Real-time synchronization and scenario linkage control method and system for multi-terminal heterogeneous data

Abstract

The application provides a real-time synchronization and scene linkage control method and system for multi-terminal heterogeneous data, relates to the technical field of linkage control, and realizes real-time collection and space-time window regulation of multi-terminal heterogeneous health care equipment through a distributed sensing network, carries out scene feature recognition and dynamic classification on the running state of the equipment and the context of the health care environment through standardized sensing information flow, carries out scene target-oriented causal correlation analysis and linkage trigger condition judgment on real-time sensing information and scene historical rules through a scene cognitive map, and comprehensively weighs and optimally matches the scene safety target and service efficiency constraint through a potential linkage response condition list, so as to obtain a linkage control scheme. The application effectively reduces the false alarm and missed alarm risks caused by data confusion or scene misjudgment, and improves the intelligent level and decision reliability of the health care service system under multi-scene deployment.

Description

Technical Field

[0001] This invention proposes a method and system for real-time synchronization and scenario-based linkage control of heterogeneous data from multiple terminals, relating to the field of linkage control technology, specifically to the field of real-time synchronization and scenario-based linkage control of heterogeneous data from multiple terminals. Background Technology

[0002] As the aging population continues to grow, the demand for smart elderly care services in various scenarios, including home, community, and institutional settings, is rapidly increasing, leading to the widespread application of IoT devices in these settings. However, existing elderly care linkage technologies face significant obstacles in data fusion and synchronization among multi-terminal heterogeneous devices. Different elderly care devices use different communication protocols (such as Modbus, MQTT, ZigBee, LoRa, etc.), resulting in varying data formats. Furthermore, the inherent deviations in the local clock oscillators of each device, coupled with inconsistent arrival times due to differences in network transmission paths, make it difficult to accurately align and correlate data from different terminals under a unified time reference. This leads to time sequence errors and causal inversions, severely impacting the accuracy of subsequent linkage decisions. Existing elderly care linkage solutions generally employ preset conditions and fixed action rules, lacking adaptability to the differences and dynamism of various scenarios. Elderly care needs vary significantly across different scenarios, including home, community, and institutional settings. Home settings prioritize privacy and minimal disturbance, community settings require consideration of population mobility, and institutional settings demand high-density, precise monitoring. A single linkage rule cannot adapt to all scenarios. Furthermore, fixed rules lack the ability to perform causal correlation analysis based on scenario context and make goal-oriented optimal decisions when dealing with complex dynamic changes. This makes it difficult to maximize safety benefits under limited system resources, thus limiting the intelligence level and response reliability of the health and wellness service system. Summary of the Invention

[0003] This invention provides a method and system for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals, in order to solve the above-mentioned problems: The present invention proposes a method and system for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals, the method comprising: S1. Real-time data collection and spatiotemporal window adjustment of multi-terminal heterogeneous health and wellness devices are performed through a distributed sensing network to obtain a standardized sensing information flow. S2. By standardizing the perception information flow, the scene features of the equipment operation status and the context of the health and wellness environment are identified and dynamically classified to obtain a scene cognitive map. S3. By using the scene cognition graph to perform causal correlation analysis and linkage trigger condition determination on real-time perceived information and scene historical patterns, a list of potential linkage response conditions is obtained. S4. By comprehensively weighing and selecting the best match between scenario security objectives and service performance constraints through a list of potential linkage response conditions, a linkage control scheme is obtained.

[0004] Furthermore, the system includes: The data acquisition and standardization module is used to collect data and perform spatiotemporal window standardization on multi-terminal heterogeneous health and wellness devices in real time through a distributed sensing network to obtain a standardized sensing information flow. The identification and classification module is used to identify and dynamically classify the scene features of the equipment operating status and the context of the health and wellness environment through a standardized perception information flow, and to obtain a scene cognitive map. The analysis and judgment module is used to perform causal correlation analysis and linkage trigger condition determination on real-time perceived information and scene historical patterns through scene cognition graph, and obtain a list of potential linkage response conditions. The weighting and matching module is used to comprehensively weigh and select the best match between scenario security objectives and service performance constraints through a list of potential linkage response conditions, so as to obtain a linkage control scheme.

[0005] The beneficial effects of this invention are as follows: This method constructs a complete intelligent linkage link from data organization to scenario cognition to causal judgment to decision output. It solves the problems of fragmented links and distorted information transmission leading to delayed decision-making in traditional health and wellness linkage solutions from a system architecture perspective. By using data governance and scenario cognition as the prerequisites for linkage decision-making, each linkage response is based on the system's true understanding of the current scenario and its safety boundaries. This effectively reduces the risk of false alarms and missed alarms caused by data chaos or scenario misjudgment, and improves the intelligence level and decision reliability of the health and wellness service system in multi-scenario deployments. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of a method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals. Detailed Implementation

[0007] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0008] In one embodiment of the present invention, the present invention proposes a method and system for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals, the method comprising: S1. Real-time data collection and spatiotemporal window adjustment of multi-terminal heterogeneous health and wellness devices are performed through a distributed sensing network to obtain a standardized sensing information flow. S2. By standardizing the perception information flow, the scene features of the equipment operation status and the context of the health and wellness environment are identified and dynamically classified to obtain a scene cognitive map. S3. By using the scene cognition graph to perform causal correlation analysis and linkage trigger condition determination on real-time perceived information and scene historical patterns, a list of potential linkage response conditions is obtained. S4. By comprehensively weighing and selecting the best match between scenario security objectives and service performance constraints through a list of potential linkage response conditions, a linkage control scheme is obtained.

