Global post-processing positioning method and system based on various types of Beidou augmentation services

Abstract

The present invention discloses a global post-processing positioning method based on various types of Beidou augmentation services, which includes obtaining three-dimensional position information of a rover at different positions using multiple positioning modes, calculating quality factors of various positioning modes at different positions, and obtaining high-precision three-dimensional positions of the rover at different positions using a reference device; classifying and training a random forest model using the three-dimensional position information, quality factors of the rover at different positions under multiple positioning modes, and the high-precision three-dimensional positions obtained by the reference device; inputting the quality factors of the rover at a certain position under multiple positioning modes into the trained random forest classification model to obtain position variances under various positioning modes, and fusing the three-dimensional position information under various positioning modes using adaptive weighting to obtain optimal position information. The present invention fuses different types of positioning results, improving the availability, reliability, and positioning accuracy of Beidou positioning.

Description

Technical Field

[0001] The present invention belongs to the technical field of satellite navigation and positioning, and particularly relates to a global post-processing positioning method and system based on various types of Beidou augmentation services. Background Art

[0002] The Beidou satellite navigation system (hereinafter referred to as the Beidou system) is constructed and developed in accordance with the "three-step" strategy. The Beidou-1 system was put into use in 2000, adopting an active positioning system, and providing pseudo-range single-point positioning, time service, wide-area differential, and short message communication services for Chinese users. The Beidou-2 system was put into use in 2012. On the basis of being compatible with the technical system of the Beidou-1 system, a passive positioning system was added, and positioning, speed measurement, time service, and short message communication services were provided for users in the Asia-Pacific region. The Beidou-3 system was put into use in 2020. On the basis of the Beidou-2 system, its performance was further improved and its functions were expanded to provide a variety of services, including: providing positioning, navigation, and timing, global short message communication, and international search and rescue services globally; providing satellite-based augmentation, satellite-based precise point positioning, ground-based augmentation, and regional short message communication services in China and its surrounding areas, supporting post-differential positioning and post-precise point positioning services.

[0003] Pseudo-range single-point positioning (SPP) is a positioning method that uses the pseudo-range observation values and broadcast ephemeris of the Beidou system, and is a basic service provided by the Beidou system. Its coverage range is global, the positioning accuracy is within 5m, it can be obtained in real time and solved post facto, and the positioning convergence time is 10 seconds.

[0004] Satellite-based precise point positioning (PPP) is to use the GEO satellites of the Beidou system to broadcast PPP-B2b satellite-based differential corrections, integrity information, and other information to provide dual-frequency observation value enhancement services for users in China and its surrounding areas. Its coverage range is China and its surrounding areas, the positioning accuracy is within 0.4m, it can be obtained in real time, and the positioning convergence time is 20 minutes.

[0005] Ground-based augmentation positioning is based on network RTK (Real Time Kinematic) positioning technology. The data processing center processes the synchronous observation data of multiple Beidou system reference stations within a certain range, generates differential data and broadcasts it through mobile communication or the Internet. Users in this area receive satellite signals and differential signals to achieve high-precision real-time dynamic positioning. Its coverage range is the area covered by mobile communication in China and its surrounding areas, the positioning accuracy is within 0.05m, it can be obtained in real time, and the positioning convergence time is 30 seconds.

[0006] Post Process Kinematic (PPK) positioning is a method where users utilize multi-frequency pseudorange and carrier phase data from synchronous observations of a single Beidou system reference station and a rover station to perform double-difference ionosphere-free combination to weaken the influence of atmospheric errors, thereby achieving high-precision post-processing positioning. Its coverage range is within a 20-kilometer radius centered on a single Beidou reference station, with a positioning accuracy of 0.02m + 1ppm × distance in kilometers. It can be solved post-eventually without a positioning convergence time.

[0007] Precise Point Positioning (PPP) is a method where users utilize multi-frequency pseudorange and carrier phase observations of a Beidou system rover station, IGS precise ephemeris, and precise satellite clock offsets during the same period to perform an undifferenced and uncombined observation model. Through a series of precise error correction models, high-precision post-processing positioning is achieved. Its coverage range is global, with a positioning accuracy within 0.2m. It can be solved post-eventually without a positioning convergence time.

