Distributed intelligent data management system and method based on cloud computing

Abstract

The invention provides a distributed intelligent data management system and method based on cloud computing, and relates to the technical field of intelligent communication. The method comprises the following steps: performing time slice segmentation and snapshot mapping analysis processing on a user terminal interaction signaling data stream to obtain an accurate and continuous signaling track data set; constructing an electronic fence logic network containing a parent region grid and a child hotspot grid, calculating a user terminal signal residence confidence coefficient, and screening out a target user terminal meeting a deep residence condition; calling target user terminal interaction information, executing rule mutual exclusion verification of the reverse sliding time window, and generating a service release instruction; and calling the cloud structured data template to render content in real time, and scheduling and issuing a personalized service message data packet. According to the invention, accurate identification of user behaviors and safe and efficient pushing of personalized information are realized, and service accuracy and user experience are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent communication technology, specifically a distributed intelligent data management system and method based on cloud computing. Background Technology

[0002] With the rapid development of mobile communication and cloud computing technologies, intelligent information management services based on user location and real-time behavior have gradually become one of the important development directions in the communications field. Especially in densely populated and complex areas such as airports, train stations, tourist attractions, and commercial centers, users' behavior patterns are diversified, and the requirements for the accuracy of information services are constantly increasing. Therefore, how to effectively manage the real-time signaling data generated by massive user terminals by combining cloud computing technology, and accurately push personalized service information based on users' real-time location and dynamic behavior characteristics, has become an important research topic in the field of intelligent information management.

[0003] Existing intelligent information push solutions mainly rely on preset geofencing or simple location triggering mechanisms to push information. However, these solutions generally have the following shortcomings in practical applications: On the one hand, traditional triggering mechanisms cannot accurately identify the real-time behavioral differences of users in a specific area, such as the difficulty in distinguishing whether a user is passing by briefly, staying temporarily, or staying for a long time; on the other hand, existing methods lack continuous and refined analysis of user terminal signaling data, making it difficult to achieve accurate matching between user location and scene, resulting in a lack of targeted information push and difficulty in ensuring user experience.

[0004] Therefore, there is an urgent need to propose a distributed intelligent data management method based on cloud computing. This method can effectively process and analyze discrete signaling data streams generated by user terminals, accurately identify users' real-time location and dwell behavior characteristics, and achieve refined user scenario positioning and personalized information push, thereby improving the accuracy of intelligent information services and user satisfaction. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a distributed intelligent data management system and method based on cloud computing.

[0006] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides a distributed intelligent data management system based on cloud computing, the system comprising: Signaling trajectory data parsing module: used to parse and process the interactive signaling data stream of user terminals by segmenting it by time slicing and combining it with snapshot mapping to obtain the signaling trajectory dataset; Electronic fence dwell analysis module: used to pre-build an electronic fence logical network containing parent area grids and child hotspot grids, calculate signal dwell confidence based on the signaling trajectory dataset, and filter target user terminals that meet the deep dwell conditions; Service release rule verification module: used to retrieve the interaction information corresponding to each target user terminal, perform rule mutual exclusion verification of the reverse sliding time window respectively, and generate the corresponding service release instruction; Personalized service content scheduling module: For each service release instruction, it calls the cloud-based structured data template for real-time content rendering, generates and schedules the distribution of the corresponding personalized service message data packet.

[0007] Furthermore, the interactive signaling data stream includes a unique terminal hardware identifier, a base station cell identifier, a reference signal received power, and a signaling interaction timestamp; the parsing process using time slicing combined with snapshot mapping includes: Multiple consecutive sampling time windows are determined according to a preset fixed period; Within each sampling time window, the interactive signaling data stream is processed by state extraction based on a snapshot mapping mechanism to obtain a multidimensional instantaneous value group corresponding to each sampling time window; Based on the time sequence of the multiple sampling time windows, the multidimensional instantaneous numerical groups are processed into time series to obtain the signaling trajectory dataset.

[0008] Furthermore, the method for constructing the electronic fence logical network comprising a parent region grid and child hotspot grids includes: The target region is divided into a first spatial grid, and multiple independent parent region grids are constructed. The parent region grid is subjected to a second spatial grid division process to construct the corresponding child hotspot grid; For each parent region grid and its corresponding child hotspot grid, associate scene attribute codes and security policy tags respectively; Based on the spatial relationship between each parent region grid and its corresponding child hotspot grid, spatial index association processing is performed to obtain an electronic fence logical network containing parent region grids and child hotspot grids.

[0009] Furthermore, the method for calculating the signal dwell confidence includes: Based on the signaling trajectory dataset, the cumulative number of base station cell identifier changes for each user terminal is counted to obtain the corresponding base station handover frequency. Based on the signaling trajectory dataset, the numerical variance of the reference signal received power of each user terminal in the time dimension is calculated to obtain the corresponding signal strength fluctuation value. The signal instability index of each user terminal is obtained by weighting the base station handover frequency and the signal strength fluctuation value based on a preset weighting coefficient. The signal instability index is subjected to inverse normalization to obtain the signal dwell confidence level of each user terminal.

[0010] Furthermore, the screening of target user terminals that meet the deep residency criteria includes: The signal dwell confidence level is compared with a preset dwell confidence threshold to obtain a dwell confidence determination result; Based on the signaling interaction timestamps of the signaling trajectory dataset, the dwell time span of each user terminal within the electronic fence logical network is determined and compared with the preset minimum dwell time to obtain the time span determination result. Based on the parent area grid of the electronic fence logical network, the base station cell identifier of the user terminal is spatially matched to obtain the parent area grid matching result; Based on the parent region grid matching result and the reference signal received power, and combined with the preset signal feature mapping library, further spatial matching is performed with the corresponding child hotspot grid to obtain the child hotspot grid matching result. Based on the retention confidence determination result, time span determination result, and sub-level hotspot grid matching result, a comprehensive judgment is made to determine the target user terminal that meets the deep retention condition.