[0009] The distributed sensing network is a collaborative detection network composed of multiple heterogeneous sensor nodes deployed in the health and wellness environment through wired or wireless communication. It can cover different spatial areas and collect multimodal data, and the nodes have the ability to synchronize time and fuse data. Multi-terminal heterogeneous health and wellness equipment refers to terminal devices used in smart health and wellness scenarios that come from different manufacturers and are based on different communication protocols and data formats, including heart rate monitors, pulse oximeters, blood pressure monitors, body movement recorders, smart mattresses, temperature and humidity sensors, smoke detectors, etc. Spatiotemporal window alignment cuts and reassembles sensor data distributed on a continuous time axis according to a uniform time window length. At the same time, it performs spatiotemporal alignment on device data from different spatial locations within the same window, eliminating data misalignment caused by clock deviation and transmission delay, so that the aligned data is comparable and correlated in time and space dimensions. The standardized sensing information stream, after being processed by protocol adaptation, format unification, time synchronization and spatiotemporal adjustment, becomes a continuous data stream with a unified data structure and semantic tags. It retains the core numerical content of the original sensing information, while adding metadata information such as device identifier, timestamp, spatial location tag and function type tag. Scene feature recognition is based on standardized perceptual information flow to extract basic features such as environmental parameters, physiological parameters, and behavioral parameters, and combines them to form a feature vector that can characterize the current state of the health and wellness scene, which is used to distinguish different scene categories; Dynamic classification uses a similarity measurement method to compare the perceived feature vector within a continuous time window with the typical patterns of known scene categories. Based on the similarity threshold, the current perceived information is assigned to an existing scene category or a new scene category is created, thus achieving online adaptive classification of scenes. Scene cognition graph is a structured knowledge representation model that stores the defining conditions, baseline of parameter change patterns, temporal dependencies between parameters, and typical linkage response patterns of various health and wellness scenarios in a graph structure. Nodes represent equipment parameters or state variables, edges represent dependencies or causal relationships, and safety protection goals and priority information are linked at the same time. In steady-state operation, under a certain scenario category, the monitored parameters fluctuate within the normal range, the parameter change trend conforms to the historical baseline, and there are no significant deviations or abnormal changes. The system is in a controllable and safe equilibrium. If the values, rates of change, or patterns of abnormal fluctuation monitoring parameters deviate significantly from the historical baseline for this scenario and exceed the normal fluctuation range, it may indicate early signs of safety risks or equipment failure. Causal correlation analysis analyzes the change nodes that deviate from historical patterns in real-time perceived information, and combines the temporal sequence relationship between parameters recorded in the scene cognition map to determine the driving and driven relationship between each deviation parameter, and identify the root cause parameter chain that leads to the current abnormal state. The linkage triggering condition determination is based on the causal relationship analysis results. It determines whether the current abnormal deviation state has reached or exceeded the preset safety warning threshold. When the comprehensive deviation index touches the safety bottom line, it determines the set of conditions that need to be activated for linkage response. The potential linkage response condition list consists of a list of linkage response methods corresponding to all abnormal deviations that have reached the safety warning line in the current scenario. Each item in the list includes the safety target category, deviation characteristic description, matching linkage response method and its adjustable execution parameters. By comprehensively weighing and selecting the best option, each candidate linkage response condition is evaluated for its contribution to the protection of the scenario's security objectives and its performance consumption indicators such as resource consumption, response latency, and equipment wear during the execution process. The optimal set of response conditions is then selected by ranking the cost-benefit ratio. The linkage control scheme generates a complete set of control instructions by binding the selected linkage response conditions with the executable actions of specific terminal devices. This set includes the target device identifier, action type, execution parameters, time window constraints, and execution status monitoring point configuration. The protocol adaptation is designed for the communication protocol stacks used by different terminal devices. The corresponding parsing driver is used to decompose the original bit stream into field-level data according to the protocol specifications, so as to realize data restoration from the physical layer to the application layer. Homogenized device data streams decode data from different protocols and convert it into a pre-defined standardized data structure, ensuring that all device data remains consistent in both semantics and format, facilitating subsequent unified processing. Clock skew calibration periodically synchronizes the local clocks of each device node in the distributed sensing network using a network time protocol or a precision time protocol, calculates and compensates for the offset of each node's clock relative to the reference clock, and eliminates time alignment errors in cross-device data. Cross-terminal synchronization alignment is performed on a calibrated clock reference, mapping data from different terminals to a unified time axis according to their actual occurrence time, ensuring that subsequent correlation analysis is based on accurate time correspondence. The spatial topology relationship is a spatial layout model between devices established based on the physical installation location, coverage area and adjacency relationship of the devices in the health and wellness scenario. It is used to determine which devices' data are spatially related within the same spatiotemporal window. Functional dependencies are described based on the logical dependency chains between equipment functions in the health and wellness business process. For example, abnormal body movement monitoring data requires verification based on heart rate data. Functional dependencies determine the strength of the logical association between equipment data. The associated anchor points are the tagging information added to each set of consolidated data in the standardized sensing information flow. The tagging content includes spatial adjacency identifiers between devices, upstream and downstream functional identifiers, and temporal sequence identifiers, which are used for rapid retrieval and correlation analysis. A sliding time window is a fixed-length time interval that moves dynamically on a continuous data stream. It is used to limit the range of data currently being processed. The window length can be automatically adjusted according to the data density. Time density is the number of data records generated by the device in this scenario per unit of time. The higher the density, the finer the monitoring granularity, which can support a shorter window length to achieve a fast response. The activity level of parameter changes is a comprehensive index of the frequency and magnitude of changes in the values ​​of each monitored parameter within the observation time range. When the activity level is high, the window length needs to be shortened to capture rapid change characteristics, while when the activity level is low, the window length can be extended to reduce computational overhead. The window boundary is dynamically optimized. Based on the time density of real-time calculation and the activity of parameter changes, the start and end positions of the sliding window are adaptively adjusted so that the window range always covers the data segment with the most analytical value at present. Within a defined sliding time window, the data set is formed by arranging all data generated by all devices in timestamp order and grouping them according to device function type, and then using the data set for subsequent correlation analysis. Basic context elements are the minimum set of information extracted from standardized perceived information flow to describe the basic state of the current health and wellness environment, including environmental physical quantities, human physiological quantities, equipment operating quantities, as well as time and space labels; The state evolution feature set is a set of quantitative features extracted after tracking and calculating the changing trend, fluctuation amplitude and rate of change of each monitored parameter within a continuous time window. It is used to characterize the dynamic change trajectory of the parameter from the past to the present. Similarity comparison uses algorithms such as dynamic time warping or Euclidean distance to match the state evolution features of the current window with the historical typical evolution trajectories of each scene category in the scene cognition graph to obtain a similarity score; The scene classification results are assigned to specific scene categories based on similarity comparison scores, and the output classification labels and their confidence scores are recorded. At the same time, the distance values ​​with typical patterns of each category are also recorded. The scene boundary definition describes the boundary conditions of the area occupied by each scene category in the parameter space, which consists of the normal value range and the rate of change limit of each monitored parameter under that scene; The admission criteria describe the minimum feature matching requirements that must be met to determine whether the current perceived information can be classified into a certain scene category, including the numerical range constraints of key parameters and the existence requirements of key feature patterns. The parameter change pattern baseline database is a database that summarizes and stores the statistical quantities of each monitored parameter under normal operating conditions for each scenario category, such as mean, standard deviation, typical fluctuation range, and diurnal rhythm characteristics, and serves as a reference benchmark for anomaly judgment. The parameter dependency graph represents the temporal dependency and numerical coupling relationship between different monitoring parameters within the same scene category in a graph structure. Directed edges indicate that a change in one parameter has a driving effect on a change in another parameter. The typical linkage response pattern library is a set of patterns formed by summarizing effective linkage response cases that have occurred in history in each scenario category. Each pattern records the abnormal characteristics, the sequence of actions executed, and the response effect score. The multidimensional mapping link is a mechanism that cross-indexes scene boundaries, parameter baselines, dependencies, and response patterns in the scene cognition graph, enabling the rapid retrieval of information from other related dimensions starting from any one dimension; Safety protection priorities are the safety dimensions that need to be prioritized in specific scenarios, such as fall prevention, stable vital signs, and environmental safety. Each target corresponds to a set of quantifiable safety thresholds. Historical pattern references are derived from the scene cognition graph, which retrieves the parameter change pattern baseline and dependency relationship graph that match the current scene category, and are used for comparison with real-time data; The safety warning line sets the critical thresholds for triggering linkage responses for each key safety protection target, including the upper limit of the absolute value of a single parameter, the upper limit of the rate of change, and the upper limit of the overall deviation. The comprehensive deviation index is a single quantitative index obtained by weighting and comprehensively calculating the degree of deviation of each current monitoring parameter from its respective baseline value, the duration of deviation, and the number of parameters deviating from the deviation, and reflects the overall severity of the anomaly. The evolution path reference information is obtained from the scene cognition map, which shows the typical time trajectory and parameter change sequence of historical anomalies from the current deviation state to the critical point of the security event under the security target category; The expected evolution time window is the remaining time range from the current moment until the deviation from the state may trigger a safety event, estimated based on evolution path reference information, and is used to determine the urgency of response actions. The historical best linkage response method is selected from the typical linkage response pattern library of the scene cognitive graph, based on the current safety target category and deviation state characteristics, and the action sequence with the highest response effect score is selected as the initial response scheme. Adaptive adjustments scale, shift, or combine the execution parameters in the historical best linkage response mode based on the specific magnitude, rate, and duration of the current real-time deviation, so that they accurately match the current abnormal state. The security target priority matrix is ​​a priority list of key security protection targets in the current scenario, sorted by importance and urgency, and records the weight coefficient and bottom-line constraints for each target. The contribution to safety assurance is assessed by evaluating the expected improvement effect of a candidate linkage response condition, which, after execution, can restore various safety indicators from their current deviations to the safe range or prevent further deterioration. This effect is expressed by a quantitative score. The total amount of system resources consumed by the candidate linkage response during execution, including the bandwidth of control command transmission, the energy consumption of equipment actions, the degree of interference of the response actions on the health care recipients, and the resource conflicts when multiple actions occur concurrently. The comprehensive cost-benefit assessment calculates and normalizes the ratio of the security contribution to the efficiency consumption of each candidate linkage response condition, and obtains a comprehensive benefit score, which serves as the basis for selection. The selection and ranking process involves sorting all candidate conditions in descending order based on the comprehensive cost and benefit assessment results, selecting the conditions with the highest scores to form the final execution candidate set, and excluding candidate conditions with scores below a preset threshold. Matching and binding converts the action descriptions in the filtered linkage response conditions into an instruction format that can be executed by specific terminal devices, including device address, operation code, parameter value and execution timestamp, forming control instructions that can be directly issued; The feedback monitoring points are preset execution status review nodes in the linkage control scheme, including instruction delivery confirmation points, action execution start points, action completion confirmation points, and effect evaluation points. Each node is equipped with timeout retransmission and abnormal alarm mechanisms. The execution status feedback information is the execution result report sent back by each terminal device after executing the linkage control command, which includes the command reception status, action execution status code, actual execution parameter value and execution completion timestamp; The evaluation of the linkage response effect compares the measured perception data after the execution is completed with the expected safety goals, calculates the actual improvement and recovery degree of each safety indicator, and forms a quantitative effect score; Based on the gap between the evaluation results of the linkage response effect and the expected goal, the deviation correction guidance information calculates the direction and magnitude of adjustment required for each baseline parameter, dependency and response mode in the current scene cognitive map; Parameter fine-tuning and pattern updates utilize bias correction guidance information and employ incremental learning or online learning methods to revise the parameter statistics, dependency weights, and response pattern scores of relevant scene categories in the scene cognition graph, enabling the cognition graph to continuously adapt to changes in the current environment and user status.

