High-availability vrs differential data broadcasting method and system

Abstract

The application discloses a high-availability VRS differential data broadcasting method and system, wherein the method comprises the following steps: a plurality of distributed algorithm nodes deployed in a container through a micro-service architecture are networked by a base station, each algorithm node generates data quality metadata by solving; each algorithm node performs an election service, calculates a multi-dimensional dynamic score based on the data quality metadata, and sorts and filters the highest-score algorithm node and the algorithm node of the last version election service result as a candidate set; if the score difference between the highest-score algorithm node and the historical cache highest-score election service result is less than a set threshold, the historical cache highest-score election service result is taken as the election service result, and corresponding differential data is broadcast to each terminal; otherwise, the highest-score algorithm node is taken as the election service result for broadcasting. The application elects differential data generated by deploying multiple algorithm nodes through a dynamic multi-dimensional evaluation algorithm, a time-space adaptive weight adjustment and a timeout fuse mechanism.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of distributed positioning, and particularly relates to a high-availability VRS differential data broadcasting method and system. BACKGROUND

[0002] The existing VRS system generally includes three parts: a reference station, a data processing and service center, and a terminal. The reference station, as the ground basis of the VRS system, needs to build a network composed of multiple continuously running reference stations. These stations are distributed in a triangular shape and collectively cover a specific area. The network contains at least three reference stations. Each reference station is connected in real time with the data processing center through a dedicated communication line and continuously uploads its observation data. The data processing and service center is the core of the entire VRS system, integrating communication control and data calculation functions. Its main work includes: establishing high-speed data connection with all reference stations using dedicated lines, receiving and fusing observation data from each station in real time; performing complex calculation and processing on these data; generating differential correction data that meet the needs of mobile users; and broadcasting these high-precision differential information to users through wireless communication networks. The terminal refers to mobile users with positioning needs, and the mobile users are equipped with GNSS receivers and wireless communication modules (modems). The receiver captures the original satellite signal and sends the user's approximate position information to the data center through the wireless network. Then, the receiver obtains the differential correction data issued by the data center, combines the original observation value and the correction information for real-time dynamic processing, and finally outputs the high-precision position result.

[0003] With the widespread application of RTK technology and the continuous growth of the number of terminals in the high-precision positioning technology field, designing and implementing a high-availability VRS broadcasting system becomes a crucial link. The existing broadcasting technology cannot meet the needs of stateless and multi-node deployment between services. The differential broadcasting system in the current industry generally uses the following architecture for differential data broadcasting:

[0004] Single-node centralized method: network solving and differential data generation are performed by a single service node. This architecture is centralized in calculation and simple in structure, but node single-point failure will directly lead to service interruption, making it difficult to meet the needs of high-availability scenarios.

[0005] Network solving separation method: network is solved by a single service node, and then each VRS node independently generates differential data based on the unified network result. Although this method generates a unified network result, the network node still has a single-point failure risk, and the overall system robustness is limited.

[0006] Therefore, the defects of the prior art include: lack of dynamic quality evaluation: key indicators such as residual consistency and network integrity of the difference data are not quantified; static election strategy defects: traditional cache election relies on time stamp or simple majority voting, ignoring the data quality dimension (such as residual consistency, network health), resulting in invalid data in the occlusion area; poor availability: single service node networking and calculation is easy to cause single point failure of the service. There is an urgent need for a method for optimizing the service availability of difference data broadcast in a multi-base station networking environment, to improve the stability of the VRS system in multiple application scenarios, especially in urban occlusion scenarios.

[0007] Chinese Patent Publication No. CN106912076A discloses a multi-terminal multi-application difference data broadcast system, which includes a difference broadcast platform, at least one reference station and at least one terminal device; the terminal device includes at least one application; the difference broadcast platform aggregates the difference data uploaded by the reference station; the difference broadcast platform performs intelligent aggregation operation on the difference data required by the terminal device, and broadcasts the difference data after intelligent aggregation to the terminal device; after receiving the broadcasted difference data, the terminal device performs scheduling and distribution according to the actual situation of the multi-application. The invention relies on a single difference broadcast platform to aggregate and process data, which has a single point failure risk, and does not involve quantitative evaluation of the quality of the difference data itself (such as residual consistency and network integrity). SUMMARY

[0008] The purpose of the present application is to provide a high-availability VRS difference data broadcast method and system to solve the above problems, which elects the difference data generated by dynamic multi-dimensional evaluation algorithm, time and space adaptive weight adjustment and timeout fuse mechanism.