[0011] Furthermore, the interaction information includes: historical interaction records of the user terminal, current system time, user identity level, and other data; the mutual exclusion check of the rules for executing the reverse sliding time window includes: Based on the sub-level hotspot grid matching results of the target user terminal, determine the associated scene attribute code and security policy label; Using scene attribute codes as index keys, construct reverse sliding time windows respectively; Based on the historical interaction records of the target user terminal, perform a comparison of records with the same code within the reverse sliding time window to obtain the result of the determination of records with the same code; The security compliance assessment result is determined based on the security policy label, current system time, and identity level. Based on the results of the same code record determination and the security compliance determination, a comprehensive judgment is made to generate the corresponding service release instruction.

[0012] Furthermore, the method for determining the security compliance assessment result includes: Pre-configure security policy labels that include blacklists, whitelists, and corresponding restricted periods; The target user's terminal identity level is compared with the blacklist and whitelist respectively to determine the list matching result; Match the current system time with the prohibited time period to determine the prohibited time period judgment result; Based on the list matching results and the results of the prohibited period determination, a comprehensive judgment is made to determine the security compliance determination result of the target user terminal.

[0013] Furthermore, the method for real-time content rendering includes: Based on the unique identifier of the target user's terminal hardware, a pre-built user profile database is retrieved to obtain the corresponding set of business attribute fields; Based on the scenario attribute code of the target user terminal, determine the business attribute fields corresponding to the current scenario; Based on cloud-based structured data templates, identify dynamic placeholders for corresponding business attribute fields in the template; Fill the business attribute fields into dynamic placeholders to complete real-time content rendering and obtain personalized service message data packets.

[0014] Furthermore, the method for scheduling and distributing personalized service message data packets includes: Based on the scenario attribute code corresponding to the target user terminal, determine the sending priority marker of the personalized service message data packet; Personalized service message data packets are divided into corresponding buffer queues according to the sending priority flag, forming a multi-level buffer queue; Monitor the load status of the communication gateway link, and dynamically adjust the data packet delivery strategy of different priority buffer queues based on the comparison result of the link load status and the preset load threshold, so as to complete the scheduling and delivery of personalized service message data packets.

[0015] Secondly, the present invention provides a distributed intelligent data management method based on cloud computing, the method comprising: The interactive signaling data stream of the user terminal is parsed and processed by time slicing and snapshot mapping to obtain the signaling trajectory dataset; A pre-constructed electronic fence logical network containing parent region grids and child hotspot grids is used to calculate the signal dwell confidence based on the signaling trajectory dataset and to filter target user terminals that meet the deep dwell conditions. Retrieve the interaction information corresponding to each target user terminal, perform rule mutual exclusion verification of the reverse sliding time window respectively, and generate the corresponding service release instruction; For each service release command, a cloud-based structured data template is invoked for real-time content rendering, generating and dispatching the corresponding personalized service message data packet.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This invention effectively solves the problem of inaccurate user behavior analysis caused by the discreteness of signaling data in the prior art by performing continuous parsing processing of user terminal interaction signaling data stream through time slicing and snapshot mapping, and realizes the accurate construction of user behavior trajectory data, thereby improving the analysis accuracy of intelligent data management.

[0017] This invention constructs an electronic fence logic network that includes parent area grids and child hotspot grids, and significantly improves the accuracy of user dwell behavior identification based on a comprehensive judgment method that combines signal dwell confidence, dwell time, and fine grid matching. This overcomes the shortcomings of existing technologies that cannot accurately distinguish between short-term and deep dwell states of users.

[0018] This invention effectively ensures the security, relevance, and efficiency of personalized service message push by performing rule-based mutual exclusion verification on user terminal interaction information through a reverse sliding time window, and by combining cloud-based structured data templates to implement real-time content rendering and message scheduling and delivery mechanisms. It solves the problems of inaccurate message push, network congestion, and delay in existing smart information services, and significantly improves user service experience and information push efficiency. Attached Figure Description

[0019] Figure 1 This is an architecture diagram of a cloud-based distributed intelligent data management system, as shown in Example 1.

[0020] Figure 2 This is a flowchart of a cloud computing-based distributed intelligent data management method, as shown in Example 2. Detailed Implementation

[0021] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1 Please see Figure 1 This invention provides a cloud computing-based distributed intelligent data management system, comprising: Signaling trajectory data parsing module: used to parse and process the interactive signaling data stream of user terminals by segmenting it by time slicing and combining it with snapshot mapping to obtain the signaling trajectory dataset; It should be noted that the interactive signaling data stream is a discrete signaling event data sequence generated by the user terminal in the mobile communication network environment due to user behavior (such as location update, base station handover, call or SMS sending, etc.); this event-driven data sequence is usually discontinuous in the time dimension, that is, there may be a lack of event records in certain time periods.

[0023] In order to effectively analyze user behavior, this embodiment adopts a combination of time slicing and snapshot mapping to transform the original discrete signaling data into a continuous and unified time-series dataset.

[0024] The interactive signaling data stream includes a unique terminal hardware identifier, a base station cell identifier, a reference signal received power, and a signaling interaction timestamp; the parsing process using time slicing combined with snapshot mapping includes: Multiple consecutive sampling time windows are determined according to a preset fixed period; It should be noted that the preset fixed period can be flexibly set according to the needs of the actual application scenario. For example, a shorter period (such as 30 seconds or 1 minute) can be selected for densely populated areas such as airports and train stations, while a longer period (such as 5 minutes or 10 minutes) can be selected for low-frequency event areas such as highways. The time window can be specifically represented as follows: ; In the formula, Represents the set of sampling time windows within the observation period. This represents the k-th time window, and n represents the total number of time windows. and These represent the start and end times of the k-th time window, respectively.

[0025] Within each sampling time window, the interactive signaling data stream is processed by state extraction based on a snapshot mapping mechanism to obtain a multidimensional instantaneous value group corresponding to each sampling time window; In the specific implementation process, based on the snapshot mapping mechanism, for each sampling time window State extraction is performed to form a multidimensional instantaneous numerical set. The specific processing method is as follows: If the current time window If at least one signaling event (such as location update or base station handover) occurs within the window, the data field of the most recently occurring event is selected to form a snapshot of the window's state, specifically represented as follows: ; In the formula, This represents the multidimensional instantaneous value set for the k-th time window. This represents the unique hardware identifier of the user terminal corresponding to the k-th time window. This represents the cell identifier of the base station where the user terminal is camped within the k-th time window. This represents the received power of the reference signal measured and reported by the terminal within the k-th time window. This represents the timestamp of the latest signaling event occurring in the k-th time window.

