Multi-parameter real-time resolving method fusing GNSS direct and reflected signals

Abstract

The invention relates to the technical field of GNSS (Global Navigation Satellite System) signal processing and remote sensing inversion, and discloses a multi-parameter real-time resolving method integrating GNSS direct and reflected signals, which comprises the following steps of: analyzing the asymmetry of correlation peaks of direct signals of a high-elevation satellite to distinguish local multipath pollution, and limiting to only select unpolluted direct signals to resolve an atmospheric parameter model; according to the method, a correlation peak self-checking step is introduced, the defect that a reference source on which atmospheric calculation depends is prone to local multipath pollution is overcome, and the purity of the atmospheric parameter model is guaranteed.

Description

Technical Field

[0001] This invention relates to a multi-parameter real-time solution method that integrates direct and reflected GNSS signals, belonging to the field of GNSS signal processing and remote sensing inversion technology. Background Technology

[0002] In the current field of GNSS (Global Navigation Satellite System) remote sensing of surface parameters (such as soil moisture and snow depth), a conventional approach is to compare the direct signal received from a zenith antenna with the reflected signal received from a nadir antenna, analyzing the differences in phase, time delay, or power to infer the physical properties of the surface medium. However, with the increasing demands for accuracy and stability in applications such as precision agriculture, hydrological monitoring, and geological disaster early warning, a fundamental premise upon which the above conventional measurement logic relies—that the direct signal received by the zenith antenna is a stable and pure reference—is beginning to reveal its limitations in the dynamic engineering environment. In fact, as the direct signal passes through the troposphere and ionosphere, its signal characteristics, especially its phase, are severely affected by dynamically changing atmospheric parameters, such as atmospheric water vapor content or the total electron content of the ionosphere, resulting in propagation delays. This means that the reference signal itself has been contaminated.

[0003] The inherent flaws in this measurement logic are amplified dramatically in specific scenarios with drastic changes in meteorological conditions (such as before a rainstorm), leading to systemic measurement failures: a sharp increase in atmospheric water vapor content directly contaminates the direct signals from all satellites (i.e., the reference source); subsequently, existing inversion methods cannot distinguish between signal changes originating from the reference source itself and signal changes from surface reflections; ultimately, the system incorrectly attributes the contamination of the reference source to changes in surface parameters, resulting in false alarms about changes in surface parameters even when the physical conditions of the ground (such as soil moisture) may not have changed. This demonstrates that the measurement logic uses a dynamically changing signal as a stable reference, and its processing flow has an inherent flaw at its logical starting point. Currently, some technical solutions attempt to use complex signal processing algorithms to distinguish... While this method separates the direct reference signal from the reflected signal, its technical approach still has limitations and fails to address the fundamental defects of reference source contamination and parameter coupling. For example, Chinese invention patent CN116338000A discloses a method for measuring the acoustic reflection coefficient of materials in a sound tube. Although this method employs complex algorithms such as Bayesian and least squares methods to process the mixed signal at the receiving end to separate the direct and reflected waves, its technical logic focuses on signal separation rather than parameter decoupling. When applied to the field of GNSS remote sensing, this approach cannot solve the fundamental problem that the direct reference signal itself is severely contaminated by the atmosphere. The separated direct wave is still a contaminated reference, and using it for comparison and inversion will inevitably lead to the incorrect attribution of atmospheric disturbances to changes in surface parameters, thus failing to avoid the systemic risks of the aforementioned defects.

[0004] Therefore, the technical problem to be solved by this invention is how to construct a signal processing flow that avoids using contaminated direct signals as static references at the signal processing method level, so that atmospheric parameters can be solved and separated before surface parameters are retrieved, and atmospheric parameters and surface parameters can be solved as two independent parameters. Summary of the Invention

[0005] This invention provides a multi-parameter real-time solution method that integrates GNSS direct and reflected signals. Its main purpose is to solve the problem in the prior art that when polluted direct signals are used as static references, the interference of atmospheric parameters cannot be separated before inverting surface parameters, and it is difficult to solve atmospheric parameters and surface parameters as independent parameters.