[0009] The technical solution of the present application is as follows:

[0010] On the one hand, the present application provides a high-availability VRS difference data broadcast method, which includes the following steps:

[0011] A plurality of distributed algorithm nodes deployed in a micro-service architecture are containerized and deployed by networking the base stations, each algorithm node calculates and generates data quality metadata including difference data, base station number, data generation timestamp, observation residual standard deviation and observation data integrity rate;

[0012] Each algorithm node performs an election service, calculates a multi-dimensional dynamic score based on the data quality metadata, and selects the algorithm node with the highest score and the algorithm node of the previous version election service result as a candidate set based on the multi-dimensional dynamic score sorting and screening;

[0013] If the difference between the highest score of the algorithm node and the highest score of the historical cache election service result is less than a set threshold value, the historical cache election service result with the highest score is taken as the election service result, and the differential data corresponding to the historical cache election service result with the highest score is broadcast to each terminal; otherwise, the algorithm node with the highest score is taken as the election service result for broadcasting.

[0014] Preferably, the differential data is calculated by a differential data generation algorithm including a differential correction number generation algorithm of network RTK; the base station number is obtained by analyzing the base station code received by the algorithm node to count the number of base stations participating in the calculation; the data generation timestamp is the timestamp of the differential data obtained at the end of the calculation; and the observation residual standard deviation calculation is specifically as follows:

[0015]

[0016] wherein, is the observation residual standard deviation; is the base station is the residual of the actual observation data and the model differential data fitting value calculated by the algorithm node, is the residual mean, is the base station number, is the base station index;

[0017] The observation data integrity rate = (effective observation epoch number / theoretical total epoch number) * 100%.

[0018] Preferably, the multi-dimensional dynamic score calculation is specifically as follows:

[0019]

[0020] wherein, is the multi-dimensional dynamic score; is the is the dimension weight, and the dimensions include network integrity, data freshness, historical stability, residual consistency and network health degree; is the is the dimension index, and the dimension index is processed by normalization; is the dimension category index;

[0021] The dimension index calculation of the five dimensions of network integrity, data freshness, historical stability, residual consistency and network health degree is as follows: the network integrity is calculated by the effective base station number / theoretical base station number; the data freshness is calculated by 1 / (current time-data generation timestamp); the historical stability is calculated by the proportion of the algorithm nodes selected as the election service result in the preset historical window; the residual consistency is calculated by 1 / base station observation residual standard deviation; and the network health degree is calculated by 1-observation data integrity rate average of the algorithm nodes and the base station.

[0022] The dimension weight is dynamically adjusted based on the GGA data uploaded by the terminal, specifically: the specific space-time scene where the terminal is located is determined based on the GGA data to automatically adjust the residual consistency dimension weight and the network integrity dimension weight, and the motion state of the terminal is determined to dynamically adjust the data freshness.

[0023] Preferably, the election service in the method is a stateless service deployed by multiple algorithm nodes, and the election service process is performed by distributed locking, specifically:

[0024] When the election service is started, each algorithm node simultaneously attempts to acquire the distributed lock, and the algorithm node that successfully acquires the distributed lock performs the election service process, publishes the data version number corresponding to the election service result to the shared storage or configuration center, and reads the data quality metadata corresponding to the data version number by the broadcast service of the shared storage or configuration center to perform differential data broadcast, and after the election service is completed, the algorithm node holding the distributed lock actively releases the distributed lock so that other algorithm nodes participate in the competition of the next round of election service.

[0025] Preferably, when the election service is performed in the method, if it is the first election service and there is no historical cache highest-score election service result, the highest-score algorithm node is directly taken as the election service result for broadcast; if there is a historical cache highest-score election service result, the election service result optimization step of difference comparison is entered.

[0026] Meanwhile, the election timeout time of the election service is preset, and if the election service result is not elected within the longest election timeout time, the timeout is fused, and the last version of the election service result is automatically rolled back for broadcast.

[0027] On the other hand, the application provides a high-availability VRS differential data broadcast system, which comprises algorithm nodes, an election service and a shared storage or configuration center.

[0028] The algorithm nodes are used for containerized deployment in a microservice architecture and networking by base stations, each algorithm node calculates and generates data quality metadata including differential data, base station number, data generation timestamp, observation residual standard deviation and observation data integrity rate.

