Household equipment control method and device, electronic equipment and medium

Abstract

The embodiment of the invention provides a household equipment control method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring multi-modal data in a household environment; carrying out desensitization processing on the multi-modal data and carrying out feature extraction to obtain multi-dimensional features; determining a preprocessing strategy matched with the multi-dimensional features in a locally deployed edge rule engine, and executing the preprocessing strategy to control the home equipment; sending the multi-dimensional features to a server, so that the server generates an active care decision based on the multi-dimensional features; and receiving and executing the active care decision issued by the server so as to control the corresponding home equipment. According to the embodiment of the invention, the edge node can instantly generate and execute the local preprocessing strategy after identifying that the multi-modal data is matched with the edge rule engine, and uploads the desensitized feature data to the server for joint decision making, thereby simplifying the process of active care decision making of household equipment. And the user privacy is guaranteed while the control efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of home appliance technology, and in particular to a control method, device, electronic device, and readable storage medium for home appliances. Background Technology

[0002] With the rapid development of IoT and AI technologies, the demand for family health management is becoming increasingly sophisticated. Smart home systems are gradually upgrading from device network control to proactive health care services, giving rise to intelligent health care systems.

[0003] During system operation, local edge nodes process multimodal data in real time and execute immediate control strategies, while simultaneously uploading feature data to the cloud for in-depth analysis and model training. In related technologies, local decision-making often relies on simple rules or is entirely dependent on the cloud, resulting in high response latency, failure upon network outages, and an inability to achieve personalized continuous optimization. When users encounter sudden health events or dynamic changes in health needs, existing solutions lack timely response and targeted care measures, failing to effectively meet users' real and dynamic health management needs. Summary of the Invention

[0004] In view of the above problems, embodiments of the present invention are proposed to provide a control method, apparatus, electronic device and readable storage medium for a home appliance that overcomes or at least partially solves the above problems.

[0005] In a first aspect, embodiments of the present invention provide a method for controlling home appliances, applied to an edge node, wherein the edge node is communicatively connected to a plurality of home appliances and a server in a home environment, the method comprising: Acquire multimodal data in the home environment; the multimodal data includes at least one of user physiological data, user behavior data, environmental data, and home appliance operation data; The multimodal data is desensitized and feature extracted to obtain multidimensional features; A preprocessing strategy matching the multidimensional features is determined in a preset edge rule engine, and the corresponding home appliances are controlled according to the preprocessing strategy. The multidimensional features are sent to the server; The system receives proactive care decisions generated by the server based on the multidimensional features and controls the corresponding home appliances to execute these decisions.

[0006] Optionally, the edge rule engine includes multiple rules, which describe the preprocessing strategy to be executed when a preset multidimensional feature is detected; The step of determining a preprocessing strategy that matches the multidimensional features in a preset edge rule engine, and controlling the corresponding home appliances according to the preprocessing strategy, includes: Among the multiple rules of the preset edge rule engine, the target rule for multidimensional feature matching is determined, and the corresponding home appliances are controlled according to the preprocessing strategy in the target rule.

[0007] Optionally, after sending the multidimensional features to the server, the method includes: Receive proactive care decisions from the server and update the edge rules engine based on the proactive care decisions.

[0008] Secondly, embodiments of the present invention provide a method for controlling home appliances, applied to a server, wherein the server is communicatively connected to at least one edge node and communicates with multiple home appliances in a home environment, the method comprising: When communicating with the server, the edge node uploads multidimensional features; these multidimensional features are generated by the edge node based on multimodal data from the home environment. Based on the aforementioned multidimensional features, user risk scenarios are determined; Based on the user risk scenarios, a proactive care decision is generated; The proactive care decision is sent to the edge node so that the edge node controls the home appliances to execute the proactive care decision.

[0009] Optionally, determining the user risk scenario based on the multidimensional features includes: A health status model is obtained; the health status model is constructed based on historical multidimensional feature vectors and a medical knowledge graph. The multidimensional features are input into the health status model to determine the user's risk scenarios.