[0010] The working principle and technical effect of the above technical solution are as follows: According to the method, real-time data collection is performed on multi-terminal heterogeneous health care devices in multiple scenarios such as homes, communities and institutions through a distributed sensing network, and space-time window normalization processing is carried out at a data entry, so as to convert original signals of different devices, different protocols and different time references into standardized sensing information flows with unified space-time coordinates. The system performs multi-dimensional scene feature recognition on device operating states and health care environment contexts in the standardized information flows, dynamically classifies continuous sensing information into corresponding health care scene categories (such as home rest, deep sleep at night, daytime activities, centralized institutional care, etc.) according to elements such as current time period, spatial position, device combination mode and personnel activity intensity, and summarizes the parameter change rule and typical linkage response mode of each scene, so as to form a reusable scene cognition map. Based on the scene cognition map as a cognition basis, the system performs scene goal-oriented causal correlation analysis on real-time sensing information and historical rules of the scene, that is, according to the key safety protection goal of the scene, selectively pays attention to the causal dependency relationship between parameters directly related to the goal, determines whether the current deviation state touches the scene safety bottom line and matches a corresponding linkage response mode, so as to generate a list of potential linkage response conditions. The system comprehensively weighs the scene safety goal priority and service efficiency consumption of each candidate linkage response condition in the list, selects and binds the conditions to specific terminal device actions according to the multi-objective optimization principle, and outputs a complete linkage control scheme.

[0011] The method constructs a complete intelligent linkage link from data normalization to scene cognition, to causal judgment and then to decision output, and solves the problem of decision lag caused by the separation of each link and information transmission distortion in the traditional health care linkage scheme from the system architecture level. By taking data governance and scene cognition as the preconditions of linkage decision-making, each linkage response is established on the premise that the system truly understands what scene is currently occurring and where the safety bottom line of the scene lies, which effectively reduces the risk of false alarms and missed alarms caused by data confusion or scene misjudgment, and improves the intelligence level and decision reliability of the health care service system under multi-scenario deployment.

[0012] In an embodiment of the present invention, the S1 includes: Real-time data capture is performed on multi-terminal heterogeneous health care devices in different scenarios through a distributed sensing network, so as to obtain multi-dimensional original device signal flows; Protocol adaptation and signal decoding are performed on each terminal device through the original device signal flows, so as to convert device data under different communication protocols into a unified expression format, and obtain isomorphic device data flows; Cross-terminal clock deviation calibration and synchronous alignment are performed on the timestamps of each piece of device data through the isomorphic device data flows, so as to obtain clock-synchronized data flows; By using clock-synchronized data streams, the data of each associated device within the same spatiotemporal window are correlated, aggregated, and encapsulated. Based on the preset spatial topology and functional dependencies between devices, connection anchors between data are established to obtain a standardized perception information stream.

[0013] The working principle and technical effects of the above solution are as follows: After the distributed sensing network completes the raw data acquisition, this method first performs protocol adaptation and signal decoding processing on each terminal device. Since different health and wellness devices (such as vital sign monitoring wristbands, security door magnets, and environmental sensors) use different communication protocols (such as Modbus, MQTT, ZigBee, LoRa, etc.), the system's built-in protocol adaptation engine can automatically identify the device type and call the corresponding decoding rules, uniformly converting all device data into the system's internal standard expression format. After format unification, the system performs cross-terminal clock deviation calibration on the timestamps carried by each device's data. Different devices have inherent frequency deviations in their local crystal oscillators, and the arrival times are inconsistent due to differences in network transmission paths. Therefore, clock deviation compensation calculations are needed for each data entry to align all data to the same global benchmark in the time dimension. After obtaining the clock synchronization data stream, the system uses a dynamically sliding spatiotemporal window as the basic processing unit to perform correlation-based organization and aggregation of all device data generated within the same window. Based on the preset spatial topology relationship between devices (such as being in the same room) and functional dependency relationship (such as the linkage monitoring relationship between vital sign monitoring devices and security devices), multiple association anchor points are established for the data, ultimately forming a standardized perception information stream.

[0014] This method addresses the issues of data silos and temporal discrepancies caused by heterogeneous data from multiple terminals due to differences in source, format, and time base at the data entry point. This ensures that all subsequent analysis stages are based on unified, accurate data carrying complete contextual relationships. Cross-terminal clock deviation calibration eliminates the risk of causal inversion caused by device clock drift and network transmission latency, preventing the incorrect identification of events that should have occurred earlier as later due to time asynchrony. By using spatial topology and functional dependencies as data association anchors, physical and logical connections between devices are established before data enters subsequent processing. This significantly reduces the computational overhead of repeatedly parsing data source relationships in subsequent scene identification and causal determination stages, improving overall data processing efficiency and response time.

[0015] In one embodiment of the present invention, the process of relevance-based data aggregation and encapsulation of data from associated devices within the same spatiotemporal window via clock-synchronized data stream, and the establishment of data association anchors based on preset spatial topological and functional dependencies between devices to obtain a standardized sensing information stream, includes: The length of the preset sliding time window is dynamically determined by synchronizing the data stream with the clock. The window boundary is adjusted according to the time density and parameter change activity of the data generated by each device within the window to obtain dynamically optimized spatiotemporal adjustment window parameters. Instead of using a fixed window, the original spatiotemporal window is dynamically processed to adapt to the differences in data density under different scenarios.

[0016] The time alignment and straightening of data generated by each device within the window are performed using the spatiotemporal straightening window parameters. The straightened data is then grouped and sorted according to the device function type to obtain straightened data groups within the same window. By grouping data within the same window, the spatial topology relationships between the devices from which the data originates are matched to obtain a description of the spatial relationships. The functional dependencies between the devices from which the data within the window originate are matched by spatial relationship description to obtain a description of functional relationships. The data set is calibrated with multidimensional association anchors by describing spatial and functional relationships, and the calibrated data is then encapsulated into a standardized perceptual information flow.