[0029] electing service, for each algorithm node to perform the electing service, based on the data quality metadata, to calculate a multi-dimensional dynamic score for the multi-algorithm node, based on the multi-dimensional dynamic score ranking, to filter the algorithm node with the highest score and the algorithm node of the previous version electing service result as a candidate set, if the score difference between the algorithm node with the highest score and the highest score electing service result in the history cache is less than a set threshold, then the highest score electing service result in the history cache is taken as the electing service result, otherwise the algorithm node with the highest score is taken as the electing service result for broadcasting;

[0030] shared storage or configuration center, for broadcasting the differential data to each terminal.

[0031] In still another aspect, the present application also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of the embodiments of the present application when executing the computer program.

[0032] In still another aspect, the present application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the method according to any one of the embodiments of the present application.

[0033] Compared with the prior art, the present application has the following technical effects:

[0034] 1. Improve system availability: the present application realizes stateless networking, calculation and broadcasting of VRS high-precision positioning service by deploying algorithm nodes, thereby improving the robustness of the system.

[0035] 2. Reduce positioning conflict rate: the present application uses a candidate set method combined with a score difference threshold to implement a cache update strategy, which not only reduces the positioning conflict rate, but also improves the stability of data broadcasting, thereby avoiding the problem of continuous switching of virtual station coordinates caused by continuous switching of differential data in the cache.

[0036] 3. Improve high occlusion area positioning success rate: the present application improves the accuracy and success rate of data broadcasting in high occlusion area positioning by dynamically adjusting the weight, thereby enhancing the complex urban area adaptability of VRS service. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 is the overall flowchart of the high-availability VRS differential data broadcasting method according to the present application. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme of the present application will be described in detail below with reference to the embodiments of the present application and the accompanying drawings.

[0039] Embodiment One

[0040] The embodiment provides a high-availability VRS differential data broadcasting method, referring to FIG. 1, comprising the following steps: Figure 1

[0041] A plurality of distributed algorithm nodes deployed in a micro-service architecture container are networked by base stations (reference stations), each algorithm node generates data quality metadata including differential data, base station number, data generation timestamp, observation residual standard deviation, and observation data integrity rate. After the algorithm node generates the data quality metadata, it is stored in the local cache pool for the election service to pull and evaluate, and a version number generated by the timestamp + node ID is added to ensure uniqueness.

[0042] As a preferred embodiment of the embodiment, the differential data is calculated by a differential data generation algorithm including a differential correction number generation algorithm of network RTK. Specifically, in a container deployment environment based on a micro-service architecture, each algorithm node (such as node 1 to node N) independently runs a solving module. The core task of the module is to process the observation data received from multiple base stations, and calculate the final differential data through a specific algorithm model (such as the differential correction number generation algorithm in network RTK, the differential GPS algorithm (DGPS) using single-frequency code observation as the main observation value, and the wide-area DGPS algorithm (WADGPS) and the like).

[0043] The base station number is obtained by analyzing the base station code received by the algorithm node to count the number of base stations currently participating in the solution. Specifically, after starting, the solving module will continuously receive observation data streams from multiple base stations. Each data packet contains a unique identifier (base station code) of its source base station. The module will maintain a list or counter of active base stations inside, and dynamically count the number of base stations currently participating in the solution by real-time analysis of these identifiers. When new base station data flows in or existing base station connection is interrupted, this count will be updated, so as to reflect the scale of the base stations participating in the current grid solution in real time.

[0044] The data generation timestamp is the timestamp of the differential data obtained at the end of the solution. Specifically, in order to measure the calculation efficiency, the solving module will enter a timestamp when a complete solution task is calculated and the differential data is ready.

[0045] The observation residual standard deviation is a core indicator for evaluating data quality. During the solution process, the algorithm model will generate a fitting solution (i.e. model predicted value) using base station observation data. The residual refers to the difference between the actual observation value of the base station and the model fitting value. The solving module will calculate the residual of all base stations or key base stations participating in the solution to form a residual sequence. The observation residual standard deviation is the standard deviation of this residual sequence, which is calculated as follows:​

[0046]

[0047] wherein, is the observation residual standard deviation; is the base station the residual of the actual observation data and the model fitting value calculated by the algorithm node, is the residual mean, is the number of base stations, is the base station index.

