parameters, propagation path delay data, and antenna phase center change data to obtain positioning error values. Based on these positioning error values, the preliminary satellite positioning results are adjusted to obtain the final positioning result. (Referencing...) Figure 4 The schematic diagram of the satellite signal error prediction model structure, corresponding to step S4 above, includes the following specific implementation process:

[0059] Using historical quantitative environmental impact parameters, historical propagation path delay data, and historical antenna phase center change data collected over the past six months at the A cross-sea bridge as the input feature set, the deviation between the positioning results obtained through high-precision processing and the preliminary satellite positioning results is used as the training output target. By capturing and utilizing the temporal dependencies between the input features through a long short-term memory network, a satellite signal error prediction model is trained.

[0060] like Figure 4As shown, the error prediction model employs a Long Short-Term Memory (LSTM) network structure, which effectively captures the nonlinear relationship and temporal dependence characteristics between environmental parameters and localization error. The model's input layer contains 42 neurons, corresponding to 42 environmental and signal feature parameters; the hidden layer contains 3 LSTM units, each with 128 neurons; and the output layer contains 3 neurons, corresponding to the three components of the 3D localization error correction vector. The model is trained using the Adam optimizer with a learning rate of 0.001, a batch size of 64, and 200 training epochs.

[0061] The currently acquired quantified environmental impact parameters, propagation path delay data, and antenna phase center change data are input into the trained satellite signal error prediction model, outputting a three-dimensional positioning error correction vector characterizing the current overall positioning error. Under strong tidal conditions, the horizontal component of this error correction vector is typically within ±1.5 cm, and the vertical component is typically within ±2.5 cm.

[0062] The initial satellite positioning results are adjusted based on the positioning error correction vector to obtain the final positioning result. The adjustment process adopts a weighted fusion method, which dynamically adjusts the weights of the error correction according to the stability of the current environmental conditions and the prediction confidence of the error model, to ensure that the optimal positioning result can be obtained under various conditions.

[0063] After error model adjustment, the final positioning results will have improved horizontal and vertical accuracy under strong tidal conditions, achieving millimeter-level high-precision positioning requirements.

[0064] This application embodiment integrates precisely quantified antenna displacement and atmospheric delay corrections into the core processing flow of the cooperative positioning algorithm in real time at the observation level. This ensures that every piece of observation information used for the final positioning calculation has eliminated known systematic errors caused by the marine environment to the greatest extent possible, thus guaranteeing the achievement and maintenance of final millimeter-level positioning accuracy in complex marine environments.

[0065] Further, the positioning results are processed to obtain the solution results; based on the threshold of the solution results, the solution results are judged to classify the first type of working condition and the second type of working condition; corresponding to the above step S5, the specific implementation process includes:

[0066] For the positioning result sequence obtained after error model adjustment, a set of quality assessment indicators that can comprehensively reflect the current positioning performance are calculated in real time. These indicators include the consistency characterization index of positioning results, the state characterization index of carrier phase ambiguity resolution, and the volatility characterization index of positioning residual sequence. Then, these indicators are used to form a multi-dimensional positioning quality state vector.

[0067] Consistency indicators include the standard deviation, maximum deviation, and average deviation of positioning results over a continuous 30-second period, used to assess the stability and consistency of positioning results. The horizontal standard deviation should be less than 3 mm, the maximum deviation less than 8 mm, and the average deviation less than 2 mm; the vertical standard deviation should be less than 5 mm, the maximum deviation less than 12 mm, and the average deviation less than 3 mm.

[0068] State characterization metrics for carrier phase ambiguity resolution are used to evaluate the quality and reliability of carrier phase ambiguity resolution. These metrics include the fixed solution ratio, ambiguity fixation time, and ambiguity reset frequency. The fixed solution ratio should be greater than 95%, the ambiguity fixation time less than 30 seconds, and the ambiguity reset frequency less than once per hour.

[0069] The volatility characteristics of the positioning residual sequence are used to evaluate the internal consistency and reliability of the positioning algorithm. These include the root mean square of the residuals, the autocorrelation coefficient of the residuals, and the spectral characteristics of the residuals. The root mean square of the residuals is less than 5 mm, the autocorrelation coefficient of the residuals is less than 0.3, and the residual spectrum exhibits white noise characteristics.

