A remote visualization monitoring cabin data management system based on a cloud platform

Abstract

The present invention relates to the technical field of data storage, and specifically to a remote visualization monitoring shelter data management system based on a cloud platform, including: The present invention obtains multi-dimensional monitoring data of the shelter, and automatically annotates context metadata including domain function attributes, business criticality levels, task IDs, and task phases; constructs a dynamic data portrait according to the metadata, routes the data to logical storage levels with different costs and performances through a rule engine, and performs life cycle management; constructs a global secondary index for task IDs, and constructs data fingerprints and fusion event indexes for key events and their context data sets; provides a unified query interface, and uses hierarchical storage information and fusion indexes to execute cross-level federated query plans. The present invention can effectively improve the overall efficiency of shelter data management.

Description

Technical Field

[0001] The present invention relates to the technical field of data storage, and particularly to a data management system for remote visual monitoring of a shelter based on a cloud platform. Background Art

[0002] Traditional shelter data storage management methods do not fully consider the combination of domain function attributes, business criticality levels, and dynamic value evolution and access patterns within the entire task cycle of shelter data for refined hierarchical storage and index optimization, resulting in limited efficiency in shelter data management.

[0003] Therefore, a data management system for remote visual monitoring of a shelter based on a cloud platform is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a data management system for remote visual monitoring of a shelter based on a cloud platform. The present invention obtains multi-dimensional monitoring data of the shelter and automatically annotates context metadata including domain function attributes, business criticality levels, task IDs, and task phases; constructs a dynamic data portrait based on the metadata, routes the data to logical storage levels with different costs and performances through a rule engine, and performs life cycle management; constructs a global secondary index for task IDs, and constructs data fingerprints and fusion event indexes for key events and their context data sets; provides a unified query interface, and executes a cross-level federated query plan using hierarchical storage information and fusion indexes. The present invention can effectively improve the overall efficiency of shelter data management.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A data management system for remote visual monitoring of a shelter based on a cloud platform, comprising:

[0007] An intelligent data access and automatic annotation module, configured to obtain multi-dimensional monitoring data of the shelter and transmit it to the cloud platform, and the cloud platform automatically annotates the metadata; the metadata includes domain function attributes, business criticality levels, task IDs, and task phases;

[0008] An adaptive hierarchical storage and dynamic routing module based on a data portrait, configured to construct a dynamic data portrait according to the metadata and the data timestamp, and dynamically route the data to logical storage levels with different storage costs and access performances in the cloud platform for storage according to the data portrait and a preset rule engine, and execute corresponding data life cycle management strategies;

[0009] A context-aware fusion index construction module for tasks and events, configured to construct a global secondary index for task IDs, and construct context data fingerprints and fusion event indexes for predefined key events and multi-source context data sets within a time window before and after the key events;

[0010] Multi-level data federated query and visualization service module, which is used to provide a unified query interface, parse user query requests, and generate and execute a federated query plan across storage levels based on hierarchical storage information and fusion event indexes.

[0011] Preferably, the multi-dimensional monitoring data includes UAV telemetry data streams, power supply system status data, equipment status data, and environmental data.

[0012] Preferably, the adaptive hierarchical storage and dynamic routing module based on data portraits includes:

[0013] Data portrait generation unit, which is used to construct a structured data portrait for each data record annotated with metadata. The portrait includes data domain identifiers, current business criticality levels, task IDs, task stage identifiers, original data timestamps, and cloud platform reception timestamps;

[0014] Rule storage unit, which is used to execute the operation of routing data to a specified layer in the logical storage levels including real-time hot storage layer, warm analysis layer, and long-term cold storage layer based on the attribute combinations in the data portrait according to a preset rule engine and in combination with preset storage policy rules.

[0015] Preferably, the data life cycle management strategy includes the following steps:

[0016] Status evaluation and decision-making: Triggered by task stage change events, perform status evaluation on data records in each logical storage level, and judge whether the data record meets the level migration conditions based on the current business criticality level, data timeliness, and the current stage of the task included in its data portrait, and with reference to preset life cycle rules;

[0017] Data optimization processing: If it meets the conditions for migrating from a storage level with high access performance to a storage level with low access performance, then according to the life cycle rules, perform predefined optimization processing on the data before migration. The optimization processing includes aggregating high-frequency raw data into statistical summaries and performing downsampling on redundant data points;

[0018] Data migration and metadata synchronization: Securely migrate the optimized data to the target logical storage level specified by the life cycle rules, and synchronously update the storage location of the data in the global index and metadata information.

[0019] Preferably, the data structure of the global secondary index of the task ID is as follows: using the task ID as the main index key, the corresponding value is a list, and this set contains the unique identifiers of all data records associated with this task ID, the indication of the current logical storage level where these data records are located, their respective timestamp ranges, data domain classifications, and current business critical level information.