[0048] The smaller the observation residual standard deviation value, the better the observation data is fitted by the model, the higher the internal consistency accuracy of the calculation result, and the more reliable the data quality.

[0049] The observation data completeness rate = (number of valid observation epochs / total theoretical epoch number) * 100%. Specifically, the calculation module relies on stable and reliable base station data stream. The network monitoring component inside or cooperated with the calculation module will record the number of epochs that should be received from each base station (based on ephemeris, base station coordinates and observation data sampling interval) and compared with the number of epochs actually successfully received. This index directly reflects the health status of the data transmission link, and low observation data completeness rate will directly affect the integrity and reliability of the calculation result.

[0050] Each algorithm node performs an election service, calculates a multi-dimensional dynamic score based on data quality metadata, and sorts and filters the highest score algorithm node and the algorithm node of the previous version election service result as a candidate set based on the multi-dimensional dynamic score. At the same time, if the data difference in the candidate set is too large, trigger BFT consensus mechanism to ensure the consistency of the result.

[0051] As a preferred embodiment of the present embodiment, the multi-dimensional dynamic score calculation is specifically:

[0052]

[0053] In the formula, is the multi-dimensional dynamic score; is dimension weight, the dimensions include network integrity, data freshness, historical stability, residual consistency and network health degree; is dimension index, the dimension index is processed by normalization to eliminate the difference in dimension and order of magnitude between different dimensions, and to ensure that each index can be weighted and summed by power operation; The dimension category index is further used in the dynamic scoring formula in this embodiment. The advantage of using the weighted power operation to calculate the product over the simple weighted summation is that if the weighted summation is used, a high score in one dimension can compensate for a low score in another dimension. When the weighted power operation to calculate the product is used to calculate the multi-dimensional dynamic score, a low score in any one dimension will significantly lower the total score, ensuring that the selected algorithm node has no obvious short board in all dimensions, and better ensuring the overall quality and reliability of the selected differential data source.

[0054] The network integrity reflects the base station integrity of the network, the data freshness reflects the real-time nature of the data, the historical stability is the proportion of the node data being selected as the optimal within a preset time window in the past, the residual consistency is used to quantify the data accuracy, and the network health degree is used to ensure the transmission reliability. The dimension index calculation of the five types of dimensions is as follows: the network integrity is calculated by the effective base station number / theoretical base station number; the data freshness is calculated by 1 / (current time-data generation timestamp); the historical stability is the proportion of the algorithm node being selected as the elected service result within a preset historical window; the residual consistency is calculated by 1 / base station observation residual standard deviation; and the network health degree is calculated by 1-algorithm node and base station observation data integrity rate average.

[0055] The embodiment performs spatio-temporal adaptive weight distribution according to the regional type and base station data quality, breaks through the limitation of static weight, first integrates the environmental scene and network state into weight distribution, and improves the adaptability in complex urban areas. The dimension weight is dynamically adjusted based on the GGA data uploaded by each terminal. Specifically, the GGA data provides the longitude and latitude of the terminal position, time stamp, number of satellites, positioning state (such as RTK fixed solution, floating point solution) and other key spatio-temporal information. After obtaining the spatio-temporal characteristics of the terminal position through the GGA data, the specific spatio-temporal scene where the terminal is located is determined to automatically adjust the residual consistency dimension weight and the network integrity dimension weight. The motion state of the terminal (the coordinates and time of the terminal can be obtained through continuous GGA data to calculate the speed of the terminal) is determined to dynamically adjust the data freshness, and the dynamic optimization of “adapting to local conditions” is realized. Specifically, after determining the space scene where the terminal is located, if the terminal is located in a high-rise area or a valley, the residual consistency weight should be increased (the terminal is in a city valley, and the number of observed satellites is small. At this time, satellite observation data with higher quality is needed. The inconsistency of residual error indicates that there may be errors (such as multipath effect and sudden change of atmospheric delay) that are not effectively eliminated by the model, which indicates that the current calculation result is unreliable). The numerical range of the adjustment can be determined according to the actual application scene, which is not limited here. The preferred range of the embodiment that can be increased is 0.2-0.5, and the default value is 0.35. If the terminal is located in the suburb, the network integrity weight should be increased. The numerical range of the adjustment can be determined according to the actual application scene, which is not limited here. The preferred range of the embodiment that can be increased is 0.35-0.45, and the default value is 0.4. After determining the time dimension where the terminal is located, when the base station disconnection rate (calculated based on base station observation data) is greater than a certain value (such as more than 20%), the network health weight should be increased. The numerical range of the adjustment can be determined according to the actual application scene, which is not limited here. The preferred range of the embodiment that can be increased is 0.15-0.25, and the default value is 0.2. If the terminal is in a fast motion state (such as vehicle navigation), or the time stamp of the GGA data shows that the data delay is large, the weight of data freshness should be increased to ensure the real-time performance of positioning.