[0010] Optionally, the method further includes: The aggregated update information is obtained by aggregating the multidimensional features sent by multiple edge nodes. The health status model is updated based on the aggregated update information.

[0011] Thirdly, embodiments of the present invention provide a device for controlling home appliances, characterized in that it is applied to an edge node, wherein the edge node is communicatively connected to a plurality of home appliances and a server in a home environment, and the device includes: The data acquisition module is used to acquire multimodal data in the home environment; the multimodal data includes at least one of user physiological data, user behavior data, environmental data, and home appliance operation data. The feature extraction module is used to desensitize the multimodal data and extract features to obtain multidimensional features; The device preprocessing module is used to determine a preprocessing strategy that matches the multidimensional features in a preset edge rule engine, and control the corresponding home devices according to the preprocessing strategy. A feature sending module is used to send the multidimensional features to the server; The decision receiving module is used to receive the proactive care decision generated by the server based on the multidimensional features, and control the corresponding home devices to execute the proactive care decision.

[0012] Optionally, the edge rule engine includes multiple rules, which describe the preprocessing strategy to be executed when a preset multidimensional feature is detected; the device preprocessing module includes: The target rule determination submodule is used to determine the target rule for multidimensional feature matching among multiple rules of the preset edge rule engine, and control the corresponding home devices according to the preprocessing strategy in the target rule.

[0013] Optionally, the device further includes: The edge rules engine update module is used to receive proactive care decisions sent from the server and update the edge rules engine according to the proactive care decisions.

[0014] Fourthly, embodiments of the present invention provide a device for controlling home appliances, characterized in that it is applied to a server, the server being communicatively connected to at least one edge node and communicating with multiple home appliances in a home environment, the device comprising: The feature receiving module is used to receive multidimensional features uploaded by the edge node when communicating with the server; the multidimensional features are generated by the edge node based on multimodal data in the home environment; The scenario determination module is used to determine the user's risk scenario based on the multidimensional features; The decision generation module is used to generate proactive care decisions based on the user's risk scenario. The decision sending module sends the proactive care decision to the edge node so that the edge node controls the home device to execute the proactive care decision.

[0015] Optionally, the scene determination module includes: The model acquisition module is used to acquire a health status model; the health status model is constructed based on historical multidimensional feature vectors and a medical knowledge graph. The multidimensional features are input into the health status model to determine the user's risk scenarios.

[0016] Optionally, the device further includes: The feature processing module is used to aggregate the multidimensional features sent by multiple edge nodes to obtain aggregated update information. The model update module is used to update the health status model based on the aggregated update information.

[0017] Fifthly, embodiments of the present invention provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method described in the first or second aspect.

[0018] In a sixth aspect, embodiments of the present invention provide a readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the control method for home appliances as described in the first or second aspect.

[0019] The embodiments of the present invention have the following advantages: This invention simplifies the proactive care decision-making process for home devices by acquiring multimodal data from a home environment; de-identifying and extracting features from the multimodal data to obtain multidimensional features; determining a preprocessing strategy matching the multidimensional features in a locally deployed edge rule engine and executing the preprocessing strategy to control home appliances; sending the multidimensional features to a server so that the server can generate proactive care decisions based on the multidimensional features; and receiving and executing the proactive care decisions issued by the server to control the corresponding home appliances. This invention simplifies the proactive care decision-making process for home appliances by enabling edge nodes to instantly generate and execute local preprocessing strategies after identifying multimodal data matching the edge rule engine, and uploading the de-identified feature data to the server for joint decision-making, thereby improving control efficiency while protecting user privacy. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating the steps of a home appliance control method provided in an embodiment of the present invention; Figure 2 This is a flowchart of another method for controlling home appliances provided in an embodiment of the present invention; Figure 3 This is a flowchart illustrating the steps of another home appliance control method provided in this embodiment of the invention; Figure 4 This is a flowchart of the steps of another home appliance control method provided in an embodiment of the present invention; Figure 5 This is a structural block diagram of a control device for home appliances provided in an embodiment of the present invention; Figure 6 This is a structural block diagram of another home appliance control device provided in an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] The terms "first," "second," etc., used in the specification and claims of this invention are used to distinguish similar objects and are not used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention can be implemented in orders other than those illustrated or described herein. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0024] With the deep integration of IoT and AI technologies in the field of smart healthcare, smart home systems are gradually evolving from automated devices that execute preset commands into intelligent care platforms capable of sensing, understanding, and proactively responding to users' health needs. The key to the successful deployment of such platforms lies in building a decision-making system that can both quickly respond to emergencies locally and continuously learn and optimize based on global data.

