Beidou pp-rtk train full-scene positioning method, system and terminal

Abstract

A Beidou PPP-RTK train full-scene positioning method, comprising the following steps: when the train is in an open or semi-sheltered scene, a train positioning terminal combines Beidou PPP-RTK with initialized INS to position the train and obtain train position information; when the train enters a full-sheltered scene, the train positioning terminal receives a UWB base station signal, then constrains the height of the train in advance, and then establishes a height-constrained UWB model, and then obtains the position information of the train through the height-constrained UWB model and the initialized INS; when the train leaves the full-sheltered scene, the train positioning terminal receives PPP-RTK correction information and a UWB base station signal, establishes a height-constrained PPP-RTK model, obtains the positioning result of the PPP-RTK, and thus obtains the position information of the train. The train positioning accuracy is high in a complex scene.

Description

Beidou PPP-RTK train full-scene positioning method, system and terminal Technical Field

[0001] This invention relates to an improvement of BeiDou train positioning technology, belonging to the field of train positioning, and particularly to a BeiDou PPP-RTK train full-scene positioning method, system and terminal. Background Technology

[0002] Precise positioning of high-speed trains is a crucial foundation for ensuring their safe operation. The BeiDou Navigation Satellite System not only provides real-time, high-precision positioning services but also boasts high autonomy and low cost, driving the transformation of train positioning from automatic to autonomous positioning technologies. However, traditional BeiDou / GNSS satellite positioning technology, pseudorange single-point positioning (SPP), has relatively low accuracy and cannot meet the requirements for lateral positioning accuracy between different tracks. Real-Time Kinematic (RTK) positioning can achieve centimeter-level accuracy but relies on dense reference stations, requiring precise control of the distance between the train and the reference stations. Precise point positioning (PPP) reduces reliance on reference stations and achieves centimeter-level accuracy, but its long initialization time makes it difficult to quickly obtain high-precision positioning information when trains enter or exit tunnels or are obstructed by canyons or buildings. Precise point positioning-real-time (PPI) is a more advanced technology. Kinematic (PPP-RTK) relies on a sparse network of stations spanning nearly 100 kilometers to achieve instantaneous centimeter-level positioning accuracy in open, unobstructed environments, making it particularly suitable for large-scale, long-distance precision positioning in railway scenarios.

[0003] Trains operate in complex environments. In canyons, forests, and other similar locations, satellite signals are blocked, causing a sharp decline in positioning accuracy and reliability. In completely obscured environments such as tunnels, satellite positioning becomes completely ineffective. Relying solely on satellite positioning cannot meet the continuous high-precision positioning requirements of trains across all scenarios. Inertial Navigation Systems (INS) are independent of external references and autonomously provide a variety of high-precision navigation parameters, which can effectively address the shortcomings of BeiDou satellite positioning in obscured areas. However, in long tunnel environments, INS accumulates navigation errors over time and cannot provide high-precision, high-reliability positioning results.

[0004] Chinese patent application CN202211424844.X, filed on November 14, 2022, discloses a train positioning optimization method, system, and rail transit vehicle. This method corrects raw satellite observation data to obtain corrected satellite observation data. Train position information is calculated using the corrected satellite observation data and RTK differential data transmitted in real-time from a ground base station system. While this technology corrects raw satellite observation data and uses the corrected satellite observation data for positioning calculation, it does not solve the problem of insufficient train positioning accuracy in complex scenarios.

[0005] The information disclosed in this background section is intended only to enhance the understanding of the overall background of this patent application and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to overcome the problem of insufficient train positioning accuracy in complex scenarios in the prior art, and to provide a Beidou PPP-RTK train all-scenario positioning method, system and terminal with high train positioning accuracy in complex scenarios.

[0007] To achieve the above objectives, the technical solution of the present invention is: a BeiDou PPP-RTK train full-scene positioning method, which includes the following steps:

[0008] Step 1: The base station network first acquires raw GNSS data and raw UWB data, then sends the raw GNSS data to the computing center, and at the same time sends the raw UWB data to the train positioning terminal.

[0009] Step 2: The computing center first processes the raw GNSS data to obtain PPP-RTK correction information, and then sends the PPP-RTK correction information to the train positioning terminal.

[0010] Step 3: The train positioning terminal first performs INS zero bias and attitude initialization while stationary to obtain the corrected INS. Then, the PPP-RTK correction information is combined with the UWB raw data for filtering. The filtered data is then input into the corrected INS to obtain the initialized INS. The initialized INS is then used to locate the train to obtain the initial position of the train.

[0011] Step 4: When the train enters the running state, the positioning method for the train includes any one or more of the following methods in any combination, as detailed below:

[0012] The first method: When the train is in an open or semi-obscured environment, the train positioning terminal combines Beidou PPP-RTK with the initialized INS to track the train's position. Then, based on the train's initial position, it recursively calculates and corrects to obtain the train's real-time position information.

[0013] The second method: When the train enters a fully shielded scene, the train positioning terminal first receives the UWB base station signal in the raw UWB data, then determines the train's prior elevation, establishes an elevation-constrained UWB model, and then tracks the train's position using the elevation-constrained UWB model and the initialized INS. Finally, combined with the train's initial position, the real-time position information of the train is obtained through recursion and correction.

[0014] Step four also includes a third scenario, as follows:

[0015] When the train leaves the fully shielded scene, the train positioning terminal first receives PPP-RTK correction information and UWB base station signals from the raw UWB data, establishes an elevation-constrained PPP-RTK model, obtains the PPP-RTK positioning results, and then inputs the PPP-RTK positioning results into the INS for updating, performs position tracking of the train, and then combines the initial position of the train to recursively obtain the real-time position information of the train.