[0008] Based on pseudorange single-point positioning, space-based precise point positioning, ground-based augmentation positioning, post-process kinematic positioning, and precise point positioning, the Beidou system can provide augmentation services and positioning capabilities of different levels such as meter-level, decimeter-level, and centimeter-level. However, due to the differences in coverage range, positioning accuracy, convergence time, and external dependence conditions of different types of augmentation services, how to make a flexible integration strategy, draw on each other's strengths to obtain the optimal comprehensive position information, and improve the availability, reliability, and accuracy of Beidou positioning services is an urgent problem to be solved. Summary of the Invention

[0009] Aiming at the problem of insufficient availability, reliability, and accuracy of Beidou positioning services caused by the differences in existing positioning technologies, the present invention provides a global post-processing positioning method based on various types of Beidou augmentation services. This method has strong practicability and high flexibility, improves the availability, reliability, and positioning accuracy of Beidou positioning, and can be widely applied in the field of Beidou navigation and positioning.

[0010] To achieve the above object, the technical solution provided by the present invention is a global post-processing positioning method based on various types of Beidou augmentation services, including the following steps:

[0011] Step 1: Based on the Beidou system, use multiple positioning modes to obtain the three-dimensional position information of the rover station at different positions, calculate and record the quality factors of various positioning modes at different positions, and at the same time, based on the Beidou system, use a reference device to obtain the high-precision three-dimensional position of the rover station at different positions;

[0012] Step 2: Use the three-dimensional position information and quality factors of the rover at different positions in various positioning modes obtained in Step 1, and the high-precision three-dimensional positions of the rover at different positions obtained by the reference device to perform classification training on the random forest model, and obtain a trained random forest classification model;

[0013] Step 3: Input the quality factors of the rover at a certain position in various positioning modes into the random forest classification model trained in Step 2 to obtain the position variances in various positioning modes, and use the adaptive weighting method to fuse the three-dimensional position information in various positioning modes to obtain the optimal positioning result at a certain position.

[0014] Furthermore, the positioning modes in Step 1 include two categories: real-time positioning and post-processing positioning. Real-time positioning includes three modes: pseudo-range single-point positioning, satellite-based precise single-point positioning, and ground-based augmentation positioning. Post-processing positioning includes two modes: post-differential positioning and post-precise single-point positioning. The quality factors of the three modes of pseudo-range single-point positioning, satellite-based precise single-point positioning, and ground-based augmentation positioning include the number of satellites used, the standard deviation of position error, the positioning solution mode, the differential time delay, and the dilution of precision. The quality factors of the post-differential positioning mode include the number of satellites used, the standard deviation of position error, the positioning solution mode, the dilution of precision, the signal-to-noise ratio, the baseline length, the integrity of the observed data, the multi-path error estimate, the carrier phase noise, and the ambiguity fixing check ratio. The quality factors of the post-precise single-point positioning mode include the number of satellites used, the standard deviation of position error, the positioning solution mode, the dilution of precision, the signal-to-noise ratio, the integrity of the observed data, the multi-path error estimate, the carrier phase noise, and the ambiguity fixing check ratio.

[0015] Furthermore, in Step 2, the differences between the three-dimensional positions of the rover at different positions in various positioning modes obtained in Step 1 and the high-precision three-dimensional positions obtained by using the reference device at this position are calculated to obtain multiple position deviations, and the position deviations are classified according to the set precision classification levels to obtain the corresponding position variances. The level classification rules are as follows:

[0016] (7)

[0017] In the formula, represents the position variance of the rover at the p th positioning mode at position i , represents the difference between the three-dimensional position of the rover at the p th positioning mode at position i and the high-precision three-dimensional position obtained by the reference device, , , , , are all set position deviation thresholds.n is the set number of precision classification levels.