[0026] If the current window If no new signaling events are generated within the specified time, a timeliness check is performed based on the event timestamp recorded in the previous window; specifically, a maximum effective lifetime is introduced. (For example, set to 15 minutes), calculate the current window's end time. The event timestamp recorded in the previous window The time difference, if If the state is still valid, then state inheritance will be performed. ; It should be noted that the above The timestamp remains unchanged to accurately reflect the time of the original signaling event and avoid false timestamp updates. like If the current state is determined to be signal loss or power off, inheritance is terminated, and the value is marked as the default value, specifically: ; In the formula, These represent the corresponding default state values. Indicates an invalid timestamp.

[0027] If at the initial stage of the observation period (e.g., window) Since there is no historical state to inherit, the current window's state snapshot is marked with the default value. Based on the time sequence of the multiple sampling time windows, the multidimensional instantaneous numerical groups are processed into time series to obtain the signaling trajectory dataset.

[0028] Finally, the time windows The obtained multidimensional instantaneous numerical data set Arranged according to a strict time window order, they form a complete and continuous signaling trajectory dataset, specifically represented as follows: ; In the formula, This represents the signaling trajectory dataset.

[0029] Electronic fence dwell analysis module: used to pre-build an electronic fence logical network containing parent area grids and child hotspot grids, calculate signal dwell confidence based on the signaling trajectory dataset, and filter target user terminals that meet the deep dwell conditions; It is understood that the electronic fence logic network is based on a two-level grid structure to achieve refined location determination and user screening. Specifically, the parent-level area grid is mainly used for large-scale area division to quickly locate the approximate area where the user is located; the child-level hotspot grid is used to locate the user's specific location hotspot at a finer granular level, thereby supporting more accurate dwell behavior analysis. Through the combination of the above two-level grid structure, the efficiency and accuracy of user behavior analysis in complex application scenarios (such as airports, commercial areas, and exhibition halls) can be effectively improved.

[0030] Specifically, the method for constructing the electronic fence logical network, which includes a parent region grid and child hotspot grids, includes: The target region is divided into a first spatial grid, and multiple independent parent region grids are constructed. It should be understood that the first spatial grid division of the target area is a relatively coarse spatial division, the purpose of which is to achieve preliminary and rapid positioning of the area in order to reduce the complexity of subsequent matching calculations.

[0031] In practice, the parent region grid can be divided using regular rectangular or hexagonal grids. Taking a regular rectangular grid as an example, it is specifically represented as follows: ; In the formula, Represents the parent region's grid set. Let M represent the i-th parent region grid, and M represent the total number of parent region grids.

[0032] The parent region grid is subjected to a second spatial grid division process to construct the corresponding child hotspot grid; Understandably, based on the parent region grid division, a more refined second spatial grid division is further adopted, subdividing each parent region grid into multiple child hotspot grids, thereby achieving fine-grained location hotspot analysis; the child hotspot grid division can be expressed as: ; In the formula, Represents the set of child hotspot grids. This represents the j-th child hotspot grid under the i-th parent region grid. This represents the number of child hotspot grids within the i-th parent region grid.

[0033] It should be noted that the first and second spatial grid divisions can be flexibly adjusted according to the needs of the actual application scenario. For example, if the target area is a large airport terminal area (e.g., about 1km long and 500m wide, with an area of ​​about 500,000 square meters), in the first spatial grid division process, the entire airport terminal area can be divided into multiple coarser-grained parent area grids. For example, the target area can be divided into rectangular grids with a length and width of 100m. At this time, about 50 parent area grids are divided, and each grid has an area of ​​about 10,000 square meters. In the second spatial grid division process, each parent area grid can be further refined. For example, each parent area grid (100m×100m) can be divided into multiple finer child hotspot grids, such as rectangular grids with a length and width of 10m. At this time, each parent grid is divided into 100 child hotspot grids, and each hotspot grid has an area of ​​100 square meters. The scale of each child hotspot grid is 10m×10m.

[0034] For each parent region grid and its corresponding child hotspot grid, associate scene attribute codes and security policy tags respectively; It should be understood that, in order to facilitate the application of information security policies and the matching of business logic, scenario attribute codes and security policy tags are defined and associated for the parent area grid and the child hotspot grid. Specifically, the scenario attribute code can indicate the specific application scenario in which each grid is located (such as waiting hall, commercial area, ticket gate, etc.); the security policy tag is used to identify the information push security level or restriction rules (such as whitelist, blacklist, etc.) corresponding to each grid.

[0035] Taking a specific parent mesh and child mesh as an example, the relationship can be represented as follows: ; ; In the formula, This represents the scene attribute code corresponding to the spatial grid. This indicates the security policy label corresponding to the spatial grid. , These represent the scene attribute codes corresponding to the i-th parent region mesh and the j-th child hotspot mesh within the i-th parent region mesh, respectively. , These represent the security policy labels corresponding to the i-th parent region grid and the j-th child hotspot grid within the i-th parent region grid, respectively.

[0036] Based on the spatial relationship between each parent region grid and its corresponding child hotspot grid, spatial index association processing is performed to obtain an electronic fence logical network containing parent region grids and child hotspot grids.

[0037] It should be noted that the spatial index association processing aims to quickly establish a clear spatial relationship between the parent region grid and the corresponding child hotspot grid, so as to quickly determine the parent region grid corresponding to any child hotspot grid; specifically, it is expressed as follows: ; In the formula, This indicates the spatial index association between a child hotspot grid and its corresponding parent region grid.

[0038] By linking scene attribute codes and security policy tags, a hierarchical electronic fence logical network with semantic information is formed.

[0039] In practical applications, based on the signaling trajectory dataset, the stability of user terminals in the electronic fence logic network is further quantitatively analyzed in order to screen target user terminals that meet the deep dwell conditions.