[0016] In step four, after the train positioning terminal receives the UWB base station signal, it further includes receiving the UWB base station elevation information. k Elevation H of UWB base station from the orbital plane k .

[0017] The specific prior elevation of the train is as follows:

[0018] U = U k -H k +H0;

[0019] Where H0 is the elevation of the combined positioning terminal antenna installed on the train, and U is taken as the pseudo-observation value h. The elevation above the Earth's ellipsoid is h < 10km, satisfying the following relationship:

[0020]

[0021] Where a is the semi-major axis of the ellipsoid, b is the semi-minor axis of the ellipsoid, h is the elevation of this point, and (x, y, z) are the geocentric coordinates of this point in the Earth-fixed coordinate system. Taking the partial derivatives of h with respect to x, y, and z respectively, we get:

[0022]

[0023] Therefore, the differential dh is:

[0024]

[0025] The elevation-constrained UWB model in step four is specifically as follows:

[0026]

[0027] Where, σ r Standard deviation This represents the standard deviation of the pseudo-observed elevation values.

[0028] The elevation-constrained PPP-RTK model in step four is specifically as follows:

[0029] The network receives GNSS pseudorange and phase observation data and constructs the original observation equations:

[0030]

[0031] Where E(·) denotes the expectation operator, and These represent pseudorange and phase observations, respectively, whose station-to-satellite distances and tropospheric interference delays have been corrected beforehand. c represents the speed of light, s represents the satellite number, r represents the receiver number, j represents the frequency band number, i represents the epoch number, and dt represents the phase distance. r and dt s It is the receiver and satellite clock bias, τ r For the wet delay of the zenith troposphere, τ r The projection function is Indicates ionospheric delay, coefficient λ is the ratio between the ionospheric delay at frequency j and the first frequency. j For the wavelength corresponding to frequency j, d r，j and δ represents the pseudorange hardware delay at frequency j between receiver r and satellite s. r，j and This represents the phase deviation between receiver r and satellite s at corresponding frequency j. The integer ambiguity corresponding to frequency j for receiver r and satellite s;

[0032] Using the reference station p at the network end as the baseline station, a weighted pseudo-observation equation for the ionosphere at the network end is constructed:

[0033]

[0034] By using the S-basis rank deficiency elimination method, the rank deficiency of both the original and pseudo-observation equations is eliminated, and a new observation equation is constructed:

[0035]

[0036] The specific form of the new estimable parameter is as follows:

[0037]

[0038] in,

[0039]

[0040] in, For the new, estimable receiver clock bias, For satellite clock bias, For the zenith tropospheric wet delay, For ionospheric delay, For receiver pseudorange hardware delay, For satellite pseudorange hardware delay, For receiver phase deviation, For satellite phase deviation, For ambiguity, This represents the hardware delay deviation between receiver r and receiver p.

[0041] The user terminal constructs the observation equation based on the received pseudorange and phase observations and the correction information broadcast by the network:

[0042]

[0043] Where the subscript u represents the user terminal receiver, To use the ionospheric delay values ​​near user stations interpolated based on the network-end ionospheric delay information as pseudo-observations, Let Δx be the unit vector from the station to the satellite. u (i) represents the correction value for the coordinates. For receiver clock bias, The tropospheric zenith wet delay is denoted as , and is the ambiguity parameter. For pseudorange hardware delay, The pseudorange is a geometrically independent combination of hardware delay values. For ionospheric slack delay, For receiver phase deviation, This refers to the receiver phase deviation;

[0044] When entering or exiting tunnels or other fully shielded areas, elevation constraints can be applied based on the received pseudo-elevation values ​​Un:

[0045]

[0046] The corresponding stochastic model is:

[0047]

[0048] Where, σ p =0.3m, σ φ =0.003m, σ I =0.3m, d n This is the ranging value from the terminal to the UWB base station n.

[0049] The PPP-RTK uses a sparsely distributed ground-based reference station network to calculate and generate a precise correction information product for satellite bias and atmospheric delay. This product is then broadcast to users via satellite communication / terrestrial networks, ultimately achieving rapid and fixed positioning with ambiguity.

[0050] A positioning system for a BeiDou PPP-RTK train full-scene positioning method, the positioning system of the BeiDou PPP-RTK train full-scene positioning method includes a base station network, a computing center and a train positioning terminal;

[0051] The reference station network includes a BeiDou / GNSS PPP-RTK reference station network and a UWB tunnel / platform base station network;

[0052] The BeiDou / GNSS PPP-RTK reference station network is deployed along the railway line with an average station spacing of 100km, receiving and sending BeiDou satellite data to the computing center;

[0053] The UWB tunnel / platform base station network provides UWB base station information and ranging information to the train terminal;

[0054] The computing center refers to the PPP-RTK network computing center, which collects BeiDou satellite data observed by the PPP-RTK reference station network, and calculates and broadcasts correction information for satellite phase deviation, satellite clock error, ionospheric delay, and tropospheric delay in real time.

[0055] The train positioning terminal is used to receive BeiDou satellite observation data, BeiDou correction information broadcast by the computing center, UWB ranging data, and INS observation data generated by itself. Through an adaptive fusion positioning algorithm, it calculates the train position in real time.

[0056] A positioning terminal for a BeiDou PPP-RTK train full-scene positioning method, the positioning terminal of the BeiDou PPP-RTK train full-scene positioning method includes a GNSS module, an INS module, a UWB module, a signal control module, a communication module and a core processing module;

[0057] The GNSS module is used to receive raw BeiDou / GNSS observation data and send 1Hz pps second pulses to the signal control module, providing GPS or BDS time to the core processing module for time calibration.