[0020] Preferably, the task- and event-oriented fusion index construction module includes:

[0021] Provide a key event type definition interface, configure key event trigger conditions, which can be based on specific numerical values, status sequences, and combinatorial logic of the data source, and set time window parameters for context data capture for each event type;

[0022] When a key event that meets the predefined conditions is detected, the system automatically records the unique ID, type, occurrence timestamp, associated shelter ID, and task ID of the event, and retrieves and aggregates all the original data records from multiple preset relevant data domains within the time window before and after the occurrence of the event from the logical storage level according to the time window parameters set for this event type to form a context data set;

[0023] Use a preset feature extraction algorithm to generate a context data fingerprint for the context data set, and the fingerprint is based on the temporal features of the data within the set;

[0024] Build a fusion event index library, which persistently associates and stores the event ID, event type, task ID, the generated context data fingerprint, and each original data segment that constitutes the context. The index library supports direct lookup based on the event ID and similarity retrieval based on the fingerprint vector.

[0025] Preferably, the execution process of the federated query plan includes:

[0026] After the query planner receives a request from the unified query API, it parses the time range, shelter ID, task ID, and event ID included in the query, and preferentially uses the global secondary index of the task ID and the fusion event index for preliminary data location;

[0027] According to the positioning result and the distribution of data in each logical storage level, decompose the original query into multiple optimized sub-queries for different physical storage backends, including time series databases, relational databases, and object storage;

[0028] Execute the sub-query and perform batch-style intelligent fusion processing on the result sets returned by each sub-query. The fusion processing includes data merging and sorting to ensure time series continuity, data deduplication based on unique keys, converting data structures from different sources into a unified output format, and performing dynamic aggregation operations on the server according to the query request.

[0029] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0030] 1. By constructing an adaptive hierarchical storage and dynamic routing mechanism based on data portraits (including domain function attributes, business key levels, task IDs, and task phases), the present invention can intelligently allocate shelter data with different values and access patterns to logical storage levels with different cost and performance characteristics (such as real-time hot storage layer, warm analysis layer, long-term cold storage layer). Combined with a refined data life cycle management strategy (including on-demand aggregation and downsampling), it avoids the massive data management problems brought by traditional "one-size-fits-all" or rough time-based hierarchical storage. This not only significantly reduces the cloud storage cost, but more importantly, by combining with task- and event-oriented context-aware fusion indexes (especially task ID global secondary indexes and fusion event indexes), when remote visualization applications request task data or retrieve key event context data, it can achieve fast and accurate positioning and retrieval, significantly shortening the data access latency and improving the overall efficiency of monitoring shelter data management.

[0031] 2. The context-aware fusion index construction module for tasks and events proposed by the present invention, especially the "context data fingerprint and fusion event index" constructed for predefined key events and their multi-source context data sets, enables the system to go beyond simple time series data queries. It supports direct context backtracking based on event IDs and more valuable similarity retrieval based on "fingerprint vectors", so as to quickly locate and analyze complex events or failure modes with similar data characteristics that occurred in history. Combined with the integration ability of the multi-level data federated query module for cross-level and cross-data source information, it provides remote users with the ability to perform in-depth correlation analysis and intelligent insights on "business system data", cabin body status data, and various event data generated during the execution of complex tasks (such as aircraft flight tests) by the "intelligent shelter monitoring integration system" (such as flight test shelters), improving the overall efficiency of monitoring shelter data management.

[0032] 3. Through the unified query interface and intelligent query planning and execution mechanism provided by the multi-level data federation query and visualization service module of the present invention, the upper-layer remote visualization application does not need to care about the actual storage location and heterogeneity of the underlying data. This module can automatically optimize the query path, execute sub-queries in parallel, and perform intelligent fusion and visualization-oriented preprocessing on the results (such as downsampling and filling points), so as to ensure that when processing large-scale and multi-dimensional shelter data, the visualization interface can still obtain fast and smooth data responses, greatly improving the user experience. At the same time, the modular design (data access and annotation, hierarchical storage, fusion index, federation query) reduces the cohesion and coupling degree of each part of the system, thereby improving the overall efficiency of monitoring shelter data management. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic structural diagram of a remote visualization monitoring shelter data management system based on a cloud platform provided by an embodiment of the present invention;

[0034] Figure 2 It is a schematic structural diagram of an adaptive hierarchical storage and dynamic routing module based on data portraits provided by an embodiment of the present invention;

[0035] Figure 3 It is a flowchart of the steps of a data life cycle management strategy provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0037] Embodiment 1

[0038] In order to improve the overall efficiency of A test flight shelter data management, a remote visualization monitoring shelter data management system based on a cloud platform is applied. Figure 1 It is a schematic structural diagram of a remote visualization monitoring shelter data management system based on a cloud platform provided by an embodiment of the present invention, including:

[0039] An intelligent data access and automatic annotation module, which is used to obtain multi-dimensional monitoring data of the shelter and transmit it to the cloud platform, and the cloud platform automatically annotates the metadata; the metadata includes domain function attributes, business key levels, task IDs, and task phases;

[0040] Furthermore, the multi-dimensional monitoring data includes unmanned aerial vehicle telemetry data streams, power supply system status data, equipment status data, and environmental data;

[0041] The UAV telemetry data stream includes real-time flight parameters (GPS position, flight altitude, flight speed, attitude angle) obtained from UAV 1, UAV 2, and UAV 3, and communication link quality data between UAVs (such as data transmission rate and packet loss rate between UAV 1 and UAV 2);

[0042] The power supply system status data includes the input voltage, input current, output voltage, and output current of the UPS in the shelter;

[0043] The device status data includes the signal receiver locking status (the status of whether the signal receiver successfully locks the signals of each UAV), the disk space remaining of the data recorder, the air conditioner operating mode, and the electric heater operating mode;

