Data space construction and quick query method and device and readable storage medium

Abstract

The invention provides a data space construction and quick query method and device and a readable storage medium. The method comprises the following steps: preprocessing 5G-A general sensing multi-modal data; constructing a 5G-A general sensing adaptive knowledge hypergraph; carrying out hypergraph matrix processing; extracting a 5G-A scene association relationship; executing relational storage mapping; and executing 5G-A intelligent query. According to the method, entity-attribute-association complete data logic is constructed through knowledge hypergraph modeling, many-to-many high-order association of multi-modal data can be accurately borne, and association relationship expression of general sensing data better meets 5G-A scene requirements; a technical path of hypergraph matrix-relational storage mapping is designed, and the existing 5G-A application ecology can be quickly fused; according to the 5G-A communication-sensing integrated core characteristic, storage logic specially associated with sensing is reserved, service data can be directly output during query, and the limitation that only single-dimension data is output in an existing scheme is avoided.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a data space construction and fast query method, apparatus and readable storage medium. Background Technology

[0002] With the evolution of 5G-A (5G-Advanced) technology, integrated communication and sensing (referred to as "integrated sensing") has become one of its core features. The network can not only provide communication services but also utilize wireless signals for high-precision environmental sensing (such as positioning, speed measurement, imaging, and target recognition). This will generate massive amounts of heterogeneous, multimodal, and strongly spatiotemporally correlated sensing fusion data, including but not limited to: raw channel state information, sensed point clouds / images, calculated target trajectories and features, network performance indicators, and user service data.

[0003] Currently, the storage and management of this data mostly uses traditional relational databases or time-series databases, which have shortcomings such as insufficient adaptability of relational database storage, weak ecological compatibility of dedicated storage systems, and failure of relational storage to match the characteristics of 5G-A sensing scenarios.

[0004] Therefore, there is an urgent need for a data space construction and query method suitable for 5G-A integrated sensing data. Summary of the Invention

[0005] The technical problem to be solved by this application is to provide a data space construction and fast query method, apparatus and readable storage medium to address the above-mentioned shortcomings of the prior art.

[0006] Firstly, this application provides a method for constructing a data space and performing fast queries, the method comprising: S1, 5G-A Sensing Multimodal Data Preprocessing: Connect to the 5G-A base station interface and perform multimodal data parsing to extract multimodal features and entity attributes with Precise Time Protocol (PTP) spatiotemporal labels; S2. Construct a 5G-A sensory-adaptive knowledge hypergraph: Generate entity nodes and hyperedges based on multimodal features and entity attributes to obtain a knowledge hypergraph containing communication-sensing linkage relationships; S3. Perform hypergraph matrixization: transform the knowledge hypergraph into an entity-feature matrix, a hyperedge-entity association matrix, and a spatiotemporal-communication perception fusion tensor. S4. Extract 5G-A contextualized relationships: Based on entity-feature matrix, hyperedge-entity association matrix and spatiotemporal-communication perception fusion tensor, the contextualized relationships between entities are mined through matrix operations and scene rule engine; S5. Perform relational storage mapping: Map entity nodes, multimodal features, and contextual relationships between entities to corresponding table structures and write them to the relational database. S6. Execute 5G-A intelligent query: Parse business query requests and convert them into structured query language SQL statements, and call the relational database indexing mechanism to complete the query and result aggregation.

[0007] In some embodiments, in S1, the multimodal data includes at least two of radar data, channel state information (CSI) data, visual data, and communication data; Multimodal data parsing includes at least two of the following processes: Radar data processing: After receiving high-resolution radar signals from 5G-A base stations, the data undergoes digital down-conversion, pulse compression, moving target display filtering, and constant false alarm rate detection to output point cloud data. Geometric features are extracted using point cloud feature extraction algorithms and bound to PTP nanosecond-level timestamps and unified coordinate system spatial coordinates. CSI data processing: After receiving the CSI matrix from the 5G-A base station, the data undergoes phase calibration, noise reduction, and time-frequency domain feature extraction to output CSI time-frequency features; it is then associated with target environment perception parameters and bound to a PTP timestamp. Visual data processing: After receiving the ultra-high-definition visual stream from the 5G-A base station, the system performs decoding, target detection, and multi-target tracking to generate stable entity identifiers across frames; after extracting visual appearance features, it binds them with PTP timestamps. Communication data processing: Receive the QoS indicators of the 5G-A base station, associate them with the corresponding targets according to the entity identifier, form communication status parameters and bind them with PTP timestamps.

[0008] In some embodiments, S2 includes: S21. Entity Node Generation: Relying on the edge computing unit of the 5G-A base station, multimodal joint tracking operation is performed. For each tracked stable target, a unique entity identifier (entity_id) is assigned, and the target type (target_type), PTP timestamp (ptp_timestamp), spatial coordinates (spatial_coord), and communication state (comm_state) are synchronously associated to form complete entity node information. S22, Hyperedge generation: Intramodal hyperedge: The correlation between features in the same modality is analyzed by edge computing units. When the calculated correlation coefficient is greater than or equal to the preset correlation threshold, the corresponding intramodal hyperedge is created and the corresponding correlation coefficient is recorded as the hyperedge attribute. Cross-modal hyperedge: The matching degree between different modal features is calculated using the 5G-A sensor fusion algorithm. If the matching degree is greater than or equal to the preset matching degree threshold, a cross-modal hyperedge is created, and the modal fusion confidence is recorded as a hyperedge attribute. Communication-Sensing Linked Hyperedge: The edge computing unit detects the correlation logic between communication parameters and sensing results in real time. When the preset linkage conditions are met, a communication-sensing linked hyperedge is created and the corresponding linkage strategy is recorded. S23. Superedge optimization: Remove low-confidence superedges and retain strong correlations with 5G-A service requirements.