[0070] When multiple preset key performance indicators in the positioning quality state vector continuously deviate from the normal operating range within the expected evaluation period, the current marine operating condition is determined to be a second type of operating condition that adversely affects the BeiDou satellite positioning calculation; otherwise, it is determined to be a first type of operating condition that meets the preset high-precision positioning conditions.

[0071] It is classified as a second type of working condition when at least two of the following conditions are met:

[0072] 1. The horizontal standard deviation is greater than 5 mm for 60 consecutive seconds;

[0073] 2. The standard deviation in the vertical direction is greater than 8 mm for 60 consecutive seconds;

[0074] 3. The fixed solution ratio remains below 90% for 120 consecutive seconds;

[0075] 4. The ambiguity reset frequency exceeds once every 10 minutes;

[0076] 5. The root mean square of the residuals is greater than 8 mm for 60 consecutive seconds;

[0077] 6. The residual autocorrelation coefficient is greater than 0.5 for 60 consecutive seconds.

[0078] Under strong tidal conditions, when the tidal height exceeds 2.5 meters and is accompanied by strong winds (wind speed greater than 15 meters per second), the system will usually be classified as a Class II operating condition. In this case, the emergency mode needs to be activated to maintain positioning accuracy.

[0079] This application's embodiments utilize a multi-factor joint judgment logic combining the signal quality indicators of BeiDou observation data with key quantitative environmental impact parameters extracted from marine environmental state data to identify marine conditions that adversely affect BeiDou satellite positioning calculations. In the context of strong tidal bores at the A-cross-sea bridge, this allows for earlier and more accurate determination of when the system enters a second type of operational condition requiring special emergency procedures. This provides valuable decision-making time and a more reliable triggering basis for subsequent adaptive mode switching and targeted error enhancement compensation, significantly improving the entire positioning system's intelligent perception capability in response to sudden and severe changes in the marine environment, the timeliness of risk warnings, and the effectiveness of subsequent emergency responses.

[0080] Furthermore, in the second type of operating condition, the satellite positioning mode is switched to emergency mode to enhance compensation for BeiDou signal errors and obtain satellite positioning results; refer to Figure 5 The diagram shows the key algorithm modules in the emergency mode, corresponding to step S6 above. The specific implementation process includes:

[0081] Based on the specific causes of the identified second type of operating conditions, the corresponding compensation algorithm modules are activated, including the multipath effect suppression algorithm, the carrier phase cycle slip detection and repair algorithm, and the dynamic compensation algorithm based on the close coupling of inertial measurement unit data and BeiDou observation data.

[0082] The multipath effect suppression algorithm employs adaptive filtering technology based on signal quality monitoring to identify and reduce the interference of multipath signals on direct signals. The algorithm first identifies multipath signals by analyzing the shape of the correlation peak and the signal-to-noise ratio variation characteristics of the signal. Then, it processes the received signal using an adaptive filter to reduce the influence of multipath components. Under strong tidal conditions, this algorithm can reduce multipath errors by approximately 70%.

[0083] The carrier phase cycle slip detection and repair algorithm employs a combination of multi-frequency combination and time series analysis to quickly and accurately detect and repair cycle slips in carrier phase observations. The algorithm first constructs a cycle slip detection combination using multi-frequency observations, then detects abrupt changes in the combination using time series analysis, and finally determines the cycle slip size and performs repair using an integer ambiguity search method. Under strong tidal conditions, the algorithm achieves a cycle slip detection success rate exceeding 98% and a repair success rate exceeding 95%.

[0084] The dynamic compensation algorithm based on the tight coupling of inertial measurement unit (IMU) data and BeiDou observation data employs a tightly coupled Kalman filter framework to fuse acceleration and angular velocity data provided by the IMU with BeiDou observation data. This algorithm can maintain high-precision positioning for a short period using inertial data when BeiDou signal quality deteriorates, and quickly reconverges after BeiDou signal recovery. Under strong tidal surge conditions, the algorithm can maintain horizontal accuracy better than 2 cm and vertical accuracy better than 3 cm even when BeiDou signal is interrupted for up to 10 seconds.

[0085] Based on the online assessment of the real-time quality of various BeiDou satellite signals, the system lowers the signal weight of satellites affected by the current unfavorable operating conditions and prioritizes signals from specific frequency bands of the BeiDou system that have stronger anti-interference capabilities to maintain the overall stability and continuity of the positioning solution. Under strong tidal conditions, the system will prioritize BeiDou B3I band signals because this band has higher power and stronger anti-interference capabilities; at the same time, the weight of low-elevation satellites will be reduced because their signals are more susceptible to multipath effects and atmospheric delay.