[0044] The environmental data includes the temperature, humidity, air pressure, and wind speed inside and outside the A test shelter;

[0045] Transmit the above data to the cloud platform via satellite communication;

[0046] The automatically annotated metadata includes annotating the domain function attributes, business criticality level, task ID, and task phase; for example, the flight data of UAV 1 is annotated as "domain = UAV telemetry_flight parameters"; the UPS in the shelter is annotated as "domain = shelter power supply_UPS"; for the business criticality level, annotations such as "criticality level = 5" (highest); "criticality level = 1" (low); the task ID annotation includes "task ID = "TF-20250508"; the annotation of the task phase is such as "task phase = in flight execution_task 1";

[0047] The adaptive hierarchical storage and dynamic routing module based on data profiling is used to construct a dynamic data profile according to the metadata and data timestamp, and based on this data profile, combined with a preset rule engine, dynamically route the data to logical storage levels with different storage costs and access performances in the cloud platform for storage, and execute corresponding data life cycle management strategies;

[0048] Figure 2 It is a schematic structural diagram of an adaptive hierarchical storage and dynamic routing module based on data profiling provided by an embodiment of the present invention;

[0049] Further, the adaptive hierarchical storage and dynamic routing module based on data profiling includes:

[0050] A data profile generation unit is used to construct a structured data profile for each data record annotated with metadata. The profile includes a data domain identifier, the current business criticality level, a task ID, a task phase identifier, an original data timestamp, and a cloud platform reception timestamp. The original data timestamp is, for example, "2025-05-08T14:10:25.500Z"; the cloud platform reception timestamp is: "2025-05-08T14:10:25.680Z".

[0051] For a piece of data from UAV 2 during the "Flight Execution_ Task 1" phase of the "TF-20250508" task, which includes key communication link quality parameters (such as a data transmission rate of 5 Mbps and a packet loss rate of 0.1%), the generated structured data profile is as follows: { data domain identifier: "UAV Telemetry_ Communication Link", current business criticality level: 5, task ID: “TF-20250508”, task phase identifier: “Flight Execution_ Task 1”, original data timestamp: “2025-05-08T14:10:25.500Z”, cloud platform reception timestamp: “2025-05-08T14:10:25.680Z”};

[0052] A rule storage unit is used to execute the operation of routing data to a specified layer in a logical storage hierarchy including a real-time hot storage layer, a warm analysis layer, and a long-term cold storage layer based on a preset rule engine according to the attribute combinations in the data profile and in combination with preset storage policy rules.

[0053] Real-time hot storage layer (Tier 1): For example, a memory database cluster (such as Redis) deployed on high-performance computing instances in the cloud;

[0054] Warm analysis layer (Tier 2): For example, a managed time-series database service provided by a cloud service provider (such as AWS Timestream);

[0055] Long-term cold storage layer (Tier 3): For example, a low-cost object storage archival service provided by a cloud service provider (such as AWS S3 Glacier Deep Archive);

[0056] Further, Rule Example 1: IF the task phase identifier in the image CONTAINS "in-flight execution" AND the current business critical level of the image >= 4 AND the data domain identifier of the image STARTS WITH "drone telemetry", THEN route to Tier 1 and set the TTL (Time-To-Live) of Tier 1 to 48 hours; Rule Example 2: IF the data domain identifier of the image == "shelter environment data", THEN route to Tier 2 and set the TTL of Tier 2 to 180 days; for example, the in-cabin and out-cabin temperature and humidity data image will match this rule, and the data will be written into the AWS Timestream of Tier 2;

[0057] Figure 3 It is a flowchart of the steps of a data life cycle management strategy provided by an embodiment of the present invention;

[0058] Further, the data life cycle management strategy includes the following steps:

[0059] Status evaluation and decision-making: Triggered by task phase change events, the status of data records in each logical storage level is evaluated, and based on the current business critical level, data timeliness, and the current phase of the task included in their data images, and with reference to the preset life cycle rules, it is judged whether the data records meet the level migration conditions;

[0060] Further, the trigger includes: When the status of the "TF-20250508" task in the task management system of the cloud platform is updated from "in-flight execution_task 1" to "preliminary analysis of post-flight data" by an operator, triggered by the task phase change event, the data life cycle management service deployed on the cloud platform is activated.

[0061] The evaluation includes: The service scans all data records in Tier 1 with the task ID = "TF-20250508".

[0062] Based on the data image (for example, the current business critical level, the data timeliness calculated from the original data timestamp has exceeded the preset Tier 1 hot storage period such as 24 hours, and the current phase of the task has changed to "preliminary analysis of post-flight data"), and with reference to the preset life cycle rules (for example: "After the task phase changes to 'preliminary analysis of post-flight data' and the hot storage period expires, the condition for migrating to Tier 2 is met").

[0063] Judgment: Identify which original drone telemetry data streams stored in Tier 1 meet the conditions for migrating to Tier 2 (warm analysis layer).

[0064] Data Optimization Processing: If the condition for migrating from a storage tier with high access performance to a storage tier with low access performance is met, then according to the lifecycle rules, predefined optimization processing is performed on the data before migration. The optimization processing includes aggregating high-frequency raw data into statistical summaries and performing downsampling on redundant data points.