[0025] In related technical solutions, the decision-making process of the system often suffers from architectural limitations. A common approach deploys the complex health decision-making logic entirely on a cloud server. All multimodal data collected from the home environment must be uploaded to the cloud for analysis, inference, and the generation of control commands before being sent to local devices for execution. This centralized processing architecture leads to significant response latency because data transmission, cloud queuing calculations, and command feedback all take time. When the system detects urgent physiological or behavioral characteristics that may indicate a fall, fainting, or sudden discomfort, this lengthy cloud-based round-trip processing flow causes dangerous delays in intervention. More seriously, if the network connection is unstable or interrupted, the entire smart care system will be paralyzed and completely lose its service capability.

[0026] Another technological approach attempts to solidify simple decision-making rules on local devices, but its drawback lies in the rigidity and isolation of these rules. Existing rule engines are typically based on static, pre-defined logic, lacking the ability to self-optimize based on actual performance and user feedback. More importantly, there is often a lack of effective collaboration mechanisms between local rules and cloud-based intelligence.

[0027] Therefore, in existing technology processes, data privacy and decision-making efficiency are often seen as a contradiction. To ensure privacy, some solutions may choose to perform thorough data anonymization or desensitization locally, but this may impair the data's value for training cloud-based models; conversely, uploading detailed data to the cloud to pursue model accuracy increases the risk of privacy leaks. Designing a smooth process that can both protect core privacy and support efficient decision-making remains a challenge that existing technologies have not adequately addressed, making it difficult for systems to achieve an ideal balance between decision-making efficiency and privacy protection.

[0028] One of the innovations of this invention lies in acquiring multimodal data from the home environment; de-identifying and extracting features from the multimodal data to obtain multidimensional features; determining a preprocessing strategy matching the multidimensional features in a locally deployed edge rule engine and executing the preprocessing strategy to control home appliances; sending the multidimensional features to a server so that the server can generate proactive care decisions based on the multidimensional features; and receiving and executing the proactive care decisions issued by the server to control the corresponding home appliances. This invention simplifies the proactive care decision-making process for home appliances by enabling edge nodes to instantly generate and execute local preprocessing strategies after identifying multimodal data matching the edge rule engine, and uploading the de-identified feature data to the server for joint decision-making, thereby improving control efficiency while protecting user privacy.

[0029] Figure 1 This is a flowchart of the steps of a home appliance control method provided in an embodiment of the present invention.

[0030] like Figure 1 As shown, the method may specifically include the following steps: Step 101, applied to edge nodes, wherein the edge nodes are respectively communicatively connected to multiple home devices and a server in the home environment; acquiring multimodal data in the home environment; the multimodal data includes at least one of user physiological data, user behavior data, environmental data, and home device operation data; This invention relates to edge computing nodes deployed in a home environment. Edge nodes can actively collect raw information streams from different physical sources and data formats through their integrated or external sensors and smart home device interfaces; these information streams are collectively referred to as multimodal data. Multimodal data can include at least one of the following: user physiological data, user behavior data, environmental data, and home device operation data. User physiological data can include user vital signs parameters; user behavior data can include user posture and movement frequency; environmental data can include environmental parameters in the home environment; and device operation data can include the operating status of the smart home devices themselves and the user's adjustment methods for the smart home devices.

[0031] By establishing and acquiring multimodal data in the home environment, a comprehensive understanding of the user's home status is achieved, providing a data foundation for subsequent intelligent analysis and breaking through the decision-making limitations brought about by a single data source.