[0086] By implementing the aforementioned targeted error compensation and data fusion weight optimization adjustments, the positioning algorithm in emergency mode outputs continuous and stable satellite positioning results under the second type of operating conditions. Under strong tidal surge conditions, emergency mode can maintain positioning accuracy within 1 cm in the horizontal direction and within 1.5 cm in the vertical direction. Although this is lower than that under normal operating conditions, it still meets the needs of most high-precision applications.

[0087] This application embodiment intelligently and adaptively switches the satellite positioning mode to a specially designed emergency mode when a second type of marine condition is identified that adversely affects the BeiDou satellite positioning calculation. It also immediately activates a series of BeiDou signal error enhancement compensation mechanisms and optimized data fusion weight adjustment strategies to address the characteristics of the current adverse condition. This improves the problem of a sharp drop in positioning accuracy or even positioning interruption when encountering harsh marine environments, and enhances the practical application effectiveness and environmental adaptability of BeiDou high-precision positioning technology in real, complex, and extreme marine environments.

[0088] This application embodiment successfully realizes a dynamic adaptive method for millimeter-level high-precision positioning of BeiDou in large-area sea areas by constructing a complete technical process, which includes intelligent perception of real-time marine environment and BeiDou observation data, refined quantitative modeling and parameter extraction of environmental impact, preliminary solution based on collaborative positioning, result optimization and adjustment through satellite signal error prediction model, and finally intelligent division of working conditions and targeted emergency response based on the quality feedback of positioning solution itself.

[0089] Example 2: BeiDou Millimeter-Level High-Precision Positioning Method for Large-Area Sea Areas under Simulated Typhoon Impact on Bridge B (Baihai Bridge)

[0090] This embodiment provides a BeiDou-based high-precision positioning method for large-area sea areas with millimeter-level accuracy. This method is suitable for high-precision positioning requirements under typhoon influence on the B-cross-sea bridge. (See reference...) Figure 1 The specific implementation steps of the method proposed in this invention include: S1. Real-time acquisition and preprocessing of marine environmental state data and BeiDou observation data that affect the propagation of BeiDou satellite signals, and extraction of environmental impact parameters related to BeiDou satellite signal propagation error and observation data quality from the marine environmental state data; S2. Establishment of a quantization model based on environmental state data, BeiDou satellite signal propagation data, and antenna phase center data; analysis of environmental impact parameters by the quantization model to obtain propagation path delay data and antenna phase center change data; S3. Processing BeiDou observation data and adaptive reference network information based on the BeiDou collaborative satellite positioning algorithm to obtain preliminary satellite positioning results; S4. Construction of a satellite signal error model to analyze environmental impact parameters, propagation path delay data, and antenna phase center change data to obtain positioning error values; adjustment of the preliminary satellite positioning results based on the positioning error values ​​to obtain positioning results; S5. Solving the positioning results to obtain solution results; judging the solution results based on the threshold of the solution results to divide them into a first type of working condition and a second type of working condition; S6. In the second type of working condition, switching the satellite positioning mode to emergency mode to enhance compensation for BeiDou signal error and obtain satellite positioning results.

[0091] A monitoring network consisting of multiple underwater pressure sensors was deployed around the B-bridge area. This network collects tidal height data covering the designated sea area in real time at a sampling frequency of 2Hz. Detailed swell characteristics are obtained by performing spectral analysis on the collected pressure data, with an analysis frequency range of 0.03Hz to 0.8Hz. Simultaneously, multiple automatic weather stations were deployed at both ends of the bridge's shoreline and on the bridge structure, collecting real-time data on temperature, atmospheric pressure, and water vapor content at multiple points within the designated area at 10-second sampling intervals. A BeiDou multi-frequency high-precision receiver was installed on the designated target. This receiver is configured to simultaneously track and output raw pseudorange observations, raw carrier phase observations, and raw Doppler shift observations for five publicly available service frequencies of the BeiDou system: B1I, B1C, B2a, B2b, and B3I. The receiver's data output rate was set to 2Hz to accommodate rapid movement and signal changes that may occur during typhoon conditions.