[0018] Input the quality factors of the rover at different positions in various positioning modes obtained in step 1 and the corresponding precision classification levels of the calculated position deviations of the rover at different positions in various positioning modes into the random forest model to train the random forest classifier, and at the same time use the five-fold cross-validation method to evaluate the classification accuracy of the random forest classifier.

[0019] Further, in step 3, input the quality factors of the rover in different types of positioning modes at a certain position into the random forest classification model trained in step 2 to obtain the precision classification levels in different types of positioning modes at this position, and then obtain the position variances corresponding to the three-dimensional position information in various positioning modes at this position; assume that at position p the i adaptive weight of the three-dimensional position information of the rover in the th positioning mode is

[0020] (8)

[0021] (9)

[0022] In the formula, represents the mathematical expectation, respectively represent the three-dimensional position information of the rover at position p in the i , j th positioning modes, N represents the number of types of positioning modes, is the optimal positioning result of the rover to be solved;

[0023] In order to minimize the value of the fusion variance , perform the extreme value calculation of the multivariate function through the Lagrange multiplier method to obtain the minimum adaptive weights of different types of positioning modes and the minimum value of the fusion variance, multiply the three-dimensional position information in various positioning modes by their corresponding minimum adaptive weights, and accumulate to obtain the optimal positioning result of the rover. The specific calculation method is as follows:

[0024] (10)

[0025] (11)

[0026] (12)

[0027] In the formula, represents the positionp The minimum adaptive weight value of the i th positioning mode, indicating the variance of the optimal positioning result, is the position p at the i th positioning mode for the position variance corresponding to the three-dimensional position information.

[0028] The present invention also provides a global post-processing positioning system based on various types of Beidou augmentation services, which is used to implement a global post-processing positioning method based on various types of Beidou augmentation services as described above.

[0029] Moreover, it includes a processor and a memory. The memory is used to store program instructions, and the processor is used to call the stored instructions in the memory to execute a global post-processing positioning method based on various types of Beidou augmentation services as described above.

[0030] Alternatively, it includes a readable storage medium, on which a computer program is stored. When the computer program is executed, it implements a global post-processing positioning method based on various types of Beidou augmentation services as described above.

[0031] Compared with the prior art, the present invention has the following advantages:

[0032] 1) For the pseudorange single-point positioning, satellite-based precise single-point positioning, ground-based augmentation three-dimensional position information obtained in real time, and the precise single-point positioning and differential positioning three-dimensional position information obtained through post-processing, comprehensively analyze the applicable range, positioning accuracy, convergence time, etc. of various types of positioning methods. According to the three-dimensional position information of each type and its quality factor, use machine learning for accuracy classification, and obtain the optimal comprehensive three-dimensional position information through adaptive weighted fusion.

[0033] 2) This method can realize the flexible fusion and comprehensive application of different types of positioning methods, improve the availability, reliability and positioning accuracy of Beidou positioning, and can be widely applied in Beidou navigation and positioning fields such as land surveying, aerial mapping, and ocean navigation surveying. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0035] Figure 1 It is a flowchart of the global post-processing positioning method based on various types of Beidou augmentation services according to an embodiment of the present invention.

[0036] Figure 2 This is the process diagram of the random forest classification modeling in the embodiment of the present invention. Specific Embodiments

[0037] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings and embodiments in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0038] Embodiment 1

[0039] As Figure 1 shown, the embodiment of the present invention provides a global post-processing positioning method based on various types of Beidou augmentation services, including the following steps:

[0040] Step 1, based on the Beidou system, according to different real-time positioning modes (pseudorange single-point positioning, satellite-based precise single-point positioning, ground-based augmentation positioning) set by the rover receiver, the three-dimensional position information and quality factors of the rover at different positions in different modes are obtained in real time. At the same time, based on the Beidou system, a reference device is used to obtain the high-precision three-dimensional position of the rover at different positions.