[0032] Step 102: Desensitize the multimodal data and extract features to obtain multidimensional features; In this embodiment of the invention, after acquiring the original multimodal data, the edge nodes sequentially perform two operations: desensitization and feature extraction. First, desensitization is performed, the core objective of which is to remove or transform sensitive information in the data that can directly or indirectly identify a specific individual or disclose personal privacy, while retaining as much data as possible for analysis. Second, feature extraction is performed, which extracts discriminative key information indicators, i.e., multidimensional features, from the original data stream. A multidimensional feature can be a mathematical vector, where each dimension represents a calculated numerical value with a clear physical or behavioral meaning.

[0033] By implementing differentiated desensitization strategies and high-dimensional feature extraction at the data source, the original sensor data is transformed into standardized feature vectors with high information density and low redundancy while strictly protecting user identity and behavioral privacy.

[0034] Step 103: Determine a preprocessing strategy that matches the multidimensional features in the preset edge rule engine, and control the corresponding home appliances according to the preprocessing strategy; In this embodiment of the invention, the edge rule engine can be a lightweight, real-time inference module embedded in the edge node, based on condition mapping to action rules. The edge rule engine may include a pre-built rule base, where each rule explicitly defines the constituent elements of a specific multi-dimensional feature and the preprocessing strategy to be triggered when the multi-dimensional feature matches the rule. For example, after matching the rule "getting up late at night," the preprocessing strategy could be "gradually turning on the path lights from the bedroom to the bathroom at 20% brightness." The edge node then sends the control command to the designated smart light fixture for execution via a local network protocol.

[0035] By deploying an edge rule engine on the edge side, localized real-time responses to preset rules are achieved, which can overcome the limitations of server decisions that cannot meet real-time requirements due to network latency, jitter, or interruption.

[0036] Step 104: Send the multidimensional features to the server; In this embodiment of the invention, a multidimensional feature vector that has undergone de-identification processing and feature extraction can be encrypted and uploaded to a server via a wide area network. The server can be a server that connects edge nodes and home devices.

[0037] By synchronizing multidimensional feature vectors to the server, an intuitive data stream that is easy to process can be provided for the server to make deeper, more proactive care decisions.

[0038] Step 105: Receive the proactive care decision generated by the server based on the multidimensional features, and control the corresponding home appliances to execute the proactive care decision.

[0039] In this embodiment of the invention, proactive care decisions can be received from the server. These proactive care decisions are control schemes generated by the server after performing a deep analysis of the uploaded multi-dimensional features using a more complex analytical model. Unlike the local preprocessing strategies of edge nodes, proactive care decisions can include long-term planning and complex decision instructions.

[0040] This invention simplifies the proactive care decision-making process for home devices by acquiring multimodal data from a home environment; de-identifying and extracting features from the multimodal data to obtain multidimensional features; determining a preprocessing strategy matching the multidimensional features in a locally deployed edge rule engine and executing the preprocessing strategy to control home appliances; sending the multidimensional features to a server so that the server can generate proactive care decisions based on the multidimensional features; and receiving and executing the proactive care decisions issued by the server to control the corresponding home appliances. This invention simplifies the proactive care decision-making process for home appliances by enabling edge nodes to instantly generate and execute local preprocessing strategies after identifying multimodal data matching the edge rule engine, and uploading the de-identified feature data to the server for joint decision-making, thereby improving control efficiency while protecting user privacy.

[0041] Figure 2 This is a flowchart of another method for controlling home appliances provided in an embodiment of the present invention; like Figure 2 As shown, the method may specifically include the following steps: Step 201 is applied to an edge node, which is communicatively connected to multiple home devices and a server in the home environment to acquire multimodal data in the home environment; the multimodal data includes at least one of user physiological data, user behavior data, environmental data, and home device operation data. This invention relates to edge computing nodes deployed in home environments. Edge nodes can actively collect raw information streams from different physical sources and data formats through their integrated or external sensors and smart home device interfaces; these information streams are collectively referred to as multimodal data. Multimodal data may include at least one of the following: user physiological data, user behavior data, environmental data, and home device operation data.