[0056] If the difference between the highest score of the algorithm node and the highest score of the historical cache election service result is less than the set threshold value, the historical cache election service result with the highest score is taken as the election service result to ensure data stability, the differential data corresponding to the historical cache election service result with the highest score is broadcast to each terminal, and the fixed state index of each algorithm node is recorded for subsequent weight update; otherwise, the algorithm node with the highest score is taken as the election service result for broadcasting. The variance threshold value can be adjusted according to the deployment environment of the current system, which is not limited here. The threshold value range set in the embodiment is preferably 10-30 points.

[0057] As the preferred embodiment of the present embodiment, in order to avoid the influence of single node accidental failure on the robustness of the result, while reducing the complexity of subsequent conflict processing. In the method, the election service is a stateless service deployed in the multi-algorithm node, and the election service process is performed through the distributed lock. If a node grabs the distributed lock, the cache election is performed. Specifically:

[0058] When the election service starts, each algorithm node simultaneously attempts to obtain the distributed lock of the specified key. The algorithm node that successfully obtains the distributed lock performs the election service process, and publishes the data version number corresponding to the election service result to the shared storage or configuration center, so that the shared storage or configuration center broadcast service reads the data quality metadata corresponding to the data version number to perform differential data broadcast. After the election service is completed, the algorithm node holding the distributed lock actively releases the distributed lock, so that other algorithm nodes participate in the competition of the next round of election service.

[0059] As the preferred embodiment of the present embodiment, when the election service is performed in the method, if it is the first election service, there is no historical cache highest score election service result, then the highest score algorithm node is directly taken as the election service result for broadcast; if there is a historical cache highest score election service result, then the election service result optimization step of difference comparison is entered;

[0060] At the same time, the election timeout time of the election service is preset. If the election service result is not elected within the longest election timeout time, then the timeout is fused, and the last version of the election service result is automatically rolled back for broadcast. The preset election timeout time can be set according to the actual application scene, which is not limited herein, and the reference range of the present embodiment is within 30-100ms.

[0061] In order to verify the effectiveness and superiority of the method provided by the present embodiment, some specific cases are provided as follows:

[0062] A typical scene with network fluctuations and partial base station shielding is simulated, and the experimental environment is configured as follows:

[0063] Hardware platform: 3 independent algorithm service nodes (configured the same), 1 broadcast service node, and a simulation terminal.

[0064] Software environment: micro-service architecture, using Redis Cluster as distributed cache and lock, and Kubernetes for container orchestration.

[0065] Test scene: simulate the urban viaduct area, set 15% of the link loss rate of one key base station to make the network inconsistent.

[0066] Data source: Real-time multi-satellite system observation data is injected into the simulation environment, including complete pseudorange, carrier phase and ephemeris information.

[0067] The core of the dynamic multi-dimensional evaluation election algorithm is its dynamic weight mechanism. The following are the key parameter settings in the experiment:

[0068] Dynamic multi-dimensional evaluation election algorithm dimension weight (ω): ω1 (network integrity) = 0.30, ω2 (data freshness) = 0.20, ω3 (historical stability) = 0.15, ω4 (residual consistency) = 0.25, ω5 (network health) = 0.10.

[0069] Dynamic adjustment rule: When the terminal latitude, longitude and elevation position extracted from the terminal GGA data is located in a canyon, automatically increase the weight of ω4 (residual consistency) to 0.35, and reduce ω2 (data freshness) to 0.15.

[0070] Election trigger mechanism: Distributed lock is used, and the node that grabs the lock performs the election, with an election period of 30 seconds.

[0071] Baseline comparison: Traditional broadcasting method: uses a simple "earliest timestamp" strategy for cache election.

[0072] In the 24-hour simulation test, the dynamic multi-dimensional evaluation election algorithm shows significant advantages over the traditional broadcasting method in all key indicators. As shown in Table 1, the comparison results of the core indicators are shown.