[0041] The quality factors include the number of satellites used, the standard deviation of position error, the positioning solution mode, the differential delay, the dilution of precision, etc. The number of satellites used is the total number of satellites participating in the positioning solution. The standard deviation of position error is the measurement error of latitude, longitude, and altitude positioning solution. If least squares positioning solution is used, the least square residual is used to represent the measurement error; if Kalman filter positioning solution is used, the position state variance is used to represent the measurement error. The positioning solution mode of the pseudorange single-point positioning mode is the pseudorange single-point positioning mode. The positioning solution modes of satellite-based precise single-point positioning include the pseudorange single-point positioning mode, the fixed solution mode of satellite-based precise single-point positioning, and the float solution mode of satellite-based precise single-point positioning. The positioning solution modes of ground-based augmentation positioning include the pseudorange single-point positioning mode, the pseudorange differential mode of ground-based augmentation positioning, the fixed solution mode of ground-based augmentation positioning, and the float solution mode of ground-based augmentation positioning. The differential delay is the difference between the time of the differential correction number used and the time of the current observation value. The dilution of precision is calculated from the horizontal dilution of precision, the vertical dilution of precision, and the time dilution of precision. The calculation formula is:

[0042] (1)

[0043] In the formula, GDOP represents the dilution of precision, HDOP represents the horizontal dilution of precision, VDOP represents the vertical dilution of precision, and TDOP represents the time dilution of precision.

[0044] Step 2, when the coverage area of the reference station includes the position of the rover station, post-processing PPK solution is performed based on the acquired raw Beidou system observation values of the reference station and the rover station to obtain the three-dimensional position information and quality factors of the rover station at different positions. At the same time, a high-precision three-dimensional position of the rover station at different positions is obtained using a reference device.

[0045] The quality factors include the number of satellites used, the standard deviation of position error, the positioning solution mode, the dilution of precision, the signal-to-noise ratio, the baseline length, the integrity of the observation data, the estimated value of the multipath error, the carrier phase noise, the ambiguity fixation check ratio, etc. The number of satellites used is the total number of satellites participating in the positioning solution. The standard deviation of position error is the measurement error of latitude, longitude, and altitude positioning solution. If the least squares positioning solution is used, the least square residual is used to represent the measurement error; if the Kalman filter positioning solution is used, the position state variance is used to represent the measurement error. The positioning solution modes of the post-differential positioning mode include the post-pseudorange single-point positioning mode, the post-differential positioning pseudorange differential mode, the post-differential positioning fixed solution mode, and the post-differential positioning float solution mode. The dilution of precision is calculated from the horizontal dilution of precision, the vertical dilution of precision, and the time dilution of precision, and the calculation method is the same as formula (1). The signal-to-noise ratio is the ratio of the carrier signal power to the noise power spectral density. The baseline length is the relative distance calculated based on the known coordinates of the reference station and the coordinates of the rover station solved.

[0046] The integrity of the observation data includes the single-frequency point observation data integrity rate and the single-system observation data integrity rate , which are calculated through the following formulas:

[0047] (2)

[0048] (3)

[0049] In the formula, represents the single-frequency point observation data integrity rate, n represents the number of satellites observed during the observation period, represents the total number of actual observation epochs of the jth satellite at a certain frequency point during the observation period, represents the total number of theoretical epochs of the jth satellite at a certain frequency point during the observation period, represents the single-system observation data integrity rate, represents the number of epochs when all frequency points of the jth satellite have valid observation data during the observation period, represents the total number of theoretical epochs of the jth satellite during the observation period.

[0050] The estimated value of the multipath error is calculated through the following formula:

[0051] (4)

[0052] In the formula, represents the estimated value of the multipath error of the observed satellite at frequency, represents the number of epochs of the sliding window, represents at epoch the computational amount of the observed satellite at frequency including multipath error and integer ambiguity information.

[0053] Carrier phase noise is calculated by the following formula:

[0054] (5)

[0055] In the formula, represents the number of triple differences of the carrier phase observations of the observed satellite at adjacent epochs at a certain frequency, represents at epoch the phase carrier phase observation of the observed satellite at a certain frequency, represents the grouped difference (triple difference) of the double differences of the carrier phase observations at adjacent epochs at a certain frequency.