[0058] The INS module is used to generate high-frequency triaxial angular velocity and triaxial acceleration raw observation data;

[0059] The UWB module is used to receive ranging information from the UWB base station and achieves time synchronization through the core processing module.

[0060] The signal control module is used to provide different frequency signals to the inertial navigation system in order to generate INS data of the corresponding frequency.

[0061] The communication module is used to transmit data via 4G / 5G / WiFi, receive PPP-RTK correction information, and output real-time combined positioning results or raw observation data from each sensor.

[0062] The core processing module is used to receive raw observation data from each sensor, synchronize the time, integrate PPP-RTK / INS / UWB adaptive fusion algorithm, and calculate the positioning results in real time.

[0063] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0064] 1. This invention discloses a BeiDou PPP-RTK train all-scenario positioning method, system, and terminal. It employs an adaptive fusion positioning scheme in different scenarios, enabling phased matching and switching between BeiDou PPP-B2b-constrained PPP-RTK technology, PPP-RTK / INS combined technology, and UWB / INS combined technology. This achieves decimeter-level positioning accuracy across all scenarios, while also possessing advantages such as high efficiency and low cost. Therefore, this invention provides high train positioning accuracy in complex scenarios.

[0065] 2. In the BeiDou PPP-RTK train full-scene positioning method, system, and terminal of this invention, the terminal positioning result and other GNSS parameters to be estimated are obtained after filtering. Based on prior track elevation constraints, the PPP-RTK convergence speed is accelerated, and the positioning accuracy is improved. Higher accuracy can be obtained even when the train exits a tunnel under conditions with fewer satellite observations. Therefore, this invention has a faster convergence speed and higher positioning accuracy.

[0066] 3. In the BeiDou PPP-RTK train all-scene positioning method, system, and terminal of this invention, the traditional method can have an elevation deviation of up to 1m, while in this solution, the elevation direction accuracy is significantly improved. Based on the prior track elevation constraint, the elevation deviation is reduced to within 0.5m. The elevation accuracy in fully obscured scenarios such as tunnels is improved by 90%, the PPP-RTK ambiguity fixing speed is accelerated when exiting tunnels, the ambiguity fixing success rate is improved, and the positioning accuracy is more accurate. Therefore, this invention has a smaller elevation deviation and more accurate positioning. Attached Figure Description

[0067] Figure 1 shows the horizontal localization results of the PPP-RTK / INS combination in Experiments 1 and 2 of this invention.

[0068] Figure 2 shows the results of PPP-RTK / INS combined elevation positioning in Experiments 1 and 2 of this invention.

[0069] Figure 3 is a diagram of the PPP-RTK / INS three-dimensional positioning error in Experiments 1 and 2 of this invention.

[0070] Figure 4 shows the results of the PPP-RTK / INS combination + INS recursive horizontal localization in the third stage of the experiment in this invention.

[0071] Figure 5 shows the results of the PPP-RTK / INS combination + INS recursive elevation positioning in the third stage of the experiment in this invention.

[0072] Figure 6 is a diagram of the 3D positioning error of the PPP-RTK / INS combination + INS recursive method in the third stage of the experiment in this invention.

[0073] Figure 7 is a diagram showing the horizontal localization results of the UWB / INS combination in the fourth stage of the experiment in this invention.

[0074] Figure 8 shows the elevation positioning results of the UWB / INS combination in the fourth stage of the experiment in this invention.

[0075] Figure 9 is a diagram of the UWB / INS three-dimensional positioning error in the fourth stage of the experiment in this invention.

[0076] Figure 10 shows the positioning error of the entire experimental stage and scene combination in this invention.

[0077] Figure 11 is a map showing the distribution and experimental locations of PPP-RTK in this invention. Detailed Implementation

[0078] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0079] Referring to Figures 1 to 11, a BeiDou PPP-RTK train full-scene positioning method includes the following steps:

[0080] Step 1: The base station network first acquires raw GNSS data and raw UWB data, then sends the raw GNSS data to the computing center, and at the same time sends the raw UWB data to the train positioning terminal.

[0081] Step 2: The computing center first processes the raw GNSS data to obtain PPP-RTK correction information, and then sends the PPP-RTK correction information to the train positioning terminal.

[0082] Step 3: The train positioning terminal first performs INS zero bias and attitude initialization while stationary to obtain the corrected INS. Then, the PPP-RTK correction information is combined with the UWB raw data for filtering. The filtered data is then input into the corrected INS to obtain the initialized INS. The initialized INS is then used to locate the train to obtain the initial position of the train.

[0083] Step 4: When the train enters the running state, the positioning method for the train includes any one or more of the following methods in any combination, as detailed below:

[0084] The first method: When the train is in an open or semi-obscured environment, the train positioning terminal combines Beidou PPP-RTK with the initialized INS to track the train's position. Then, based on the train's initial position, it recursively calculates and corrects to obtain the train's real-time position information.

[0085] The second method: When the train enters a fully shielded scene, the train positioning terminal first receives the UWB base station signal in the raw UWB data, then determines the train's prior elevation, establishes an elevation-constrained UWB model, and then tracks the train's position using the elevation-constrained UWB model and the initialized INS. Finally, combined with the train's initial position, the real-time position information of the train is obtained through recursion and correction.

[0086] In step four, after the train positioning terminal receives the UWB base station signal, it further includes receiving the UWB base station elevation information. k Elevation H of UWB base station from the orbital plane k .