[0009] In some embodiments, in S3 For the entity-feature matrix: the row dimension corresponds to the number of 5G-A synesthetic entities, the column dimension corresponds to the total dimension of the multimodal features, and the matrix elements are the feature values ​​corresponding to the entities; For the hyperedge-entity association matrix: the row dimension corresponds to the type of 5G-A scenario-based hyperedge, the column dimension corresponds to the entity combination, and the matrix elements are boolean values ​​used to mark whether the entity combination has a corresponding hyperedge association; For the spatiotemporal-communication sensing fusion tensor: according to the time slice of 5G-A sensing data, the entity-feature matrix and the hyperedge-entity association matrix are stacked to form a 3D spatiotemporal-communication sensing fusion tensor.

[0010] In some embodiments, in S4, the rule engine has built-in 5G-A specific association rules, which specifically include: Low-altitude UAV formation rules: When two UAV entities meet the following conditions, such as spatial distance less than or equal to formation spacing threshold, speed difference less than or equal to cooperative speed threshold, stable communication link status, and continuous coverage greater than or equal to the preset number of time slices, they are marked as "formation following" associations. The association attributes include formation identification information. Vehicle-road cooperative rules: When the spatial distance between a vehicle and a roadside unit is less than or equal to the communication coverage radius, the Communication Channel Quality Indicator (CQI) is greater than or equal to the service assurance threshold, and the radar detects the vehicle's turning / overtaking intention, a "vehicle-road cooperative warning" association is marked. The association attributes include the specific warning type. Communication-Sensing Linkage Rule: When the entity's communication parameters show "CQI less than the sensing optimization threshold" or "Reference Signal Received Power (RSRP) less than the link guarantee threshold", mark it as "Sensing Strategy Enhancement" and the associated attribute contains the sensing algorithm adjustment instruction.

[0011] In some embodiments, S5 includes: S51. Map the entity nodes to an entity base table used to store entity node information. S52. Map the multimodal features into modal attribute tables for storing the high-dimensional features and key parameters of the corresponding modalities; S53. Map the scenario-based relationships to hyper-edge association tables, including intra-modal association tables, cross-modal association tables, and communication-perception linkage association tables.

[0012] In some embodiments, S6 includes: S61. Query parsing process: Extract key dimensions, relationship types, entity attribute filtering conditions, and result fields from the business query request to obtain the query logic framework. S62, SQL Conversion Operation: Maps the query logic framework to a multi-table JOIN SQL statement supported by a relational database; S63. Index call optimization: When executing a query, enable the pre-designed composite index in the relational database to quickly locate the target data through the index; S64. Result Aggregation Processing: Based on entity identifier and timestamp, the results obtained from multi-table queries are correlated and matched to verify whether each entity meets the filtering conditions; the result data of entities that meet the conditions are integrated and the query results are output in the format specified by business requirements.

[0013] Secondly, this application provides a data space construction and fast query device, the device comprising: The 5G-A sensing multimodal data preprocessing module is configured to perform 5G-A sensing multimodal data preprocessing: it connects to the 5G-A base station interface and performs multimodal data parsing to extract multimodal features and entity attributes with Precise Time Protocol (PTP) spatiotemporal tags; The 5G-A sensory-adaptive knowledge hypergraph construction module is configured to construct a 5G-A sensory-adaptive knowledge hypergraph by generating entity nodes and hyperedges based on multimodal features and entity attributes, thereby obtaining a knowledge hypergraph containing communication-sensing linkage relationships. The Hypergraph Matrixing module is configured to perform hypergraph matrixing: transforming the knowledge hypergraph into an entity-feature matrix, a hyperedge-entity association matrix, and a spatiotemporal-communication perception fusion tensor; The 5G-A scenario-based association extraction module is configured to extract 5G-A scenario-based association relationships: based on entity-feature matrix, hyperedge-entity association matrix and spatiotemporal-communication perception fusion tensor, it mines the scenario-based association relationships between entities through matrix operations and scenario rule engine; The relational storage mapping module is configured to perform relational storage mapping: mapping entity nodes, multimodal features, and contextual relationships between entities to corresponding table structures and writing them to the relational database; The 5G-A intelligent query adaptation module is configured to execute 5G-A intelligent queries: parse business query requests and convert them into structured query language (SQL) statements, and call relational database indexing mechanisms to complete the query and result aggregation.

[0014] Thirdly, this application provides a data space construction and fast query apparatus, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to implement the data space construction and fast query method described in the first aspect above.

[0015] Fourthly, this application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the data space construction and fast query method described in the first aspect.