[0056] The ambiguity fixing check ratio is calculated by the following formula:

[0057] (6)

[0058] In the formula, ratio represents the ambiguity fixing check ratio, represents the sum of the squares of the second smallest residuals in the fixed solution, represents the sum of the squares of the smallest residuals in the fixed solution.

[0059] Step 3: Perform post - processing PPP solution according to the acquired raw Beidou system observations of the rover, precise satellite ephemeris, clock difference file, and corrections to obtain the three - dimensional position information and quality factor of the rover at different positions. Meanwhile, the Beidou system uses a reference device to obtain the high - precision three - dimensional position of the rover at different positions.

[0060] The quality factor includes the number of satellites used, the standard deviation of position error, the positioning solution mode, the dilution of precision, the signal-to-noise ratio, the integrity of observation data, the estimated value of multipath error, the carrier phase noise, the ambiguity fixing verification ratio, etc. The number of satellites used is the total number of satellites participating in the positioning solution. The standard deviation of position error is the measurement error of latitude, longitude, and altitude positioning solution. If the least squares positioning solution is used, the measurement error is represented by the least square residual; if the Kalman filter positioning solution is used, the measurement error is represented by the position state variance. The positioning solution modes of the post-precision point positioning mode include the post-pseudorange single-point positioning mode, the post-precision point positioning fixed solution mode, and the post-precision point positioning float solution mode. The dilution of precision is calculated from the horizontal dilution of precision, the vertical dilution of precision, and the time dilution of precision, and the calculation method is the same as formula (1). The signal-to-noise ratio is the ratio of the carrier signal power to the noise power spectral density. The evaluation method of the integrity of observation data is the same as formulas (2) and (3). The evaluation method of multipath error is the same as formula (4). The calculation method of carrier phase noise is the same as formula (5). The evaluation method of the ambiguity fixing verification ratio is the same as formula (6).

[0061] Step 4: Use the three-dimensional position information and quality factors of the rover at different positions in various positioning modes obtained in Steps 1 - 3, and the high-precision three-dimensional positions of the rover at different positions obtained by the reference device to perform classification training on the random forest model, and obtain the trained random forest classification model.

[0062] The three-dimensional position information of the five positioning modes (pseudorange single-point positioning, satellite-based precise point positioning, ground-based augmentation positioning, post-differential positioning, post-precision point positioning) obtained in Steps 1 - 3 is subtracted from the high-precision three-dimensional position obtained by the reference device to obtain multiple position deviations. The position deviations are classified into 8 precision classification levels according to the precision level classification table set in Table 1. Among them, the first class has the highest precision classification level, and the eighth class has the lowest precision classification level. Each precision classification level has a corresponding position variance.

[0063] Table 1 Precision Level Classification List

[0064]

[0065] Random forest is a supervised machine learning ensemble algorithm with decision trees as the base learners. During the construction of decision trees, it adds random selection of attributes, which can effectively solve the overfitting problem and improve the anti-noise performance. The random forest classification method is to randomly sample the feature variables and labeled samples of the training data set when constructing multiple decision trees. Each sampling result obtains a decision tree, and each tree can generate decision rules and classification results that conform to its own attributes. The implementation of the random forest classification algorithm is to aggregate the decision rules and classification results of all trees.

[0066] Input the quality factors of the rover at different positions in various positioning modes obtained in Steps 1 - 3 and the corresponding accuracy classification levels of the calculated position deviations of the rover at different positions in various positioning modes into the random forest model to train the random forest classifier. At the same time, use the five - fold cross - validation method to evaluate the classification accuracy of the random forest classifier, that is, in the five - fold cross - validation, four - fifths of the data are repeatedly used for model training in turn, and one - fifth of the data is used to test the classification accuracy of the random forest classifier, such as Figure 2 , so as to improve the accuracy of the overall random forest classification model.