[0042] In some embodiments, user physiological data may include vital signs parameters such as heart rate, respiratory rate, and blood oxygen saturation monitored in a non-intrusive or minimally invasive manner via millimeter-wave radar, smart wearable devices, etc., to assess the user's basic health status. User behavior data may include user activity information captured by visual sensors, infrared arrays, acoustic sensors, etc., such as location movement trajectories and specific postures, such as sitting, lying down, or falling, to understand the user's behavioral patterns and intentions. Environmental data may include physical space parameters collected through an environmental sensor network, such as temperature, humidity, and light intensity, to characterize the micro-environment in which the user is located. Device operation data may include status feedback from smart home devices themselves, such as the set temperature and operating mode of air conditioners, the on / off status and brightness of lights, and the percentage of curtains open or closed, to reflect the current execution status of the system.

[0043] By establishing and acquiring multimodal data in the home environment, a comprehensive understanding of the user's home status is achieved, providing a data foundation for subsequent intelligent analysis and breaking through the decision-making limitations brought about by a single data source.

[0044] Step 202: Desensitize the multimodal data and extract features to obtain multidimensional features; In this embodiment of the invention, after acquiring the original multimodal data, the edge nodes sequentially perform two operations: desensitization and feature extraction. First, desensitization is performed, the core objective of which is to remove or transform sensitive information in the data that can directly or indirectly identify a specific individual or disclose personal privacy, while retaining as much data as possible for analysis. Second, feature extraction is performed, which extracts discriminative key information indicators, i.e., multidimensional features, from the original data stream. A multidimensional feature can be a mathematical vector, where each dimension represents a calculated numerical value with a clear physical or behavioral meaning.

[0045] In some embodiments, the desensitization processing method may include: applying real-time face blurring and background preservation technology to the camera video stream; first converting audio data to text, and then filtering out entity words such as names and addresses; and performing anonymized hashing processing on the device identifier.

[0046] In some embodiments, the feature extraction method may include: extracting "respiratory frequency waveform" and "intermittent variability of heartbeat" as physiological features from the original radar point cloud sequence using digital signal processing algorithms; calculating "duration of stillness in the bedroom area" and "movement speed from the bedroom to the bathroom" as behavioral features from the user's movement location sequence; and calculating "temperature rise slope over the past hour" as environmental features from the ambient temperature sequence. Through feature extraction, the system transforms massive amounts of unstructured raw data into structured, machine-readable feature vectors, greatly reducing the complexity of subsequent calculations and focusing on core information.

[0047] Step 203: The edge rule engine includes multiple rules, which describe the preprocessing strategy to be executed when a preset multidimensional feature is detected; among the multiple rules of the preset edge rule engine, the target rule for matching the multidimensional feature is determined, and the corresponding home appliances are controlled according to the preprocessing strategy in the target rule. In this embodiment of the invention, each rule in the edge rule engine is essentially a mapping relationship from multi-dimensional features to preprocessing strategies. Preset multi-dimensional features refer to specific combinations of features and threshold conditions that engineers or the system pre-set to be detected when defining the rules.

[0048] In some embodiments, the edge rule engine can compare multidimensional features within the current time window with the rule base one by one. The comparison process may also involve short-term analysis of the time-series characteristics of the multidimensional features. For example, a rule targeting "gas leak risk" might have preset feature patterns including: "combustible gas concentration sensor reading > 10 ppm" for 5 seconds, and simultaneously "no human activity detected in the kitchen area". Only when the multidimensional features simultaneously meet both conditions will the rule be matched as a candidate target rule. If multiple rules are activated simultaneously, the engine will resolve conflicts based on metadata such as rule priority, specificity, or recent usage, ultimately selecting a single target rule to execute its preprocessing strategy, such as "immediately shut off the smart gas valve and turn on the range hood to its maximum setting".

[0049] Step 204: Receive the proactive care decision sent from the server, and update the edge rule engine according to the proactive care decision; In this embodiment of the invention, while the proactive care decision is being issued and executed, the proactive care decision can also be regarded as a rule by the edge node, and this rule can be updated in the edge rule engine.