[0006] To achieve the above objectives, this invention provides a multi-parameter real-time solution method that integrates direct and reflected GNSS signals, comprising: Step 101, the establishment and verification of the atmospheric parameter model, involves acquiring the direct signals of GNSS satellites within a set high elevation angle range; calculating an asymmetry metric value for the correlation peak corresponding to the direct signals of GNSS satellites within the set high elevation angle range in the receiver correlator, the asymmetry metric value being used to characterize the symmetry of the correlation peak; determining whether the asymmetry metric value is lower than a preset first threshold, and limiting the selection to only direct signals from GNSS satellites within the set high elevation angle range whose asymmetry metric value is lower than the preset first threshold, and based on their observation values, calculating the atmospheric parameter model characterizing the atmospheric state; Step 102, signal purification step: Select GNSS satellites within the set low elevation angle range, and based on the atmospheric parameter model calculated in step 101, predict and remove the atmospheric propagation effects contained in the direct and reflected signals of the low elevation angle satellites to obtain pure direct reference signals and pure reflected signals. Step 103, surface feature isolation step, by comparing the pure reflected signal with the pure direct reference signal, isolates the surface reflection features introduced only by surface reflection; Step 104, Surface parameter inversion step, based on surface reflection characteristics, at least one surface physical parameter is obtained by inversion.

[0007] Preferably, in step 103, the surface reflection characteristics are continuously acquired as a time series, and step 104 includes: step 201, performing time-domain analysis on the time series of surface reflection characteristics to separate a low-frequency component characterizing the slow time-varying characteristics of the surface medium and a high-frequency component characterizing the fast time-varying characteristics of the surface medium; and step 202, calculating at least two different surface physical parameters corresponding to the slow time-varying characteristics and the fast time-varying characteristics, respectively, based on the low-frequency component and the high-frequency component.

[0008] Preferably, in step 101, the atmospheric parameter model calculation includes: based on the carrier phase observation values ​​of the direct signals of GNSS satellites within a set high elevation angle range, where the asymmetry metric value is lower than a preset first threshold, the total tropospheric zenith delay parameter is calculated.

[0009] Preferably, in step 101, the atmospheric parameter model calculation includes: calculating the total electron content parameter of the ionosphere based on the carrier phase observation value of the direct signal of the GNSS satellite within a set high elevation angle range, where the asymmetry metric value is lower than a preset first threshold.

[0010] Preferably, in step 101, calculating the asymmetry metric includes: obtaining the early correlator power and the late correlator power corresponding to the correlation peak, and calculating the normalized difference between the early correlator power and the late correlator power as the asymmetry metric.

[0011] Preferably, before solving the atmospheric parameter model, step 101 further includes: step 401, real-time monitoring of the availability of GNSS satellites within a set high elevation angle range, the availability being determined based on the number of GNSS satellites and geometric distribution factor within the set high elevation angle range; and determining whether the availability meets a preset confidence threshold; and step 402, if the availability does not meet the preset confidence threshold, pausing the solution based on the real-time direct signal, and instead using a backup atmospheric parameter model generated by time extrapolation based on historical solution results to replace the atmospheric parameter model in step 102.

[0012] Preferably, the method further includes: outputting the atmospheric parameter model obtained in step 101 as an independent atmospheric monitoring parameter.

[0013] Preferably, in step 102, obtaining the pure direct reference signal and the pure reflected signal includes: subtracting the corresponding atmospheric delay predicted based on the atmospheric parameter model and the geometric path of the low-elevation satellite from the original received signals of the direct and reflected signals of the low-elevation satellite, respectively.

[0014] Preferably, in step 202, at least two different surface physical parameters include: at least one parameter calculated based on the low-frequency component, selected from the group consisting of: soil moisture parameter, snow cover parameter, and vegetation parameter calculated based on the high-frequency component.

[0015] Preferably, in step 201, the time-domain analysis includes: using a low-pass filter to obtain the time series of surface reflection characteristics to obtain the low-frequency component, and using a high-pass filter to obtain the time series of surface reflection characteristics to obtain the high-frequency component.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. This method establishes a sequential solution processing flow, which uses the direct signal from high-elevation satellites to solve the atmospheric state and uses this atmospheric state as a known quantity to correct the direct and reflected signals from low-elevation satellites. It processes the reconstruction of the time sequence and decomposes the coupling attribution problem of atmospheric changes and surface changes in signals in traditional measurements into two independent and sequential solution steps. This avoids systematic misjudgment caused by atmospheric disturbances such as water vapor changes being incorrectly attributed to surface parameters such as soil moisture changes.

[0017] 2. Before calculating the atmospheric parameters, a quality self-verification step for high-elevation angle direct signals is introduced. By analyzing the morphological characteristics of the signal correlation peaks, signals contaminated by local multipath pollution can be identified and eliminated. This pre-verification mechanism ensures the purity of the reference source used to calculate atmospheric parameters, thereby guaranteeing the accuracy of subsequent signal purification operations using the model. This allows the method to maintain the stability and reliability of the calculation logic even in actual deployment environments with near-field reflections.