[0067] Step 5: Input the quality factors of the rover in multiple positioning modes at a certain position into the random forest classification model trained in Step 4 to obtain the position variances in various positioning modes. Use the adaptive weighted method to fuse the three - dimensional position information in various positioning modes to obtain the optimal positioning result at a certain position.

[0068] The redundant measurement information of multi - source data can reduce the measurement error of a single data source. The adaptive weighted fusion method is based on the multi - source measurement data and their corresponding characteristic standard deviations. Taking the fusion variance as the minimum mean - square error as the principle, the weighted factors of the multi - source measurement data are automatically estimated, so that the result of the fused measurement data is the optimal result.

[0069] Input the quality factors of the rover in different types of positioning modes at a certain position into the random forest classification model trained in Step 4 to obtain the accuracy classification levels in different types of positioning modes at this position, and then obtain the position variances corresponding to the three - dimensional position information in different types of positioning modes at this position. Assume that at position p the adaptive weight of the three - dimensional position information of the rover in the i th positioning mode is , and the three - dimensional position information in various positioning modes is mutually independent and an unbiased estimate of the optimal position information, then:

[0070] (8)

[0071] (9)

[0072] In the formula, represents the mathematical expectation, respectively represent the three - dimensional position information of the rover at position p in the i , j th positioning modes, N represents the number of types of positioning modes, is the optimal positioning result of the rover to be solved.

[0073] In order to make the fusion variance The value is the smallest. By using the Lagrange multiplier method to solve the extreme value of a multivariate function, the minimum adaptive weight of various positioning modes and the minimum value of the fusion variance are obtained. Multiply the three-dimensional position information in various positioning modes by their corresponding minimum adaptive weights and accumulate them to obtain the optimal positioning result of the rover station. , and the specific calculation method is as follows:

[0074] (9)

[0075] (10)

[0076] (11)

[0077] In the formula, represents the position p at the i -th minimum adaptive weight of the -th positioning mode, represents the variance of the optimal positioning result, p is the position variance corresponding to the three-dimensional position information in the i -th positioning mode at the position

[0078] The use of the adaptive weighting method can achieve the flexible fusion of various types of Beidou positioning methods, improve the availability, reliability, and positioning accuracy of Beidou positioning, and expand the comprehensive application of Beidou.

[0079] Embodiment 2

[0080] Based on the same inventive concept, the present invention also provides a global post-processing positioning system based on various types of Beidou enhanced services, including a processor and a memory. The memory is used to store program instructions, and the processor is used to call the program instructions in the memory to execute a global post-processing positioning method based on various types of Beidou enhanced services as described above.

[0081] Embodiment 3

[0082] Based on the same inventive concept, the present invention also provides a global post-processing positioning system based on various types of Beidou enhanced services, including a readable storage medium. A computer program is stored on the readable storage medium, and when the computer program is executed, it implements a global post-processing positioning method based on various types of Beidou enhanced services as described above.

[0083] Specifically in implementation, the method proposed by the technical solution of the present invention can be automatically run by those skilled in the art using computer software technology. The system device for implementing the method, such as a computer-readable storage medium storing the corresponding computer program of the technical solution of the present invention and a computer device including running the corresponding computer program, should also be within the protection scope of the present invention.

[0084] The specific embodiments described herein are merely illustrative of the spirit of the present invention. Those skilled in the art to which the present invention pertains can make various modifications, supplements, or use similar methods of substitution to the specific embodiments described, but will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

