Beidou water vapor inversion method and system based on buoy platform

Abstract

The embodiment of the invention provides a Beidou water vapor inversion method and system based on a buoy platform. The method is applied to the technical field of data processing, and comprises the following steps: receiving a Beidou dual-frequency signal through a direct antenna to obtain original observation data; the multi-path effect and the platform shake are inhibited by combining buoy motion characteristics obtained by the MEMS sensor to obtain pre-processed data; calculating a troposphere delay estimated value; establishing a delay-water vapor conversion model; suppressing short message interference to obtain effective observation data; and inputting the data into a model to calculate the amount of precipitable water and the relative humidity. According to the method, the key technical problems of sea surface multi-path effect suppression, platform shake compensation, marine atmosphere modeling, short message interference elimination and the like are solved, so that reliable water vapor parameter inversion on the buoy platform is realized.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing, and in particular, to a Beidou water vapor inversion method and system based on a buoy platform. Background Art

[0002] The Beidou satellite navigation system has been widely used in the field of atmospheric water vapor remote sensing. In the prior art, land-based GNSS water vapor remote sensing has been relatively mature, and the atmospheric water vapor content is monitored with high precision through a receiver network. Marine water vapor remote sensing mainly relies on meteorological satellites and radiosondes, and the combination of a buoy platform and the Beidou system for water vapor remote sensing is a new research direction. The buoy platform has advantages such as flexible observation points and wide coverage, providing a new technical means for marine atmospheric water vapor monitoring.

[0003] However, there are multiple technical difficulties in the existing Beidou water vapor remote sensing of buoy platforms. First, the sea surface multipath effect and the swaying of the buoy platform seriously affect the quality of the observed data; second, the spatio-temporal variation characteristics of the marine atmospheric environment are significantly different from those of the land environment, and traditional water vapor inversion models are difficult to accurately describe; third, the Beidou short message service is prone to produce frequency superposition interference in marine applications, affecting the observation accuracy. Summary of the Invention

[0004] The present disclosure provides a Beidou water vapor inversion method and system based on a buoy platform. The present disclosure solves key technical problems such as suppression of the sea surface multipath effect, platform swaying compensation, marine atmosphere modeling, and short message interference elimination, so as to realize reliable water vapor parameter inversion on the buoy platform.

[0005] According to a first aspect of the present disclosure, there is provided a Beidou water vapor inversion method based on a buoy platform, including: receiving and processing the Beidou satellite dual-frequency signal through a direct antenna to obtain Beidou original observation data including pseudorange observation values, carrier phase observation values, and satellite orbit parameters; according to the Beidou original observation data and in combination with the buoy motion characteristic information provided by the MEMS sensor, suppressing the multipath effect and platform swaying through wavelet transform and adaptive filtering to obtain preprocessed Beidou observation data; based on the preprocessed Beidou observation data, calculating and processing the wet and dry delay components to obtain a tropospheric delay estimation value; according to the tropospheric delay estimation value, using a trained neural network to model the delay-water vapor conversion relationship to obtain a water vapor parameter inversion model; for the preprocessed Beidou observation data, identifying and suppressing the short message frequency superposition interference through an adaptive notch filtering technique to obtain interference-removed observation data; inputting the interference-removed observation data into the water vapor parameter inversion model, and obtaining water vapor parameter estimation values including precipitable water and relative humidity through real-time calculation.

[0006] According to a second aspect of the present disclosure, a Beidou water vapor inversion system based on a buoy platform is provided, including:

[0007] A receiving module, configured to receive and process the Beidou satellite dual-frequency signal through a direct antenna to obtain Beidou original observation data including pseudorange observation values, carrier phase observation values, and satellite orbit parameters;

[0008] A suppression module, configured to suppress multipath effects and platform jitter through wavelet transform and adaptive filtering according to the Beidou original observation data and in combination with the buoy motion characteristic information provided by the MEMS sensor, to obtain preprocessed Beidou observation data;

[0009] A calculation module, configured to calculate and process the wet and dry delay components based on the preprocessed Beidou observation data to obtain a tropospheric delay estimate value;

[0010] A modeling module, configured to model the delay-water vapor conversion relationship by using a trained neural network according to the tropospheric delay estimate value to obtain a water vapor parameter inversion model;

[0011] An identification module, configured to identify and suppress the short message overlapping frequency interference for the preprocessed Beidou observation data through an adaptive notch filtering technique to obtain interference-removed observation data;

[0012] An input module, configured to input the interference-removed observation data into the water vapor parameter inversion model to obtain water vapor parameter estimate values including precipitable water and relative humidity through real-time calculation.