[0092] All collected raw marine environmental status data and BeiDou observation data underwent rigorous time synchronization processing to be unified to the GPST time reference, and preliminary quality control was performed, including removing obvious gross errors exceeding reasonable limits, missing data, and interrupted periods. Based on the preprocessed real-time tidal height and swell spectrum data, a platform hydrodynamic response inversion algorithm based on linear wave theory was applied. This algorithm considers the direct effect of wind on the platform under typhoon conditions and calculates the three-dimensional instantaneous offset of the phase center of the three-dimensional GNSS receiving antenna relative to its nominal or static reference position caused by the combined hydrodynamic disturbances of wind, waves, and currents on the positioning platform. Using preprocessed multi-point temperature, atmospheric pressure, and water vapor content data within the positioning area, combined with a pre-stored three-dimensional meteorological grid information covering the area with high spatiotemporal resolution, an improved bilinear spatial interpolation method was employed, and a correction factor based on elevation differences was applied to generate and update high-precision real-time correction values ​​for the wet delay component (ZWD) in the tropospheric propagation delay in the zenith direction above the positioning target.

[0093] A coupled quantization model was constructed, utilizing historical marine environmental data collected over the past six months from actual observations in the B-bridge area, statistical data on the actual delay characteristics of BeiDou satellite signals during propagation in this area due to environmental influences, and the precise calibration results of the phase center position of the GNSS receiving antenna installed on the positioning platform as prior information or the basis for model parameter calibration. This coupled quantization model integrates two core mechanisms to handle antenna phase center changes and satellite signal propagation path delays, respectively:

[0094] The first core mechanism is a physical response mechanism based on fully coupled dynamic analysis of wind, waves, current, and structure, used to accurately simulate and predict the dynamic changes in the three-dimensional spatial position of the GNSS receiving antenna phase center. This mechanism strongly couples real-time forecasted typhoon wind field models, complex surge field models based on nonlinear wave theory, and regional flow field change models potentially driven by strong winds or storm surges with a refined finite element structural dynamics model of the positioning platform. This coupled solution process considers the interaction of multiple physical fields and the nonlinear response characteristics of the platform structure under extreme loads. By strongly coupling the wind field, nonlinear wave field, flow field, and structural dynamics, this mechanism solves the problem of failure in traditional linear models.

[0095] The second core mechanism is a real-time meteorological gradient correction mechanism based on four-dimensional variational data assimilation, used to accurately estimate the propagation path delay of satellite signals in the troposphere. This mechanism fuses multi-point meteorological observation data collected in real time from multiple automatic weather stations deployed in the region with the forecast field results output by a high-resolution numerical weather prediction model optimized for severe weather such as typhoons. Through four-dimensional variational assimilation, a four-dimensional atmospheric parameter field with high spatiotemporal resolution covering the location area is constructed, thereby enabling more accurate capture of the rapid spatiotemporal changes and significant spatial gradient distribution of atmospheric parameters under typhoon conditions.

[0096] During real-time positioning, the instantaneous three-dimensional offset of the receiving antenna phase center and the real-time high-precision correction value of the zenith tropospheric wet delay are used as current dynamic environmental impact parameters and input into the aforementioned coupled quantization model. Through the computational module built into this coupled quantization model, driven by the aforementioned physical response mechanism and real-time meteorological gradient correction mechanism, the propagation path delay data of each BeiDou satellite signal and the precise three-dimensional spatial position change data of the receiving antenna phase center are continuously output in real time.

[0097] To ensure the stability of the reference network during typhoon conditions, this embodiment constructs an enhanced adaptive reference network consisting of multiple fixed reference stations in the B-cross-sea bridge area. The distance between each reference station is approximately 5 to 8 kilometers, and all stations are equipped with highly reliable communication links and long-term backup power systems. An enhanced BeiDou collaborative satellite positioning algorithm processes BeiDou observation data and adaptive reference network information. This algorithm first performs rigorous enhancement quality detection and control on the input raw BeiDou observation data, employing multiple statistical tests to eliminate gross errors and outlier observations. Subsequently, network correction data obtained from the adaptive reference network is used to finely correct systematic errors in the observation data. After correction, the algorithm constructs observation equations using optimized double-difference carrier phase observations and estimates parameters using an extended Kalman filter method that adaptively adjusts its process noise covariance matrix to calculate the preliminary three-dimensional position of the positioning target.