Claims

1. A global post-processing positioning method based on various types of Beidou augmentation services, characterized in that, It includes the following steps: Step 1: Based on the Beidou system, use multiple positioning modes to obtain the three-dimensional position information of the rover at different positions, calculate and record the quality factors of various positioning modes at different positions. Meanwhile, based on the Beidou system, use a reference device to obtain the high-precision three-dimensional position of the rover at different positions; The positioning modes include two categories: real-time positioning and post-processing positioning. Real-time positioning includes three modes: pseudorange single-point positioning, space-based precise single-point positioning, and ground-based augmentation positioning. Post-processing positioning includes two modes: post-differential positioning and post-precise single-point positioning. The quality factors of the three modes of pseudorange single-point positioning, space-based precise single-point positioning, and ground-based augmentation positioning include the number of satellites used, the standard deviation of position error, the positioning solution mode, the differential delay, and the dilution of precision. The quality factors of the post-differential positioning mode include the number of satellites used, the standard deviation of position error, the positioning solution mode, the dilution of precision, the signal-to-noise ratio, the baseline length, the integrity of the observed data, the estimated value of the multipath error, the carrier phase noise, and the ambiguity fixing check ratio. The quality factors of the post-precise single-point positioning mode include the number of satellites used, the standard deviation of position error, the positioning solution mode, the dilution of precision, the signal-to-noise ratio, the integrity of the observed data, the estimated value of the multipath error, the carrier phase noise, and the ambiguity fixing check ratio; Step 2: Use the three-dimensional position information, quality factors of various positioning modes of the rover at different positions obtained in Step 1, and the high-precision three-dimensional position of the rover at different positions obtained by the reference device to perform classification training on the random forest model, and obtain a trained random forest classification model; Step 3: Input the quality factors of various positioning modes of the rover at a certain position into the trained random forest classification model obtained in Step 2 to obtain the position variances of various positioning modes, and use the adaptive weighting method to fuse the three-dimensional position information of various positioning modes to obtain the optimal positioning result at a certain position.

2. The global post-processing positioning method based on various types of Beidou augmentation services according to claim 1, characterized in that: The number of satellites used in step 1 is the total number of satellites participating in the positioning solution; the standard deviation of the position error is the measurement error of the latitude, longitude, and altitude positioning solutions. If the least squares positioning solution is used, the measurement error is represented by the least square residual. If the Kalman filter positioning solution is used, the measurement error is represented by the position state variance; the positioning solution mode of the pseudorange single point positioning mode is the pseudorange single point positioning mode; the positioning solution modes of the satellite-based precise point positioning include the pseudorange single point positioning mode, the fixed solution mode of the satellite-based precise point positioning, and the float solution mode of the satellite-based precise point positioning; the positioning solution modes of the ground-based augmentation positioning include the pseudorange single point positioning mode, the pseudorange differential mode of the ground-based augmentation positioning, the fixed solution mode of the ground-based augmentation positioning, and the float solution mode of the ground-based augmentation positioning; the positioning solution modes of the post-differential positioning mode include the post-pseudorange single point positioning mode, the post-differential positioning pseudorange differential mode, the post-differential positioning fixed solution mode, and the post-differential positioning float solution mode; the positioning solution modes of the post-precise point positioning mode include the post-pseudorange single point positioning mode, the post-precise point positioning fixed solution mode, and the post-precise point positioning float solution mode; the differential time delay is the difference between the time of the differential correction number used and the time of the current observation value.

3. A global post-processing positioning method based on various types of Beidou augmentation services as claimed in claim 1, characterized in that: In step 1, the dilution of precision is calculated from the horizontal dilution of precision, the vertical dilution of precision, and the time dilution of precision. The calculation formula is: （1） In the formula, GDOP represents the dilution of precision, HDOP represents the horizontal dilution of precision, VDOP represents the vertical dilution of precision, and TDOP represents the time dilution of precision; The signal-to-noise ratio is the ratio of the carrier signal power to the noise power spectral density; the baseline length is the relative distance calculated based on the known reference station coordinates and the solved rover coordinates.