[0016] This application provides a data space construction and fast query method, apparatus, and readable storage medium. The method includes: 5G-A sensing multimodal data preprocessing: interfacing with a 5G-A base station interface and performing multimodal data parsing to extract multimodal features and entity attributes with Precise Time Protocol (PTP) spatiotemporal tags; constructing a 5G-A sensing-adaptive knowledge hypergraph: generating entity nodes and hyperedges based on multimodal features and entity attributes to obtain a knowledge hypergraph containing communication-sensing linkage relationships; performing hypergraph matrixing: converting the knowledge hypergraph into an entity-feature matrix, a hyperedge-entity association matrix, and a spatiotemporal-communication-sensing fusion tensor; extracting 5G-A scenario-based association relationships: based on the entity-feature matrix, hyperedge-entity association matrix, and spatiotemporal-communication-sensing fusion tensor, mining scenario-based association relationships between entities through matrix operations and a scenario rule engine; performing relational storage mapping: mapping entity nodes, multimodal features, and scenario-based association relationships between entities to corresponding table structures and writing them into a relational database; and performing 5G-A intelligent query: parsing business query requests and converting them into Structured Query Language (SQL). The statement invokes the relational database indexing mechanism to complete the query and result aggregation. This application constructs a complete data logic of "entity-attribute-relationship" through knowledge hypergraph modeling, which can accurately carry many-to-many high-order relationships of multimodal data such as radar, CSI, and vision. It avoids the inefficiency and semantic disconnection problems caused by the reliance on complex table joins in traditional relational databases, making the expression of the relationship of sensory data more in line with the needs of 5G-A scenarios. This application designs a technical path of "hypergraph matrixization → relational storage mapping", which transforms the knowledge hypergraph into a standard relational table structure. It can be directly compatible with existing SQL business systems without the need for additional development of adaptation interfaces or modification of existing architecture, which greatly reduces the deployment cost and maintenance difficulty of the 5G-A sensory data storage system and can be quickly integrated into the existing 5G-A application ecosystem. This application specifically reserves storage logic for the "communication-sensing integration" core feature of 5G-A, such as "communication parameters-sensing strategies", which can completely retain key related information. When querying, there is no need to splice data across databases. It can directly output data that supports low-altitude security, vehicle-road cooperation and other services, avoiding the limitation of existing solutions that only output single-dimensional data. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] Figure 1 A flowchart illustrating a data space construction and fast query method provided in this application embodiment; Figure 2 A flowchart illustrating yet another data space construction and fast query method provided in this application embodiment; Figure 3 A schematic diagram illustrating the construction of a 5G-A sensory-adaptive knowledge hypergraph provided in an embodiment of this application; Figure 4 This is a schematic diagram of hypergraph matrixing processing provided in an embodiment of this application; Figure 5 A schematic diagram of the structure of a data space construction and fast query device provided in an embodiment of this application; Figure 6 This is a schematic diagram of another data space construction and fast query device provided in an embodiment of this application.

[0019] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0020] To enable those skilled in the art to better understand the technical solution of this application, the embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0021] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining this application and are not intended to limit this application.

[0022] It is understood that, without conflict, the various embodiments and features in the embodiments of this application can be combined with each other.

[0023] It is understood that, for ease of description, only the parts relevant to this application are shown in the accompanying drawings, while parts unrelated to this application are not shown in the drawings.

[0024] It is understood that each unit or module involved in the embodiments of this application may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple units or modules may be integrated into one entity structure.

[0025] It is understood that the terms "first," "second," etc., used in the embodiments of this application are used to distinguish different objects or to distinguish different treatments of the same object, rather than to describe a specific order of objects.

[0026] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this application may occur in a different order than those marked in the accompanying drawings.

[0027] It is understood that the flowcharts and block diagrams of this application illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this application. Each block in a flowchart or block diagram may represent a unit, module, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagrams and flowcharts may be implemented using a hardware-based system to implement the specified function, or using a combination of hardware and computer instructions.

[0028] It is understood that the units and modules involved in the embodiments of this application can be implemented by software or by hardware. For example, the units and modules can be located in the processor.

[0029] It is understood that the specific values ​​of each parameter in this application are merely illustrative examples, and in practical applications, the parameters can be optimized and adjusted based on specific requirements.

[0030] In existing technologies, the 5G-A integrated sensing data storage and management solution mainly suffers from the following drawbacks: a) Insufficient adaptability of relational database storage: Traditional relational databases rely on two-dimensional table structures, which can only carry flat associations of "entity-attribute", and cannot efficiently map the multimodal many-to-many high-order relationships unique to 5G-A sensory data; if cross-modal data queries are required, they must be completed through multiple complex table join operations, which is not only logically complex, but also difficult to meet the stringent requirements of 5G-A services for query latency.

[0031] b) Weak ecosystem compatibility of dedicated storage systems: Although dedicated storage solutions such as graph databases are good at handling relationships, their syntax system is significantly different from existing SQL business systems. They require the development of dedicated adaptation interfaces to achieve data interoperability, which increases the system deployment cost and the difficulty of later operation and maintenance, making it difficult to quickly integrate into the existing 5G-A network application ecosystem.

[0032] c) The storage of association relationships does not match the characteristics of 5G-A sensing scenarios: The existing storage solution is not designed for the core characteristics of 5G-A "communication-sensing integration" and cannot effectively store high-level associations specific to sensing; this results in query results that can only provide single-dimensional data and cannot fully output the association information that supports 5G-A sensing business decisions, making it difficult to meet the actual needs of 5G-A sensing scenarios.

[0033] To address the aforementioned shortcomings, this application provides a data space construction and rapid query method suitable for 5G-A integrated sensing data. The main concepts of this application include: (1) By using knowledge hypergraph modeling, the synesthetic data is transformed into a structured hypergraph containing “entity-attribute-association”, and then mapped to a standard relational table structure through matrix processing. This solves the pain points of relational databases’ insufficient storage adaptability to high-order relational data and low efficiency of complex queries.

[0034] (2) By using the technical path of “hypergraph matrixization → relational storage mapping”, the constructed database is compatible with the existing relational database query mechanism. It can be seamlessly connected without modifying the existing business system, which solves the problems of high cost and difficulty in implementation of dedicated storage and existing ecosystem.

[0035] (3) By accurately modeling the exclusive association of synesthesia through knowledge hypergraph, and combining the multimodal fusion mechanism to realize the spatiotemporal alignment and semantic association of heterogeneous data, the gap in the adaptability of existing solutions in 5G-A synesthesia scenarios is filled, and the query results can fully output the associated information to support business decisions.