[0050] In some embodiments, the update method of the edge rules engine can be one of the following two approaches: One approach is rule generation: analyze the triggering conditions behind proactive care decisions, formalize them into a new "if-then" rule, and add it to the edge rule library. For example, if the server decides "automatically maintain the living room temperature at 25°C during a user's cold," a local rule can be generated: "If the 'user's health status label' is 'cold' and the 'location' is in the 'living room,' then set the target temperature of the temperature control device to 25°C."

[0051] Another approach is rule tuning: adjusting the parameters of existing rules based on feedback from the implementation of decisions. For example, if after implementing a decision it is found that users still feel uncomfortable with the brightness of "automatically dimming lights at night," the brightness threshold in that rule can be automatically adjusted from 30% to 20%. Through this continuous updating, the real-time responsiveness of the edge can be continuously optimized and personalized as the server learns.

[0052] Step 205: Send the multidimensional features to the server; In this embodiment of the invention, a multidimensional feature vector that has undergone de-identification processing and feature extraction can be encrypted and uploaded to a server via a wide area network. The server can be a server that connects edge nodes and home devices.

[0053] By synchronizing multidimensional feature vectors to the server, an intuitive data stream that is easy to process can be provided for the server to make deeper, more proactive care decisions.

[0054] Step 206: Receive the proactive care decision generated by the server based on the multidimensional features, and control the corresponding home appliances to execute the proactive care decision.

[0055] In this embodiment of the invention, proactive care decisions can be received from the server. These proactive care decisions are control schemes generated by the server after performing a deep analysis of the uploaded multi-dimensional features using a more complex analytical model. Unlike the local preprocessing strategies of edge nodes, proactive care decisions can include long-term planning and complex decision instructions.

[0056] This invention achieves a closed-loop collaboration between rapid local response and intelligent server analysis by acquiring multimodal data from a home environment; desensitizing and extracting features from the multimodal data to obtain multidimensional features; matching target rules corresponding to the multidimensional features in a preset edge rule engine; controlling corresponding home appliances based on preprocessing strategies in the target rules; receiving proactive care decisions from the server and updating the rules in the edge rule engine accordingly; sending the multidimensional features to the server; receiving proactive care decisions generated by the server based on the multidimensional features; and controlling the corresponding home appliances to execute these decisions. This invention enables real-time execution of preprocessing strategies based on feature matching at edge nodes, while simultaneously dynamically optimizing the local rule base using server-issued decisions and uploading feature data to the server to generate proactive care decisions. This ensures low-latency control and continuously improves the system's adaptability and care accuracy through rule iteration and server decisions.

[0057] Figure 3 This is a flowchart illustrating the steps of another home appliance control method provided in this embodiment of the invention; like Figure 3 As shown, the method may specifically include the following steps: Step 301, applied to a server, wherein the server communicates with at least one edge node and multiple home devices in the home environment, and when communicating with the server, receives multidimensional features uploaded by the edge node; the multidimensional features are generated by the edge node based on multimodal data in the home environment; In this embodiment of the invention, the server can act as a decision center connecting multiple edge nodes and receive multidimensional features that have been standardized by the edge nodes through a secure communication link.

[0058] In some embodiments, the server may also receive the edge node ID, timestamp, and multi-dimensional features representing the state of the home environment under the node's jurisdiction at a specific time, uploaded by the edge node. The server may also perform legality verification and decryption on the received data, and store it in the corresponding user context data queue.

[0059] By receiving processed multidimensional features, the system achieves aggregated processing of edge node data, providing a data foundation for multi-user, more macro-level server analysis and decision-making while protecting user privacy.

[0060] Step 302: Determine the user risk scenario based on the multidimensional features; In this embodiment of the invention, user risk scenarios can be determined based on multidimensional features and a preset analysis strategy. User risk scenarios can characterize potential adverse situations that users may face now or in the near future, requiring proactive intervention from the system.

[0061] By applying advanced anomaly detection, pattern recognition, and multi-feature fusion analysis algorithms, a precise mapping from low-level sensor features to high-level semantic risk scenarios is achieved. This helps the system go beyond simple threshold alarms, enabling it to identify complex, progressive risks (such as health decline trends and accumulated safety hazards), thus allowing for earlier and more effective intervention.