[0017] Specifically, multi-dimensional association anchor points are calibrated on the consolidated data set through spatial and functional relationship descriptions, and the calibrated data is encapsulated into a standardized perceptual information flow, including: By describing spatial relationships, the physical proximity and spatial inclusion of the devices from which the data in the consolidated data group are sourced are quantified to determine which data belong to the same room, the same area, or the same monitoring unit, and spatial affiliation labels between data are obtained. By describing the functional relationships, the monitoring collaboration type and causal dependence direction between the devices from which the data in the consolidated data group are categorized, and it is determined which data constitutes the vital sign monitoring chain or the environment-behavior linkage monitoring equivalent functional collaboration group in the health and wellness service, and the functional collaboration relationship label between the data is obtained. By performing cross-correlation analysis on the consolidated data groups using spatial affiliation labels and functional synergy labels, data combinations that simultaneously satisfy spatial proximity and functional synergy are identified, and these combinations are marked as high-priority linkage association groups to obtain high-value association data combination identifiers. By using high-value associated data combination identifiers, each data point in the consolidated data group is assigned a unified spatiotemporal coordinate identifier and a functional collaboration group identifier. This allows the data to directly access its complete context information without re-analyzing its source relationship when entering subsequent processing stages, thus obtaining a standardized perceptual information flow carrying three-dimensional associated anchor points.

[0018] The working principle and technical effects of the above solution are as follows: This method dynamically adjusts the time length of the sliding window based on the actual data flow density and parameter change activity of each device within the window. During periods of high data density (such as during the day when elderly people are frequently active), the window is contracted to ensure time accuracy; during periods of low data density (such as during nighttime sleep), the window is appropriately widened to ensure data integrity, thus obtaining dynamically optimized window parameters. The system precisely aligns and organizes the data from each device within the window according to the time axis, and groups and sorts them according to device function type (such as vital sign monitoring, environmental sensing, and security alarm), forming structured and organized data groups. The system matches the spatial affiliation between devices in the data set (determining which devices belong to the same physical space unit) and the functional collaboration between devices (determining which devices form collaborative working pairs in health and wellness monitoring). Based on this dual association, the system performs multi-dimensional anchor point labeling on the data set, adding a spatial location identifier (e.g., bedroom / living room) and a functional collaboration group identifier (e.g., vital signs-behavior linkage monitoring group) to each data point. Data combinations that simultaneously satisfy spatial proximity and functional collaboration are identified and marked as high-priority linkage association groups. Finally, the data is encapsulated into a standardized perceptual information flow carrying three-dimensional association anchor points (time coordinates, spatial location, and functional collaboration relationship).

[0019] This method elevates data alignment from simple time alignment to a multi-dimensional calibration level that combines spatiotemporal unification with functional association, ensuring that each piece of perceived data carries complete contextual information from the moment it is generated. Dynamic window optimization avoids the dual drawbacks of fixed windows—loss of detail in dense data and redundancy in sparse data—improving the scenario adaptability of data alignment. Through dual anchor point calibration of spatial affiliation and functional synergy, the system can proactively identify high-value potential correlations between data points. For example, heart rate monitoring data and mattress movement data within the same bedroom can be marked as a high-priority correlation group, allowing for direct causal determination by invoking the correlation of this group of data without subsequent re-analysis. This design significantly reduces the burden of repeatedly parsing data source relationships in subsequent stages, lowers processing latency, and ensures that data from the collection end possesses informational completeness suitable for direct use in intelligent decision-making.

[0020] In one embodiment of the present invention, S2 includes: By extracting and combining basic contextual elements through standardized perceptual information flow, a basic feature set of the scene is obtained; By tracking and analyzing the changing trends, fluctuation amplitudes, and characteristic patterns of each monitoring parameter through the basic feature set of the scenario, we can distinguish between steady-state operation and abnormal fluctuation states and obtain the state evolution feature set. performing similarity comparison and dynamic classification on the typical patterns of each scene category within a continuous time window through a state evolution feature set, and classifying perception information with similar context attributes and evolution patterns into the same scene category to obtain a scene classification result; inductively summarizing the parameter variation rules and typical linkage response patterns of each category of scenes based on the scene classification result to obtain a scene cognitive graph.

[0021] The working principle and technical effects of the above technical solution are as follows: this method extracts basic context elements such as the current time period (e.g., early morning / daytime / night), spatial location (e.g., bedroom / living room / corridor), device working mode (e.g., normal monitoring / standby / alarm), and personnel activity intensity, which are combined to form a scene basic feature set describing the basic attributes of the current health care environment. The system tracks and analyzes the variation trend, fluctuation amplitude and fluctuation pattern of each monitoring parameter (heart rate, respiratory rate, body movement frequency, indoor temperature, door and window status, etc.) within a continuous time window, distinguishes each parameter as being in a steady-state operation state or an abnormal fluctuation state, and forms a feature set reflecting the dynamic evolution state of each monitoring indicator. Based on the state evolution feature set, the system performs similarity comparison on the typical patterns of each scene category within the continuous time window, and classifies perception information with similar context attributes (e.g., both in the night time period, both in the bedroom space, both with low activity intensity) and similar evolution patterns (e.g., the parameter fluctuation amplitudes are all in a low level range) into the same scene category, so as to realize dynamic scene classification. For each type of scene in the classification result, the system respectively summarizes the variation rule of each monitoring parameter under the normal operation state, the typical manifestation of various deviation situations, and the linkage response modes adopted in history to deal with different deviations under this type of scene, so as to form a structured scene cognitive graph.

[0022] This method solves the cognitive deficiency problem of the traditional health care systems that only collect data without understanding scenes, and transforms the system from a mechanical data recorder into an intelligent agent with scene perception capability. Through dynamic scene classification, the system can identify the specific current health care scene category (e.g., determining whether the current scene is a daytime home activity scene or a night deep sleep monitoring scene), so as to provide scene context reference for subsequent causal correlation analysis and linkage decision-making. By summarizing the parameter variation rules and linkage response patterns of various scenes, the system can establish differentiated normal-abnormal judgment benchmarks for different scenes. The same heart rate value may belong to normal fluctuation in a daytime activity scene, but may constitute an abnormality in a night deep sleep scene, which effectively avoids misjudgment caused by processing all scenes with a unified threshold, and improves the scene adaptation accuracy and reliability of linkage triggering.

[0023] In an embodiment of the present invention, the step of inductively summarizing the parameter variation rules and typical linkage response patterns of each category of scenes based on the scene classification result to obtain a scene cognitive graph comprises: The scene classification results are used to extract and organize the defining conditions of each scene category to obtain the scene boundary definitions and admission condition descriptions for each type of scene. By defining scene boundaries and describing admission conditions, statistical summarization of each monitoring parameter in each scene category under normal operating conditions is performed to obtain a baseline library of parameter change patterns for various scenes. By using the baseline library of parameter change patterns, we sort out and summarize the typical temporal dependencies and historical records of linkage responses between different parameters in each scenario category, and obtain the parameter dependency map and typical linkage response pattern library for various scenarios. By using a baseline library of parameter variation patterns, a graph of dependencies between parameters, and a library of typical linkage response patterns, a multi-dimensional mapping link is established through structured association and unified encapsulation to obtain a scene cognition graph.

[0024] Specifically, by using a baseline library of parameter variation patterns, a graph of inter-parameter dependencies, and a library of typical linkage response patterns for structured association and unified encapsulation, a multi-dimensional mapping link is established to obtain a scene cognition graph, including: The baseline library of parameter change patterns standardizes the value range and fluctuation range of each monitoring parameter in each scenario category under normal operating conditions, and stores the parameter identifier and the corresponding baseline data as key-value pairs to obtain the parameter-baseline mapping index table. The temporal sequence, direction of influence, and response delay time window of each parameter are quantitatively described by the parameter dependency graph. The dependency is expressed in the form of a triplet of predecessor parameter-follower parameter-delay window to obtain a causal dependency mapping table between parameters. By sorting out the matching relationship between different abnormal deviation situations and corresponding linkage response methods in various scenario categories through the typical linkage response pattern library, the triggering conditions, response actions, execution devices and expected effects corresponding to each deviation situation are associated and stored to obtain the deviation-response mapping pattern table. By establishing a bidirectional index and cross-link between the parameter-baseline mapping index table, the parameter causal dependency mapping table, and the deviation-response mapping pattern table with the scenario category as the primary key, a scenario cognitive map with multi-dimensional bidirectional query capability is obtained by using any parameter to query its associated scenario, related parameters, corresponding baseline, and possible linkage response methods.