[0036] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0037] This application provides a data space construction and fast query method. The working process of this method can be implemented by electronic devices, such as computers, handheld smart terminals, etc. For ease of explanation, the implementation subject of the method in each embodiment of this application is described as a computer.

[0038] Figure 1 This is a schematic diagram illustrating the data space construction and fast query method provided in the embodiments of this application. Figure 2 Another schematic diagram illustrating the data space construction and fast query method provided in the embodiments of this application, as shown below. Figure 1 as well as Figure 2 As shown, this application provides a data space construction and fast query method, which includes S1-S6, as follows: S1, 5G-A Sensing Multimodal Data Preprocessing: Connect to the 5G-A base station interface and perform multimodal data parsing to extract multimodal features and entity attributes with Precision Time Protocol (PTP) spatiotemporal tags; In this step, by connecting to the integrated sensing interface of the 5G-A base station, the parsing of multimodal data, including high-resolution radar point clouds, high-precision Channel State Information (CSI) signals, ultra-high-definition visual streams, and communication QoS indicators, is completed. During the parsing process, multimodal features with IEEE 1588v2 PTP nanosecond-level spatiotemporal tags are extracted simultaneously, ultimately outputting the basic attributes and time-series feature sequences of the target entity, providing standardized data input for subsequent processes.

[0039] In some embodiments, in S1, the multimodal data includes at least two of radar data, CSI data, visual data, and communication data; In this application, multimodal data parsing includes at least two of the following processes: (1) Radar data processing: After receiving the high-resolution radar signal from the 5G-A base station, the signal is processed sequentially through digital down-conversion, pulse compression, moving target display filtering, and constant false alarm rate detection to output point cloud data; geometric features are extracted through point cloud feature extraction algorithm and bound to PTP nanosecond-level timestamps and unified coordinate system spatial coordinates; (2) CSI data processing: After receiving the CSI matrix from the 5G-A base station, the CSI time-frequency features are output after phase calibration, noise reduction, and time-frequency domain feature extraction; the target environment perception parameters are associated and bound to the PTP timestamp; (3) Visual data processing: After receiving the ultra-high-definition visual stream from the 5G-A base station, the system performs decoding, target detection, and multi-target tracking to generate stable entity identifiers across frames; after extracting visual appearance features, the system binds them to PTP timestamps. (4) Communication data processing: Receive the QoS (Quality of Service) index of the 5G-A base station, associate it with the corresponding target according to the entity identifier, form communication status parameters and bind PTP timestamp.

[0040] S2. Construct a 5G-A sensory-adaptive knowledge hypergraph: Generate entity nodes and hyperedges based on multimodal features and entity attributes to obtain a knowledge hypergraph containing communication-sensing linkage relationships; This step uses the target in the 5G-A sensing scenario as the entity node of the knowledge hypergraph, and defines multimodal features and communication parameters as attributes of the entity node. On this basis, the knowledge hypergraph is constructed through three types of hyperedges: intramodal hyperedges, cross-modal hyperedges, and communication-sensing linkage hyperedges.

[0041] Figure 3 This is a schematic diagram illustrating the construction of a 5G-A sensing-adaptive knowledge hypergraph provided in an embodiment of this application, such as... Figure 3 As shown, in some embodiments, S2 includes: S21. Entity Node Generation: Relying on the edge computing unit of the 5G-A base station, multimodal joint tracking operation is performed. For each tracked stable target, a unique entity identifier (entity_id) is assigned, and the target type (target_type), PTP timestamp (ptp_timestamp), spatial coordinates (spatial_coord), and communication state (comm_state) are synchronously associated to form complete entity node information. In this application, the definition of entity nodes is specifically adapted for the 5G-A sensing integration scenario, and the core attribute settings of entity nodes are as follows: Entity identifier (entity_id): A unique identifier assigned by the 5G-A base station, containing the base station ID prefix; target_type: Uses an enumeration value, specifically including drone / vehicle / pedestrian / base_station / rsu, etc.; PTP timestamp ptp_timestamp: A nanosecond-level timestamp based on IEEE 1588v2 PTP synchronization; Spatial coordinates (spatial_coord): Spatial coordinates in a unified coordinate system, such as UTM (Universal Transverse Mercator) and latitude and longitude. Communication status comm_state: Stores communication QoS parameters in JSON format, including CQI, RSRP, link status and other information.

[0042] S22, Hyperedge generation: Intramodal hyperedge: The correlation between features in the same modality is analyzed by edge computing units. When the calculated correlation coefficient is greater than or equal to the preset correlation threshold, the corresponding intramodal hyperedge is created and the corresponding correlation coefficient is recorded as the hyperedge attribute. Cross-modal hyperedge: The matching degree between different modal features is calculated using the 5G-A sensor fusion algorithm. If the matching degree is greater than or equal to the preset matching degree threshold, a cross-modal hyperedge is created, and the modal fusion confidence is recorded as a hyperedge attribute. Communication-Sensing Linked Hyperedge: The edge computing unit detects the correlation logic between communication parameters and sensing results in real time. When the preset linkage conditions are met, a communication-sensing linked hyperedge is created and the corresponding linkage strategy is recorded. In this application, the definition of the hyperedge is specifically adapted for the 5G-A sensing integration scenario, and the core attributes of the hyperedge are set as follows: (1) Intramodal hyperedge: used to connect associated features under the same mode, such as the "range-velocity-RCS" association of radar and the "phase shift-Doppler frequency shift" association of CSI. The attribute of this type of hyperedge is the feature correlation coefficient; (2) Cross-modal hyperedge: used to connect complementary features of different modalities, such as "radar location-visual appearance" association and "CSI environmental perception-radar target detection" association, whose attribute is modal fusion confidence; (3) Communication-Perception Linkage Hyperedge: Used to connect communication parameters and perception strategies, such as "Communication CQI Threshold - Radar Detection Algorithm" association, "Communication Link Status - Visual Frame Rate Adjustment" association, with the attribute being the linkage strategy label, such as "Algorithm Switching" "Parameter Adjustment" etc.