[0098] Under typhoon conditions, the dynamic adaptive weighting strategy within this collaborative positioning algorithm dynamically adjusts the weight of each observation in the solution model in real time based on the real-time carrier-to-noise ratio, elevation angle, multipath interference level assessed through signal cross-correlation characteristics, and path environment reliability factors and signal reception stability factors indirectly output by the coupled quantization model, which characterize the degree of influence of typhoon atmospheric disturbances on the propagation paths of each satellite signal. This is achieved through a preset optimization objective function. Simultaneously, the algorithm optimizes the frequency combination method for BeiDou multi-frequency observation data, prioritizing specific frequency combinations that are more resistant to interference and less affected by multipath effects in strong winds and rain.

[0099] A satellite signal error prediction model was trained using historical quantitative environmental impact parameters, propagation path delay data, antenna phase center change data, and the deviation between the corresponding high-precision reference positioning results and the preliminary positioning results collected in the B-bridge area over the past year. This model employs a deep learning architecture combining a bidirectional long short-term memory (Bi-LSTM) network and an attention mechanism. The Bi-LSTM network captures long-term dependencies in the temporal data, while the attention mechanism identifies and focuses on the input features that contribute most to the current positioning error. The model's input feature set contains 64 dimensions, and the hidden layers consist of four Bi-LSTM units, each with 256 neurons, along with an attention layer.

[0100] During real-time positioning, the currently acquired quantified environmental impact parameters, propagation path delay data, and antenna phase center change data are input into a trained error prediction model. This model outputs a three-dimensional positioning error correction vector characterizing the current overall positioning error. Based on this positioning error correction vector, the preliminary satellite positioning results are adjusted. An adaptive weighted fusion method is used in the adjustment process. This method, when determining the weights for error correction, additionally considers the rate of change of key environmental parameters, so as to dynamically adjust the smoothness or update frequency of error correction application when environmental parameters change rapidly.

[0101] To shorten error correction latency and enhance system autonomy in offline or weak network environments, the satellite signal error prediction model is distributed across edge computing nodes. The trained Bi-LSTM and attention-based deep learning models are processed using lightweight techniques such as model quantization and pruning, and then deployed to edge computing devices near the BeiDou receiver, such as high-performance ARM embedded processors or FPGA accelerator cards. These edge nodes receive local sensor data and raw BeiDou observation data in real time, independently run the error prediction model, and generate local error correction vectors. This reduces the latency of data transmission back to the central server for computation, improving real-time performance. Furthermore, a lightweight collaborative mechanism can be established between edge nodes. By exchanging local error information and key parameters, a certain degree of error correction collaboration can be achieved without relying on the central server, thus providing high-precision positioning services even when the network is limited or interrupted. This distributed deployment reduces error correction latency to 50 milliseconds, thereby improving the reliability of autonomous positioning in field or offshore environments.

[0102] The adjusted positioning results are calculated in real time to obtain a set of quality assessment indicators that comprehensively reflect the current positioning performance. These include consistency indicators for positioning results, state indicators for carrier phase ambiguity resolution, and volatility indicators for positioning residual sequences. The operating condition judgment criteria for each indicator are adaptively adjusted based on the actual signal performance in the typhoon environment. When multiple preset key indicators in these quality assessment indicators continuously deviate from the normal operating range within a predetermined assessment period—that is, when several of the following conditions are simultaneously met: horizontal standard deviation greater than 8 mm for 30 consecutive seconds, fixed solution ratio less than 85% for 60 consecutive seconds, and the number of available satellites less than 6 for 6 consecutive seconds—the current marine operating conditions are comprehensively judged as marine operating conditions that adversely affect BeiDou satellite positioning calculations.

[0103] In emergency positioning mode, the system activates corresponding compensation algorithm modules that are deeply optimized for typhoon environments based on the specific triggers of adverse marine conditions identified. These algorithm modules include: an enhanced multipath effect suppression algorithm unit that combines signal time-spectrum feature recognition based on convolutional neural networks (CNN) with adaptive Kalman filtering; a high-dynamic carrier phase cycle slip detection and repair algorithm unit that combines multi-frequency, multi-system observations with wavelet transform analysis; and a dynamic compensation algorithm unit that tightly couples high-precision inertial measurement unit (IMU) data with BeiDou observation data as the core emergency support. The tightly coupled algorithm unit employs an optimized Kalman filtering framework that can estimate and compensate for various random errors of the IMU online. In cases of extremely poor BeiDou signal quality or short-term complete interruption, it mainly relies on the high-precision IMU for short-term high-precision navigation calculations and achieves rapid and seamless relocking and data fusion when the BeiDou signal is restored.