[0013] The present disclosure processes the Beidou satellite dual-frequency signal received by the direct antenna to obtain high-quality original observation data, effectively improving the accuracy of data acquisition. At the same time, by using the buoy motion characteristic information provided by the MEMS sensor and combining wavelet transform and adaptive filtering techniques, the multipath effects and platform jitter are effectively suppressed, significantly reducing the interference of the sea surface environment on the observation data. And the wet and dry delay components are accurately calculated through an improved Saastamoinen model to accurately obtain the tropospheric delay estimate value, providing a reliable data basis for subsequent water vapor parameter inversion. Moreover, a delay-water vapor conversion relationship model is established by using a trained neural network, fully considering the particularity of the ocean environment and improving the adaptability of water vapor parameter inversion. In particular, the short message overlapping frequency interference in the marine environment is identified and suppressed through an adaptive notch filtering technique, solving the interference problem of the Beidou short message service on the observation data in the marine environment. Finally, the processed observation data is input into the water vapor parameter inversion model to achieve high-precision estimation of precipitable water and relative humidity.

[0014] It should be understood that the content described in the Summary of the Invention section is not intended to limit the key or important features of the embodiments of the present disclosure, nor to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In conjunction with the accompanying drawings and with reference to the following detailed description, the above and other features, advantages, and aspects of the embodiments of the present disclosure will become more apparent. The drawings are used to better understand the solution and do not limit the present disclosure. In the drawings, the same or similar reference numerals denote the same or similar elements, where:

[0016] Figure 1 FIG. shows a flowchart of a Beidou water vapor inversion method based on a buoy platform according to an embodiment of the present disclosure;

[0017] Figure 2 FIG. shows a block diagram of a Beidou water vapor inversion system based on a buoy platform according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present disclosure clearer, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Apparently, the described embodiments are some, but not all, of the embodiments of the present disclosure. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present disclosure without creative efforts shall fall within the scope of protection of the present disclosure.

[0019] In addition, the term "and / or" in this article is merely a description of the associated relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after.

[0020] Figure 1 FIG. shows a schematic flow diagram of a Beidou water vapor inversion method 100 in an embodiment of the present disclosure, as Figure 1 shown, the method 100 includes:

[0021] S110: Receive and process the Beidou satellite dual-frequency signal through a direct antenna to obtain Beidou original observation data including pseudorange observation values, carrier phase observation values, and satellite orbit parameters;

[0022] Optionally, the Beidou satellite B1I / B3I dual-frequency signals are received and selected by setting the elevation angle, and the dual-frequency signals of 6-12 visible Beidou satellites are obtained; the Beidou satellite dual-frequency signals are input into the carrier phase smoothed pseudorange processing unit to obtain pseudorange observations and carrier phase observations; the pseudorange observations and carrier phase observations are combined with the ephemeris data, and satellite orbit parameters are obtained through satellite orbit calculation processing; the pseudorange observations, carrier phase observations, and satellite orbit parameters are combined and sorted according to the RINEX format standard to obtain the original Beidou observation data.

[0023] Among them, the direct antenna is the key hardware for signal reception. The satellite signals are screened by setting the elevation mask angle of the antenna. The mask angle is set to 10°, and this angle value is determined based on the multipath effect and atmospheric refraction characteristics in the marine environment. During the signal reception process, the Beidou B1I and B3I dual-frequency signals are synchronously received. The selection of the dual-frequency signals is based on the consideration of signal stability and anti-interference ability. Visible satellites are selected during the signal reception process. The mask angle setting limits the reception range to the area above the elevation angle of 10°. Through satellite orbit calculation and antenna pointing control, 6-12 satellites with better signal quality are selected from all visible Beidou satellites. The dual-frequency signals of these satellites constitute the basic observation data, and the data contains the original measurement values of pseudorange and carrier phase.

[0024] Carrier phase smoothing of pseudorange is a key data processing step. This processing improves the accuracy of pseudorange measurement through the high-precision characteristics of the carrier phase. During the processing, the change amount of the carrier phase in consecutive epochs is used as the pseudorange correction amount to smooth the original pseudorange observations. The pseudorange observations reflect the geometric distance from the satellite to the receiver, while the carrier phase observations provide more accurate distance change amount information. Satellite orbit determination is an important link to obtain the precise satellite position. This process uses the orbital elements contained in the ephemeris data, combines the pseudorange observations and carrier phase observations to perform precise calculation of the orbit parameters. The perturbation factors of satellite motion, including the influence of the non-spherical gravity of the earth, solar gravity, lunar gravity, etc., are considered in the calculation process, and the precise orbit parameters of the satellite are obtained through iterative calculation.

[0025] In the data combination and arrangement stage, the RINEX format standard is adopted, which is an internationally common GNSS observation data format. During the arrangement process, information such as the observation epoch, satellite PRN number, signal type, and observation value type is organized in accordance with the standard format. The data file in RINEX format consists of two parts: the file header information and the observation data. The file header records basic parameters such as the observation station information, receiver information, and antenna information, and the observation data part records the observation values of each epoch in chronological order. The data processing unit samples and quantizes the received dual-frequency signals, converting the analog signals into digital signals. The original observation values obtained after signal demodulation are processed by carrier phase smoothing, and the pseudorange measurement accuracy is improved from the meter level to the decimeter level. During the orbit determination process, satellite velocity information is obtained through the measurement of Doppler frequency shift, and the satellite position is calculated in combination with the ephemeris parameters. The entire data acquisition and processing process forms a complete observation data chain, providing basic data support for subsequent water vapor inversion.