[0018] 3. Further analysis of the time series of pure surface reflectance characteristics was conducted, utilizing the differences in the physical characteristics of different surface parameters in response to time scales. For example, the slow time-varying components in the signal time series were assigned to relatively stable surface medium parameters such as soil moisture, while the fast time-varying components such as fluctuations or variances were assigned to surface cover parameters affected by disturbances such as vegetation dynamics. This processing method enables the separation and inversion of multiple different surface physical parameters from a mixed reflectance signal, thereby enhancing the information dimension of surface remote sensing. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the sequential process for multi-parameter solution of the present invention; Figure 2 This is a schematic diagram illustrating the self-verification of the asymmetry of the relevant peaks in this invention; Figure 3 This is a logic diagram for satellite availability diagnosis and mode switching in this invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention.

[0021] The multi-parameter real-time solution method for fusion of direct and reflected GNSS signals disclosed in this invention is based on a sequential decoupled processing flow. This flow performs an atmospheric parameter model establishment and verification step at the system level to obtain a high-reliability atmospheric state reference. Using this reference, a signal purification step corrects the low-elevation angle observation signal, and a surface feature isolation step extracts the clean surface response. Finally, a surface parameter inversion step solves the target physical parameters, thus logically separating the coupling effects of atmospheric disturbances and surface parameters. In the actual deployment of the system, the high-elevation angle signal of the zenith antenna used to receive the direct signal serves as the reference source for atmospheric parameter solution. The signal is susceptible to local multipath contamination from the near-field or surrounding environment of the equipment, such as support rods and buildings. If this contamination is not identified, it will directly interfere with the calculation benchmark of the atmospheric parameter model, leading to systematic deviations in subsequent purification steps. To address this issue, in step 101, this method performs a quality self-check on the direct signal from high-elevation satellites. Specifically, before calculating atmospheric parameters, the system accesses the underlying output of the correlator inside the GNSS receiver. Ideally, the correlation peak of a pure GNSS signal should present a symmetrical triangle, but the introduction of local multipath inevitably leads to distortion of the correlation peak shape. This method utilizes this characteristic by obtaining the early correlator power corresponding to the correlation peak (…). ) and late correlator power ( ), and calculate a normalized asymmetric metric, which can be achieved using quantified in the form of, where This is used to characterize the symmetry of the relevant peaks; then, the system will... With a preset first threshold ( The first threshold is compared. It is not a fixed empirical value, but rather a value determined through offline calibration of the GNSS receiver in an open, unobstructed environment with low multipath propagation. This involves collecting 24 hours of clean signal data and calculating its... The statistical distribution is calculated, and its mean is taken plus three times the standard deviation. The result is used as the first threshold of the receiver hardware; during the solution process, step 101 is limited to selecting only those results. The absolute value is below the first threshold The pure, high-elevation satellite direct signal is obtained, while all signals judged to be contaminated by local multipath are dynamically removed from the solution pool, thus eliminating the influence of local multipath contamination signals at the solution source.

[0022] In certain extreme conditions, such as deep mountain valleys or urban high-rise environments, high-elevation satellites may be blocked over large areas or for extended periods, resulting in an insufficient number of clean high-elevation satellites available for calculation. This makes it impossible to reliably calculate the atmospheric parameter model in real time, thus posing a risk of interruption to the entire sequential logic chain. To address this signal blockage situation, this method introduces a pre-diagnostic mechanism before step 101. The system monitors the availability of GNSS satellites within the set high-elevation range in real time. This availability is determined by comprehensively considering the number of satellites that have passed the aforementioned asymmetric verification and the geometric distribution factor (PDOP) of these satellites. The system then compares this availability with a preset confidence threshold, which can be specifically defined as having at least 4 available satellites and a PDOP value no greater than 6. If the availability meets the confidence threshold, the system is in real-time calculation mode, executes step 101 normally, and stores the calculation result as a high-confidence anchor point. If the availability does not meet the confidence threshold, the system automatically switches to time-domain hold-up mode, suspends the calculation based on the real-time direct signal, and adopts a backup atmospheric parameter model. This backup model is based on... The system switches to the latest high-confidence anchor point stored before switching to this mode, and combines it with a time propagation model, such as a constant-preserving or linear extrapolation model. The generated backup model replaces the real-time atmospheric parameter model to perform subsequent steps 102, ensuring the continuity of the core calculation process under harsh signal environments. In real-time calculation mode, after the system has screened out clean high-elevation angle direct signals that meet the first threshold, the core calculation process of step 101 is started. This process can reuse known precision point positioning (PPP) or related residual analysis techniques. One implementation... The method involves calculating the tropospheric zenith total delay (ZTD) parameter generated when the direct signal from a clean, high-elevation satellite passes through a neutral atmosphere, based on carrier phase observations. This ZTD value constitutes the atmospheric parameter model. In another application scenario, such as when the ionosphere is active, this calculation step can be replaced by or supplemented with calculating the ionospheric total electron content (TEC) parameter. Therefore, this method also includes a parallel output step, which outputs the atmospheric parameter model as an independent atmospheric monitoring parameter, thus enabling a single device to have the dual functions of atmospheric remote sensing and surface remote sensing.