[0104] Meanwhile, the data fusion weight adjustment strategy will be further strengthened in this emergency mode. In addition to prioritizing the use of frequency band signals with strong anti-interference capabilities such as Beidou B3I and B2b and completely excluding satellites with extremely low elevation angles, the weight of satellites with medium elevation angles will be adjusted more precisely and dynamically based on the specific degree of signal attenuation caused by wind and rain in real time assessments and by analyzing indicators such as the rapid change gradient of signal strength.

[0105] To enable more intelligent operation condition calculation and emergency mode switching, a multi-layer perceptron (MLP)-based operation condition classifier is introduced on top of the existing quality assessment indicators. The input feature vector of this classifier not only includes the consistency of positioning results, ambiguity resolution status, and residual volatility, but also integrates the abnormal change rates of the extracted environmental impact parameters, such as the sudden increase rate of tidal surge intensity, abnormal atmospheric refractive index gradient, residual information of the antenna phase center change model, as well as the convergence speed and iteration count of the cooperative positioning algorithm. Through training on a large amount of historical data under normal, tidal surge, typhoon, and signal obstruction conditions, the MLP classifier can output the probability distribution of the current positioning condition in real time. When the probability of a specific abnormal operation condition exceeds a preset threshold, the system can not only switch to emergency mode but also trigger an early warning mechanism for that specific risk, providing the operator with a detailed risk report. This enables more refined and proactive operation condition management and risk control, significantly improving system safety, decision-making efficiency, and proactive response to emergencies.

[0106] In this embodiment, for the extreme condition of typhoon, the data acquisition strategy, model construction, algorithm parameters and emergency response mechanism in each step are optimized. Dynamic response prediction based on the coupling effect of wind field and wave field is introduced, and a more robust multi-sensor deep fusion strategy is adopted in emergency mode to maximize the positioning accuracy under the influence of typhoon.

[0107] Although specific examples of the invention have been shown, those skilled in the art will understand that various adjustments, modifications, substitutions, and improvements can be made to these examples without departing from the core principles and spirit of the invention. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A BeiDou method for millimeter-level high-precision positioning over large sea areas, characterized in that, Includes the following steps: Collect and process marine environmental status data and BeiDou observation data, and extract environmental impact parameters related to BeiDou satellite signal propagation error and observation data quality from the marine environmental status data; A coupled quantization model is established based on the environmental state data, BeiDou satellite signal propagation data, and antenna phase center data. The quantization model analyzes environmental impact parameters through the synergistic effect of the surge disturbance physical response mechanism and the atmospheric gradient real-time correction mechanism, and obtains propagation path delay data and antenna phase center change data. The BeiDou observation data and adaptive reference network information are processed based on the BeiDou collaborative satellite positioning algorithm to obtain preliminary satellite positioning results. A satellite signal error model is constructed to analyze environmental impact parameters, propagation path delay data, and antenna phase center change data to obtain positioning error values ​​and adjust the preliminary satellite positioning results. The positioning results are processed and, based on the threshold of the processing results, the operating conditions are divided into two categories. The second category is when the current marine conditions have an adverse impact on the BeiDou satellite positioning processing, while the first category is when the preset high-precision positioning conditions are met. In the second category, the satellite positioning mode is switched to emergency mode to enhance BeiDou signal error compensation and obtain satellite positioning results.

2. The BeiDou method for millimeter-level high-precision positioning over large sea areas according to claim 1, characterized in that, The marine environmental status data and BeiDou observation data include: The system includes real-time data on tidal height and swell spectrum collected by an underwater pressure sensor network deployed in the designated sea area, and real-time data on temperature, air pressure, and water vapor content at multiple points in the designated area collected by an automatic weather station network deployed on shore-based and sea-surface platforms; and BeiDou observation data including raw pseudorange, raw carrier phase, and raw Doppler observations at multiple frequencies output in real time by a BeiDou receiver mounted on the designated target.

3. The BeiDou method for millimeter-level high-precision positioning over large sea areas according to claim 1, characterized in that, The process of obtaining environmental impact parameters includes: Based on real-time tidal height and surge spectrum data, a platform hydrodynamic response inversion algorithm based on linear wave theory is used to calculate the three-dimensional instantaneous offset of the phase center of the three-dimensional GNSS receiving antenna relative to the nominal position caused by hydrodynamic disturbance of the platform. Using multi-point temperature, air pressure and water vapor content data of the positioning area, a bilinear interpolation method combining regional three-dimensional meteorological grid information is used, and an elevation difference correction factor is applied to generate a high-precision real-time correction value for the wet delay component of the tropospheric propagation delay in the zenith direction above the positioning target.