[0087] The specific prior elevation of the train is as follows:

[0088] U = U k -H k +H0;

[0089] Where H0 is the elevation of the combined positioning terminal antenna installed on the train, and U is used as the pseudo-observation value h. This approximate elevation value U is used as a pseudo-observation value to constrain the PPP-RTK solution outside the tunnel and the UWB solution inside the tunnel. This will improve the positioning speed and accuracy of PPP-RTK and significantly improve the accuracy of UWB elevation positioning. The elevation above the Earth's ellipsoid is h < 10km, satisfying the following relationship:

[0090]

[0091] Where a is the semi-major axis of the ellipsoid, b is the semi-minor axis of the ellipsoid, h is the elevation of this point, and (x, y, z) are the geocentric coordinates of this point in the Earth-fixed coordinate system. Taking the partial derivatives of h with respect to x, y, and z respectively, we get:

[0092]

[0093] Therefore, the differential dh is:

[0094]

[0095] The elevation-constrained UWB model in step five is specifically as follows:

[0096]

[0097] Where, σ r Standard deviation This represents the standard deviation of the pseudo-observed elevation values.

[0098] The elevation-constrained PPP-RTK model in step four is specifically as follows:

[0099] The network receives GNSS pseudorange and phase observation data and constructs the original observation equations:

[0100]

[0101] Where E(·) denotes the expectation operator, and These represent pseudorange and phase observations, respectively, whose station-to-satellite distances and tropospheric interference delays have been corrected beforehand. c represents the speed of light, s represents the satellite number, r represents the receiver number, j represents the frequency band number, i represents the epoch number, and dt represents the phase distance. r and dt s It is the receiver and satellite clock bias, τ r For the wet delay of the zenith troposphere, τ r The projection function is Indicates ionospheric delay, coefficient λ is the ratio between the ionospheric delay at frequency j and the first frequency. j For the wavelength corresponding to frequency j, d r,j and δ represents the pseudorange hardware delay at frequency j between receiver r and satellite s. r,j and This represents the phase deviation between receiver r and satellite s at corresponding frequency j. The integer ambiguity corresponding to frequency j for receiver r and satellite s;

[0102] Using the reference station p at the network end as the baseline station, a weighted pseudo-observation equation for the ionosphere at the network end is constructed:

[0103]

[0104] By using the S-basis rank deficiency elimination method, the rank deficiency of both the original and pseudo-observation equations is eliminated, and a new observation equation is constructed:

[0105]

[0106] The specific form of the new estimable parameter is as follows:

[0107]

[0108] in,

[0109]

[0110] d r，DCB =d r，2 -d r，1 ;

[0111] in, For the new, estimable receiver clock bias, For satellite clock bias, For the zenith tropospheric wet delay, For ionospheric delay, For receiver pseudorange hardware delay, For satellite pseudorange hardware delay, For receiver phase deviation, For satellite phase deviation, For ambiguity, This represents the hardware delay deviation between receiver r and receiver p.

[0112] The user terminal constructs the observation equation based on the received pseudorange and phase observations and the correction information broadcast by the network:

[0113]

[0114] Where the subscript u represents the user terminal receiver, To use the ionospheric delay values ​​near user stations interpolated based on the network-end ionospheric delay information as pseudo-observations, Let Δx be the unit vector from the station to the satellite. u (i) represents the correction value for the coordinates. For receiver clock bias, The tropospheric zenith wet delay is denoted as , and is the ambiguity parameter. For pseudorange hardware delay, The pseudorange is a geometrically independent combination of hardware delay values. For ionospheric slack delay, For receiver phase deviation, This refers to the receiver phase deviation;

[0115] When entering or exiting tunnels or other fully shielded areas, elevation constraints can be applied based on the received pseudo-elevation values ​​Un:

[0116]

[0117] The corresponding stochastic model is:

[0118]

[0119] Where, σ p =0.3m, σ φ =0.003m, σ I =0.3, d n This is the ranging value from the terminal to the UWB base station n.

[0120] The PPP-RTK uses a sparsely distributed ground-based reference station network to calculate and generate a precise correction information product for satellite bias and atmospheric delay. This product is then broadcast to users via satellite communication / terrestrial networks, ultimately achieving rapid and fixed positioning with ambiguity.

[0121] The process of filtering the PPP-RTK correction information and the original UWB data before inputting them into the corrected INS also includes filtering the UWB positioning data. The UWB positioning data is obtained by measuring the propagation time of the signal to calculate the distance between the UWB base station and the train positioning terminal, thereby determining the position of the train positioning terminal.

[0122] A positioning system for a BeiDou PPP-RTK train full-scene positioning method, the positioning system of the BeiDou PPP-RTK train full-scene positioning method includes a base station network, a computing center and a train positioning terminal;

[0123] The reference station network includes a BeiDou / GNSS PPP-RTK reference station network and a UWB tunnel / platform base station network;

[0124] The BeiDou / GNSS PPP-RTK reference station network is deployed along the railway line with an average station spacing of 100km, receiving and sending BeiDou satellite data to the computing center;

[0125] The UWB tunnel / platform base station network provides UWB base station information and ranging information to the train terminal;

[0126] The computing center refers to the PPP-RTK network computing center, which collects BeiDou satellite data observed by the PPP-RTK reference station network, and calculates and broadcasts correction information for satellite phase deviation, satellite clock error, ionospheric delay, and tropospheric delay in real time.

[0127] The train positioning terminal is used to receive BeiDou satellite observation data, BeiDou correction information broadcast by the computing center, UWB ranging data, and INS observation data generated by itself. Through an adaptive fusion positioning algorithm, it calculates the train position in real time.

[0128] A positioning terminal for a BeiDou PPP-RTK train full-scene positioning method, the positioning terminal of the BeiDou PPP-RTK train full-scene positioning method includes a GNSS module, an INS module, a UWB module, a signal control module, a communication module and a core processing module;

[0129] The GNSS module is used to receive raw BeiDou / GNSS observation data and send 1Hz pps second pulses to the signal control module, providing GPS or BDS time to the core processing module for time calibration.