[0023] After successfully acquiring the atmospheric parameter model characterizing the atmospheric state, such as the calculated ZTD value, in step 101, the system enters the signal purification stage in step 102. The purpose of this stage is to address the core dilemma: low-elevation satellite signals used for surface measurements are also contaminated by atmospheric propagation effects similar to those of high-elevation signals. In step 102, the system selects GNSS satellites within a set low-elevation range. A typical elevation range for observation can be... to Based on the atmospheric parameter model calculated in step 101, and combined with the real-time geometric position of the low-elevation satellite, the system uses a standard tropospheric projection function to positively predict the direct atmospheric signal to the low-elevation satellite. The precise atmospheric delay caused ( Similarly, its reflected signal was also predicted. The corresponding atmospheric delay generated on its different geometric paths ( Finally, the system performs a purification operation, subtracting the predicted atmospheric delay from the original received signals of both the direct and reflected signals from the low-elevation satellite. and This process yields a clean direct reference signal and a clean reflected signal at the algorithm level, eliminating the common atmospheric propagation effects of both signals. In step 102, the tropospheric projection function is used to calculate the atmospheric delay along a specific elevation path. This projection function can be a well-known empirical model, such as the Vienna Mapping Function (VMF) or the Global Mapping Function (GMF). In a specific engineering implementation, the processor within the GNSS receiver can store the grid coefficients or their mathematical expressions for the VMF model. During purification, the processor uses the receiver's own geodetic coordinates and the geometric elevation angle of the current low-elevation satellite as inputs into the VMF model to calculate the projection coefficient values ​​corresponding to the direct and reflected paths, thereby obtaining the projection delay of the ZTD value along the two different paths. In the subsequent step 103, the system performs surface feature isolation, i.e., by comparing the clean reflected signal (… ) and pure direct reference signal ( One approach is to calculate the difference between the two; since the common air pollution has been purified, the difference... It will only include phase difference, time delay difference, or power variation introduced by surface reflection; this difference This refers to the surface reflection characteristics.

[0024] In traditional inversion step 104, the surface reflection characteristics are directly used as the basis for the inversion. Inverting single surface parameters, but in real-world scenarios with vegetation cover, this... The signal is actually the dielectric constant of the surface soil, which is a slowly time-varying parameter, and the surface vegetation layer disturbed by wind, which is a rapidly time-varying parameter. The superposition of these two leads to distortion of the single soil moisture inversion result. To address this, this method provides an enhanced implementation of step 104, in which the surface reflectance characteristics are... As a time series The data is acquired continuously; then, in step 201, the system... A simple and efficient way to perform time-domain analysis on time series data is to use digital filters: a low-pass filter with a cutoff frequency set to 0.01Hz is used for processing. The time series is used to obtain its low-frequency components characterizing the slow time-varying properties of the surface medium; simultaneously, it is processed through a high-pass filter with a cutoff frequency set to 0.1Hz. The system calculates the variance of the time series data within a short time window to obtain its high-frequency components characterizing the rapid time-varying properties of the surface medium. Finally, in step 202, the system solves for various surface parameters based on the separated components. Specifically, it calculates at least one parameter selected from the following groups based on the low-frequency components and the corresponding surface inversion model: soil moisture parameters and snow cover parameters. It also calculates vegetation parameters based on the high-frequency components and another inversion model. For the surface physical parameter inversion described in steps 104 and 202, the inversion model can be established after equipment deployment through a field calibration procedure. This procedure includes: deploying at least one auxiliary measurement unit within the monitoring range of the GNSS receiving equipment to provide ground truth values, such as a time-domain reflectometer (TDR) probe for measuring soil volumetric water content or a probe for characterizing vegetation biomass. The method employs a quantitative optical observation device; within a continuous observation period, such as 30 days, it simultaneously acquires the true readings of the auxiliary measurement unit and the low-frequency and high-frequency component time series output in step 201 of this method; finally, it performs regression analysis in the processor, for example, performing polynomial fitting between the TDR soil moisture true series and the low-frequency component series to determine the model coefficients used for inverting soil moisture, and simultaneously performing exponential fitting between the vegetation index true series and the statistical variance of the high-frequency component series to determine the model coefficients used for inverting vegetation parameters; all the above method steps can be executed on a single GNSS receiving device, which uses an antenna facing the zenith to receive direct signals and an antenna facing the nadir to receive reflected signals. This method reconstructs the signal processing flow and utilizes existing hardware to achieve multi-parameter calculation.