4. The BeiDou method for millimeter-level high-precision positioning over large sea areas according to claim 3, characterized in that, A coupled quantization model is established based on the environmental state data, BeiDou satellite signal propagation data, and antenna phase center data. This quantization model analyzes environmental impact parameters through the synergistic effect of the surge disturbance physical response mechanism and the real-time atmospheric gradient correction mechanism, obtaining propagation path delay data and antenna phase center change data. The process includes: Based on historical ocean environment data and actual delay characteristics of BeiDou satellite signal propagation, and combined with the calibration results of the receiving antenna phase center, a coupled quantization model is constructed. This model combines a physical response mechanism based on the influence of surge disturbance on the position of the receiving antenna phase center with a real-time meteorological gradient correction mechanism based on the influence of atmospheric parameters on the satellite signal propagation path delay. The high-precision real-time correction values ​​of the three-dimensional instantaneous offset of the receiving antenna phase center relative to its nominal position and the wet delay component in the tropospheric propagation delay in the zenith direction are used as current dynamic parameters and input into the coupled quantization model. Through the calculation module driven by the physical mechanism built into the coupled quantization model, the model outputs in real time the current BeiDou satellite signal propagation path delay data and the precise three-dimensional spatial position change data of the receiving antenna phase center.

5. The BeiDou method for millimeter-level high-precision positioning over large sea areas according to claim 1, characterized in that, The process of constructing a satellite signal error model by analyzing environmental impact parameters, propagation path delay data, and antenna phase center change data to obtain positioning error values ​​includes: Using historically collected quantified environmental impact parameters, historical propagation path delay data, and historical antenna phase center change data as input feature sets, the deviation between the positioning result obtained through high-precision processing and the preliminary satellite positioning result is used as the training output target. A hybrid architecture machine learning algorithm combining long short-term memory networks and attention mechanisms is used to capture and utilize the temporal dependencies between input features, and a satellite signal error prediction model is trained. The currently acquired quantified environmental impact parameters, propagation path delay data, and antenna phase center change data are input into the trained satellite signal error prediction model, and the model outputs a three-dimensional positioning error correction vector characterizing the current comprehensive positioning error.

6. The BeiDou method for millimeter-level high-precision positioning over large sea areas according to claim 1, characterized in that, The process of calculating the positioning results and classifying the working conditions into first and second categories based on the threshold values ​​of the calculation results includes: For the positioning result sequence obtained after error model adjustment, a set of quality assessment indicators that can comprehensively reflect the current positioning performance are calculated in real time. The indicators include the consistency characterization index of the positioning results, the state characterization index of carrier phase ambiguity resolution, and the volatility characterization index of the positioning residual sequence. These indicators constitute a multi-dimensional positioning quality state vector. When multiple preset key performance indicators in the positioning quality state vector continuously deviate from their normal operating range under preset high-precision positioning conditions within a predetermined evaluation period, the current marine working condition is determined to be the second type of working condition that adversely affects the BeiDou satellite positioning solution; otherwise, it is determined to be the first type of working condition that meets the preset high-precision positioning conditions.

7. The BeiDou method for millimeter-level high-precision positioning over large sea areas according to claim 1, characterized in that, The second type of operating condition switches to emergency mode, enhances BeiDou signal error compensation, and obtains satellite positioning results, including: Based on the specific causes of the identified second type of operating condition, the corresponding compensation algorithm modules are activated, including a multipath effect suppression algorithm, a carrier phase cycle slip detection and repair algorithm, and a dynamic compensation algorithm based on the close coupling of inertial measurement unit data and BeiDou observation data. Based on the online evaluation results of the real-time quality of each BeiDou satellite signal, the satellite signals affected by the current unfavorable operating condition are reduced, and frequency band signals from the BeiDou system with anti-interference capabilities are given priority to maintain the overall stability and continuity of the positioning solution. By performing the above error compensation and data fusion weight optimization adjustment, the positioning algorithm in the emergency mode outputs a satellite positioning result that remains continuous and stable under the second type of operating condition.

Images