[0025] The working principle and technical effects of the above technical solution are as follows: After the scenario classification is completed, the defining conditions of each scenario category are organized (clarifying the time range, spatial range, equipment combination method, and personnel activity intensity boundary of the corresponding scenario), forming scenario boundary definitions and access condition descriptions to ensure that real-time sensing information can accurately match its corresponding scenario. The system statistically summarizes the value range, fluctuation range, and change rate of each monitoring parameter in each scenario category under normal operating conditions, forming a baseline library of parameter change patterns; at the same time, it sorts out the typical temporal dependencies between different parameters in each scenario category (such as which parameter changes usually before which parameter) and linkage response history, summarizing them into a parameter dependency graph and a typical linkage response pattern library. The system uses the scenario category as the primary key and performs bidirectional indexing and cross-linking of the parameter-baseline mapping index table, the parameter causal dependency mapping table, and the deviation-response mapping pattern table. Through any parameter, one can reverse query its corresponding scenario, associated parameters, corresponding normal operating baseline, and linkage response methods that may be triggered in that scenario, constructing a complete scenario cognition graph with multi-dimensional bidirectional query capabilities.

[0026] This method deepens the scene cognition graph from a conceptual cognitive structure into a structured knowledge base that can be invoked by computing systems in real time. By establishing a bidirectional index between three mapping tables with scene category as the primary key, the system can quickly reverse locate the scene to which any currently perceived parameter belongs, retrieve the baseline of related parameters for that scene, identify the degree of deviation of the current deviation from the baseline, and directly match the corresponding linkage response mode under that scene in real-time causal determination, without having to perform data retrieval and correlation analysis again for each determination. This structured design significantly reduces the query time in the causal determination process, improves the decision-making efficiency of the system under millisecond-level response requirements, and also gives the scene cognition graph the attribute of a dynamically updatable knowledge base.

[0027] In one embodiment of the present invention, S3 includes: By matching and locating the scene category to which the current real-time perceived information belongs through the scene cognition map, the historical change patterns, parameter dependencies, and key security protection targets corresponding to the scene category are retrieved from the map to obtain the historical pattern reference benchmark and scene security focus information for the current scene. By referencing historical patterns and comparing real-time perceived information with the focus of scene safety concerns, the system performs time-series alignment and deviation comparison to identify change nodes and deviation magnitudes in real-time data that deviate from historical patterns. Based on the focus of scene safety concerns, the system performs key causal correlation determination on the temporal relationships between parameters that cause deviations, and obtains the causal correlation topology and abnormal deviation characteristics of the current scene. By combining causal relationship topology, abnormal deviation features, known evolution patterns in the scene cognitive map, and historical linkage response experience, it is determined whether the current deviation state has reached the scene safety limit, and the corresponding linkage response method is matched for the deviation situation that requires linkage response, so as to obtain a list of potential linkage response conditions.

[0028] The working principle and technical effects of the above-mentioned technical solution are as follows: After obtaining the scene cognitive map, this method first performs scene category matching and positioning on the current real-time perceived information. Based on the current time period, spatial location, and perceived parameter characteristics, it retrieves the corresponding scene category entries from the map and retrieves the historical change patterns of the scene category, the dependencies between various parameters, and the preset key safety protection targets for this type of scene, forming a historical pattern reference benchmark and safety focus for the current scene. The system performs time-series alignment and deviation comparison between the real-time perceived information and the historical patterns of the scene to identify which parameters deviated from their normal operating baseline at what time and the magnitude of the deviation. In this process, the system focuses on determining the causal relationship between the parameters that caused the deviation based on the scene's safety focus. That is, it does not perform uniform causal modeling on all parameters, but prioritizes analyzing key causal chains directly related to the current scene's safety targets (such as focusing on the causal relationship between changes in body movement signals and changes in heart rate in a fall prevention scenario), forming a causal relationship topology and abnormal deviation characteristics focused on the safety target. The system matches the current deviation status with the known evolution patterns of the scene category in the scene cognition graph, determines whether the deviation has reached the safety limit, and matches the corresponding linkage response method for situations that require linkage response, generating a list of potential linkage response conditions.

[0029] This method addresses the problems of traditional linkage systems that perform general data analysis without a clear objective, leading to wasted computational resources and delayed responses. By using a scenario cognitive graph as a constraint for objective-oriented causal correlation analysis, the system can focus on key causal chains directly related to the current scenario's safety objectives. This avoids indiscriminate general probability analysis of all parameters, significantly reducing computational load while maintaining accuracy. By comparing deviations from historical patterns rather than generalizing predictions of unknown states to determine linkage trigger requirements, the generation of linkage conditions is based on experience regarding what has already happened in the scenario and how to respond most effectively. This effectively avoids the prediction inaccuracies caused by data sparsity in general risk prediction models in health and wellness scenarios, improving the scenario-specificity and reliability of linkage trigger determination.

[0030] In one embodiment of the present invention, the method of determining whether the current deviation state has reached the scene safety limit by combining causal correlation topology, abnormal deviation features, known evolution patterns in the scene cognitive map, and historical linkage response experience, and matching corresponding linkage response methods for deviation situations requiring linkage response, thereby obtaining a list of potential linkage response conditions, including: The degree and duration of the current deviation from the normal baseline are calculated by using causal relationship topology and abnormal deviation characteristics to obtain the comprehensive deviation index value of the current deviation. By comparing the comprehensive deviation index value with the safety protection threshold in the scene cognition map item by item, it is determined whether the current deviation state has reached or exceeded the safety warning line that triggers linkage response, and the judgment result and the corresponding safety target category are obtained. If the determination result is that the safety bottom line has been reached, the category of the safety target reached is identified by the scene safety focus information, and the known evolution pattern corresponding to the safety target is retrieved from the scene cognition map to obtain the evolution path reference information and the expected evolution time window. By referencing evolution path information and expected evolution time windows, the system matches the historical best linkage response mode corresponding to the corresponding evolution path from the typical linkage response mode library of the scene cognition graph, and adaptively adjusts the execution parameters of the matched response mode according to the specific state of the current real-time deviation to obtain the specific linkage response mode. For multiple simultaneous deviation scenarios, a specific linkage response method is matched to each scenario. The linkage response methods are then sorted and integrated according to the priority of their respective safety objectives to obtain a list of potential linkage response conditions.

[0031] Specifically, by referencing evolution path information and the expected evolution time window, the system matches the historically optimal linkage response mode corresponding to the corresponding evolution path from the typical linkage response mode library of the scene cognitive map. Then, based on the specific state of the current real-time deviation, the execution parameters of the matched response mode are adaptively adjusted to obtain the specific linkage response mode, including: By referencing evolution path information and the expected evolution time window, historical linkage response records matching the evolution path are retrieved from the typical linkage response pattern library of the scene cognitive graph. The retrieval results are sorted according to the response success rate and effect satisfaction to obtain the historical optimal linkage response method and its execution parameter template. By analyzing the deviation type, deviation magnitude, rate of change, and duration of the current real-time deviation, the trigger sensitivity, response strength, and execution duration in the historical optimal linkage response mode execution parameter template are adaptively adjusted to match the execution strength of the response mode with the actual severity of the current deviation, thus obtaining a preliminary linkage response mode after parameter adjustment. By describing the spatial relationships and functional collaboration relationships of the scene category in the scene cognition map, the execution devices involved in the preliminary linkage response method are identified and supplemented. Device actions that are not within the spatial range or cannot collaborate with other devices in the scene are eliminated, and related device actions that can collaborate in response are supplemented to obtain the linkage response method after spatial adaptation. By using the spatially adapted linkage response method to perform conflict detection and resource coordination with other linkage tasks that are already in execution in the current scenario, it can identify whether there are device occupancy conflicts, instruction timing conflicts, or security logic conflicts, and perform priority arbitration and execution timing adjustment according to the conflict resolution rules to obtain a specific linkage response method that can be issued immediately.