[0025] Example 1: In a specific deployment of a monitoring station for flood early warning in a key watershed, the equipment is installed in a canyon environment with complex signals. Its zenith antenna is not only interfered with by the near-field equipment support pole, but also adjacent to a steep rock face. This inevitably leads to the receiving high-elevation-angle satellite direct signals being contaminated by local multipath interference. Under typical meteorological conditions before a rainstorm, the local atmospheric water vapor content in the area begins to rise sharply, causing a change in the tropospheric zenith total delay parameter (ZTD). This condition makes it difficult for traditional GNSS-R methods to distinguish whether the signal change originates from atmospheric (ZTD), local multipath, or from a real change in surface soil moisture. In step 101 of this invention, all satellite direct signals within a set high elevation angle range are acquired, and the receiver correlator is accessed to calculate the correlation peak asymmetry metric of each signal in real time. The system processor detected a high-elevation satellite signal from the side adjacent to the rock face. The value reached 0.31, while another set of satellite signals, reflected near the support rod, had a value of 0.31. The value also reached 0.18, both of which are higher than the first threshold preset by the system through offline calibration. (Set to 0.10).

[0026] Accordingly, the solution logic in step 101 is triggered. The system dynamically removes the two sets of satellite signals contaminated by local multipath contamination from the atmospheric parameter solution pool, and only selects the remaining clean direct signals that have passed the verification and are not contaminated. Based on their carrier phase observations, a high-confidence atmospheric parameter model is calculated, which reflects that the local ZTD value is rising rapidly. This high-confidence atmospheric parameter model provides the basis for subsequent decoupling. Then, in step 102, the system uses this ZTD model as a known quantity to positively predict and remove all atmospheric propagation effects contained in the direct and reflected signals of satellites within the set low elevation angle range, thereby obtaining the clean direct reference signal and the clean reflected reference signal. The system first transmits the signal; then, in step 103, by comparing the two purified signals, it isolates the surface reflection characteristics introduced solely by surface reflection; finally, in step 104, the system performs inversion based on these surface reflection characteristics and simultaneously outputs two sets of decoupled, independent physical parameters: one is the atmospheric monitoring parameter (ZTD value) reflecting drastic changes in water vapor from step 101, and the other is the surface physical parameter (soil moisture value) that remains stable from step 104 before the rainstorm reaches the ground; the sequential decoupling and self-verification process of this method avoids false alarms caused by atmospheric disturbances or local multipath being incorrectly attributed to changes in surface parameters in this complex interference scenario.

[0027] Example 2: In a controlled field test environment, a GNSS receiving system including zenith and nadir antennas was deployed above a large, homogeneous soil test bed. A high-precision time-domain reflectometer (TDR) array was deployed within the test bed to provide continuous ground truth values ​​of surface soil moisture. The experiment compared the method of this invention with the complete process of steps 101 to 104, examining the stability of the retrieved surface parameters when two key remote sensing interferences—local multipath pollution and drastic atmospheric changes—occurred simultaneously. The experiment design included three parallel data processing groups: Control group A used the traditional method, directly comparing the original low-elevation direct signal and reflected signal; Control group B used the sequential decoupling process of this invention, but disabled the correlation peak asymmetry verification function in step 101, using all high-elevation signals to build an atmospheric model; The sample group of this invention used the complete method of this invention, including the correlation peak asymmetry verification in step 101. Before the experiment, the first threshold of the receiver was determined through offline calibration. The value was 0.12. At T=0, the TDR array readings of the soil test bed showed an average volumetric water content of 22.3%, and the atmospheric conditions were stable with no obvious local multipath sources. The experiment was conducted according to the following time points: at T=30 minutes, a large metal reflector was introduced near the zenith antenna to create local multipath pollution; at T=60 minutes, while keeping the local multipath pollution constant, a water vapor generator was used to cause an increase in atmospheric water vapor content above the monitoring station, leading to a sharp change in ZTD. The TDR array monitored and confirmed throughout the process that the physical water content of the soil test bed remained at 22.3%. Within a stable range of 0.5%, each treatment group continuously outputs its retrieved soil moisture values, and the data records at key time points are shown in Table 1.