[0065] Further, for the drone telemetry data stream determined to meet the migration condition from Tier 1 to Tier 2, especially some ultra-high-frequency raw parameters (such as the raw data of the drone's attitude angle every 10 milliseconds).

[0066] According to the predefined optimization processing strategy defined in the lifecycle rules, before migration, the data processing service (such as a Spark cluster) on the cloud platform performs the following on this data: Aggregate high-frequency raw data (such as attitude angles) into statistical summaries (for example, calculate their average, maximum, minimum, and variance every 100 ms or per second); For some continuous but redundant data points with a large number of values that do not change much (such as the altitude data of the drone during a steady cruise phase), perform downsampling (for example, selectively retain 1-2 points that can represent the trend from 10 points per second).

[0067] Data Migration and Metadata Synchronization: Securely migrate the optimized data to the target logical storage tier specified by the lifecycle rules, and synchronously update the storage location of this data in the global index and metadata information.

[0068] Migrate the aggregated / downsampled data after optimization processing from Redis in Tier 1 to AWS Timestream in Tier 2 securely. This migration process occurs within the internal network of the cloud platform to ensure speed and security.

[0069] Task and Event Oriented Context-Aware Fusion Index Building Module, used to build a global secondary index for task IDs and build context data fingerprints and fusion event indexes for multi-source context data sets within predefined key events and their surrounding time windows.

[0070] When any data with task ID = "TF-20250508" (such as the GPS location data of drone 1, the UPS input voltage data of the field hospital) is successfully stored in any logical storage tier (Tier 1, Tier 2, or Tier 3) of the cloud platform, an asynchronous message triggered by the data persistence service will notify the "Global Secondary Index Service for Task IDs" (such as an Elasticsearch cluster) deployed on the cloud platform.

[0071] Furthermore, the data structure of the global secondary index for task IDs is as follows: using the task ID as the main index key, the corresponding value is a list, and this set contains the unique identifiers of all data records associated with this task ID, the indication of the current logical storage level where these data records are located, their respective timestamp ranges, data domain classifications, and current business criticality level information;

[0072] Example of data structure: For the main index key with task ID = "TF-20250508", its corresponding value in Elasticsearch is an ever-updating list of documents, and the structure of each document (representing an element in the list) is as follows:

[0073] { "Unique identifier": "UAV1_GPS_pos_20250508T141025500Z", "Indication of the current logical storage level": "Tier1_InfluxDB_RawFlightData", "Timestamp range": {"gte": "2025-05-08T14:10:25.500Z", "lte": "2025-05-08T14:10:25.500Z"}, "Data domain classification": "UAV telemetry_flight parameters", "Current business criticality level information": 5}

[0074] Furthermore, the task- and event-oriented fusion index construction module includes:

[0075] Provide an interface for defining key event types, configure key event triggering conditions, which can be based on specific numerical values, status sequences, and combinatorial logic of the data source, and set time window parameters for capturing context data for each event type;

[0076] For example, the system administrator configures key event types for tasks through the management interface of the cloud platform. For example, define an event type named "UAV_Critical_Link_Degradation" (deterioration of the UAV's critical link quality):

[0077] Configure key event triggering conditions: The data source is the "communication link quality data between UAVs" in the "UAV telemetry data stream". When the data transmission rate < 2Mbps and the packet loss rate > 5% for more than 3 seconds.

[0078] Set time window parameters for capturing context data: From T - 30 seconds to T + 15 seconds before and after the event occurrence time (T).

[0079] Associated data fields: "UAV Telemetry_Communication Link" (link data of the UAV and another UAV communicating with it), "UAV Telemetry_Flight Parameters" (GPS positions, altitudes, and attitudes of the two UAVs), "Mobile Cabin Equipment_Signal Receiver" (lock status of the signal receiver);

[0080] When a critical event that meets predefined conditions is detected, the system automatically records the unique ID, type, occurrence timestamp, associated mobile cabin ID, and task ID of the event, and retrieves and aggregates all the original data records from multiple predefined relevant data fields within the time window before and after the occurrence of the event from the logical storage level according to the time window parameters set for the event type, forming a context data set;

[0081] Furthermore, during the execution of the task "TF-20250508", the communication link between UAV 1 and UAV 2 triggered the "UAV_Critical_Link_Degradation" event (denoted as EVT_005) at 15:20:30Z;

[0082] The real-time CEP (Complex Event Processing) engine of the cloud platform detected this situation and immediately automatically recorded the unique ID = "EVT_005", type, occurrence timestamp, associated mobile cabin ID = "FC-001", and task ID = "TF-20250508" of the event, and recorded the UAVs involved as UAV 1 and UAV 2;

[0083] The "Context Data Aggregation Service" (cloud platform microservice) retrieves and aggregates the communication link data and flight parameters of UAV 1 and UAV 2 during this period, as well as the status data of the FC-001 signal receiver during this period from each logical storage level (assisted by global secondary indexing through the task ID) according to the time window (15:20:00Z - 15:20:45Z) set for the event type, forming the "context data set" of EVT_005;

[0084] A preset feature extraction algorithm is used to generate a context data fingerprint for the context data set, and the fingerprint is based on the temporal features of the data within the set;

[0085] Furthermore, for the "context data set" of EVT_005, the "feature extraction service" of the cloud platform uses a preset feature extraction algorithm (LSTM) to process the time series features of the data therein: For example, for the time series of "data transmission rate" (data within the window) between UAV 1 and UAV 2, calculate its minimum value, mean value, last value, and linear regression slope. For the time series of "packet loss rate", calculate its maximum value, mean value, and number of sudden increases. For the time series of the relative distance and relative speed between the two UAVs, calculate their mean values and ranges of variation.