[0043] S23. Hyperedge Optimization: Remove low-confidence hyperedges and retain strong correlations with 5G-A service requirements. To ensure the effectiveness of the knowledge hypergraph, the generated hyperedges are filtered and optimized. By removing low-confidence hyperedges and retaining strong correlations with 5G-A service requirements, the efficiency of subsequent data processing is improved.

[0044] S3. Perform hypergraph matrixization: transform the knowledge hypergraph into an entity-feature matrix, a hyperedge-entity association matrix, and a spatiotemporal-communication perception fusion tensor. This step uses matrix transformation technology to structurally transform the constructed knowledge hypergraph, ultimately generating two types of core data: one is a structured matrix, which includes entity-feature matrix and hyperedge-entity association matrix; the other is a 3D spatiotemporal-communication perception fusion tensor, which enables deep fusion of spatiotemporal information and synesthetic data in the hypergraph, providing structured data support for subsequent association extraction.

[0045] Figure 4 This is a schematic diagram of the hypergraph matrixing process provided in the embodiments of this application, such as... Figure 4 As shown, in some embodiments, in S3, For the entity-feature matrix: the row dimension corresponds to the number of 5G-A synesthetic entities, the column dimension corresponds to the total dimension of the multimodal features, and the matrix elements are the feature values ​​corresponding to the entities; Specifically, in the entity-feature matrix, the row dimension corresponds to the number of 5G-A sensing entities, and the column dimension corresponds to the total dimension of multimodal features, covering radar geometric features, CSI time-frequency features, visual appearance features, and communication parameters, etc.; the matrix elements are the feature values ​​corresponding to the entities, and the storage format adopts a type adapted to high-dimensional data (such as half-precision floating point).

[0046] For the hyperedge-entity association matrix: the row dimension corresponds to the type of 5G-A scenario-based hyperedge, the column dimension corresponds to the entity combination, and the matrix elements are boolean values ​​used to mark whether the entity combination has a corresponding hyperedge association; Specifically, in the hyperedge-entity association matrix, the row dimension corresponds to the type of 5G-A scenario-based hyperedge, and the column dimension corresponds to the entity combination (including binary combination and multi-element combination); the matrix elements are Boolean values ​​used to indicate whether the entity combination has a corresponding hyperedge association.

[0047] For the spatiotemporal-communication sensing fusion tensor: according to the time slice of 5G-A sensing data, the entity-feature matrix and the hyperedge-entity association matrix are stacked to form a 3D spatiotemporal-communication sensing fusion tensor.

[0048] Specifically, based on the time slices of 5G-A sensing data, the entity-feature matrix and the hyperedge-entity association matrix are stacked to form a 3D tensor, namely the spatiotemporal-communication sensing fusion tensor, thereby preserving the spatiotemporal association between entities, features, and hyperedges.

[0049] S4. Extract 5G-A contextualized relationships: Based on entity-feature matrix, hyperedge-entity association matrix and spatiotemporal-communication perception fusion tensor, the contextualized relationships between entities are mined through matrix operations and scene rule engine; This step, based on matrix operation logic and the 5G-A service rule engine, extracts targeted relationships from the structured data after hypergraph matrix transformation, which meet the requirements of 5G-A scenarios. The extracted relationships must contain complete descriptive information, including relationship type, relationship validity period, and service adaptation tags, ensuring that the relationship data can directly serve 5G-A service scenarios.

[0050] In some embodiments, in S4, the rule engine has built-in 5G-A specific association rules, which specifically include: Low-altitude UAV formation rules: When two UAV entities meet the following conditions, such as spatial distance less than or equal to formation spacing threshold, speed difference less than or equal to cooperative speed threshold, stable communication link status, and continuous coverage greater than or equal to the preset number of time slices, they are marked as "formation following" associations. The association attributes include formation identification information. Vehicle-road cooperative rules: When the spatial distance between a vehicle and a roadside unit is less than or equal to the communication coverage radius, the Channel Quality Indicator (CQI) is greater than or equal to the service assurance threshold, and the radar detects the vehicle's turning / overtaking intention, a "vehicle-road cooperative warning" is marked. The association attributes include the specific warning type. Communication-Sensing Linkage Rules: When the entity communication parameters show "CQI less than the sensing optimization threshold" or "Reference Signal Received Power (RSRP) less than the link guarantee threshold", the "Sensing Strategy Enhancement" association is marked. The association attributes contain sensing algorithm adjustment instructions, such as switching the radar CFAR (Constant False Alarm Rate) algorithm or increasing the CSI feature extraction dimension.

[0051] S5. Perform relational storage mapping: Map entity nodes, multimodal features, and contextual relationships between entities to corresponding table structures and write them to the relational database. This step establishes mapping rules between relational databases and the three core data types obtained from the knowledge hypergraph parsing: entity nodes, multimodal features, and contextualized relationships. Entity nodes are mapped to entity base tables, multimodal features to modal attribute tables, and hyperedge relationships to hyperedge association tables. All table structures strictly adhere to relational database specifications, support SQL syntax queries, and ensure the standardization and accessibility of data storage.