[0130] The INS module is used to generate high-frequency triaxial angular velocity and triaxial acceleration raw observation data;

[0131] The UWB module is used to receive ranging information from the UWB base station and achieves time synchronization through the core processing module.

[0132] The signal control module is used to provide different frequency signals to the inertial navigation system in order to generate INS data of the corresponding frequency.

[0133] The communication module is used to transmit data via 4G / 5G / WiFi, receive PPP-RTK correction information, and output real-time combined positioning results or raw observation data from each sensor.

[0134] The core processing module is used to receive raw observation data from each sensor, synchronize the time, integrate PPP-RTK / INS / UWB adaptive fusion algorithm, and calculate the positioning results in real time.

[0135] The following are supplementary descriptions of the present invention:

[0136] The differential-non-polarity combined PPP-RTK technology can freely select the reference station network within the region. Compared with the traditional ionospheric de-polarity combined PPP-RTK technology that relies on the global reference station network, it is more suitable for building a nationwide high-precision satellite positioning system for trains that only uses BeiDou and is based on domestic benchmarks, which meets the needs of modern railway safety application services.

[0137] Example 1:

[0138] A BeiDou PPP-RTK train full-scene positioning method, comprising the following steps:

[0139] Step 1: The base station network first acquires raw GNSS data and raw UWB data, then sends the raw GNSS data to the computing center, and at the same time sends the raw UWB data to the train positioning terminal.

[0140] Step 2: The computing center first processes the raw GNSS data to obtain PPP-RTK correction information, and then sends the PPP-RTK correction information to the train positioning terminal.

[0141] Step 3: The train positioning terminal first performs INS zero bias and attitude initialization while stationary to obtain the corrected INS. Then, the PPP-RTK correction information is combined with the UWB raw data for filtering. The filtered data is then input into the corrected INS to obtain the initialized INS. The initialized INS is then used to locate the train to obtain the initial position of the train.

[0142] Step 4: When the train enters the running state, the positioning method for the train includes any one or more of the following methods in any combination, as detailed below:

[0143] The first method: When the train is in an open or semi-obscured environment, the train positioning terminal combines Beidou PPP-RTK with the initialized INS to track the train's position. Then, based on the train's initial position, it recursively calculates and corrects to obtain the train's real-time position information.

[0144] The second method: When the train enters a fully shielded scene, the train positioning terminal first receives the UWB base station signal in the raw UWB data, then determines the train's prior elevation, establishes an elevation-constrained UWB model, and then tracks the train's position using the elevation-constrained UWB model and the initialized INS. Finally, combined with the train's initial position, the real-time position information of the train is obtained through recursion and correction.

[0145] Example 2:

[0146] Example 2 is basically the same as Example 1, except that:

[0147] Taking the partial derivatives of h with respect to x, y, and z respectively, we get:

[0148]

[0149] Therefore, the differential dh is:

[0150]

[0151] The above equation is the prior track elevation constraint equation, where dh=U-h0, U is the elevation pseudo-observation formula, and h0 is the elevation corresponding to the approximate point (x0, y0, z0);

[0152] Using UWB observations r n With elevation pseudo-observation U n Combined to form the observation equation:

[0153]

[0154] Where (Xn, Yn, Zn) are the coordinates of the nth UWB base station, and the position parameters (x, y, z) based on the prior track elevation constraints can be solved by the least squares adjustment method.

[0155] UWB ranging accuracy is 0.2m, corresponding to the prior standard deviation:

[0156] σ r =0.2;

[0157] For the accuracy of pseudo-observations of elevation, the track gradient is usually less than 20‰, based on the distance d between the train positioning terminal and the UWB base station. n 20‰ is used as the accuracy of the pseudo-observation value of elevation. 10m corresponds to an elevation change of less than 0.2m, corresponding to an accuracy of 0.2m. The following is the formula for the standard deviation of the pseudo-observation value of elevation:

[0158]

[0159] The elevation-constrained UWB model in step five is specifically as follows:

[0160]

[0161] Where, σ r Standard deviation This represents the standard deviation of the pseudo-observed elevation values.

[0162] The elevation-constrained PPP-RTK model is as follows:

[0163] The network receives GNSS pseudorange and phase observation data and constructs the original observation equations:

[0164]

[0165] Where E(·) denotes the expectation operator, and These represent pseudorange and phase observations, respectively, whose station-to-satellite distances and tropospheric interference delays have been corrected beforehand. c represents the speed of light, s represents the satellite number, r represents the receiver number, j represents the frequency band number, i represents the epoch number, and dt represents the phase distance. r and dt s It is the receiver and satellite clock bias, τ r For the wet delay of the zenith troposphere, τ r The projection function is Indicates ionospheric delay, coefficient λ is the ratio between the ionospheric delay at frequency j and the first frequency. j For the wavelength corresponding to frequency j, d r，j and δ represents the pseudorange hardware delay at frequency j between receiver r and satellite s. r，j and This represents the phase deviation between receiver r and satellite s at corresponding frequency j. The integer ambiguity corresponding to frequency j for receiver r and satellite s;

[0166] Using the reference station p at the network end as the baseline station, a weighted pseudo-observation equation for the ionosphere at the network end is constructed:

[0167]

[0168] By using the S-basis rank deficiency elimination method, the rank deficiency of both the original and pseudo-observation equations is eliminated, and a new observation equation is constructed:

[0169]

[0170] The specific form of the new estimable parameter is as follows:

[0171]

[0172] in,

[0173]

[0174] d r，DCB =d r，2 -d r，1 ;

[0175] in, For the new, estimable receiver clock bias, For satellite clock bias, For the zenith tropospheric wet delay, For ionospheric delay, For receiver pseudorange hardware delay, For satellite pseudorange hardware delay, For receiver phase deviation, For satellite phase deviation, For ambiguity, This represents the hardware delay deviation between receiver r and receiver p.