[0040] In implementation, the method for calculating the signal dwell confidence includes: Based on the signaling trajectory dataset, the cumulative number of base station cell identifier changes for each user terminal is counted to obtain the corresponding base station handover frequency. It is understandable that in an electronic fence logical network, the base station handover behavior of a user terminal can usually reflect its location movement. For example, if a user stays in a certain area continuously, the cell identifier of the base station it is connected to usually remains stable, while if the user is moving, the base station handover frequency is often higher. Therefore, this embodiment uses the base station handover frequency as an important basic indicator for the stability analysis of the user's stay, and the specific calculation formula is as follows: ; In the formula, This represents the base station handover frequency of the p-th user terminal within the observation period. This represents the base station cell identifier of the p-th user terminal in the k-th time window. The base station identifier change indication function is specifically defined as follows: .

[0041] Based on the signaling trajectory dataset, the numerical variance of the reference signal received power of each user terminal in the time dimension is calculated to obtain the corresponding signal strength fluctuation value. It should be understood that after obtaining the user's base station handover frequency, further analysis of the user's camping situation from the perspective of signal quality stability is needed to improve the accuracy of camping determination. Specifically, for terminals in stable locations, the reference signal received power usually fluctuates less; conversely, if the terminal's location changes frequently, the reference signal received power often shows significant fluctuations.

[0042] Based on this, this embodiment uses the time variance of the reference signal received power as an indicator of signal strength fluctuation, and the specific formula is as follows: ; In the formula, This represents the signal strength fluctuation value of the p-th user terminal. This represents the reference signal received power of the p-th user terminal in the k-th time window. This represents the average received power of the reference signal for the p-th user terminal.

[0043] The signal instability index of each user terminal is obtained by weighting the base station handover frequency and the signal strength fluctuation value based on a preset weighting coefficient. In the specific implementation process, based on the aforementioned two indicators, a signal instability index is constructed through a weighted combination method, specifically expressed as: ; In the formula, This represents the signal instability index of the p-th user terminal. This indicates the preset weighting coefficient. This represents the maximum value among all user base station handover frequencies. This represents the maximum value among all user terminal signal strength fluctuations.

[0044] It should be understood that, considering extreme circumstances or There is a risk that the normalization ratio may exceed the historical maximum value, resulting in a normalization ratio greater than 1; to avoid the occurrence of a signal instability index. To address the mathematical anomalies, this embodiment adds a saturation constraint to the weighted results, limiting the signal instability index to no more than 1, thereby ensuring that all relevant indicators remain within the legal value range.

[0045] It should be noted that the preset weighting coefficient Based on multiple experiments, the value range is 0 to 1, and the specific value is determined according to actual needs.

[0046] It should also be noted that, for the purpose of calculating stability, the maximum value among all user base station handover frequencies is used. and the maximum value among all user terminal signal strength fluctuation values. You can take the historical maximum value or the maximum value within the current batch.

[0047] The signal instability index is subjected to inverse normalization to obtain the signal dwell confidence level of each user terminal.

[0048] Understandably, to facilitate the determination of dwelling behavior, the signal instability index is converted into a more easily understood signal dwelling confidence level through reverse processing, that is, the probability of the terminal's stable dwelling in the current area, specifically expressed as: ; In the formula, This represents the signal dwell confidence level of the p-th user terminal.

[0049] It should be noted that the closer the dwell confidence score is to 1, the more likely the user is to stay stably in a certain spatial area; conversely, the smaller the value, the more likely the user's behavior is to move around.

[0050] In practical applications, the signal dwell confidence index alone is not sufficient to fully determine whether a user truly meets the deep dwell conditions. Therefore, this embodiment further combines multiple factors such as the user's signal dwell confidence, dwell time span, and spatial grid matching relationship to implement a comprehensive judgment in order to accurately screen out target user terminals that meet the deep dwell conditions.

[0051] Specifically, the screening of target user terminals that meet the deep residency criteria includes: The signal dwell confidence level is compared with a preset dwell confidence threshold to obtain a dwell confidence determination result; In practice, a pre-set dwell confidence threshold is used to compare the signal dwell confidence of each user terminal, thereby achieving preliminary screening. The specific judgment method is as follows: ; In the formula, This represents the retention confidence determination result of the p-th user terminal, where 1 indicates that the threshold condition is met and 0 indicates that the threshold condition is not met.

[0052] Based on the signaling interaction timestamps of the signaling trajectory dataset, the dwell time span of each user terminal within the electronic fence logical network is determined and compared with the preset minimum dwell time to obtain the time span determination result. Understandably, based on the initial confidence level of user dwell time, further analysis is conducted using dwell time duration to definitively determine whether a user exhibits deep dwell behavior. Specifically, the process of calculating and comparing dwell time spans can be represented as follows: ; In the formula, This represents the time span determination result for the p-th user terminal, where 1 indicates that the dwell time requirement is met, and 0 indicates that the dwell time requirement is not met. This represents the cumulative time span during which the p-th user terminal continuously resides in the electronic fence logical network within the observation period. This indicates the preset minimum dwell time threshold, which ranges from 5 to 10 minutes. The specific value can be set according to the actual needs of the scenario.

[0053] Based on the parent area grid of the electronic fence logical network, the base station cell identifier of the user terminal is spatially matched to obtain the parent area grid matching result; It should be noted that, based on the determination of dwell confidence and time span, spatial location matching is further implemented; specifically, spatial correspondence matching is performed between the base station cell identifier of the user terminal and the parent area grid to determine the approximate area range of the user. The specific matching method can be expressed as follows: ; In the formula, This represents the parent region grid matched by the p-th user terminal. This represents the parent region grid that matches the spatial location of the terminal base station cell identifier.

[0054] Based on the parent region grid matching result and the reference signal received power, and combined with the preset signal feature mapping library, further spatial matching is performed with the corresponding child hotspot grid to obtain the child hotspot grid matching result. In the specific implementation process, in the parent-level regional grid Given that the location is already determined, all its corresponding sub-level hotspot grids. The reference signal strength benchmark is used to perform distance comparison, and a fine matching is completed using the Euclidean distance minimization algorithm, specifically expressed as follows: ; In the formula, This indicates the signal difference distance between the user's current RSRP and the target grid reference RSRP. This represents the reference signal received power measured by the user terminal within the current time window. Represents the sub-hotspot grid recorded in the signal feature mapping library. The baseline RSRP characteristic value.