4. A global post - processing positioning method based on various types of Beidou augmentation services as claimed in claim 1, characterized in that: The integrity of the observed data in Step 1 includes the integrity rate of single-frequency point observed data and the integrity rate of single-system observed data , which is calculated by the following formula: （2） （3） In the formula, represents the integrity rate of single-frequency point observation data, n represents the number of satellites observed during the observation period, represents the total number of actual observation epochs of the j-th satellite at a certain frequency point during the observation period, represents the total number of theoretical epochs of the j-th satellite at a certain frequency point during the observation period, represents the integrity rate of single-system observation data, represents the number of epochs in which the j-th satellite has valid observation data at all frequency points during the observation period, represents the total number of theoretical epochs of the j-th satellite during the observation period; Multipath error estimate value It is calculated by the following formula: （4） In the formula, represents the estimated value of the multipath error of the observed satellite at frequency, represents the number of epochs of the sliding window, represents the computational amount containing multipath error and integer ambiguity information of the observed satellite at frequency at epoch ​ 5. The global post-processing positioning method based on various types of Beidou augmentation services according to claim 1, characterized in that: Carrier phase noise in Step 1 is calculated by the following formula: （5） In the formula, represents the number of triple differences of the carrier phase observations of a satellite at a certain frequency point for adjacent epochs, represents at epoch the carrier phase observation of a satellite at a certain frequency point is observed, represents the group difference of the double differences of the carrier phase observations for adjacent epochs at a certain frequency point; The ambiguity fixing check ratio is calculated using the following formula: （6） where ratio represents the ambiguity fixing verification ratio, represents the second smallest residual sum of squares in the fixed solution, represents the smallest residual sum of squares in the fixed solution.

6. The global post-processing positioning method based on various types of Beidou augmentation services according to claim 1, characterized in that: In step 2, the three-dimensional positions of the rover at different positions in various positioning modes obtained in step 1 are subtracted from the high-precision three-dimensional positions obtained using the reference device at that position to obtain multiple position deviations. The position deviations are divided according to the set precision classification levels to obtain the corresponding position variances. The level division rules are as follows: （7） In the formula, represents the position variance of the rover at position p in the i th positioning mode, represents the difference between the three-dimensional position of the rover and the high-precision three-dimensional position obtained by the reference device at position p in the i th positioning mode, , , , , are all set position deviation threshold values, n is the set number of precision classification levels; The quality factors of the rover at different positions in various positioning modes obtained in step 1 and the precision classification levels corresponding to the position deviations of the rover at different positions in various positioning modes calculated are input into the random forest model to train the random forest classifier, and at the same time, the five-fold cross-validation method is used to evaluate the classification accuracy of the random forest classifier.

7. A global post-processing positioning method based on various types of Beidou augmentation services according to claim 1, characterized in that: In step 3, the quality factors of the rover at a certain position under different types of positioning modes are input into the random forest classification model trained in step 2 to obtain the accuracy classification levels of different types of positioning modes at this position, and then the position variances corresponding to the three-dimensional position information of different types of positioning modes at this position are obtained; assume that at the position p the adaptive weight of the three-dimensional position information of the rover under the i th positioning mode is . If the three-dimensional position information under various positioning modes is independent of each other and is an unbiased estimate of the optimal position information, then: （8） （9） In the formula, represents the mathematical expectation, Respectively indicate the location p The mobile station is i , j The three-dimensional position information under the positioning modes, N represents the number of positioning modes. is the optimal positioning result of the mobile station to be solved; To minimize the fusion variance value, the extreme value of the multivariate function is calculated by the Lagrange multiplier method to obtain the minimum adaptive weights of various positioning modes and the minimum value of the fusion variance. Multiply the three-dimensional position information in each positioning mode by its corresponding minimum adaptive weight and accumulate to obtain the optimal positioning result of the rover , and the specific calculation method is as follows: （10） （11） （12） In the formula, represents the minimum adaptive weight of the p th i positioning mode at the position, represents the variance of the optimal positioning result, is the position variance corresponding to the three-dimensional position information under the p th i positioning mode at the position.

8. An all-region post-processing positioning system based on various types of Beidou augmentation services, characterized in that, It includes a processor and a memory. The memory is used to store program instructions, and the processor is used to call the program instructions in the memory to execute a global post-processing positioning method based on various types of Beidou augmentation services as described in any one of claims 1-7.

9. An all-region post-processing positioning system based on various types of Beidou augmentation services, characterized in that, It includes a readable storage medium, and a computer program is stored on the readable storage medium. When the computer program is executed, it implements a global post-processing positioning method based on various types of Beidou augmentation services as described in any one of claims 1-7.

Images