[0176] The user terminal constructs the observation equation based on the received pseudorange and phase observations and the correction information broadcast by the network:

[0177]

[0178] Where the subscript u represents the user terminal receiver, To use the ionospheric delay values ​​near user stations interpolated based on the network-end ionospheric delay information as pseudo-observations, Let Δx be the unit vector from the station to the satellite. u (i) represents the correction value for the coordinates. For receiver clock bias, The tropospheric zenith wet delay is denoted as , and is the ambiguity parameter. For pseudorange hardware delay, The pseudorange is a geometrically independent combination of hardware delay values. For ionospheric slack delay, For receiver phase deviation, This refers to the receiver phase deviation;

[0179] When entering or exiting tunnels or other fully shielded areas, elevation constraints can be applied based on the received pseudo-elevation observation value Un:

[0180]

[0181] The corresponding stochastic model is:

[0182]

[0183] Where, σ p =0.3m, σ φ =0.003m, σ I =0.3, d n This is the ranging value from the terminal to the UWB base station n.

[0184] Example 3:

[0185] Example 3 is basically the same as Example 1, except that:

[0186] A positioning system for a BeiDou PPP-RTK train all-scenario positioning method is disclosed. The positioning system includes a reference station network, a computing center, and a train positioning terminal. The reference station network comprises a BeiDou / GNSS PPP-RTK reference station network and a UWB tunnel / platform base station network. The BeiDou / GNSS PPP-RTK reference station network is deployed along the railway line with an average station spacing of 100km, receiving and transmitting BeiDou satellite data to the computing center. The UWB tunnel / platform base station network provides UWB base station information and ranging information to the train terminal. The computing center refers to the PPP-RTK network-side computing center, which collects BeiDou satellite data observed by the PPP-RTK reference station network, and calculates and broadcasts correction information for satellite phase deviation, satellite clock error, ionospheric delay, and tropospheric delay in real time. The train positioning terminal receives BeiDou satellite observation data, BeiDou correction information broadcast by the computing center, UWB ranging data, and its own generated INS observation data, and calculates the train position in real time using an adaptive fusion positioning algorithm.

[0187] Example 4:

[0188] Example 4 is basically the same as Example 1, except that:

[0189] A positioning terminal for a BeiDou PPP-RTK train all-scenario positioning method includes a GNSS module, an INS module, a UWB module, a signal control module, a communication module, and a core processing module. The GNSS module receives raw BeiDou / GNSS observation data and sends 1Hz pps pulses per second to the signal control module, providing GPS or BDS time for time calibration of the core processing module. The INS module generates high-frequency triaxial angular velocity and triaxial acceleration raw observation data. The UWB module receives UWB base station ranging information and achieves time synchronization through the core processing module. The signal control module provides different frequency signals to the inertial navigation system to generate corresponding frequency INS data. The communication module transmits data via 4G / 5G / WiFi, receives PPP-RTK correction information, and outputs real-time combined positioning results or raw observation data from each sensor. The core processing module... The module receives raw observation data from various sensors and performs time synchronization. It integrates a PPP-RTK / INS / UWB adaptive fusion algorithm to calculate positioning results in real time. High-speed trains typically consist of eight carriages forming a power unit group with a total length of approximately 210 meters. Beidou PPP-RTK / INS / UWB combined equipment can be deployed at both the head and tail of the train. Multiple terminals jointly provide positioning results, thereby improving system redundancy and positioning availability. Simultaneously, by utilizing known distance constraints between terminals, multi-station collaborative positioning can be achieved, enhancing the error detection capability of observation data and effectively improving the system's positioning accuracy and reliability.

[0190] Example 6:

[0191] Example 6 is basically the same as Example 1, except that:

[0192] A dynamic experimental scenario for PPP-RTK / INS / UWB was set up using a track-guided trolley, covering a 15m*45m rooftop area. A GNSS / INS / UWB combined positioning terminal was fixed to the upper surface of the trolley. The initial INS attitude was (0, 0, 290°), the GNSS antenna arm error was (-0.18m, -0.04m, -0.51m), and the UWB antenna arm error was (-0.18m, -0.04m, -0.03m). Figure 11 shows the PPP-RTK network distribution in Wuhan, Hubei Province, and the actual... The test site map shows that the PPP-RTK reference network used is the Wuhan, Hubei reference network, as shown in Figure 11. There are a total of 11 base stations distributed around Wuhan, with an average station spacing of 38.24 km. The nearest reference station is 11.80 km from the test site. The UWB base stations are distributed in a rectangle at 4 points outside the track, with an east-west spacing of about 45 m and a north-south spacing of about 15 m. Base stations 1 and 3 are at the same height as the UWB antenna of the combined positioning terminal, while base stations 2 and 4 are 1 m higher. The positions of the UWB base stations are accurately measured using RTK. The experiment began at 135470 seconds on October 14th (GPS weekday), lasting for one hour until 139070 seconds, and was divided into four phases: 1. The first 600 seconds were kept stationary for inertial navigation zero-bias compensation and attitude calculation; 2. From 2600s to 2600s, PPP-RTK+INS was used to verify the combined positioning performance in open scenes; 3. From 2600s to 2600s, GNSS data was disconnected for 10 seconds every 10 seconds, and single INS inference was performed for 10 seconds, repeated 50 times, to verify the combined positioning performance in complex occlusion scenarios; 4. From 2600s to 3600s, UWB+INS was used to verify the combined positioning performance in fully occluded scenarios. During the experiment, a UBUC reference station with an average distance of 50m was used as the base station to calculate the millimeter-level RTK / INS combined positioning true value, in order to compare the PPP-RTK / INS / UWB combined positioning error at each stage.