[0052] In some embodiments, S5 includes: S51. Map the entity nodes to an entity base table used to store entity node information. The entity base table mainly stores core fields such as entity_id, target_type, ptp_timestamp, spatial_coord, and comm_state. It uses entity_id+ptp_timestamp as a composite primary key and supports basic filtering operations based on "entity + time range".

[0053] S52. Map the multimodal features into modal attribute tables for storing the high-dimensional features and key parameters of the corresponding modalities; The modal attribute tables are set up according to the different modalities of 5G-A sensing, including radar feature tables, CSI feature tables, visual feature tables, and communication feature tables. These tables are all associated with the entity base table with entity_id+ptp_timestamp as a foreign key, and are used to store the high-dimensional features and key parameters of the corresponding modality.

[0054] S53. Map the scenario-based relationships to hyper-edge association tables, including intra-modal association tables, cross-modal association tables, and communication-perception linkage association tables.

[0055] The hyperedge association tables are set up according to the hyperedge type, including intra-modal association tables, cross-modal association tables, and communication-perception linkage association tables. The tables contain fields such as relation_id, entity_combination (stores the combination of associated entity identifiers), relation_type (scenario-based association type), and relation_attr (stores association attributes in JSON format), and use entity_combination+relation_type as a combined index.

[0056] S6. Execute 5G-A intelligent query: Parse business query requests and convert them into Structured Query Language (SQL) statements, and call the relational database indexing mechanism to complete the query and result aggregation.

[0057] This step is responsible for receiving cross-modal query requests in 5G-A sensing service scenarios. First, it parses the query request into an abstract query syntax tree, then generates SQL statements that can be directly executed by a relational database through adaptation and conversion logic. Subsequently, it uses the database's built-in indexing mechanism to efficiently execute the query operation, and then aggregates the query results before returning them to the requesting client. The entire query process requires no additional development of a dedicated query interface.

[0058] In some embodiments, S6 includes: S61. Query parsing process: Extract key dimensions, relationship types, entity attribute filtering conditions, and result fields from the business query request to obtain the query logic framework. S62, SQL Conversion Operation: Maps the query logic framework to a multi-table JOIN SQL statement supported by a relational database; S63. Index call optimization: When executing a query, enable the pre-designed composite index in the relational database to quickly locate the target data through the index; S64. Result Aggregation Processing: Based on entity identifier and timestamp, the results obtained from multi-table queries are correlated and matched to verify whether each entity meets the filtering conditions; the result data of entities that meet the conditions are integrated and the query results are output in the format specified by business requirements.

[0059] The "5G-A Intelligent Query Execution" step in this process is fully adapted to the actual query needs of 5G-A sensing services. The following example illustrates this in detail: (1) Example of query request: The query requirement proposed by the business side is to "query UAV entities that have a 'following formation' relationship with the target entity within a specified time slice and a specified spatial coordinate range, and whose communication link status is stable, and return the radar characteristics and communication parameters of these entities".

[0060] (2) Query parsing process: First, extract the key dimensions in the query requirements, including the two spatiotemporal dimensions of time slice range and spatial coordinate range, the two filtering dimensions of association relationship type "formation following" and entity attribute filtering condition "communication link stability", and the output dimension of result fields (i.e. radar features and communication parameters), to form a clear query logic framework.

[0061] (3) SQL conversion operation: The parsed query dimensions are mapped to multi-table JOINSQL statements supported by relational databases to ensure that the query logic can be accurately converted into database executable instructions.

[0062] (4) Index call optimization: When executing a query, the pre-designed composite index entity_id+ptp_timestamp+spatial_coord in the relational database is enabled to quickly locate the target data through the index, optimize the query efficiency in multi-table JOIN scenarios, and reduce query time.

[0063] (5) Result aggregation processing: According to the entity identifier and timestamp, the results obtained from the multi-table query are correlated and matched, and at the same time, it is verified whether each entity meets the filtering conditions such as "stable communication link"; the radar characteristics and communication parameters of the entities that meet the conditions are integrated, and the query results are finally output in the format specified by the business requirements to ensure the integrity and availability of the results.

[0064] This application provides a method for data space construction and fast querying. This method is based on knowledge hypergraph technology and is applicable to 5G-A integrated sensing data. Compared with existing technologies, this application has the following advantages: 1. This application constructs a complete data logic of "entity-attribute-relationship" through knowledge hypergraph modeling, which can accurately carry the many-to-many high-order relationship of multimodal data such as radar, CSI, and vision. It avoids the low efficiency and semantic disconnection problems caused by the reliance on complex table joins in traditional relational databases, and makes the expression of the relationship of synesthetic data more in line with the needs of 5G-A scenario.

[0065] 2. This application designs a technical path of "hypergraph matrixization → relational storage mapping", which transforms the knowledge hypergraph into a standard relational table structure. It is directly compatible with existing SQL business systems, without the need for additional development of adaptation interfaces or modification of existing architecture. This significantly reduces the deployment cost and maintenance difficulty of the 5G-A sensor data storage system, and can be quickly integrated into the existing 5G-A application ecosystem.

[0066] 3. This application specifically reserves storage logic for the "communication-sensing integration" core feature of 5G-A, such as "communication parameters-sensing strategy", which can completely retain key related information. When querying, there is no need to splice data across databases. It can directly output data that supports low-altitude security, vehicle-road cooperation and other services, avoiding the limitation of existing solutions that only output single-dimensional data.