[0073] Table 1 Comparison results table

[0074]

[0075] Result analysis:

[0076] Convergence time improvement: Multi-dimensional dynamic scoring prioritizes data with optimal observation quality through residual consistency, and uses dynamic weights to strengthen precision indicators when terminals move to canyons, enabling terminals to obtain stable and reliable corrections faster, thereby accelerating convergence.

[0077] Conflict rate is greatly reduced: Traditional methods are not sensitive to network differences caused by base station disconnection. Multi-dimensional dynamic scoring can effectively identify abnormal data caused by network problems, greatly improving the consistency of broadcast data.

[0078] Performance improvement in high occlusion area: In areas with poor satellite signal, the multi-dimensional dynamic scoring algorithm can adaptively select differential data with more complete network and more consistent residuals, providing stronger support for terminals in harsh environments and ensuring the continuity and reliability of positioning.

[0079] The embodiment shows that the multi-dimensional dynamic scoring and the election service applied to the differential data broadcasting system can effectively deal with the challenges brought by node service heterogeneity and network environment fluctuation. The embodiment improves the reliability of broadcast data and the positioning performance of terminals through intelligent data quality evaluation and election strategy.

[0080] Embodiment Two

[0081] Correspondingly, the embodiment provides a high-availability VRS differential data broadcasting system for implementing the method of the embodiment one, including an algorithm node, an election service and a shared storage or configuration center.

[0082] The algorithm node is used for containerized deployment in a micro-service architecture and networking through base stations, each algorithm node calculates and generates data quality metadata including differential data, base station number, data generation timestamp, observation residual standard deviation and observation data integrity rate.

[0083] The election service is used for each algorithm node to perform election service, multi-dimensional dynamic scoring of multiple algorithm nodes is calculated based on the data quality metadata, the highest-scored algorithm node and the algorithm node of the last version election service result are selected as a candidate set based on multi-dimensional dynamic scoring sorting and screening, if the score difference between the highest-scored algorithm node and the historical cache highest-scored election service result is less than a set threshold, the historical cache highest-scored election service result is taken as the election service result, otherwise the highest-scored algorithm node is taken as the election service result for broadcasting.

[0084] The shared storage or configuration center is used for broadcasting differential data to each terminal.

[0085] Embodiment Three

[0086] The embodiment provides an electronic device, including a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the method of the embodiment one when executing the computer program.

[0087] Embodiment Four

[0088] The embodiment provides a computer readable storage medium, which stores a computer program, and the computer program is executable on a processor to implement the method of the embodiment one.

[0089] In the embodiments of the present application, “at least one” means one or more, and “multiple” means two or more. “And / or” describes the association relationship of the associated objects, which means that there can be three kinds of relationships, for example, A and / or B can represent the cases of A alone, A and B together, and B alone. Wherein A and B can be singular or plural. The character “ / ” generally represents an “or” relationship between the front and rear associated objects. “At least one of the following” and the like means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, wherein a, b, and c can be single or multiple.

[0090] Those skilled in the art can appreciate that the units and algorithm steps described in the embodiments disclosed herein can be realized by electronic hardware, computer software and combination of electronic hardware and computer software. Whether the functions are realized in hardware or software mode depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0091] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0092] In several embodiments provided by the present application, any function realized in the form of a software function unit and sold or used as an independent product can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the parts that make contributions to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM for short), a random access memory (Random Access Memory, RAM for short), a magnetic disk or an optical disk, and various media that can store program codes.

[0093] The above only describes the embodiments of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent flow transformation based on the content of the specification and drawings of the present application, or direct or indirect application in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A highly available VRS differential data broadcasting method, characterized in that, Includes the following steps: Several distributed algorithm nodes deployed in a containerized manner using a microservice architecture are networked with base stations. Each algorithm node calculates and generates data quality metadata, including differential data, number of base stations, data generation timestamp, observation residual standard deviation, and observation data integrity rate. The number of base stations is obtained by statistically analyzing the base station codes received by the parsing algorithm nodes to count the number of base stations currently participating in the solution; the data generation timestamp is the timestamp of obtaining the differential data after the solution is completed; the calculation of the observation residual standard deviation is specifically as follows: in, The standard deviation of the observed residuals; For base stations The residual between the actual observed data and the model difference data fitted values ​​solved by the algorithm nodes. The mean of the residuals, For the number of base stations, For base station indexing; The observation data integrity rate = (number of valid observation epochs / total theoretical epochs) * 100%; Each algorithm node performs an election service, calculates multi-dimensional dynamic scores for multiple algorithm nodes based on data quality metadata, and sorts and selects the algorithm node with the highest score and the algorithm node from the previous version's election service results as a candidate set based on the multi-dimensional dynamic scores. If the difference between the highest-scoring algorithm node and the historical cached highest-scoring election service result is less than the set threshold, then the historical cached highest-scoring election service result will be used as the election service result, and the difference data corresponding to the historical cached highest-scoring election service result will be broadcast to each terminal; otherwise, the highest-scoring algorithm node will be used as the election service result and broadcast.