[0026] The processing unit performs signal tracking to lock the B1I and B3I signals transmitted by Beidou satellites. The tracking loop includes a code tracking loop and a carrier tracking loop, and the measurement of the pseudocode phase and carrier phase is achieved through correlators and phase discriminators. The original observation values obtained from the measurement are processed by carrier phase smoothing, and the selection of the smoothing time window is based on the signal continuity and ionospheric variation characteristics. The extended Kalman filter algorithm is used for satellite orbit determination. The state vector includes position, velocity, and clock offset parameters. The observation equation is constructed using pseudorange and carrier phase observation values, and the dynamic estimation of parameters is achieved through two stages: prediction and update.

[0027] The motion characteristics of the buoy platform have a significant impact on the quality of the observation data, and the movement of the antenna phase center will introduce additional measurement errors. Therefore, the influence of the platform attitude change on the observation values needs to be considered during the data processing. The generated Beidou original observation data is stored in RINEX format, with a data recording interval of 1 second, including dual-frequency pseudorange observation values, carrier phase observation values, Doppler observation values, and signal strength and other information. These data constitute the basic data set for water vapor inversion.

[0028] S120: According to the Beidou original observation data and combined with the buoy motion characteristic information provided by the MEMS sensor, the multipath effect and platform jitter are suppressed through wavelet transform and adaptive filtering to obtain the preprocessed Beidou observation data;

[0029] Optionally, acceleration and angular velocity data are obtained from the MEMS sensor for attitude calculation to obtain the buoy motion characteristic information; the Beidou raw observation data are subjected to wavelet multi-scale decomposition for noise characteristic extraction to obtain the multipath characteristic components; the buoy motion characteristic information and the multipath characteristic components are combined for adaptive filtering processing to obtain the filtered observation data; the filtered observation data are subjected to cycle slip detection for continuity detection to obtain the cycle slip detection result; the filtered observation data are repaired according to the cycle slip detection result to obtain the repaired observation data; variance component calculation and weight allocation are performed on the repaired observation data to obtain the preprocessed Beidou observation data.

[0030] Among them, the MEMS sensor is used to obtain the buoy motion characteristic information. The MEMS sensor includes a triaxial accelerometer and a triaxial gyroscope, and the buoy attitude is calculated by measuring the acceleration and angular velocity data. The acceleration data reflect the linear acceleration changes of the buoy in three directions, and the angular velocity data characterize the rotation of the buoy around three axes. The attitude calculation uses the quaternion algorithm to fuse the acceleration and angular velocity data to obtain the pitch angle, roll angle and heading angle information of the buoy. For the Beidou raw observation data, applying wavelet transform for multi-scale decomposition is the key step to remove the multipath effect. The db4 wavelet basis is selected for wavelet transform, and the observation data are decomposed into multiple layers, and the signal is decomposed into different frequency components. The low-frequency part reflects the main trend of the signal, and the high-frequency part contains the multipath effect and noise information. The high-frequency coefficients are screened by the threshold processing method to extract the multipath characteristic components.

[0031] The combined processing of the buoy motion characteristic information and the multipath characteristic components uses adaptive filtering technology. A buoy motion state space model is established, with the attitude information provided by the MEMS sensor as the state quantity and the multipath characteristic components as the observation quantity. The adaptive filter dynamically adjusts the filter coefficients according to the signal characteristics to suppress the multipath effect. The periodic characteristics of the sea wave motion are considered during the filtering process, and different filtering parameters are used for different frequency interferences.

[0032] Performing cycle slip detection on the filtered observation data is an important link to ensure data continuity. The cycle slip detection uses a comprehensive detection method of multiple criteria, including the ionospheric residual method, the Doppler integration method and the phase combination method. The detection result contains the data continuity identifier for each epoch, and the identifier value reflects whether a cycle slip occurs and the magnitude of the cycle slip. For the detected cycle slip points, a method combining interpolation and extrapolation between epochs is used for repair. For small cycle slips, linear interpolation repair is performed using the carrier phase observation values before and after the cycle slip; for large cycle slips, the cycle slip is repaired in combination with the pseudorange observation values. The influence of the buoy motion on the carrier phase is considered during the repair process to ensure the physical rationality of the repair result.

[0033] The last step is to calculate the variance components of the observed data and assign weights, and the following mathematical model is adopted for this process:

[0034]

[0035] Where: represents the variance of the overall observed value; represents the variance component caused by the multipath effect; represents the variance component caused by the platform movement; represents the variance component of the observation noise; α1, α2, α3 are the weight coefficients of each component.