[0032] The working principle and technical effect of the above technical solution are as follows: After determining that the current deviation has reached the safety limit of the scene, this method first compares the comprehensive deviation index value (comprehensive deviation degree and duration) with the safety protection threshold in the scene cognition map to confirm the specific safety target category reached (such as fall risk, abnormal heart rate, environmental hazard, etc.). After confirming the category, the system retrieves the known evolution pattern corresponding to this type of safety target from the scene cognition map, that is, how the subsequent development usually occurs under similar scenarios and similar deviations in the past, to obtain evolution path reference information and expected evolution time window. The system retrieves historical linkage response records that match the evolution path from the typical linkage response pattern library, sorts them according to historical response success rate and effect satisfaction, and selects the historical optimal response method and its execution parameter template. On this basis, the system adaptively adjusts the trigger sensitivity, response intensity, and execution duration in the historical optimal response method execution parameter template according to the specific state of the current deviation (deviation type, deviation magnitude, rate of change, and duration), so that the response intensity is accurately matched with the actual deviation severity. Then, the system combines the spatial association and functional collaboration description of the scene in the scene cognition map to spatially adapt the execution devices involved in the response method (removing devices that are not in the same space or cannot collaborate, and supplementing the associated devices that can collaborate). It also performs conflict detection (device occupancy conflict, instruction timing conflict, security logic conflict) between the adjusted response method and other linkage tasks that are already in the execution state in the current scene. According to the conflict resolution rules, it performs priority arbitration and execution timing adjustment, and finally outputs a specific linkage response method that can be issued immediately.

[0033] This method achieves a leap in the linkage response mechanism from fixed rule triggering to a three-layer flexible decision-making approach based on historical best practices, adaptive adjustment to the current real-time state, and conflict resolution. By matching the optimal response method from historical experience, the system can quickly learn from successful handling experiences in similar scenarios when facing sudden security incidents, avoiding the time overhead of complex logical reasoning from scratch. By adaptively adjusting the execution parameters to the current state, the response intensity is precisely matched with the severity of the deviation, avoiding overreaction for minor deviations and delays in handling for severe deviations, thus avoiding the problems of over-response for small deviations and under-response for large deviations under traditional fixed rules. Through spatial adaptation and conflict detection, the executability and temporal logic correctness of linkage commands in multi-device collaborative scenarios are ensured, effectively avoiding response failures or delays caused by command conflicts or device occupancy, thereby improving the success rate and security of linkage execution.

[0034] In one embodiment of the present invention, S4 includes: By identifying the security protection priorities of the current scenario through a list of potential linkage response conditions, a scenario security target priority matrix is ​​obtained; The security contribution of each candidate linkage response condition is evaluated by using the scenario security target priority matrix, and the efficiency consumption of each condition during execution is analyzed simultaneously to obtain the comprehensive cost and benefit evaluation results of each candidate condition. The candidate conditions are selected and ranked based on the comprehensive cost-benefit evaluation results, and the selected conditions are matched and bound with the specific terminal device actions to obtain the optimal linkage response scheme. By optimizing the linkage response scheme and setting feedback monitoring points for the execution process, and updating the scene cognition map and correcting the condition judgment based on the execution results, a complete linkage control scheme is obtained.

[0035] The working principle and technical effects of the above technical solution are as follows: the present method identifies the security protection priority of the current scenario (clarifying the priority ranking among different security objectives in this scenario, such as fall protection, vital sign abnormality warning, and environmental hazard handling), so as to form a scenario security objective priority matrix. The system conducts a two-dimensional evaluation on each candidate linkage response condition: on the one hand, the contribution degree of the condition to achieving the scenario security objective is evaluated, that is, to what extent the execution of the response can contain or eliminate the current abnormal state; on the other hand, the consumption of system service performance during the execution of the condition is analyzed synchronously, including response delay occupation, device load occupation, communication bandwidth occupation, energy consumption, etc. Based on the two-dimensional evaluation results of comprehensive cost and benefit, the system performs multi-objective optimal selection and screening on each candidate condition, preferentially retains the conditions with high security contribution and low performance consumption, eliminates the conditions with low security contribution or excessively high performance consumption, and sorts the screened conditions according to the priority of the security objective. The system instantiates, matches and binds the screened and sorted linkage response conditions with specific terminal device actions (such as activating camera recording, sending a graded alarm to the nursing station, adjusting the air conditioner temperature, pushing a notification to family members, etc.), so as to form a preferred linkage response scheme. The system sets an execution feedback monitoring point and a result return mechanism for each control instruction issued in the scheme, so that the execution result can be returned to the system as a basis for subsequent correction.

[0036] The present method upgrades the linkage decision-making from the traditional single trigger rule matching to the optimal decision-making level of multi-objective comprehensive trade-off. Through the identification and matrix expression of the priority of scenario security objectives, the system can execute in an orderly manner according to the security importance when facing multiple coexisting linkage response demands, which avoids the problem of resource contention or delay of key responses caused by disordered execution when multiple conditions are triggered simultaneously. Through the two-dimensional comprehensive evaluation and optimal screening of the security contribution degree and performance consumption of each candidate condition, the maximization of security benefits is realized under the condition of limited system resources (device load, communication bandwidth, energy consumption), which not only ensures rapid response in high-risk scenarios, but also avoids the waste of device energy consumption and the increase of the burden on nursing staff caused by excessive linkage in low-risk scenarios. By setting the execution feedback monitoring point, the system evolves from one-time decision-making to a continuous optimization mode that is traceable, evaluable and improvable.

[0037] In an embodiment of the present invention, the setting of feedback monitoring points for the execution process through the preferred linkage response scheme, and the updating of the scenario cognitive map and the correction of condition judgment according to the execution result to obtain a complete linkage control scheme includes: Setting execution status monitoring points and a result return mechanism for the control instructions issued to each terminal device through the preferred linkage response scheme, so as to obtain execution status feedback information of the linkage instructions; Confirming the execution completion degree of the linkage instruction and evaluating the effect through the execution status feedback information, so as to obtain linkage response effect evaluation information; By comparing and guiding the deviation between the linkage response effect evaluation information and the expected scenario safety target, deviation correction guidance information is obtained; By using deviation correction guidance information, the baseline of parameter change patterns, the graph of inter-parameter dependencies, or the library of typical linkage response patterns in the scene cognitive graph are fine-tuned or updated. The optimized cognitive graph is then fed back to the updated reference benchmark to obtain a complete linkage control scheme.