[0086] These extracted time series feature values are combined into a standardized digital vector (for example, by PCA dimensionality reduction or direct splicing followed by normalization) as the "context data fingerprint" of EVT_005.

[0087] Build a fusion event index library, which persistently associates and stores the event ID, event type, task ID with the generated context data fingerprint and each original data segment that constitutes the context. The index library supports direct lookup based on the event ID and similarity retrieval based on the fingerprint vector.

[0088] Persistently store the event ID = "EVT_005", event type, task ID, the generated context data fingerprint vector, and the exact location information or access handle of each original data segment that constitutes the context in the corresponding storage layers of the cloud platform as an entry in the fusion event index library on the cloud platform (for example, a dedicated database configured with a Faiss vector retrieval library).

[0089] If the communication link between UAVs becomes abnormal again in a subsequent task, the system can calculate its context fingerprint and search the index library for the event that was most similar to its "context data fingerprint" (i.e., the time series feature vector) in history, so as to possibly find similar events such as EVT_005 and their detailed data at that time, providing a reference for troubleshooting or performance analysis.

[0090] In this embodiment, by constructing an adaptive hierarchical storage and dynamic routing mechanism based on data portraits (including domain function attributes, business criticality levels, task IDs, and task phases), data in the shelter can be intelligently allocated to logical storage levels with different cost and performance characteristics (such as real-time hot storage layer, warm analysis layer, long-term cold storage layer) according to different values and access patterns. Combined with refined data life cycle management strategies (including on-demand aggregation and downsampling), it avoids the problems of massive data management brought about by traditional "one-size-fits-all" or rough hierarchical storage based only on time. This not only significantly reduces the cloud storage cost, but more importantly, by combining with task- and event-oriented context-aware fusion indexes (especially the global secondary index of task IDs and fusion event indexes), when remote visualization applications request task data or retrieve context data of key events, fast and accurate positioning and retrieval can be achieved, significantly shortening the data access latency and improving the overall efficiency of monitoring shelter data management.

[0091] The context-aware fusion index construction module for tasks and events proposed in this embodiment, especially the "context data fingerprint and fusion event index" constructed for predefined key events and their multi-source context data sets, enables the system to go beyond simple time-series data queries. It supports direct context backtracking based on event IDs and more valuable similarity retrieval based on "fingerprint vectors", so as to quickly locate and analyze complex events or fault patterns with similar data characteristics that occurred in history. Combining the integration capabilities of the multi-level data federated query module for cross-level and cross-data source information, it provides remote users with the ability to perform in-depth correlation analysis and intelligent insights on "business system data", cabin status data, and various event data generated during the execution of complex tasks (such as aircraft flight tests) by the "intelligent shelter monitoring integration system" (such as flight test shelters), improving the overall efficiency of monitoring shelter data management.

[0092] The multi-level data federated query and visualization service module is used to provide a unified query interface, parse user query requests, and generate and execute a federated query plan across storage levels based on hierarchical storage information and fusion event indexes.

[0093] Furthermore, the execution process of the federated query plan includes:

[0094] After receiving a request from the unified query API, the query planner parses the time range, shelter ID, task ID, and event ID included in the query, and preferentially uses the global secondary index of task IDs and fusion event indexes for preliminary data positioning;

[0095] The remote test flight mission command initiated a query request through the remote visualization monitoring application provided by the cloud platform: View the flight altitude, speed and communication link quality parameters of all UAVs 1 and 2 related to the event "EVT_005" (link deterioration between UAVs 1 and 2) in the mission "TF-KZ01-20250508" within 30 seconds before and after the event, and request that the data be displayed at a granularity of 100 milliseconds.

[0096] The cloud platform's unified query API (such as a GraphQL endpoint) receives this request. The query planner parses the request and extracts the time range contained in the query (calculated by the timestamp of EVT_005 and 30 seconds before and after), the cabin ID = "FC-001" (implicit or explicit), the task ID = "TF-20250508", and the event ID = "EVT_005".

[0097] Prioritize indexes: The query planner first uses the "Fusion Event Index Library" to obtain the location list of the original data fragments that constitute its context through event ID="EVT_005".

[0098] According to the positioning results and the distribution of data at each logical storage level, the original query is decomposed into multiple optimized sub-queries for different physical storage backends, including time series databases, relational databases, and object storage. For example, it is found that the latest flight parameters of drone 1 are still in Tier 1's InfluxDB, while drone 2's data from the same period and some link quality historical data have entered Tier 2's TimescaleDB, and earlier FC-001-related device logs may have been archived in Tier 3's S3 object storage. The query planner decomposes the original user's macro query request into multiple optimized sub-queries for different physical storage backends (including time series databases such as InfluxDB, relational databases such as TimescaleDB (whose core is PostgreSQL), and object storage such as S3).