[0055] It should be noted that the distance minimization algorithm can also be extended to a comprehensive vector matching algorithm based on weighted distance and multi-dimensional signal features (such as RSRQ, SINR, etc.), depending on the actual application process.

[0056] It should be noted that the main purpose of the signal feature mapping library is to establish accurate and reliable signal strength benchmark features for each sub-level hotspot grid, thereby providing reliable support for subsequent fine-grained spatial location matching. The specific construction logic is as follows: Multiple reference sampling points are selected within the target area. These reference sampling points are evenly distributed in each sub-level hotspot grid, and signal characteristic data corresponding to each sampling point are collected at different time periods. The data of each sampling point includes, but is not limited to, multi-dimensional signal characteristic parameters such as reference signal received power (RSRP), reference signal received quality (RSRQ), and signal-to-interference-plus-noise ratio (SINR). After the acquisition of the original signal feature data is completed, the signal data acquired by each sampling point in different time periods are statistically processed to obtain the signal feature distribution of each sub-level hotspot grid. Further analysis of multiple sets of signal feature data obtained from sampling points within each sub-level hotspot grid is performed using clustering analysis methods (such as K-means algorithm or Gaussian mixture model GMM) to identify stable signal feature cluster centers within the grid, and the values ​​of these cluster centers are used as the baseline signal features of the grid. Finally, the reference signal features of each sub-level hotspot grid are associated and stored with the spatial identifier of the corresponding grid to form a signal feature mapping library, which is used for subsequent real-time location matching and dwell determination.

[0057] It is understandable that, in order to continuously improve the adaptability and accuracy of the signal feature mapping library, this embodiment can also periodically or dynamically update and optimize the signal feature mapping library according to changes in the scenario, so as to adapt to the impact of environmental changes (such as base station parameter adjustments, new buildings, etc.) on signal propagation characteristics, and ensure that the signal feature mapping library can always meet the actual application requirements.

[0058] Subsequently, the matching grid is selected according to the minimum distance principle, specifically as follows: ; In the formula, This represents the sub-level hotspot grid matched by the p-th user terminal; It should be noted that this only applies to the parent region grid. Perform the above calculations within the sub-grids of all j.

[0059] Based on the retention confidence determination result, time span determination result, and sub-level hotspot grid matching result, a comprehensive judgment is made to determine the target user terminal that meets the deep retention condition.

[0060] Finally, based on the above multi-dimensional judgment results, this embodiment performs a comprehensive judgment to determine the target user terminal that ultimately meets the deep residency conditions. The specific judgment method is exemplified as follows: ; In the formula, This represents the result of the deep residency condition determination for the p-th user terminal, where 1 indicates that the deep residency condition is met and 0 indicates that the deep residency condition is not met.

[0061] In practical applications, spatial dwell time determination alone is insufficient to fully confirm whether a user terminal is suitable for instant personalized service information push. Therefore, this embodiment further retrieves the interaction information corresponding to each target user terminal and performs rule mutual exclusion verification of reverse sliding time window to ensure the security, compliance and targeting of service push.

[0062] Service release rule verification module: used to retrieve the interaction information corresponding to each target user terminal, perform rule mutual exclusion verification of the reverse sliding time window respectively, and generate the corresponding service release instruction; It should be noted that the interactive information includes data such as the user terminal's historical interaction records, current system time, and user's identity level; through the joint analysis of this data, it is possible to avoid frequently or repeatedly pushing the same type of service information, prevent user resentment, and ensure the security and compliance of information push.

[0063] In implementation, the mutual exclusion check of the rules for executing the reverse sliding time window includes: Based on the sub-level hotspot grid matching results of the target user terminal, determine the associated scene attribute code and security policy label; It is understandable that, after determining the specific sub-level hotspot grid that the target user terminal matches, this embodiment further extracts the scene attribute code and security policy tag associated with the hotspot grid in order to determine the corresponding business scenario and information push rules.

[0064] Using scene attribute codes as index keys, construct reverse sliding time windows respectively; It should be noted that the aforementioned reverse sliding time window refers to a data window that slides and analyzes data along a fixed time length in the past, using the current system time as a reference point. Specifically, it refers to the data window analyzed based on the scene attribute code. The reverse sliding time window can be represented as: ; In the formula, This represents a reverse sliding time window based on scene attribute codes. Indicates the current system time. This indicates the length of the sliding time window, which can be 30 minutes or 1 hour, depending on the specific business requirements.

[0065] Based on the historical interaction records of the target user terminal, perform a comparison of records with the same code within the reverse sliding time window to obtain the result of the determination of records with the same code; It should be understood that, in order to avoid repeatedly pushing similar information within a short period of time, this embodiment performs a code comparison analysis on the user terminal's historical interaction records. The specific determination method is as follows: ; In the formula, This represents the result of determining the historical interaction code matching of the p-th user terminal. Represents scene attribute code In the reverse sliding time window The number of times it appears in the text.

[0066] It should be noted that the historical interaction same code record determination result A value of 1 indicates that there are no records with the same code within the current time window, and the message can be pushed; the result of the historical interaction record determination. A value of 0 indicates that a record already exists within the current time window and cannot be pushed again.

[0067] The security compliance assessment result is determined based on the security policy label, current system time, and identity level. Understandably, in addition to historical interaction records, security and compliance are also important factors in push notification decisions.

[0068] Specifically, the method for determining the security compliance assessment result includes: Pre-configure security policy labels that include blacklists, whitelists, and corresponding restricted periods; It should be noted that, to ensure the security and compliance of push notifications, this embodiment pre-establishes security policy tags for different scenarios. These tags explicitly include blacklists, whitelists, and prohibited sending periods for subsequent security determination, specifically as follows: ; In the formula, This represents the security policy label corresponding to the j-th child hotspot grid within the i-th parent region grid. This represents the set of user identity levels from which push notifications are prohibited. This represents the set of user identity levels that are allowed to send push messages. This represents the set of time periods during which push notifications are prohibited.