[0036] The weight assignment adopts the following formula:

[0037]

[0038] Where: W i represents the weight of the i-th observed value; β i represents the quality factor of the observed value; γ i represents the weight decay coefficient; represents the variance of the observed value; n is the total number of observed values.

[0039] Taking the processing of buoy observation data at a certain time as an example, the attitude data collected by the MEMS sensor shows that the buoy swings periodically under the action of sea waves. The wavelet transform decomposes the original observed data into four layers, and the multipath characteristics are identified from the high-frequency coefficients. The adaptive filter dynamically adjusts the filtering parameters according to the buoy motion period to suppress the multipath effect. During the cycle slip detection process, it is found that the carrier phase observed values have discontinuous points, and the continuity of the observed values is restored through the data repair algorithm. Through the calculation of variance components and weight assignment, the observed data is reasonably weighted to generate reliable preprocessed Beidou observation data.

[0040] S130: Based on the preprocessed Beidou observation data, calculate and process the wet and dry delay components to obtain the tropospheric delay estimation value;

[0041] Optionally, combine the elevation angle information in the preprocessed Beidou observation data with the meteorological parameters to obtain the zenith delay calculation parameters; separate the dry delay component through the zenith delay calculation parameters to obtain the zenith dry delay value; correct the zenith dry delay value according to the sea level pressure and temperature data to obtain the corrected zenith dry delay value; separate the wet delay component with the corrected zenith dry delay value as the reference quantity to obtain the zenith wet delay value; perform tilt conversion on the zenith dry delay value and the zenith wet delay value through spherical harmonic function mapping to obtain the tilt delay value; perform regional correction on the tilt delay value according to the marine atmospheric profile data to obtain the tropospheric delay estimation value.

[0042] Among them, the elevation angle information is extracted from the preprocessed Beidou observation data, and combined with real-time meteorological parameters, including air pressure, temperature, and relative humidity, to form the basic data for zenith delay calculation. During the acquisition process of zenith delay calculation parameters, the particularity of the marine atmospheric environment is taken into account, and parameters such as sea surface height, air pressure, and temperature are used as key inputs. In the separation process of the zenith dry delay component, an improved Saastamoinen model is used for calculation. This model considers the variation laws of atmospheric pressure and temperature with height, and combines the actual meteorological parameters at the location of the buoy to preliminarily estimate the zenith dry delay value. In this process, the atmospheric pressure data directly affects the calculation result of the dry delay.

[0043] The correction process of the zenith dry delay value is the key link to improve the accuracy. The sea-level air pressure data is obtained, and the real-time air pressure observation value at the buoy location is used, and reduced in combination with the standard atmospheric pressure height change model. The acquisition of temperature data includes two parts: sea surface temperature and vertical temperature gradient, which are collected in real time by temperature detection equipment. The correction process considers the particularity of the air-sea interface and makes a refined correction to the dry delay value. The separation process of the wet delay component is based on the corrected zenith dry delay value. By analyzing the vertical distribution characteristics of water vapor content in the atmosphere, a wet delay calculation model is established. This process makes full use of the sea surface relative humidity data and the vertical distribution law of water vapor to quantitatively estimate the wet delay. Among them, the vertical water vapor distribution model is constructed based on the characteristics of the marine atmospheric boundary layer.

[0044] The conversion from zenith delay to slant delay adopts the spherical harmonic function mapping method. The mapping function considers the earth's curvature and the variation of atmospheric refractive index with height, and converts the dry delay value and wet delay value in the zenith direction into the delay values in the satellite line-of-sight direction. The spherical harmonic function expansion uses 8th-order expansion coefficients to ensure the mapping accuracy. The regional correction link uses the marine atmospheric profile data for refined processing. The marine atmospheric profile data contains the vertical distribution information of temperature, humidity, and air pressure, and the correction coefficient is calculated by the numerical integration method to correct the slant delay value. This process pays special attention to the spatial variation characteristics of atmospheric parameters in the marine environment.

[0045] Taking the data processing of a certain observation period as an example: the satellite elevation angle information is extracted from the Beidou observation data, and at the same time, the air pressure, temperature, and relative humidity data observed by the buoy meteorological station are obtained. These parameters are input into the zenith delay calculation model to obtain the initial zenith delay value. Subsequently, according to the sea-level air pressure observation value and temperature gradient data, the zenith dry delay is corrected. In the process of wet delay separation, the sea surface water vapor observation data and the vertical distribution model are used to calculate the zenith wet delay value. The spherical harmonic function mapping converts the zenith delay into the slant delay in different satellite directions, and finally, through the regional correction using the marine atmospheric profile data, the tropospheric delay estimation value is obtained.