[0099] Execute the subqueries and perform batch intelligent fusion processing on the result sets returned by each subquery, wherein the fusion processing includes data merging and sorting to ensure the continuity of time series, data deduplication based on unique keys, conversion of data structures from different sources into a unified output format, and dynamic aggregation operations on the server side according to query requests.

[0100] The "federated query execution engine" of the cloud platform (e.g., a query engine built on Apache Calcite, or a Presto / Trino cluster) distributes these subqueries concurrently (or according to dependencies) to the corresponding InfluxDB service, TimescaleDB service, and S3 service on the cloud platform through its built-in or pluggable database adapters (possibly querying Parquet files on S3 through Athena or Spark SQL).

[0101] After collecting the result sets returned by all subqueries, perform intelligent fusion processing, including:

[0102] Data merging and sorting: Globally merge and precisely sort the data of UAV 1 and UAV 2 from different sources (InfluxDB, TimescaleDB) strictly according to timestamps to ensure the continuity of the time series and the comparability between different parameters.

[0103] Data deduplication based on unique keys: Ensure that there are no duplicate data points.

[0104] Unified output format conversion: Convert all data into the unified JSON or other standardized formats expected by the front-end visualization components (or subsequent analysis tools).

[0105] Server-side dynamic aggregation operation: If the front-end requests data at a granularity of 100 milliseconds, and a certain subquery returns finer raw data (such as 10 milliseconds) from Tier 2, then perform dynamic aggregation (e.g., taking the average or the first value of every 10 points) on this part of the data on the server side to match the required granularity.

[0106] In this embodiment, through the unified query interface and intelligent query planning and execution mechanism provided by the multi-level data federated query and visualization service module, the upper-layer remote visualization application does not need to care about the actual storage location and heterogeneity of the underlying data. This module can automatically optimize the query path, execute subqueries in parallel, and perform intelligent fusion and visualization-oriented preprocessing (such as downsampling, filling points) on the results, so as to ensure that when processing large-scale and multi-dimensional shelter data, the visualization interface can still obtain fast and smooth data response, greatly improving the user experience. At the same time, the modular design (data access and annotation, hierarchical storage, fusion index, federated query) makes the functions of each part of the system cohesive and reduces the coupling degree, thus improving the overall efficiency of the monitoring shelter data management.

[0107] In this embodiment, multi-dimensional monitoring data of the mobile cabin is obtained, and context metadata including domain function attributes, business criticality levels, task IDs, and task stages is automatically labeled; a dynamic data portrait is constructed based on the metadata, and the data is routed to logical storage levels with different costs and performances through a rule engine, and lifecycle management is executed; a global secondary index for task IDs is constructed, and data fingerprints and fusion event indexes are constructed for key events and their context data sets; a unified query interface is provided, and a cross-level federated query plan is executed using hierarchical storage information and fusion indexes. The present invention can effectively improve the overall efficiency of mobile cabin data management.

[0108] Embodiment 2

[0109] In order to improve the overall efficiency of data management for the B test flight mobile cabin, a remote visualization monitoring mobile cabin data management system based on a cloud platform is applied. Figure 1 As shown in the structural schematic diagram of a remote visualization monitoring mobile cabin data management system provided by an embodiment of the present invention, it includes:

[0110] An intelligent data access and automatic annotation module, which is used to obtain multi-dimensional monitoring data of the mobile cabin and transmit it to the cloud platform, and the cloud platform automatically annotates the metadata; the metadata includes domain function attributes, business criticality levels, task IDs, and task stages;

[0111] Further, the multi-dimensional monitoring data includes UAV telemetry data streams, power supply system status data, equipment status data, and environmental data.

[0112] An adaptive hierarchical storage and dynamic routing module based on data portraits, which is used to construct a dynamic data portrait according to the metadata and data timestamps, and based on this data portrait, combined with a preset rule engine, dynamically route the data to logical storage levels with different storage costs and access performances in the cloud platform for storage, and execute corresponding data lifecycle management policies;

[0113] Further, the adaptive hierarchical storage and dynamic routing module based on data portraits includes:

[0114] A data portrait generation unit, which is used to clearly construct a structured data portrait for each data record labeled with metadata. This portrait includes data domain identifiers, current business criticality levels, task IDs, task stage identifiers, original data timestamps, and cloud platform reception timestamps;

[0115] A rule storage unit, which is used to execute the operation of routing the data to a specified layer in the logical storage levels including a real-time hot storage layer, a warm analysis layer, and a long-term cold storage layer based on the attribute combinations in the data portrait according to a preset rule engine and combined with a preset storage policy rule.

[0116] Further, the data life cycle management policy includes the following steps:

[0117] Status evaluation and decision-making: Triggered by task phase change events, perform status evaluation on data records in each logical storage level, and based on the current business critical level, data timeliness, and the current phase of the task included in their data portraits, and refer to the preset life cycle rules to determine whether the data records meet the level migration conditions;

[0118] Data optimization processing: If the condition for migrating from a storage level with high access performance to a storage level with low access performance is met, then according to the life cycle rules, perform predefined optimization processing on the data before migration. The optimization processing includes aggregating high-frequency raw data into statistical summaries and performing downsampling on redundant data points;

[0119] Data migration and metadata synchronization: Securely migrate the optimized data to the target logical storage level specified by the life cycle rules, and synchronously update the storage locations of this data in the global index and metadata information.