[0069] The target user's terminal identity level is compared with the blacklist and whitelist respectively to determine the list matching result; It should be noted that, to avoid sensitive or special user groups (such as users with high security levels or those specially marked) receiving inappropriate information pushes, this embodiment compares the user terminal's identity level with pre-set blacklists and whitelists to determine whether to allow service messages to be sent to the user terminal. The specific list matching determination method is as follows: ; In the formula, This represents the match result for the p-th user terminal, where 1 indicates push notifications are allowed and 0 indicates push notifications are disabled. This represents the identity level of the p-th user terminal.

[0070] Understandably, the aforementioned blacklist and whitelist mechanism can effectively control the message push permissions of specific user terminals and avoid compliance risks.

[0071] Match the current system time with the prohibited time period to determine the prohibited time period judgment result; It should be understood that, in addition to matching user identity levels, this embodiment further considers the time factor to ensure that service messages are not pushed during specific time periods (such as nighttime, sensitive periods during holidays, etc.), thereby avoiding interference to users.

[0072] Specifically, this embodiment matches the current system time with a preset set of prohibited time periods. The specific matching method is as follows: ; In the formula, This represents the result of the prohibited period determination for the p-th user terminal, where 1 indicates that push notifications are allowed and 0 indicates that push notifications are prohibited.

[0073] Based on the list matching results and the results of the prohibited period determination, a comprehensive judgment is made to determine the security compliance determination result of the target user terminal.

[0074] It is understandable that, based on the aforementioned matching of identity level lists and prohibited time periods, this embodiment implements a comprehensive judgment to clearly determine the security compliance judgment result of the target user terminal, thereby providing a reliable basis for subsequent service pushes. The specific judgment method is expressed as follows: ; In the formula, This represents the final security compliance determination result for the p-th user terminal; where 1 indicates compliance and service push is allowed; and 0 indicates non-compliance and service push is prohibited.

[0075] Based on the results of the same code record determination and the security compliance determination, a comprehensive judgment is made to generate the corresponding service release instruction.

[0076] Finally, this embodiment makes a comprehensive decision based on both historical interaction records and security compliance assessment results to generate a clear service release instruction, which is exemplified as follows: ; In the formula, This represents the final service release decision result for the p-th user terminal; where 1 indicates that service push is allowed and 0 indicates that service push is prohibited.

[0077] It should be noted that when the above comprehensive judgment result is 1, this embodiment generates a service release instruction that allows the service message to be issued, which is used for the real-time rendering and push process of subsequent personalized service information; when the above comprehensive judgment result is 0, the service message is refused to be pushed to the target user terminal, and a corresponding service interception instruction is generated. At the same time, it is possible to record this interception event for subsequent auditing or user experience analysis.

[0078] In order to effectively implement these rigorously screened and judged service release instructions into the actual information push process, this embodiment further calls the cloud-based structured data template in real time for each service release instruction to render the content, generate personalized service message data packets for the target user terminal, and effectively schedule and distribute them.

[0079] Personalized service content scheduling module: For each service release instruction, it calls the cloud-based structured data template for real-time content rendering, generates and schedules the distribution of the corresponding personalized service message data packet.

[0080] It should be understood that this embodiment uses a pre-built cloud-based structured data template to dynamically generate personalized service message data packets in real time, thereby effectively meeting the personalized service needs of different users. Simultaneously, through real-time rendering, this embodiment can send personalized content that conforms to the actual scenario and user preferences to the user terminal in a timely and accurate manner, significantly improving user experience and information push efficiency.

[0081] Specifically, the method for real-time content rendering includes: The user profile database is retrieved based on the unique identifier of the target user's terminal hardware to obtain the corresponding set of business attribute fields; It should be noted that, in order to accurately achieve precise matching and rendering of personalized content, this embodiment uses the user terminal's unique hardware identifier (such as IMEI, MAC address, or other device unique identifier) ​​to retrieve a pre-built user profile database in the cloud in real time, and extracts a set of business attribute fields specific to the target user terminal; specifically, the set of business attribute fields can be represented as: ; in, This represents the set of business attribute fields corresponding to the p-th target user terminal. This represents the e-th business attribute field corresponding to the p-th user terminal profile, and q represents the total number of business attribute fields, which depends on the scenario type. For example, in an airport boarding gate scenario, it may include specific fields such as boarding time and flight number.

[0082] It should be understood that the business attribute fields may cover personalized data such as user preferences, historical interaction records, identity level, consumption records or interest tags, in order to achieve accurate content matching.

[0083] Based on the scenario attribute code of the target user terminal, determine the business attribute fields corresponding to the current scenario; Understandably, after obtaining the set of business attribute fields for the target user terminal profile, the system further identifies the business attribute fields required for the current specific business scenario based on the scenario attribute code of the target user terminal determined in the aforementioned steps. This ensures the effectiveness and scenario relevance of the rendered content, specifically as follows: ; In the formula, This represents the set of business attribute fields corresponding to the j-th child hotspot grid (i.e., the current scene) under the i-th parent grid; It should be noted that, All are derived from the target user terminal business attribute field set. subscript This indicates the index of the specific attribute field required for the current scenario within the user profile set; This indicates that the set of business attribute fields used in the current scenario is a subset of the set of business attribute fields in the target user terminal profile.

[0084] Based on cloud-based structured data templates, identify dynamic placeholders for corresponding business attribute fields in the template; It should be noted that, to effectively integrate business attribute fields with template content, a pre-built cloud-based structured data template is used. This template contains multiple dynamic placeholders, which are used to match and populate different business attribute fields in real time, enabling rapid and automated rendering of personalized information. Specifically, this is represented as follows: ; In the formula, This represents the set of dynamic placeholders contained in the template. This represents the h-th dynamic placeholder in the template.

[0085] Fill the business attribute fields into dynamic placeholders to complete real-time content rendering and obtain personalized service message data packets.

[0086] Understandably, the aforementioned set of user profile business attribute fields is effectively matched with the set of business attribute fields required for the current scenario. Then, through the aforementioned dynamic placeholder mechanism, the corresponding business attribute fields are populated into the template in real time to complete content rendering and generate a personalized service message data packet for the target user terminal. Specifically, this is represented as follows: ; really, This represents the personalized service message data packet generated for the p-th user terminal. This indicates a real-time rendering function based on a template and a set of attribute fields. This represents a structured data template in the cloud.