[0028] Table 1: Comparison of Soil Moisture (Volume Water Content %) Retrieved by Each Treatment Group under Combined Disturbance Data analysis is shown in Table 1. At T=45 minutes, when only local multipath is introduced, control group A is basically unaffected because its inversion does not depend on high elevation angle signals; control group B, because the verification in step 101 is disabled, will... The inclusion of a polluted high-elevation angle signal with a value as high as 0.45 in the calculation resulted in contamination of the established atmospheric parameter model, causing the retrieved humidity to jump to 30.2%. The sample group of this invention, through asymmetric verification in step 101, identified and removed this polluted signal, and its established atmospheric parameter model remained uncontaminated, with the retrieved humidity stabilizing at 22.7%. At T=75 minutes, when a dramatic change in atmospheric water vapor occurred, the original reference signal of control group A was contaminated by ZTD changes, and the system attributed the atmospheric change to surface changes, causing the retrieved humidity to jump to 3. The soil moisture content was 8.6%, which deviated significantly from the true value. Control group B, based on a model already polluted by local multipath sources, further amplified the error by processing the dramatic changes in atmospheric water vapor, resulting in a rise in the retrieved humidity to 46.5%. In contrast, the sample group of this invention, having excluded the polluted signal in step 101, had an unpolluted atmospheric parameter model, enabling it to accurately predict and eliminate the atmospheric propagation effect caused by the dramatic changes in ZTD in step 102. The soil moisture content retrieved in step 104 was 22.8%, consistent with the true ground value of 22.3%.

[0029] Example 3: To further clarify the necessity of steps 101 and 102 in the sequential decoupling process of the present invention, the following comparative example is set. This comparative example adopts a conventional surface parameter inversion method in the art. This method does not include the establishment and verification step of the atmospheric parameter model in step 101 of the present invention, nor does it perform the signal purification step in step 102. Instead, it directly performs steps 103 and 104, that is, directly compares the original direct signal and the original reflected signal of the received low-elevation satellite to isolate the surface reflection characteristics and invert the surface parameters. This method corresponds to the processing method used by control group A in Example 2. This comparative example adopts the same method as Example 2. The test platform, test conditions, and ground true data were used to invert the soil moisture results, which are listed in column A of the control group in Table 1. The data shows that in the initial stable phase (T=15min), the inversion result of 22.7% was basically consistent with the ground true value of 22.4%. When local multipath was introduced at T=45min, since this conventional method does not rely on high elevation angle signals, the inversion result of 22.8% was not affected by this specific multipath source. However, at T=75min, when the atmospheric water vapor drastic change ZTD rapidly increased, this conventional method exposed its technical defects. This method incorporated the atmospheric ZTD changes caused by the addition of low elevation angle direct reference signals (…). The propagation delay on the surface was incorrectly attributed to changes in soil moisture, leading to severe distortion in the inversion results, which jumped to 38.6%, deviating from the true ground value of 22.3%. This caused a systemic false alarm. The results of this comparison confirm that conventional techniques lacking priority calculation of atmospheric parameters and signal purification steps are ineffective when faced with rapid changes in atmospheric conditions commonly seen in remote sensing. Because the processing logic uses direct signals polluted by the atmosphere as a measurement reference, it cannot separate the coupling effect between atmospheric parameters and surface parameters, leading to misjudgment of surface parameters.