[0120] A task- and event-oriented context-aware fusion index construction module for constructing a global secondary index of task IDs and constructing context data fingerprints and fusion event indexes for a multi-source context data set within a predefined critical event and its time window before and after;

[0121] Further, the data structure of the global secondary index of task IDs is: using the task ID as the main index key, and its corresponding value is a list. The set contains the unique identifiers of all data records associated with this task ID, the indication of the logical storage level where these data records are currently located, their respective timestamp ranges, data domain classifications, and current business critical level information.

[0122] Further, the task- and event-oriented fusion index construction module includes:

[0123] Provide a key event type definition interface to configure key event trigger conditions. The conditions can be based on specific numerical values, status sequences, and combinatorial logic of data sources, and set time window parameters for context data capture for each event type;

[0124] When a key event that meets the predefined conditions is detected, the system automatically records the unique ID, type, occurrence timestamp, associated shelter ID, and task ID of the event, and retrieves and aggregates all raw data records from multiple preset relevant data domains within the time window before and after the occurrence of the event from the logical storage level according to the time window parameters set for the event type to form a context data set;

[0125] Generate a context data fingerprint for the context data set using a preset feature extraction algorithm, where the fingerprint is based on the temporal features of the data within the set;

[0126] Construct a fusion event index library that persistently associates and stores event IDs, event types, task IDs with the generated context data fingerprints and each original data segment that constitutes the context. The index library supports direct lookup based on event IDs and similarity retrieval based on fingerprint vectors.

[0127] In this embodiment, by constructing an adaptive hierarchical storage and dynamic routing mechanism based on data portraits (including domain function attributes, business criticality levels, task IDs, and task phases), the data in the shelter can be intelligently allocated to logical storage levels with different costs and performance characteristics (such as real-time hot storage layer, warm analysis layer, long-term cold storage layer) according to different values and access patterns. Combined with a refined data life cycle management strategy (including on-demand aggregation and downsampling), it avoids the problem of massive data management brought about by the traditional "one-size-fits-all" or rough time-based hierarchical storage. This not only significantly reduces the cloud storage cost, but more importantly, by combining with task- and event-oriented context-aware fusion indexes (especially the global secondary index of task IDs and the fusion event index), when the remote visualization application requests task data or retraces the context data of key events, it can achieve fast and accurate positioning and retrieval, significantly shortening the data access latency and improving the overall efficiency of monitoring the shelter data management.

[0128] The context-aware fusion index construction module for tasks and events proposed in this embodiment, especially the "context data fingerprint and fusion event index" constructed for predefined key events and their multi-source context data sets, enables the system to go beyond simple time-series data queries. It supports direct context backtracking based on event IDs and more valuable similarity retrieval based on "fingerprint vectors", so as to quickly locate and analyze complex events or fault patterns with similar data characteristics that occurred in history. Combined with the integration ability of the multi-level data federated query module for cross-level and cross-data source information, it provides remote users with the ability to deeply associate and analyze and gain intelligent insights into the "business system data", the state data of the cabin itself, and various event data generated during the execution of complex tasks (such as aircraft flight tests) by the "intelligent shelter monitoring integration system" (such as the flight test shelter), improving the overall efficiency of monitoring the shelter data management.

[0129] The multi-level data federated query and visualization service module is used to provide a unified query interface, parse user query requests, and generate and execute a federated query plan across storage levels based on hierarchical storage information and fusion event indexes.

[0130] Furthermore, the process of executing the federated query plan includes:

[0131] After receiving a request from the unified query API, the query planner parses the time range, shelter ID, task ID, and event ID included in the query, and preferentially uses the global secondary index of the task ID and the fused event index for preliminary data location;

[0132] According to the location result and the distribution of data at each logical storage level, the original query is decomposed into multiple optimized sub-queries for different physical storage backends, including time series databases, relational databases, and object storage;

[0133] Execute the sub-queries, and perform batch-style intelligent fusion processing on the result sets returned by each sub-query. The fusion processing includes data merging and sorting to ensure time series continuity, data deduplication based on unique keys, converting data structures from different sources into a unified output format, and performing dynamic aggregation operations on the server side according to the query request.

[0134] In this embodiment, through the unified query interface and the intelligent query planning and execution mechanism provided by the multi-level data federation query and visualization service module, the upper-layer remote visualization application does not need to care about the actual storage location and heterogeneity of the underlying data. This module can automatically optimize the query path, execute sub-queries in parallel, and perform intelligent fusion and visualization-oriented preprocessing (such as downsampling, filling points) on the results, so as to ensure that when processing large-scale and multi-dimensional shelter data, the visualization interface can still obtain fast and smooth data responses, greatly improving the user experience. At the same time, the modular design (data access and annotation, hierarchical storage, fusion index, federation query) makes the functions of each part of the system cohesive and reduces the coupling degree, thus improving the overall efficiency of monitoring shelter data management.

[0135] In this embodiment, by obtaining multi-dimensional monitoring data of the shelter and automatically annotating the context metadata including domain function attributes, business key levels, task IDs, and task phases; constructing a dynamic data portrait according to the metadata, routing the data to logical storage levels with different costs and performances through a rule engine, and performing life cycle management; constructing a global secondary index of the task ID, and constructing data fingerprints and fused event indexes for key events and their context data sets; providing a unified query interface, and using hierarchical storage information and fused indexes to execute a cross-level federation query plan. The present invention can effectively improve the overall efficiency of shelter data management.