[0038] Specifically, by using deviation correction guidance information to fine-tune parameters or update patterns in the corresponding parameter change baseline, parameter dependency graph, or typical linkage response pattern library in the scene cognitive map, the optimized cognitive map is fed back to the updated reference baseline to obtain a complete linkage control scheme, including: By using deviation correction guidance information, the reasons for deviations that fail to achieve the expected safety goals in actual execution are classified, distinguishing between parameter baseline setting deviations, dependency judgment deviations, and response mode selection deviations, and obtaining deviation cause type determination information. By comparing the deviation cause type information with the corresponding parameter change pattern baseline in the scene cognition map, when the deviation is determined to be due to inaccurate baseline setting, the baseline value is progressively corrected according to the statistical characteristics of the actual deviation, so that the corrected baseline is closer to the actual operation pattern in the scene, and the updated parameter change pattern baseline is obtained. When the judgment deviation originates from the judgment deviation of the parameter dependency relationship, the weights of the temporal sequence relationship and delay window of the parameters of this scene category in the scene cognition graph are adjusted or the relationship is updated based on the measured data of the order of parameter changes and response delay time during the actual execution process, so as to obtain the updated parameter dependency relationship graph. When the deviation is determined to be due to improper selection of response mode or inaccurate adjustment of execution parameters, the matching confidence between the deviation and the response mode is updated based on the actual evolution trajectory of the current deviation and the actual effect of the matched response mode. The parameter values ​​in the execution parameter template of the response mode are fine-tuned and optimized to obtain the updated typical linkage response mode library. The updated parameter change baseline, the updated parameter dependency graph, and the updated typical linkage response pattern library obtained through the above steps are re-encapsulated into the corresponding scene category entries of the scene cognition graph. This allows the scene cognition graph to be continuously iterated and optimized under the drive of actual execution experience and applied to causal relationship determination and linkage decision-making to obtain a complete linkage control scheme.

[0039] The working principle and technical effects of the above technical solution are as follows: This method collects the execution time, completion status, and response status information returned by the device for each linkage command through pre-set execution status monitoring points and result feedback mechanisms, forming execution status feedback information. The system confirms the completion degree of the linkage command and tracks and monitors the changes in the elderly's subsequent health status and environmental status after the linkage response is completed, assessing whether the linkage response effectively curbs the further deterioration of the abnormal state or eliminates the safety threat, forming linkage response effect evaluation information. The system compares the actual effect evaluation information with the expected scenario safety target: when the deviation between the actual effect and the expected target exceeds the preset acceptable range, the system classifies and judges the cause of the deviation, distinguishing it as benchmark reference deviation caused by inaccurate parameter baseline setting, causal path deviation caused by inaccurate judgment of the dependency relationship between parameters, or execution effect deviation caused by improper response mode selection or inaccurate adjustment of execution parameters. Based on the different types of deviation causes, the system makes targeted corrections to the corresponding components in the scene cognition graph: it progressively corrects baseline values ​​for baseline setting deviations; it adjusts and updates the weights and delay windows of temporal precedence relationships for dependency judgment deviations; and it updates the confidence level of the matching between deviation scenarios and response methods, and fine-tunes the execution parameter templates for response mode deviations. All updated elements are re-encapsulated into the corresponding scene category entries in the scene cognition graph, enabling the graph to continuously iterate and optimize under the drive of actual execution experience, and serving as a reference benchmark for subsequent causal relationship determination and linkage decision-making.

[0040] This method fundamentally solves the static defects of traditional collaborative solutions—which only execute, do not reflect, and do not evolve—by implementing a complete closed-loop mechanism from execution feedback to effect evaluation, deviation cause classification, targeted correction of map elements, and re-encapsulation and storage. By feeding back deviations between actual execution results and expected goals to each component of the scenario cognitive map for targeted correction—adjusting the baseline if it's inaccurate, adjusting relationships if they are inaccurate, and adjusting the response mode if it's not optimal—the system can accurately identify and precisely strengthen its own cognitive structure, rather than making general adjustments that could lead to errors interfering with already accurate parts. By continuously iterating the scenario cognitive map driven by actual execution experience, the system gradually approaches the actual operating rules of the health and wellness environment as its operating time increases and cases accumulate. This continuously improves the accuracy of causal relationship determination and collaborative decision-making, achieving a fundamental leap from static deployment to continuous self-evolution of the health and wellness collaborative system, effectively reducing the burden and risk of frequent manual parameter adjustments required in long-term operation and maintenance.

[0041] According to one embodiment of the present invention, the system includes: The data acquisition and standardization module is used to collect data and perform spatiotemporal window standardization on multi-terminal heterogeneous health and wellness devices in real time through a distributed sensing network to obtain a standardized sensing information flow. An identification and classification module, configured to perform scene feature identification and dynamic classification on device operating states and health care environment contexts through a standardized perceived information flow, so as to obtain a scene cognitive graph; An analysis and determination module, configured to perform scene target-oriented causal correlation analysis and linkage trigger condition determination on real-time perceived information and scene historical rules through the scene cognitive graph, so as to obtain a potential linkage response condition list; A trade-off and matching module, configured to perform comprehensive trade-off and optimal matching on scene safety targets and service efficiency constraints through the potential linkage response condition list, so as to obtain a linkage control scheme.

[0042] The working principle and technical effects of the above technical solution are as follows: The system performs real-time data collection on multi-terminal heterogeneous health care devices in multiple scenes such as homes, communities and institutions through a distributed sensing network, and performs spatiotemporal window normalization processing at the data entry, converting original signals of different devices, different protocols and different time references into a standardized perceived information flow with unified spatiotemporal coordinates. The system performs multi-dimensional scene feature identification on the device operating states and the health care environment contexts in the standardized information flow, dynamically classifies continuous perceived information into corresponding health care scene categories (such as home rest, deep sleep at night, daytime activities, centralized institutional care, etc.) according to elements such as current time period, spatial position, device combination mode and personnel activity intensity, and summarizes the parameter variation rules and typical linkage response modes of various scenes, so as to form a reusable scene cognitive graph. Based on the scene cognitive graph as the cognitive basis, the system performs scene target-oriented causal correlation analysis on the real-time perceived information and the historical rules of the scene, that is, according to the key safety protection target of the scene, selectively focuses on the causal dependency relationship between parameters directly related to the target, determines whether the current deviation state touches the scene safety bottom line and matches the corresponding linkage response mode, and generates a potential linkage response condition list. The system performs comprehensive trade-off between scene safety target priority and service efficiency consumption on each candidate linkage response condition in the list, optimally screens and binds them to specific terminal device actions according to the multi-objective optimization principle, and outputs a complete linkage control scheme.

[0043] This system constructs a complete intelligent linkage chain from data normalization to scene cognition, to causal determination, and to decision output, and solves the problem of decision lag caused by the separation of each link and information transmission distortion in traditional health care linkage solutions from the system architecture level. By taking data governance and scene cognition as the preconditions of linkage decision-making, each linkage response is established on the premise that the system truly understands what scene is currently occurring and where the safety bottom line of the scene lies, which effectively reduces the risk of false positives and missed reports caused by data confusion or scene misjudgment, and improves the intelligence level and decision-making reliability of the health care service system deployed in multiple scenes.

[0044] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals, characterized in that, The method includes: S1. Real-time data collection and spatiotemporal window adjustment of multi-terminal heterogeneous health and wellness devices are performed through a distributed sensing network to obtain a standardized sensing information flow. S2. By standardizing the perception information flow, the scene features of the equipment operation status and the context of the health and wellness environment are identified and dynamically classified to obtain a scene cognitive map. S3. By using the scene cognition graph to perform causal correlation analysis and linkage trigger condition determination on real-time perceived information and scene historical patterns, a list of potential linkage response conditions is obtained. S4. By comprehensively weighing and selecting the best match between scenario security objectives and service performance constraints through a list of potential linkage response conditions, a linkage control scheme is obtained.

2. The method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals according to claim 1, characterized in that, S1 includes: By using a distributed sensing network, real-time data capture is performed on multi-terminal heterogeneous health and wellness devices in different scenarios to obtain multi-dimensional raw device signal streams. By performing protocol adaptation and signal decoding on each terminal device through the original device signal stream, device data under different communication protocols is converted into a unified expression format to obtain a homogenized device data stream. By homogenizing the device data stream, the timestamps of each device's data are calibrated and synchronized across terminals to obtain a clock synchronization data stream. By using clock-synchronized data streams, the data of each associated device within the same spatiotemporal window are correlated, aggregated, and encapsulated. Based on the preset spatial topology and functional dependencies between devices, connection anchors between data are established to obtain a standardized perception information stream.