[0087] For example, taking a tourist attraction as an example, when user A enters Qingyun Mountain, a national 5A-level scenic spot, the system has determined that user A meets the conditions for deep stay and has obtained a service release instruction.

[0088] The system first retrieves the set of business attribute fields for user A from the cloud-based user profile database in real time, based on the unique hardware identifier of user A's mobile terminal. for: {Name: A, Travel Preference: Natural Scenery, Membership Level: Silver Member, Last Visit Record: Waterfall Attraction}; It should be noted that the above user profile information comes from user A's historical browsing records, online registration information, and other interactive behavior data.

[0089] The system further determines the specific subset of business attribute fields corresponding to the current scene based on the scene attribute code of user A's current location in the scenic area (e.g., the main peak viewing platform); the set of business attribute fields corresponding to this scene. For example: {Name: A, Travel Preference: Natural Scenery, Membership Level: Silver Card Member}; The system calls a pre-set structured SMS template from the cloud. Example template content: Dear [Name], as a [Member Level], based on your [Travel Preferences], we recommend you visit the sea of ​​clouds near the main peak of Qingyun Mountain. Visibility is excellent today, so don't miss it! Finally, the system populates the business attribute fields from User A's profile into the dynamic placeholders of the template in real time, completes the real-time rendering of the content, and generates a personalized service message data packet for User A. The specific content is: Dear A, as a Silver Card member, based on your favorite natural scenery, we recommend that you visit the sea of ​​clouds near the main peak of Qingyun Mountain. The visibility is excellent today, so don't miss it! In practical applications, if the system generates a large number of personalized service message data packets at the same time and sends them directly, it may lead to network congestion and message delays, thereby reducing the user experience. Therefore, this embodiment further provides a scheduling and delivery mechanism for personalized service message data packets. By marking the sending priority, managing the buffer queue, and dynamically monitoring the link load, it can effectively ensure that personalized service message data packets can be sent efficiently and in an orderly manner.

[0090] Specifically, the method for scheduling and distributing personalized service message data packets includes: Based on the scenario attribute code corresponding to the target user terminal, determine the sending priority marker of the personalized service message data packet; It should be noted that, in order to more effectively manage and allocate system sending resources, a message sending priority flag is determined based on the specific scenario attribute code of the target user terminal, so as to achieve differentiated message processing methods. The sending priority flag is represented as follows: ; In the formula, This indicates the sending priority flag for the personalized service message data packet corresponding to the p-th target user terminal. This represents a mapping function that determines priority based on scene attribute codes; Understandably, urgent notifications or important reminders have a higher priority than ordinary introductions or recommendations.

[0091] Personalized service message data packets are divided into corresponding buffer queues according to the sending priority flag, forming a multi-level buffer queue; Understandably, after obtaining the sending priority flag for each message data packet, the system further divides the personalized service message data packets into different levels of buffer queues based on different priority flags, forming a multi-level buffer queue management structure, specifically represented as follows: ; In the formula, This represents the buffer queue for sending personalized service message data packets with priority w. This indicates the total number of priority levels of the buffer queues set in the system.

[0092] It should be noted that W = 1 indicates the highest priority, and the priority decreases sequentially.

[0093] It should be understood that through the above-mentioned multi-level buffer queue management structure, the system can prioritize sending urgent or high-priority messages, thereby improving the overall information push efficiency.

[0094] Monitor the load status of the communication gateway link, and dynamically adjust the data packet delivery strategy of different priority buffer queues based on the comparison result of the link load status and the preset load threshold, so as to complete the scheduling and delivery of personalized service message data packets.

[0095] Finally, to further ensure the stability and efficiency of message sending, the load status of the communication gateway link is monitored in real time, and the message sending strategy is dynamically adjusted in combination with the preset load threshold.

[0096] In the specific implementation process, firstly, the current load status of the link is determined through real-time monitoring, specifically as follows: ; In the formula, This indicates the real-time load status monitored by the current communication gateway link. This represents a function for real-time monitoring of link load. Secondly, the current load is compared with a preset threshold. The specific comparison process is as follows: ; In the formula, This indicates a plan for adjusting the message delivery strategy based on load conditions. This indicates the preset link load threshold.

[0097] It should be noted that when the link load is below or equal to the threshold, the system prioritizes increasing the frequency and number of high-priority message data packets sent; when the link load exceeds the threshold, the system explicitly reduces or even suspends message sending in the low-priority buffer queue to ensure that data packets in the high-priority queue can be delivered in a timely manner, avoiding network congestion and delays, thereby effectively improving user satisfaction and experience.

[0098] Through the aforementioned dynamic adjustment strategy mechanism, this embodiment clearly ensures that the system's scheduling and delivery of personalized service message data packets are always in the optimal state, thereby significantly improving service quality and user experience.

[0099] Example 2: like Figure 2 As shown in the example, the parts not detailed in this embodiment are as shown in Example 1. This embodiment discloses a distributed intelligent data management method based on cloud computing, including: The interactive signaling data stream of the user terminal is parsed and processed by time slicing and snapshot mapping to obtain the signaling trajectory dataset; A pre-constructed electronic fence logical network containing parent region grids and child hotspot grids is used to calculate the signal dwell confidence based on the signaling trajectory dataset and to filter target user terminals that meet the deep dwell conditions. Retrieve the interaction information corresponding to each target user terminal, perform rule mutual exclusion verification of the reverse sliding time window respectively, and generate the corresponding service release instruction; For each service release command, a cloud-based structured data template is invoked for real-time content rendering, generating and dispatching the corresponding personalized service message data packet.

[0100] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0101] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A distributed intelligent data management system based on cloud computing, characterized in that, The system includes: Signaling trajectory data parsing module: used to parse and process the interactive signaling data stream of user terminals by segmenting it by time slicing and combining it with snapshot mapping to obtain the signaling trajectory dataset; Electronic fence dwell analysis module: used to pre-build an electronic fence logical network containing parent area grids and child hotspot grids, calculate signal dwell confidence based on the signaling trajectory dataset, and filter target user terminals that meet the deep dwell conditions; Service release rule verification module: used to retrieve the interaction information corresponding to each target user terminal, perform rule mutual exclusion verification of the reverse sliding time window respectively, and generate the corresponding service release instruction; Personalized service content scheduling module: For each service release instruction, it calls the cloud-based structured data template for real-time content rendering, generates and schedules the distribution of the corresponding personalized service message data packet.