[0046] S140: Based on the estimated tropospheric delay, use the trained neural network to model the delay-water vapor conversion relationship to obtain a water vapor parameter inversion model;

[0047] Optionally, match the estimated tropospheric delay with multi-year regional sounding data to obtain delay-water vapor corresponding samples; extract the temperature and humidity data of the marine boundary layer from the delay-water vapor corresponding samples to obtain vertical distribution characteristic quantities; combine the vertical distribution characteristic quantities with the delay-water vapor corresponding samples to obtain neural network input features; classify and label the neural network input features according to seasons and time to obtain training samples with time-varying features; optimize the network structure of the training samples with time-varying features through cross-validation to obtain the network topology structure; train the network topology structure with the training samples with time-varying features to obtain a water vapor parameter inversion model.

[0048] Among them, for the matching process of delay-water vapor corresponding samples, the estimated tropospheric delay is matched with long-term regional sounding data. A spatio-temporal correspondence relationship is established during the matching process, and meteorological element data with similar spatio-temporal proximity to buoy observations is selected from the sounding data to form delay-water vapor corresponding samples. Extracting the marine boundary layer data from the delay-water vapor corresponding samples is the basis for constructing an accurate water vapor inversion model. The marine boundary layer refers to the atmosphere from the sea surface to a height of about 2000 meters. The temperature and humidity distribution within this layer directly affect the accuracy of water vapor inversion. During the extraction process, the temperature and humidity data of different height layers are processed in layers, and the average value, standard deviation, and vertical gradient of each layer are calculated to obtain vertical distribution characteristic quantities.

[0049] The combination of the vertical distribution characteristic quantities and the delay-water vapor corresponding samples constitutes the neural network input features. During the combination process, the vertical distribution characteristic quantities are used as auxiliary variables, enhancing the model's ability to represent the marine boundary layer structure. The input features include multi-dimensional information such as tropospheric delay values, vertical temperature distribution, and vertical humidity distribution. The seasonal and time classification and labeling of the neural network input features consider the time-varying characteristics of atmospheric water vapor. The labeling process is classified according to the four seasons of spring, summer, autumn, and winter, and the 24 hours of a day are divided into several time periods corresponding to different atmospheric states. Each data sample is labeled with seasonal and time information to form training samples with time-varying features.

[0050] Cross-validation is used to optimize the network structure for training samples with time-varying features. The K-fold cross-validation method is adopted to divide the dataset into a training set and a validation set. Through multiple trainings and validations, the performance of different network structures is evaluated. The network structure parameters include the number of hidden layers, the number of neurons in each layer, the type of activation function, etc. The optimization process determines the optimal network topology by comparing indicators such as mean square error and correlation coefficient. After determining the network topology, the neural network is trained. The training process uses the backpropagation algorithm to iteratively optimize and adjust the network weights and bias parameters. In each iteration, the error between the network output and the true value is calculated using the training samples with time-varying features, and the network parameters are updated through the gradient descent method.

[0051] Taking the buoy observations in a certain sea area as an example, radiosonde data for the past five years were collected and spatiotemporally matched with the tropospheric delay estimates. Temperature and humidity data every 100 meters in the height range of 0 - 2000 meters were extracted from the matched data, and the vertical distribution characteristics were calculated. These characteristics were combined with the delay values and classified according to seasons and observation times. Through cross-validation, a neural network structure with 3 hidden layers was determined, with 16, 32, and 16 neurons in each layer respectively. After multiple rounds of training, the network parameters gradually converged, forming a water vapor parameter inversion model that can accurately reflect the delay-water vapor conversion relationship.

[0052] S150: For the preprocessed Beidou observation data, the short message overlapping frequency interference is identified and suppressed through the adaptive notch filtering technique to obtain the interference-removed observation data;

[0053] Optionally, the preprocessed Beidou observation data is used to extract interference characteristics through carrier-to-noise ratio monitoring to obtain the interference characteristic signal; the center frequency of the adaptive notch filter is adjusted according to the interference characteristic signal to obtain the notch filter parameters; the preprocessed Beidou observation data is filtered using the notch filter parameters to obtain the preliminary filtered data; the tracking loop bandwidth of the preliminary filtered data is adjusted to obtain the bandwidth adjustment parameters; the correlation interval is optimized using the bandwidth adjustment parameters to obtain the correlator output data; the orbit and clock offset of the correlator output data are predicted through Kalman filtering to obtain the interference-removed observation data.

[0054] Among them, carrier-to-noise ratio (CNR) monitoring is the basis for interference feature extraction. By analyzing the CNR of the preprocessed Beidou observation data, the ratio of signal power to noise power is calculated. When there are abnormal changes in the CNR, it indicates that the signal is affected by multi-frequency interference. The CNR monitoring uses the sliding window method to continuously evaluate the CNR of each observation epoch and identify the interference feature signals. The adjustment of the center frequency of the adaptive notch filter is the core of interference suppression. According to the frequency characteristics of the interference feature signals, the center frequency of the notch filter is dynamically adjusted. The parameters of the notch filter include the center frequency, bandwidth, and attenuation depth. The adjustment of the center frequency is based on the main frequency analysis of the interference signal, the bandwidth is set according to the spectral width of the interference signal, and the attenuation depth is determined according to the interference intensity.