3. The method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals according to claim 2, characterized in that, The process involves relevance-based data aggregation and encapsulation of data from associated devices within the same spatiotemporal window via clock-synchronized data streams, and establishing data association anchors based on preset spatial topological and functional dependencies between devices to obtain a standardized perceptual information stream, including: The length of the preset sliding time window is dynamically determined by synchronizing the data stream with the clock. The window boundary is adjusted according to the time density and parameter change activity of the data generated by each device within the window to obtain dynamically optimized spatiotemporal adjustment window parameters. The time alignment and straightening of data generated by each device within the window are performed using the spatiotemporal straightening window parameters. The straightened data is then grouped and sorted according to the device function type to obtain straightened data groups within the same window. By grouping data within the same window, the spatial topology relationships between the devices from which the data originates are matched to obtain a description of the spatial relationships. The functional dependencies between the devices from which the data within the window originate are matched by spatial relationship description to obtain a description of functional relationships. The data set is calibrated with multidimensional association anchors by describing spatial and functional relationships, and the calibrated data is then encapsulated into a standardized perceptual information flow.

4. The method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals according to claim 1, characterized in that, S2 includes: By extracting and combining basic contextual elements through standardized perceptual information flow, a basic feature set of the scene is obtained; By tracking and analyzing the changing trends, fluctuation amplitudes and characteristic patterns of each monitoring parameter through the basic feature set of the scenario, the steady-state operation state and the abnormal fluctuation state are distinguished, and the state evolution feature set is obtained. By using state evolution feature sets to compare the similarity of typical patterns of each scene category within a continuous time window and dynamically classify them, perceptual information with similar contextual attributes and evolution patterns is grouped into the same scene category to obtain scene classification results. By summarizing the parameter variation patterns and typical linkage response modes of each type of scenario based on the scenario classification results, a scenario cognitive map is obtained.

5. The method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals according to claim 4, characterized in that, The process involves summarizing and generalizing the parameter variation patterns and typical linkage response modes of each scene category based on the scene classification results to obtain a scene cognitive map, including: The scene classification results are used to extract and organize the defining conditions of each scene category to obtain the scene boundary definitions and admission condition descriptions for each type of scene. By defining scene boundaries and describing admission conditions, statistical summarization of each monitoring parameter in each scene category under normal operating conditions is performed to obtain a baseline library of parameter change patterns for various scenes. By using the baseline library of parameter change patterns, we sort out and summarize the typical temporal dependencies and historical records of linkage responses between different parameters in each scenario category, and obtain the parameter dependency map and typical linkage response pattern library for various scenarios. By using a baseline library of parameter variation patterns, a graph of dependencies between parameters, and a library of typical linkage response patterns, a multi-dimensional mapping link is established through structured association and unified encapsulation to obtain a scene cognition graph.

6. The method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals according to claim 1, characterized in that, S3 includes: By matching and locating the scene category to which the current real-time perceived information belongs through the scene cognition map, the historical change patterns, parameter dependencies, and key security protection targets corresponding to the scene category are retrieved from the map to obtain the historical pattern reference benchmark and scene security focus information for the current scene. By referencing historical patterns and comparing real-time perceived information with the focus of scene safety concerns, the system performs time-series alignment and deviation comparison to identify change nodes and deviation magnitudes in real-time data that deviate from historical patterns. Based on the focus of scene safety concerns, the system performs key causal correlation determination on the temporal relationships between parameters that cause deviations, and obtains the causal correlation topology and abnormal deviation characteristics of the current scene. By combining causal relationship topology, abnormal deviation features, known evolution patterns in the scene cognitive map, and historical linkage response experience, it is determined whether the current deviation state has reached the scene safety limit, and the corresponding linkage response method is matched for the deviation situation that requires linkage response, so as to obtain a list of potential linkage response conditions.

7. The method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals according to claim 6, characterized in that, The process involves combining causal correlation topology, abnormal deviation features, known evolution patterns in the scene cognitive map, and historical linkage response experience to determine whether the current deviation state has reached the scene's safety threshold. It then matches corresponding linkage response methods to deviation scenarios requiring linkage response, obtaining a list of potential linkage response conditions, including: The degree and duration of the current deviation from the normal baseline are calculated by using causal relationship topology and abnormal deviation characteristics to obtain the comprehensive deviation index value of the current deviation. By comparing the comprehensive deviation index value with the safety protection threshold in the scene cognition map item by item, it is determined whether the current deviation state has reached or exceeded the safety warning line that triggers linkage response, and the judgment result and the corresponding safety target category are obtained. If the determination result is that the safety bottom line has been reached, the category of the safety target reached is identified by the scene safety focus information, and the known evolution pattern corresponding to the safety target is retrieved from the scene cognition map to obtain the evolution path reference information and the expected evolution time window. By referencing evolution path information and expected evolution time windows, the system matches the historical best linkage response mode corresponding to the corresponding evolution path from the typical linkage response mode library of the scene cognition graph, and adaptively adjusts the execution parameters of the matched response mode according to the specific state of the current real-time deviation to obtain the specific linkage response mode. For multiple simultaneous deviation scenarios, a specific linkage response method is matched to each scenario. The linkage response methods are then sorted and integrated according to the priority of their respective safety objectives to obtain a list of potential linkage response conditions.

8. The method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals according to claim 1, characterized in that, S4 includes: By identifying the security protection priorities of the current scenario through a list of potential linkage response conditions, a scenario security target priority matrix is ​​obtained; The security contribution of each candidate linkage response condition is evaluated by using the scenario security target priority matrix, and the efficiency consumption of each condition during execution is analyzed simultaneously to obtain the comprehensive cost and benefit evaluation results of each candidate condition. The candidate conditions are selected and ranked based on the comprehensive cost and benefit evaluation results. The selected conditions are then matched and bound with the specific terminal device actions to obtain the optimal linkage response scheme. By optimizing the linkage response scheme and setting feedback monitoring points in the execution process, and updating the scene cognition map and correcting the condition judgment based on the execution results, a complete linkage control scheme is obtained.

9. The method for real-time synchronization and scenario-based linkage control of heterogeneous data across multiple terminals according to claim 8, characterized in that, The process involves setting feedback monitoring points for the execution process through optimized linkage response schemes, updating the scene cognition map and correcting condition judgments based on the execution results, thereby obtaining a complete linkage control scheme, including: By optimizing the linkage response scheme, the execution status monitoring points and result feedback mechanism are set for the control commands issued to each terminal device to obtain the execution status feedback information of the linkage commands. By using execution status feedback information, the completion rate of the linkage command is confirmed and the effect is evaluated, thereby obtaining linkage response effect evaluation information; By comparing and guiding the deviation between the linkage response effect evaluation information and the expected scenario safety target, deviation correction guidance information is obtained; By using deviation correction guidance information, the baseline of parameter change patterns, the graph of inter-parameter dependencies, or the library of typical linkage response patterns in the scene cognitive graph are fine-tuned or updated. The optimized cognitive graph is then fed back to the updated reference benchmark to obtain a complete linkage control scheme.

10. A real-time synchronization and scenario-based linkage control system for heterogeneous data from multiple terminals, characterized in that, The system includes: The data acquisition and standardization module is used to collect data and perform spatiotemporal window standardization on multi-terminal heterogeneous health and wellness devices in real time through a distributed sensing network to obtain a standardized sensing information flow. The identification and classification module is used to identify and dynamically classify the scene features of the equipment operating status and the context of the health and wellness environment through a standardized perception information flow, and to obtain a scene cognitive map. The analysis and judgment module is used to perform causal correlation analysis and linkage trigger condition determination on real-time perceived information and scene historical patterns through scene cognition graph, and obtain a list of potential linkage response conditions. The weighting and matching module is used to comprehensively weigh and select the best match between scenario security objectives and service performance constraints through a list of potential linkage response conditions, so as to obtain a linkage control scheme.

Images