[0193] Open scene simulation:

[0194] Figure 1 shows the PPP-RTK / INS combined positioning results for experimental phases 1 and 2. In phase 1, the vehicle remained stationary; in phase 2, the vehicle's 12-lap trajectory overlapped significantly, and the elevation results fluctuated with changes in track elevation. Figure 2 shows the PPP-RTK / INS combined positioning error diagrams for phases 1 and 2. The results show that the horizontal positioning error is better than 2 cm, the elevation positioning error is better than 5 cm, and the root mean square errors (RMSE) in the north, east, and celestial directions are 0.51 cm, 0.90 cm, and 2.47 cm, respectively, verifying the combined system's ability to achieve centimeter-level positioning performance in open environments.

[0195] Semi-occluded scene simulation:

[0196] As shown in Figure 3, during Experiment 3, when GNSS data was disconnected and restored every 10 seconds, the PPP-RTK and INS combined positioning results showed a slight trajectory deviation relative to the orbit during the 10-second inertial navigation calculation. Simultaneously, the elevation also deviated from the orbital plane. As can be seen from the error diagrams in Figure 4 for the north, east, and sky directions, the maximum deviations in the north and east directions were both within 35cm, while the maximum deviation in the elevation direction was 55cm. The RMSE values ​​for the north, east, and sky directions were 4.77cm, 3.40cm, and 7.19cm, respectively, verifying that the combined system possesses decimeter-level positioning performance in semi-obscured scenarios.

[0197] Fully occluded scene simulation:

[0198] Figure 5 shows the results of the UWB / INS combined positioning in Experiment Phase 4. A slight curvature appeared on the horizontal positioning trajectory, with a deviation within 20cm, mainly due to the influence of the stone piers beside the track, causing UWB observations to be blocked and resulting in multipath effects. The elevation fluctuation in Figure 5 is within 1m, primarily because the elevation difference between the four UWB base stations is small (1m), leading to a poor elevation geometry and thus affecting elevation positioning capability. The statistical analysis in Figure 6 shows the three-dimensional errors of the UWB / INS combined positioning. The RMSE values ​​in the north, east, and sky directions are 5.96cm, 3.74cm, and 43.36cm, respectively. The horizontal error is less than 20cm, and the elevation error is less than 1m, demonstrating that the combined system possesses decimeter-level positioning performance even in fully obscured scenarios.

[0199] Figure 7 shows the 3D positioning errors in all four stages of the experiment. In stage 1, the vehicle remained stationary. In stage 2, the positioning accuracy of the PPP-RTK / INS combination was better than 2cm, proving that the system has centimeter-level positioning capability under open conditions. In stage 3, under the simulated complex short-term occlusion scenario, the horizontal offset calculated by INS was less than 35cm within 10 seconds, and the RMSE in the north, east, and sky directions were 4.77cm, 3.40cm, and 7.19cm, respectively, indicating that the combined positioning system has centimeter-to-decimeter-level positioning capability in complex occlusion scenarios. In stage 4, the horizontal error of the UWB / INS combined positioning was less than 20cm, and the RMSE in the north, east, and sky directions were 5.96cm, 3.74cm, and 43.36cm, respectively, proving that the combined system has decimeter-level positioning capability in fully occluded scenarios.

[0200] The above description is only a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. Any equivalent modifications or changes made by those skilled in the art based on the content disclosed in the present invention should be included within the scope of protection set forth in the claims.

Claims

1. A BeiDou PPP-RTK train full-scene positioning method, characterized in that: The BeiDou PPP-RTK train full-scene positioning method includes the following steps: Step 1: The base station network first acquires raw GNSS data and raw UWB data, then sends the raw GNSS data to the computing center, and simultaneously sends the raw UWB data to the train positioning terminal; Step 2: The computing center first processes the raw GNSS data to obtain PPP-RTK correction information, and then sends the PPP-RTK correction information to the train positioning terminal; Step 3: The train positioning terminal first performs INS zero bias and attitude initialization while stationary to obtain the corrected INS, then combines the PPP-RTK correction information with the raw UWB data for filtering, and then inputs the filtered data into the corrected INS to obtain the initialized INS, and then uses the initialized INS to locate the train and obtain the initial position of the train; Step 4: When the train enters the driving state, the train positioning method is as follows: When the train is in an open or semi-obscured scene, the train positioning terminal combines BeiDou PPP-RTK with the initialized INS. The system performs position tracking on the train and then, based on the train's initial position, recursively corrects and obtains the train's real-time position information. When the train enters a fully occluded scene, the train positioning terminal first receives the UWB base station signal from the raw UWB data, then establishes the train's prior elevation, and finally establishes an elevation-constrained UWB model. The train's position is then tracked using the elevation-constrained UWB model and an initialized INS. The train's real-time position information is then obtained by recursively correcting and combining this with the train's initial position. When the train leaves the fully occluded scene, the train positioning terminal first receives PPP-RTK correction information and the UWB base station signal from the raw UWB data, establishes an elevation-constrained PPP-RTK model, obtains the PPP-RTK positioning result, and then inputs the PPP-RTK positioning result into the INS for updating. The system then performs position tracking on the train and, based on the train's initial position, recursively corrects and obtains the train's real-time position information. The system also includes receiving UWB base station elevation information after the train positioning terminal receives the UWB base station signal from the raw UWB data. Elevation of UWB base station from the orbital plane The specific prior elevation of the train is as follows: ;in, The elevation of the combined positioning terminal antenna installed on the train, As a pseudo-observation of elevation, the elevation above the Earth's ellipsoid is h < 10 km, satisfying the following relationship: ;in, For the semi-major axis of the ellipsoid, For the minor semi-axis of the ellipsoid, Let (x0, y0, z0) be the elevation of this point, and (x0, y0, z0) be the geocentric coordinates of this point. Taking the partial derivatives with respect to x0, y0, and z0 respectively, we get: Thus, the differential for: 。 2. The BeiDou PPP-RTK train full-scene positioning method according to claim 1, characterized in that: The elevation-constrained UWB model in step four is specifically as follows: ;in, Standard deviation, This represents the standard deviation of the pseudo-observed elevation values.