[0055] In the filtering process, the notch filter parameters are applied to the preprocessed Beidou observation data. The filter selectively attenuates the interference frequency band in the frequency domain while retaining the useful signal components. The interference in the preliminarily filtered data is effectively suppressed, but further parameter optimization is required. The adjustment of the tracking loop bandwidth is an important parameter optimization step. The bandwidth of the preliminarily filtered data is adaptively adjusted, and the bandwidth adjustment parameters change with the dynamic characteristics of the signal. When the signal is interfered, the tracking loop bandwidth is appropriately reduced to enhance the anti-interference ability; when the signal returns to normal, the standard bandwidth setting is restored to maintain the tracking performance.

[0056] The optimization of the correlator spacing is a key step in signal tracking. According to the bandwidth adjustment parameters, the spacing of the correlators is dynamically adjusted. The optimization of the correlator spacing directly affects the measurement accuracy of the code phase and carrier phase. The optimization process takes into account factors such as signal strength, multipath effect, and dynamic characteristics. The correlator output data contains the optimized code phase and carrier phase measurement values. The prediction of the orbit and clock offset by Kalman filtering is the final data processing step. The correlator output data is input into the Kalman filter to establish the state equation and observation equation, and predict the satellite orbit parameters and receiver clock offset. The prediction results are used to correct the observation data and compensate for the errors caused by interference.

[0057] Taking a buoy observation as an example, through CNR monitoring, it is found that there is multi-frequency interference in the short message of the Beidou observation data. The interference feature signal is manifested as a periodic decrease in the CNR. The center frequency of the interference is determined through spectral analysis. The adaptive notch filter adjusts its parameters according to this frequency characteristic, achieving directional suppression of the interference signal. After filtering, the tracking loop bandwidth is temporarily reduced to enhance the anti-interference performance, and at the same time, the correlator spacing is optimized to maintain the tracking accuracy. Finally, the orbit and clock offset parameters are predicted and corrected through Kalman filtering, and reliable observation data is obtained.

[0058] S160: Input the observation data after interference removal into the water vapor parameter inversion model, and obtain the estimated values of water vapor parameters including precipitable water and relative humidity through real-time calculation.

[0059] Optionally, update the delay amount of the observation data after interference removal through a tropospheric delay calculator to obtain the real-time delay value; perform current meteorological parameter correction processing on the real-time delay value to obtain the corrected delay value; input the corrected delay value into the water vapor parameter inversion model to obtain the preliminary water vapor parameters; extract the precipitation component from the preliminary water vapor parameters to obtain the precipitable water; extract the humidity component from the preliminary water vapor parameters to obtain the relative humidity; compare and verify the precipitable water and relative humidity with the data of the adjacent radiosonde station to obtain the estimated values of water vapor parameters.

[0060] Among them, the tropospheric delay calculator first receives the observation data after interference removal and updates the delay amount in real time based on the dynamic atmosphere model. The update process of the delay amount takes into account the time-varying characteristics of atmospheric parameters, and the generated real-time delay value reflects the current atmospheric state. The current meteorological parameter correction is the key to improving the inversion accuracy. When performing correction processing on the real-time delay value, make full use of the meteorological data collected by the buoy platform, including air pressure, temperature, and relative humidity. The correction process takes into account the particularity of meteorological parameters in the ocean environment, especially paying attention to the changes in meteorological parameters at the air-sea interface. The corrected delay value more accurately reflects the atmospheric water vapor content. When the corrected delay value is input into the water vapor parameter inversion model, the model will perform parameter estimation by combining historical training data. The preliminary water vapor parameters contain multiple characterization quantities of the water vapor content in the atmosphere, and specific parameters need to be further decomposed and extracted.

[0061] The precipitation component extraction process specifically processes the precipitation information in the preliminary water vapor parameters. The precipitable water represents the vertical integral value of the water vapor content in the atmospheric column, and its extraction process is based on the vertical distribution characteristics of water vapor density. The extraction of the precipitation component fully considers the humidity structure of the ocean boundary layer. The humidity component extraction focuses on the calculation of air relative humidity. When extracting the relative humidity from the preliminary water vapor parameters, the influences of temperature and air pressure need to be considered. The calculation of relative humidity is based on the ratio of the actual water vapor pressure to the saturated water vapor pressure, which reflects the saturation degree of the water vapor content in the air.

[0062] In the comparison and verification link, compare the extracted precipitable water and relative humidity with the data of the adjacent radiosonde station. The radiosonde data is used as the reference standard, and the closest sounding data is selected for comparison through space-time matching. The influences of spatial differences and time delays are considered during the comparison process, and the estimated values of water vapor parameters are obtained through reasonable weight allocation.