[0067] This application utilizes knowledge hypergraph technology as its core support and addresses the three core characteristics of 5G-A sensing data—multimodal heterogeneity, close sensing correlation, and high real-time query requirements—by constructing a complete system covering integrated multimodal sensing data acquisition, data space construction, and cross-modal spatiotemporal joint rapid querying. This application effectively solves the problems commonly found in traditional technical solutions, such as asynchronous acquisition, lack of effective data space correlation, and low query efficiency, providing a reliable technical path for efficient data processing in 5G-A sensing scenarios.

[0068] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0069] Figure 5 A schematic diagram of the data space construction and fast query device provided in the embodiments of this application, such as... Figure 5 As shown, this application provides a data space construction and fast query device, the device comprising: The 5G-A sensing multimodal data preprocessing module 11 is configured to perform 5G-A sensing multimodal data preprocessing: connect to the 5G-A base station interface and perform multimodal data parsing to extract multimodal features and entity attributes with PTP spatiotemporal labels; The 5G-A sensory adaptation knowledge hypergraph construction module 12 is configured to construct a 5G-A sensory adaptation knowledge hypergraph: generate entity nodes and hyperedges based on multimodal features and entity attributes to obtain a knowledge hypergraph containing communication-sensing linkage relationships; Hypergraph matrixing module 13 is configured to perform hypergraph matrixing processing: converting the knowledge hypergraph into an entity-feature matrix, a hyperedge-entity association matrix, and a spatiotemporal-communication perception fusion tensor; The 5G-A scenario-based association extraction module 14 is configured to extract 5G-A scenario-based association relationships: based on the entity-feature matrix, the hyperedge-entity association matrix, and the spatiotemporal-communication perception fusion tensor, the scenario-based association relationships between entities are mined through matrix operations and the scenario rule engine; The relational storage mapping module 15 is configured to perform relational storage mapping: mapping entity nodes, multimodal features, and contextual relationships between entities to corresponding table structures and writing them to the relational database; The 5G-A intelligent query adaptation module 16 is configured to execute 5G-A intelligent queries: parse business query requests and convert them into SQL statements, and call the relational database indexing mechanism to complete the query and result aggregation.

[0070] Regarding the limitations on the data space construction and fast query device, please refer to the limitations on the data space construction and fast query method in the above embodiments of this application, which will not be repeated here.

[0071] Figure 6 Another schematic diagram of the data space construction and fast query device provided in the embodiments of this application, as shown below. Figure 6 As shown, the device includes a memory 22 and a processor 21. The memory stores a computer program, and the processor is configured to run the computer program to perform the methods described in the above embodiments of this application.

[0072] The memory is connected to the processor. The memory can be flash memory, read-only memory or other types of memory. The processor can be a central processing unit or a microcontroller.

[0073] In some embodiments, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described in the above embodiments of this application.

[0074] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, computer program modules or other data. Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), DVD or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0075] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of this application, and this application is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this application, and these modifications and improvements are also considered to be within the scope of protection of this application.

Claims

1. A method for constructing a data space and performing fast queries, characterized in that, The method includes: S1, 5G-A Sensing Multimodal Data Preprocessing: Connect to the 5G-A base station interface and perform multimodal data parsing to extract multimodal features and entity attributes with Precise Time Protocol (PTP) spatiotemporal labels; S2. Construct a 5G-A sensory-adaptive knowledge hypergraph: Generate entity nodes and hyperedges based on multimodal features and entity attributes to obtain a knowledge hypergraph containing communication-sensing linkage relationships; S3. Perform hypergraph matrixization: transform the knowledge hypergraph into an entity-feature matrix, a hyperedge-entity association matrix, and a spatiotemporal-communication perception fusion tensor. S4. Extract 5G-A contextualized relationships: Based on entity-feature matrix, hyperedge-entity association matrix and spatiotemporal-communication perception fusion tensor, the contextualized relationships between entities are mined through matrix operations and scene rule engine; S5. Perform relational storage mapping: Map entity nodes, multimodal features, and contextual relationships between entities to corresponding table structures and write them to the relational database. S6. Execute 5G-A intelligent query: Parse business query requests and convert them into structured query language SQL statements, and call the relational database indexing mechanism to complete the query and result aggregation.

2. The data space construction and fast query method according to claim 1, characterized in that, In S1, the multimodal data includes at least two of the following: radar data, channel state information (CSI) data, visual data, and communication data. Multimodal data parsing includes at least two of the following processes: Radar data processing: After receiving high-resolution radar signals from 5G-A base stations, the data undergoes digital down-conversion, pulse compression, moving target display filtering, and constant false alarm rate detection to output point cloud data. Geometric features are extracted using point cloud feature extraction algorithms and bound to PTP nanosecond-level timestamps and unified coordinate system spatial coordinates. CSI data processing: After receiving the CSI matrix from the 5G-A base station, the data undergoes phase calibration, noise reduction, and time-frequency domain feature extraction to output the CSI time-frequency features. Associate the target environment awareness parameters and bind them to the PTP timestamp; Visual data processing: After receiving the ultra-high-definition visual stream from the 5G-A base station, the system performs decoding, target detection, and multi-target tracking to generate stable entity identifiers across frames; after extracting visual appearance features, it binds them with PTP timestamps. Communication data processing: Receive the QoS indicators of the 5G-A base station, associate them with the corresponding targets according to the entity identifier, form communication status parameters and bind them with PTP timestamps.