3. The BeiDou PPP-RTK train full-scene positioning method according to claim 2, characterized in that: The elevation-constrained PPP-RTK model in step four specifically involves: receiving GNSS pseudorange and phase observation data at the network end and constructing the original observation equations. ;in, Represents the expectation operator, and These represent pseudorange and phase observations, respectively, whose station-to-satellite distances and tropospheric dry delays have been corrected beforehand. Represents the speed of light. Indicates the satellite number, Indicates the receiver number. Indicates the frequency band number. Indicates the epoch number. and It's the receiver and satellite clock bias. For the zenith tropospheric wet delay, The projection function is ， Indicates ionospheric delay, coefficient For the corresponding frequency The ratio between the first frequency ionospheric delay and the first frequency ionospheric delay. For the corresponding frequency wavelength, and Indicates receiver and satellite Corresponding frequency pseudorange hardware delay, and Indicates receiver and satellite Corresponding frequency Phase deviation, For receiver and satellite Corresponding frequency The corresponding integer ambiguity; based on the receiver Using the baseline station, a network-end ionospheric weighted pseudo-observation equation is constructed: By using the S-basis rank deficiency elimination method, the rank deficiency of the original observation equation and the pseudo-observation equation is eliminated, and a new observation equation is constructed: The specific form of the new estimable parameter is as follows: ;in, ， ， ， ， ;in, For the new, estimable receiver clock bias, For satellite clock bias, For the zenith tropospheric wet delay, For ionospheric delay, For receiver pseudorange hardware delay, For satellite pseudorange hardware delay, For receiver phase deviation, For satellite phase deviation, For ambiguity, For receiver and receiver Hardware latency discrepancies between them.

4. The BeiDou PPP-RTK train full-scene positioning method according to claim 1, characterized in that: The user terminal constructs the observation equation based on the received pseudorange and phase observations and the correction information broadcast by the network: ; where subscript Indicates the user terminal receiver. To use the ionospheric delay values ​​near user stations interpolated based on the network-end ionospheric delay information as pseudo-observations, The unit vector from the station to the satellite. These are the correction values ​​for the coordinates. For receiver clock bias, For tropospheric zenith wet delay, For ambiguity parameters, For pseudo-range hardware delay, The pseudorange is a geometrically independent combination of hardware delay values. For ionospheric slack delay, To account for receiver phase deviation; when entering or exiting tunnels or other fully shielded areas, elevation constraints can be applied based on the received pseudo-elevation observation value Un: The corresponding stochastic model is: ;in, ， ， ， ， This is the ranging value from the terminal to the UWB base station n.

5. The BeiDou PPP-RTK train full-scene positioning method according to claim 1, characterized in that: The raw GNSS data includes satellite observation information and PPP-B2b information, while the raw UWB data includes UWB base station coordinates, ranging values, and UWB base station signals.

6. A positioning system for the BeiDou PPP-RTK train full-scene positioning method as described in claim 1, characterized in that: The positioning system of the BeiDou PPP-RTK train full-scene positioning method includes a reference station network, a computing center, and a train positioning terminal. The reference station network includes a BeiDou / GNSS PPP-RTK reference station network and a UWB tunnel / platform base station network. The BeiDou / GNSS PPP-RTK reference station network is deployed along the railway line with an average station spacing of 100km, receiving and transmitting BeiDou satellite data to the computing center. The UWB tunnel / platform base station network provides UWB base station information and ranging information to the train terminal. The computing center refers to the PPP-RTK network-end computing center, which collects BeiDou satellite data observed by the PPP-RTK reference station network, calculates and broadcasts correction information for satellite phase deviation, satellite clock error, ionospheric delay, and tropospheric delay in real time. The train positioning terminal receives BeiDou satellite observation data, BeiDou correction information broadcast by the computing center, UWB ranging data, and its own generated INS observation data, and calculates the train position in real time through an adaptive fusion positioning algorithm.

7. A positioning terminal for the BeiDou PPP-RTK train full-scene positioning method as described in claim 1, characterized in that: The positioning terminal of the BeiDou PPP-RTK train full-scene positioning method includes a GNSS module, an INS module, a UWB module, a signal control module, a communication module, and a core processing module. The GNSS module receives raw BeiDou / GNSS observation data and sends 1Hz pps pulses per second to the signal control module, providing GPS or BDS time for time calibration of the core processing module. The INS module generates high-frequency triaxial angular velocity and triaxial acceleration raw observation data. The UWB module receives UWB base station ranging information and achieves time synchronization through the core processing module. The signal control module provides different frequency signals to the inertial navigation system to generate corresponding frequency INS data. The communication module transmits data via 4G / 5G / WiFi, receives PPP-RTK correction information, and outputs real-time combined positioning results or raw observation data from each sensor. The core processing module receives raw observation data from each sensor, performs time synchronization, integrates a PPP-RTK / INS / UWB adaptive fusion algorithm, and calculates the positioning result in real time.

Images