[0136] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A remote visualization monitoring mobile cabin data management system based on a cloud platform, characterized in that, It includes: An intelligent data access and automatic annotation module, which is used to obtain multi-dimensional monitoring data of the mobile cabin and transmit it to the cloud platform, and the cloud platform automatically annotates the metadata; The metadata includes domain function attributes, business criticality levels, task IDs, and task phases; An adaptive hierarchical storage and dynamic routing module based on data portraits, which is used to construct a dynamic data portrait according to the metadata and data timestamps, and based on this data portrait, combined with a preset rule engine, dynamically route the data to logical storage levels with different storage costs and access performances in the cloud platform for storage, and execute corresponding data life cycle management strategies; A context-aware fusion index construction module for tasks and events, which is used to construct a global secondary index for task IDs, and construct context data fingerprints and fusion event indexes for multi-source context data sets within predefined key events and their surrounding time windows; A multi-level data federated query and visualization service module, which is used to provide a unified query interface, parse user query requests, and generate and execute a federated query plan across storage levels based on hierarchical storage information and fusion event indexes.

2. The data management system for remote visualization monitoring of a mobile cabin based on a cloud platform according to claim 1, wherein: The multi-dimensional monitoring data includes unmanned aerial vehicle telemetry data streams, power supply system status data, equipment status data, and environmental data.

3. The data management system for remote visual monitoring cabin based on cloud platform according to claim 1, characterized in that: The adaptive hierarchical storage and dynamic routing module based on data portraits includes: A data portrait generation unit, which is used to construct a structured data portrait for each data record annotated with metadata. The portrait includes a data domain identifier, the current business criticality level, a task ID, a task phase identifier, an original data timestamp, and a cloud platform reception timestamp; A rule storage unit, which is used to execute the operation of routing the data to a specified layer in the logical storage levels including a real-time hot storage layer, a warm analysis layer, and a long-term cold storage layer based on the attribute combinations in the data portrait according to a preset rule engine and combined with a preset storage policy rule.

4. The remote visualization monitoring shelter data management system based on a cloud platform according to claim 1, characterized in that: The data life cycle management strategy includes the following steps: Status evaluation and decision-making: Triggered by task phase change events, perform status evaluation on data records in each logical storage level, and based on the current business criticality level, data timeliness, and the current phase of the task included in their data portraits, and referring to the preset life cycle rules, determine whether the data record meets the hierarchical migration conditions; Data optimization processing: If it meets the conditions for migrating from a storage level with high access performance to a storage level with low access performance, then according to the life cycle rules, perform predefined optimization processing on the data before migration. The optimization processing includes aggregating high-frequency raw data into statistical summaries and performing downsampling on redundant data points; Data migration and metadata synchronization: Securely migrate the optimized data to the target logical storage level specified by the life cycle rules, and synchronously update the storage location of the data in the global index and metadata information.

5. The data management system for remote visual monitoring of a shelter based on a cloud platform according to claim 1, wherein: The data structure of the global secondary index of the task ID is as follows: using the task ID as the main index key, and the corresponding value is a list. The set contains the unique identifiers of all data records associated with the task ID, the indication of the logical storage level where these data records are currently located, their respective timestamp ranges, data domain classifications, and current business critical level information.

6. The data management system for remote visualization monitoring cabin based on cloud platform according to claim 1, characterized in that: The task- and event-oriented fusion index construction module includes: Providing a key event type definition interface, configuring key event trigger conditions, which can be based on the numerical values, status sequences, and combinational logic of data sources, and setting time window parameters for context data capture for each event type; When a key event that meets the predefined conditions is detected, the system automatically records the unique ID, type, occurrence timestamp, associated shelter ID, and task ID of the event, and retrieves and aggregates all the original data records from multiple preset relevant data domains within the time window before and after the occurrence of the event from the logical storage level according to the time window parameters set for the event type to form a context data set; Using a preset feature extraction algorithm to generate a context data fingerprint for the context data set, and the fingerprint is obtained based on the temporal features of the data in the set; Constructing a fusion event index library, which persistently associates and stores the event ID, event type, task ID, the generated context data fingerprint, and each original data segment that constitutes the context. The index library supports direct lookup based on the event ID and similarity retrieval based on the fingerprint vector.

7. The data management system for remote visual monitoring cabin based on cloud platform according to claim 1, characterized in that: The execution process of the federated query plan includes: After the query planner receives a request from the unified query API, it parses the time range, shelter ID, task ID, and event ID included in the query, and preferentially uses the global secondary index of the task ID and the fusion event index for preliminary data positioning; According to the positioning results and the distribution of data in each logical storage level, decompose the original query into multiple optimized sub-queries for different physical storage backends, including time series databases, relational databases, and object storage; Execute the sub-queries, and perform batch-style intelligent fusion processing on the result sets returned by each sub-query. The fusion processing includes data merging and sorting to ensure the continuity of the time series, data deduplication based on unique keys, converting data structures from different sources into a unified output format, and performing dynamic aggregation operations on the server side according to the query request.

Images