[0030] Example 4: This example combines Figures 1 to 3 This paper describes a multi-parameter real-time solution method that integrates direct and reflected GNSS signals, such as... Figure 1 As shown, it uses a zenith antenna to acquire high-elevation direct signals and a nadir antenna to acquire low-elevation direct / reflected signals. The high-elevation direct signals are sent to steps 101, 401, and 402 to establish and verify an atmospheric parameter model. The results of this model are written into the D1 database as an atmospheric parameter model. The low-elevation direct / reflected signals are sent to step 102 for purification. Step 102 reads the atmospheric parameter model from the D1 database to process the signals. The purified pure direct reference signal and pure reflected signal are sent to step 102. 03. Surface feature isolation is performed. The isolated surface reflectance features are written into D2: Surface Reflectance Feature Time Series. This time series is read by step 201 to perform time domain analysis. The time domain analysis separates the signal into high-frequency components and low-frequency components. Both are sent to steps 104 and 202 for surface physical parameter inversion. The finally inverted surface physical parameters, such as slow time-varying soil moisture and snow cover, and fast time-varying surface physical parameters, such as vegetation parameters, can be provided to E2: External System / User along with atmospheric monitoring parameters from D1.

[0031] like Figure 2 As shown, the horizontal axis represents time in minutes, and the vertical axis represents asymmetric measurements. In the diagram, a dotted line indicates the preset first threshold value, for example... =0.10, a solid line represents a pure signal whose metric value is always below this first threshold, while a dashed line represents a contaminated signal whose metric value is above this first threshold during the period from the 2nd to the 18th minute, reaching a peak around the 7th minute. The method of this invention identifies whether a signal is contaminated based on whether this metric value is below the first threshold; for example... Figure 3As shown, this logic is used to deal with the situation where satellite availability is insufficient at high elevation angles. After the system starts, it enters the satellite availability monitoring state, which is determined based on the number of satellites and PDOP. If the availability is determined to meet the confidence threshold, the system enters the real-time calculation mode, executes steps 101-104 to calculate atmospheric and surface parameters, and returns to the satellite availability monitoring state after one calculation cycle is completed. If the availability is determined not to meet the confidence threshold, the system enters the time domain hold mode. At this time, the real-time calculation will be paused and a backup atmospheric model will be used. After one hold cycle is completed and a retry is performed, the system returns to the satellite availability monitoring state.

[0032] Example 5: The preset first threshold used in step 101 of the method of the present invention ( Its calibration requires a reproducible procedure, and the threshold may need adjustment during long-term system operation due to radome aging or slow changes in the near-field environment. The method of the present invention may include a method for... The system employs an online adaptive update procedure, which is periodically executed by the processor within the GNSS receiver: the system continuously caches a long-term window, such as a 48-hour sliding window, containing all high-elevation satellites acquired within that window (set to have an elevation angle greater than...). Asymmetric measure of ) Historical data; the system updates and calculates according to a preset cycle, such as every 24 hours, and stores all data within the 48-hour sliding window. The data undergoes statistical histogram analysis to identify the main lobe of its probability density distribution. This main lobe reflects the statistical baseline of the signal morphology under the current receiver hardware and deployment environment. Furthermore, the system calculates the mean of this main lobe data after statistically eliminating 5% of extreme outliers. and standard deviation Finally, based on the statistical results, the system automatically recalculates and updates the first threshold required in step 101. It can be set to ,in This is a confidence coefficient, which can be 3 or 4; through this online adaptive procedure, the first threshold... It can automatically track the slow drift of receiver status and near-field environment, ensuring the effectiveness of the correlation peak self-verification step in step 101 during long-term operation.

[0033] Furthermore, when the method of the present invention performs time-domain analysis in step 201 to separate low-frequency and high-frequency components, the cutoff frequencies of the low-pass and high-pass filters on which it relies require an engineering calibration standard. This calibration procedure can be determined through a one-time field calibration experiment after equipment deployment: select two representative test points, test point one being bare soil (without vegetation cover) and test point two being an area with typical vegetation cover (such as grassland); secondly, perform a complete wet-dry cycle at test point one (bare soil), such as irrigation followed by natural drying, for 48 hours, while continuously acquiring the time series of its surface reflectance characteristics using the method of the present invention (steps 101-103). Simultaneously, at test site two (vegetated area), a relatively stable soil moisture period, such as 24 hours, was selected, and under conditions of natural wind disturbance, the time series of its surface reflectance characteristics were continuously acquired. Next, the system processes the acquired data... and Power spectral density (PSD) analysis was performed on the two time series. The PSD spectrum is used to determine the upper frequency limit where more than 95% of the signal energy is concentrated, denoted as . A typical observation value can be 0.02 Hz; through analysis From the PSD spectrum, determine the lower frequency limit at which the signal energy begins to appear, denoted as . A typical observation value can be 0.08 Hz; finally, the system selects a value located at... and Frequency values ​​between, such as Hz, serving as the cutoff frequency for the low-pass and high-pass filters in step 201; this calibration procedure provides a physical observation basis for the selection of filter cutoff frequencies, ensuring that the time-domain analysis in step 201 can match the physical response characteristics of the local land cover.