2. The high-availability VRS differential data broadcasting method according to claim 1, characterized in that, The differential data is calculated using a differential data generation algorithm, including a differential correction generation algorithm for network RTK.

3. The high-availability VRS differential data broadcasting method according to claim 1, characterized in that, The multidimensional dynamic scoring calculation is specifically as follows: In the formula, Multidimensional dynamic scoring; for The weights are categorized by dimensions, including network integrity, data freshness, historical stability, residual consistency, and network health. for Category dimension metrics, wherein the dimension metrics are processed by normalization; Indexed by dimension categories; The five dimensions of network integrity, data freshness, historical stability, residual consistency, and network health are calculated as follows: network integrity is calculated by dividing the number of effective base stations by the theoretical number of base stations; data freshness is calculated by dividing 1 by (current time - data generation timestamp); and historical stability is calculated by the proportion of algorithm nodes selected as election service results within a preset historical window. The residual consistency is calculated by 1 / standard deviation of the base station observation residuals; the network health is calculated by 1 - the mean of the observation data integrity rates of the algorithm nodes and the base stations. The dimensional weights are dynamically adjusted based on the GGA data uploaded by the terminal. Specifically, the specific spatiotemporal scene of the terminal is determined by the GGA data to automatically adjust the residual consistency dimension weight and the network integrity dimension weight, and the motion state of the terminal is determined to dynamically adjust the data freshness.

4. The high-availability VRS differential data broadcasting method according to claim 1, characterized in that, The election service in the method is a stateless service deployed across multiple algorithm nodes. The election service process is executed through distributed locking, specifically as follows: When the election service starts, each algorithm node simultaneously attempts to acquire the distributed lock. The algorithm node that successfully acquires the distributed lock executes the election service process and publishes the data version number corresponding to the election service result to the shared storage or configuration center. The broadcast service of the shared storage or configuration center then reads the data quality metadata corresponding to the data version number for differential data broadcasting. After the election service is completed, the algorithm node holding the distributed lock actively releases the distributed lock so that other algorithm nodes can participate in the competition for the next round of election service.

5. The high-availability VRS differential data broadcasting method according to claim 1, characterized in that, When performing the election service in the method, if it is the first election service and there is no historical cache of the election service result with the highest score, the algorithm node with the highest score is directly broadcast as the election service result. If there is a historical cached election service result with the highest score, then proceed to the election service result optimization step of score difference comparison; At the same time, a timeout period for the election service is preset. If no election result is obtained within the longest timeout period, the timeout circuit breaker will be triggered, and the election result of the previous version will be automatically rolled back for broadcast.

6. A highly available VRS differential data broadcasting system, characterized in that, The system is used to implement the method as described in any one of claims 1 to 5, and includes an algorithm node, an election service module, and a shared storage or configuration center; Algorithm nodes are used for containerized deployment in a microservice architecture and for networking base stations. Each algorithm node calculates and generates data quality metadata, including differential data, number of base stations, data generation timestamp, observation residual standard deviation, and observation data integrity rate. The election service module is used for each algorithm node to perform election services. It calculates multi-dimensional dynamic scores for multiple algorithm nodes based on data quality metadata. Based on the multi-dimensional dynamic scores, it sorts and selects the algorithm node with the highest score and the algorithm node with the previous version's election service result as a candidate set. If the difference between the highest score algorithm node and the historical cached highest score election service result is less than a set threshold, the historical cached highest score election service result is used as the election service result; otherwise, the algorithm node with the highest score is used as the election service result and broadcast. A shared storage or configuration center is used to broadcast differential data to each terminal.

7. An electronic device, the electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 5.

Images