2. The distributed intelligent data management system based on cloud computing according to claim 1, characterized in that, The interactive signaling data stream includes a unique terminal hardware identifier, a base station cell identifier, a reference signal receiving power, and a signaling interaction timestamp. The parsing process, which combines time-slicing with snapshot mapping, includes: Multiple consecutive sampling time windows are determined according to a preset fixed period; Within each sampling time window, the interactive signaling data stream is processed by state extraction based on a snapshot mapping mechanism to obtain a multidimensional instantaneous value group corresponding to each sampling time window; Based on the time sequence of the multiple sampling time windows, the multidimensional instantaneous numerical groups are processed into time series to obtain the signaling trajectory dataset.

3. The distributed intelligent data management system based on cloud computing according to claim 1, characterized in that, The method for constructing the electronic fence logical network, which includes a parent region grid and child hotspot grids, includes: The target region is divided into a first spatial grid, and multiple independent parent region grids are constructed. The parent region grid is subjected to a second spatial grid division process to construct the corresponding child hotspot grid; For each parent region grid and its corresponding child hotspot grid, associate scene attribute codes and security policy tags respectively; Based on the spatial relationship between each parent region grid and its corresponding child hotspot grid, spatial index association processing is performed to obtain an electronic fence logical network containing parent region grids and child hotspot grids.

4. The distributed intelligent data management system based on cloud computing according to claim 1, characterized in that, The method for calculating the signal dwell confidence includes: Based on the signaling trajectory dataset, the cumulative number of base station cell identifier changes for each user terminal is counted to obtain the corresponding base station handover frequency. Based on the signaling trajectory dataset, the numerical variance of the reference signal received power of each user terminal in the time dimension is calculated to obtain the corresponding signal strength fluctuation value. The signal instability index of each user terminal is obtained by weighting the base station handover frequency and the signal strength fluctuation value based on a preset weighting coefficient. The signal instability index is subjected to inverse normalization to obtain the signal dwell confidence level of each user terminal.

5. A distributed intelligent data management system based on cloud computing according to claim 1, characterized in that, The screening of target user terminals that meet the deep residency criteria includes: The signal dwell confidence level is compared with a preset dwell confidence threshold to obtain a dwell confidence determination result; Based on the signaling interaction timestamps of the signaling trajectory dataset, the dwell time span of each user terminal within the electronic fence logical network is determined and compared with the preset minimum dwell time to obtain the time span determination result. Based on the parent area grid of the electronic fence logical network, the base station cell identifier of the user terminal is spatially matched to obtain the parent area grid matching result; Based on the parent region grid matching result and the reference signal received power, and combined with the preset signal feature mapping library, further spatial matching is performed with the corresponding child hotspot grid to obtain the child hotspot grid matching result. Based on the retention confidence determination result, time span determination result, and sub-level hotspot grid matching result, a comprehensive judgment is made to determine the target user terminal that meets the deep retention condition.

6. A distributed intelligent data management system based on cloud computing according to claim 1, characterized in that, The interactive information includes: historical interaction records of the user terminal, current system time, user identity level, and other data; the mutual exclusion check of the rules for executing the reverse sliding time window includes: Based on the sub-level hotspot grid matching results of the target user terminal, determine the associated scene attribute code and security policy label; Using scene attribute codes as index keys, construct reverse sliding time windows respectively; Based on the historical interaction records of the target user terminal, perform a comparison of records with the same code within the reverse sliding time window to obtain the result of the determination of records with the same code; The security compliance assessment result is determined based on the security policy label, current system time, and identity level. Based on the results of the same code record determination and the security compliance determination, a comprehensive judgment is made to generate the corresponding service release instruction.

7. A distributed intelligent data management system based on cloud computing according to claim 6, characterized in that, The methods for determining the security compliance assessment results include: Pre-configure security policy labels that include blacklists, whitelists, and corresponding restricted periods; The target user's terminal identity level is compared with the blacklist and whitelist respectively to determine the list matching result; Match the current system time with the prohibited time period to determine the prohibited time period judgment result; Based on the list matching results and the results of the prohibited period determination, a comprehensive judgment is made to determine the security compliance determination result of the target user terminal.

8. A distributed intelligent data management system based on cloud computing according to claim 1, characterized in that, The method for real-time content rendering includes: Based on the unique identifier of the target user's terminal hardware, retrieve and pre-built user profile database to obtain the corresponding set of business attribute fields; Based on the scenario attribute code of the target user terminal, determine the business attribute fields corresponding to the current scenario; Based on cloud-based structured data templates, identify dynamic placeholders for corresponding business attribute fields in the template; Fill the business attribute fields into dynamic placeholders to complete real-time content rendering and obtain personalized service message data packets.

9. A distributed intelligent data management system based on cloud computing according to claim 1, characterized in that, The method for scheduling and distributing personalized service message data packets includes: Based on the scenario attribute code corresponding to the target user terminal, determine the sending priority marker of the personalized service message data packet; Personalized service message data packets are divided into corresponding buffer queues according to the sending priority flag, forming a multi-level buffer queue; Monitor the load status of the communication gateway link, and dynamically adjust the data packet delivery strategy of different priority buffer queues based on the comparison result of the link load status and the preset load threshold, so as to complete the scheduling and delivery of personalized service message data packets.

10. A distributed intelligent data management method based on cloud computing, characterized in that, The method includes: The interactive signaling data stream of the user terminal is parsed and processed by time slicing and snapshot mapping to obtain the signaling trajectory dataset; A pre-constructed electronic fence logical network containing parent region grids and child hotspot grids is used to calculate the signal dwell confidence based on the signaling trajectory dataset and to filter target user terminals that meet the deep dwell conditions. Retrieve the interaction information corresponding to each target user terminal, perform rule mutual exclusion verification of the reverse sliding time window respectively, and generate the corresponding service release instruction; For each service release command, a cloud-based structured data template is invoked for real-time content rendering, generating and dispatching the corresponding personalized service message data packet.

Images