[0063] Taking a buoy observation as an example, after the interference-removed observed data is input into the tropospheric delay calculator, a real-time delay value reflecting the current atmospheric state is obtained. The delay value is corrected by combining the air pressure and temperature data measured by the buoy meteorological station, and then the corrected delay value is input into the inversion model. The preliminary water vapor parameters output by the model are subjected to component extraction to obtain the precipitable water vapor characterizing the atmospheric water vapor content and the relative humidity reflecting the humidity condition. These parameters are compared and verified with the sounding data of the neighboring sounding station, and through spatio-temporal matching and weight assignment, an accurate estimated value of the water vapor parameters is obtained.

[0064] Figure 2 FIG. shows a block diagram of a Beidou water vapor inversion system 200 based on a buoy platform according to an embodiment of the present disclosure. As Figure 2 shown, the system 200 includes:

[0065] A receiving module 210, configured to receive and process the Beidou satellite dual-frequency signal through a direct antenna to obtain Beidou original observed data including pseudorange observations, carrier phase observations, and satellite orbit parameters;

[0066] A suppression module 220, configured to suppress multipath effects and platform jitter through wavelet transform and adaptive filtering according to the Beidou original observed data and in combination with the buoy motion characteristic information provided by the MEMS sensor, to obtain preprocessed Beidou observed data;

[0067] A calculation module 230, configured to calculate and process the wet and dry delay components based on the preprocessed Beidou observed data to obtain a tropospheric delay estimate;

[0068] A modeling module 240, configured to model the delay-water vapor conversion relationship by using a trained neural network according to the tropospheric delay estimate to obtain a water vapor parameter inversion model;

[0069] An identification module 250, configured to identify and suppress the short message overlapping frequency interference through an adaptive notch filtering technique for the preprocessed Beidou observed data to obtain interference-removed observed data;

[0070] An input module 260, configured to input the interference-removed observed data into the water vapor parameter inversion model to obtain an estimated value of water vapor parameters including precipitable water vapor and relative humidity through real-time calculation. Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the described modules can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0071] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A Beidou water vapor inversion method based on a buoy platform, characterized in that Including: Receiving and processing the Beidou satellite dual-frequency signal through a direct antenna to obtain the original Beidou observation data including pseudorange observation values, carrier phase observation values, and satellite orbit parameters; According to the original Beidou observation data and combined with the buoy motion characteristic information provided by the MEMS sensor, suppressing multipath effects and platform shaking through wavelet transform and adaptive filtering to obtain the preprocessed Beidou observation data; Based on the preprocessed Beidou observation data, calculating and processing the wet and dry delay components to obtain the tropospheric delay estimation value; According to the tropospheric delay estimation value, using the trained neural network to model the delay-water vapor conversion relationship to obtain the water vapor parameter inversion model; For the preprocessed Beidou observation data, identifying and suppressing the short message overlapping frequency interference through the adaptive notch filtering technology to obtain the observation data after removing interference; Inputting the observation data after removing interference into the water vapor parameter inversion model, and obtaining the water vapor parameter estimation values including precipitable water and relative humidity through real-time calculation.

2. The Beidou water vapor inversion method based on a buoy platform according to claim 1, wherein The step of receiving and processing the Beidou satellite dual-frequency signal through a direct antenna to obtain the original Beidou observation data including pseudorange observation values, carrier phase observation values, and satellite orbit parameters includes: Selecting the Beidou satellite B1I / B3I dual-frequency signal through elevation angle setting to obtain the dual-frequency signals of 6-12 visible Beidou satellites; Inputting the Beidou satellite dual-frequency signal into the carrier phase smoothed pseudorange processing unit to obtain the pseudorange observation value and the carrier phase observation value; Combining the pseudorange observation value and the carrier phase observation value with the ephemeris data, and obtaining the satellite orbit parameters through satellite orbit solution processing; Combining and arranging the pseudorange observation value, the carrier phase observation value, and the satellite orbit parameters according to the RINEX format standard to obtain the original Beidou observation data.

3. The Beidou water vapor inversion method based on a buoy platform according to claim 1, wherein The step of, according to the original Beidou observation data and combined with the buoy motion characteristic information provided by the MEMS sensor, suppressing multipath effects and platform shaking through wavelet transform and adaptive filtering to obtain the preprocessed Beidou observation data includes: Obtaining the buoy motion characteristic information by performing attitude solution on the acceleration and angular velocity data obtained from the MEMS sensor; Extracting the noise characteristics of the original Beidou observation data through wavelet multi-scale decomposition to obtain the multipath characteristic components; Combining the buoy motion characteristic information and the multipath characteristic components for adaptive filtering processing to obtain the filtered observation data; Performing continuity detection on the filtered observation data through cycle slip detection to obtain the cycle slip detection result; Repairing the filtered observation data according to the cycle slip detection result to obtain the repaired observation data; Calculating the variance component and weight distribution for the repaired observation data to obtain the preprocessed Beidou observation data.