[0034] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0035] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A multi-parameter real-time solution method integrating direct and reflected GNSS signals, characterized in that, include: Step 101, the establishment and verification of the atmospheric parameter model, involves obtaining the direct signal from GNSS satellites within the set high elevation angle range; The system calculates an asymmetry metric for the correlation peak of the direct signal from GNSS satellites within a set high elevation angle range in the receiver correlator. The asymmetry metric is used to characterize the symmetry of the correlation peak. It determines whether the asymmetry metric is lower than a preset first threshold and limits the selection to direct signals from GNSS satellites within the set high elevation angle range whose asymmetry metric is lower than the preset first threshold. Based on their observations, an atmospheric parameter model characterizing the atmospheric state is calculated. Step 102, signal purification step: Select GNSS satellites within the set low elevation angle range, and based on the atmospheric parameter model calculated in step 101, predict and remove the atmospheric propagation effects contained in the direct and reflected signals of the low elevation angle satellites to obtain pure direct reference signals and pure reflected signals. Step 103, surface feature isolation step, by comparing the pure reflected signal with the pure direct reference signal, isolates the surface reflection features introduced only by surface reflection; Step 104, Surface parameter inversion step, based on surface reflection characteristics, at least one surface physical parameter is obtained by inversion.

2. The multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 1, characterized in that, In step 103, the surface reflection characteristics are continuously acquired as a time series, and step 104 includes: step 201, performing time-domain analysis on the time series of surface reflection characteristics to separate a low-frequency component characterizing the slow time-varying characteristics of the surface medium and a high-frequency component characterizing the fast time-varying characteristics of the surface medium; and step 202, calculating at least two different surface physical parameters corresponding to the slow time-varying characteristics and the fast time-varying characteristics, respectively, based on the low-frequency component and the high-frequency component.

3. The multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 1, characterized in that, In step 101, the atmospheric parameter model is solved by: based on the carrier phase observation value of the direct signal of the GNSS satellite within the set high elevation angle range, where the asymmetry metric value is lower than the preset first threshold, the total tropospheric zenith delay parameter is calculated.

4. The multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 1, characterized in that, In step 101, the atmospheric parameter model is solved by: based on the carrier phase observation value of the direct signal of the GNSS satellite within the set high elevation angle range, the asymmetry metric value of which is lower than the preset first threshold, the total electron content parameter of the ionosphere is calculated.

5. The multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 1, characterized in that, In step 101, calculating the asymmetry metric includes: obtaining the early correlator power and late correlator power corresponding to the correlation peak, and calculating the normalized difference between the early correlator power and the late correlator power as the asymmetry metric.

6. The multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 1, characterized in that, Before solving the atmospheric parameter model, step 101 also includes: step 401, real-time monitoring of the availability of GNSS satellites within a set high elevation angle range, the availability being determined based on the number of GNSS satellites and geometric distribution factor within the set high elevation angle range; and determining whether the availability meets a preset confidence threshold; and step 402, if the availability does not meet the preset confidence threshold, then pausing the solution based on the real-time direct signal, and instead using a backup atmospheric parameter model generated by time extrapolation based on historical solution results to replace the atmospheric parameter model in step 102.

7. The multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 1, characterized in that, The method also includes: outputting the atmospheric parameter model obtained in step 101 as an independent atmospheric monitoring parameter.

8. The multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 1, characterized in that, In step 102, obtaining the pure direct reference signal and the pure reflected signal includes subtracting the corresponding atmospheric delay predicted based on the atmospheric parameter model and the geometric path of the low-elevation satellite from the original received signals of the direct and reflected signals of the low-elevation satellite, respectively.

9. A multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 2, characterized in that, In step 202, at least two different surface physical parameters include: at least one parameter calculated based on the low-frequency component, selected from the group consisting of: soil moisture parameter, snow cover parameter, and vegetation parameter calculated based on the high-frequency component.

10. A multi-parameter real-time solution method for fusing direct and reflected GNSS signals according to claim 2, characterized in that, In step 201, the time-domain analysis includes: using a low-pass filter to obtain the time series of surface reflection characteristics to obtain the low-frequency component, and using a high-pass filter to obtain the time series of surface reflection characteristics to obtain the high-frequency component.

Images