3. The data space construction and fast query method according to claim 1, characterized in that, S2 includes: S21. Entity Node Generation: Relying on the edge computing unit of the 5G-A base station, multimodal joint tracking operation is performed. For each tracked stable target, a unique entity identifier (entity_id) is assigned, and the target type (target_type), PTP timestamp (ptp_timestamp), spatial coordinates (spatial_coord), and communication state (comm_state) are synchronously associated to form complete entity node information. S22, Hyperedge generation: Intramodal hyperedge: The correlation between features in the same modality is analyzed by edge computing units. When the calculated correlation coefficient is greater than or equal to the preset correlation threshold, the corresponding intramodal hyperedge is created and the corresponding correlation coefficient is recorded as the hyperedge attribute. Cross-modal hyperedge: The matching degree between different modal features is calculated using the 5G-A sensor fusion algorithm. If the matching degree is greater than or equal to the preset matching degree threshold, a cross-modal hyperedge is created, and the modal fusion confidence is recorded as a hyperedge attribute. Communication-Sensing Linked Hyperedge: The edge computing unit detects the correlation logic between communication parameters and sensing results in real time. When the preset linkage conditions are met, a communication-sensing linked hyperedge is created and the corresponding linkage strategy is recorded. S23. Superedge optimization: Remove low-confidence superedges and retain strong correlations with 5G-A service requirements.

4. The data space construction and fast query method according to claim 1, characterized in that, In S3 For the entity-feature matrix: the row dimension corresponds to the number of 5G-A synesthetic entities, the column dimension corresponds to the total dimension of the multimodal features, and the matrix elements are the feature values ​​corresponding to the entities; For the hyperedge-entity association matrix: the row dimension corresponds to the type of 5G-A scenario-based hyperedge, the column dimension corresponds to the entity combination, and the matrix elements are boolean values ​​used to mark whether the entity combination has a corresponding hyperedge association; For the spatiotemporal-communication sensing fusion tensor: according to the time slice of 5G-A sensing data, the entity-feature matrix and the hyperedge-entity association matrix are stacked to form a 3D spatiotemporal-communication sensing fusion tensor.

5. The data space construction and fast query method according to claim 1, characterized in that, In S4, the rule engine has built-in 5G-A exclusive association rules, specifically including: Low-altitude UAV formation rules: When two UAV entities meet the following conditions, such as spatial distance less than or equal to formation spacing threshold, speed difference less than or equal to cooperative speed threshold, stable communication link status, and continuous coverage greater than or equal to the preset number of time slices, they are marked as "formation following" associations. The association attributes include formation identification information. Vehicle-road cooperative rules: When the spatial distance between a vehicle and a roadside unit is less than or equal to the communication coverage radius, the Communication Channel Quality Indicator (CQI) is greater than or equal to the service assurance threshold, and the radar detects the vehicle's turning / overtaking intention, a "vehicle-road cooperative warning" association is marked. The association attributes include the specific warning type. Communication-Sensing Linkage Rule: When the entity's communication parameters show "CQI less than the sensing optimization threshold" or "Reference Signal Received Power (RSRP) less than the link guarantee threshold", mark it as "Sensing Strategy Enhancement" and the associated attribute contains the sensing algorithm adjustment instruction.

6. The data space construction and fast query method according to claim 1, characterized in that, S5 includes: S51. Map the entity nodes to an entity base table used to store entity node information. S52. Map the multimodal features into modal attribute tables for storing the high-dimensional features and key parameters of the corresponding modalities; S53. Map the scenario-based relationships to hyper-edge association tables, including intra-modal association tables, cross-modal association tables, and communication-perception linkage association tables.

7. The data space construction and fast query method according to claim 1, characterized in that, S6 includes: S61. Query parsing process: Extract key dimensions, relationship types, entity attribute filtering conditions, and result fields from the business query request to obtain the query logic framework. S62, SQL Conversion Operation: Maps the query logic framework to a multi-table JOIN SQL statement supported by a relational database; S63. Index call optimization: When executing a query, enable the pre-designed composite index in the relational database to quickly locate the target data through the index; S64. Result Aggregation Processing: Based on entity identifier and timestamp, the results obtained from multi-table queries are correlated and matched to verify whether each entity meets the filtering conditions; the result data of entities that meet the conditions are integrated and the query results are output in the format specified by business requirements.

8. A data space construction and fast query device, characterized in that, The device includes: The 5G-A sensing multimodal data preprocessing module is configured to perform 5G-A sensing multimodal data preprocessing: it connects to the 5G-A base station interface and performs multimodal data parsing to extract multimodal features and entity attributes with Precise Time Protocol (PTP) spatiotemporal tags; The 5G-A sensory-adaptive knowledge hypergraph construction module is configured to construct a 5G-A sensory-adaptive knowledge hypergraph by generating entity nodes and hyperedges based on multimodal features and entity attributes, thereby obtaining a knowledge hypergraph containing communication-sensing linkage relationships. The Hypergraph Matrixing module is configured to perform hypergraph matrixing: transforming the knowledge hypergraph into an entity-feature matrix, a hyperedge-entity association matrix, and a spatiotemporal-communication perception fusion tensor; The 5G-A scenario-based association extraction module is configured to extract 5G-A scenario-based association relationships: based on entity-feature matrix, hyperedge-entity association matrix and spatiotemporal-communication perception fusion tensor, it mines the scenario-based association relationships between entities through matrix operations and scenario rule engine; The relational storage mapping module is configured to perform relational storage mapping: mapping entity nodes, multimodal features, and contextual relationships between entities to corresponding table structures and writing them to the relational database; The 5G-A intelligent query adaptation module is configured to execute 5G-A intelligent queries: parse business query requests and convert them into structured query language (SQL) statements, and call relational database indexing mechanisms to complete the query and result aggregation.

9. A data space construction and fast query device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to implement the data space construction and fast query method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the data space construction and fast query method as described in any one of claims 1-7.

Images