4. The Beidou water vapor inversion method based on a buoy platform according to claim 1, characterized in that The step of, based on the preprocessed Beidou observation data, calculating and processing the wet and dry delay components to obtain the tropospheric delay estimation value includes: Combine the elevation angle information in the preprocessed Beidou observation data with meteorological parameters to obtain zenith delay calculation parameters; Perform separation processing on the dry delay component through the zenith delay calculation parameters to obtain the zenith dry delay value; Perform correction processing on the zenith dry delay value according to sea-level air pressure and temperature data to obtain the corrected zenith dry delay value; Perform separation processing on the wet delay component with the corrected zenith dry delay value as the reference quantity to obtain the zenith wet delay value; Perform tilt conversion on the zenith dry delay value and the zenith wet delay value through spherical harmonic function mapping to obtain the tilt delay value; Perform regional correction on the tilt delay value according to the marine atmospheric profile data to obtain the tropospheric delay estimation value.

5. The Beidou water vapor inversion method based on a buoy platform according to claim 1, wherein According to the tropospheric delay estimation value, use the trained neural network to model the delay-water vapor conversion relationship to obtain the water vapor parameter inversion model, including: Match the tropospheric delay estimation value with multi-year regional sounding data to obtain delay-water vapor corresponding samples; Extract the temperature and humidity data of the marine boundary layer from the delay-water vapor corresponding samples to obtain vertical distribution characteristic quantities; Combine the vertical distribution characteristic quantities with the delay-water vapor corresponding samples to obtain neural network input features; Classify and label the neural network input features according to season and time to obtain training samples with time-varying features; Optimize the network structure of the training samples with time-varying features through cross-validation to obtain the network topology structure; Perform training processing on the network topology structure and the training samples with time-varying features to obtain the water vapor parameter inversion model.

6. The Beidou water vapor inversion method based on a buoy platform according to claim 1, wherein, For the preprocessed Beidou observation data, identify and suppress the short message overlapping frequency interference through the adaptive notch filtering technology to obtain the interference-removed observation data, including: Extract interference characteristic signals from the preprocessed Beidou observation data through carrier-to-noise ratio monitoring to obtain interference characteristic signals; Adjust the center frequency of the adaptive notch filter according to the interference characteristic signals to obtain notch filter parameters; Perform filtering processing on the preprocessed Beidou observation data through the notch filter parameters to obtain preliminary filtered data; Adjust the tracking loop bandwidth of the preliminary filtered data to obtain bandwidth adjustment parameters; Optimize the correlator interval with the bandwidth adjustment parameters to obtain correlator output data; Perform orbit and clock error prediction on the correlator output data through Kalman filtering to obtain the interference-removed observation data.

7. The Beidou water vapor inversion method based on a buoy platform according to claim 1, characterized in that Input the interference-removed observation data into the water vapor parameter inversion model, and obtain the water vapor parameter estimation values including precipitable water and relative humidity through real-time calculation, including: Update the delay amount of the interference-removed observation data through a tropospheric delay calculator to obtain the real-time delay value; Perform current meteorological parameter correction processing on the real-time delay value to obtain the corrected delay value; Input the corrected delay value into the water vapor parameter inversion model to obtain preliminary water vapor parameters; Extract the precipitation component from the preliminary water vapor parameters to obtain the precipitable water; Extract the humidity component from the preliminary water vapor parameters to obtain the relative humidity; Compare and verify the precipitable water and the relative humidity with the data of the adjacent radiosonde station to obtain the estimated value of the water vapor parameters.

8. A Beidou water vapor inversion system based on a buoy platform, which is used to implement the Beidou water vapor inversion method based on a buoy platform as described in any one of claims 1-7, characterized in that, The Beidou water vapor inversion system based on the buoy platform includes: A receiving module, configured to receive and process the Beidou satellite dual-frequency signal through a direct antenna to obtain the original Beidou observation data including pseudorange observations, carrier phase observations, and satellite orbit parameters; A suppression module, configured to suppress multipath effects and platform jitter through wavelet transform and adaptive filtering according to the original Beidou observation data and in combination with the buoy motion characteristic information provided by the MEMS sensor to obtain the preprocessed Beidou observation data; A calculation module, configured to calculate the wet and dry delay components based on the preprocessed Beidou observation data to obtain the tropospheric delay estimate; A modeling module, configured to model the delay-water vapor conversion relationship using the trained neural network according to the tropospheric delay estimate to obtain a water vapor parameter inversion model; An identification module, configured to identify and suppress the short message overlapping frequency interference for the preprocessed Beidou observation data through an adaptive notch filtering technique to obtain the observation data after removing the interference; An input module, configured to input the observation data after removing the interference into the water vapor parameter inversion model to obtain the estimated value of the water vapor parameters including the precipitable water and the relative humidity through real-time calculation.

Images