[0062] Step 303: Generate a proactive care decision based on the user risk scenario; In this embodiment of the invention, after clarifying the user's risk scenario, the process can proceed to the decision-making and planning stage to generate a proactive care decision. The logic for generating the proactive care decision can be based on a query of the policy tree in the server, or it can be based on calculations using an optimization algorithm.

[0063] By incorporating user risk scenario assessment, proactive care decisions can be made that comprehensively consider multiple factors such as user preferences, equipment capabilities, and environmental conditions.

[0064] Step 304: Send the proactive care decision to the edge node so that the edge node controls the home device to execute the proactive care decision.

[0065] In this embodiment of the invention, the server can send an instruction packet containing the proactive care decision to the target edge node based on the target edge node ID specified in the proactive care decision. The instruction packet may include a sequence of actions to be executed, execution constraints, expiration time, and requirements for reporting the execution results.

[0066] By establishing a traceable command issuance and confirmation mechanism, it is ensured that important control commands can be ultimately executed regardless of the network state of the edge node.

[0067] This invention, in its embodiments, receives multi-dimensional features generated from multimodal data uploaded by edge nodes during communication with a server; determines user risk scenarios based on these features; generates proactive care decisions based on the determined risk scenarios; and sends the decisions to the edge nodes to control home appliances to execute them. This invention, by centrally processing the feature-rich data from edge nodes, achieves accurate determination of user risk scenarios and generation of care decisions on the server, effectively leveraging the server's global analysis advantages. It provides the edge side with proactive care decisions that differ from local rules, forming a collaborative architecture between server-side decision-making and edge execution.

[0068] Figure 4 This is a flowchart of the steps of another home appliance control method provided in an embodiment of the present invention; like Figure 4 As shown, the method may specifically include the following steps: Step 401, applied to a server, the server communicating with at least one edge node and multiple home devices in the home environment, receiving multidimensional features uploaded by the edge node when communicating with the server; the multidimensional features are generated by the edge node based on multimodal data in the home environment; In this embodiment of the invention, the server can act as a decision center connecting multiple edge nodes and receive multidimensional features that have been standardized by the edge nodes through a secure communication link.

[0069] In some embodiments, the server may also receive the edge node ID, timestamp, and multi-dimensional features representing the state of the home environment under the node's jurisdiction at a specific time, uploaded by the edge node. The server may also perform legality verification and decryption on the received data, and store it in the corresponding user context data queue.

[0070] By receiving processed multidimensional features, the system achieves aggregated processing of edge node data, providing a data foundation for multi-user, more macro-level server analysis and decision-making while protecting user privacy.

[0071] Step 402, obtain the health status model; the health status model is constructed based on historical multidimensional feature vectors and medical knowledge graph; In this embodiment of the invention, the health status model can be an artificial intelligence model that integrates data-driven and knowledge-driven approaches. Its construction relies on two main pillars: first, massive, cross-user historical multimodal data, which enables the model to automatically learn normal patterns, individual differences, and statistical correlations between various health-related indicators; second, a structured medical knowledge graph, which encodes prior knowledge in the field of human medicine and defines the logical and probabilistic relationships between symptoms, signs, lifestyle habits, diseases, and risk factors. For example, "snoring and sleep apnea at night" is a typical symptom of "obstructive sleep apnea."

[0072] By loading a specialized model that integrates big data patterns and medical knowledge, the server gains near-professional health risk assessment capabilities, enabling more accurate proactive care decisions.

[0073] Step 403: Input the multidimensional features into the health status model to determine the user risk scenario; In this embodiment of the invention, the server can input the multi-dimensional feature vectors received simultaneously into the health status model and process them together to generate user risk scenarios.

[0074] In some examples, the health status model calculates the deviation of each feature value and performs joint inference based on multiple features under the semantic network constraints of the medical knowledge graph. For example, if the health status model simultaneously receives multiple features such as "average nighttime heart rate increased by 5% in the past week compared to the previous week", "deep sleep rate decreased by 10%", "daytime activity decreased", and "ambient noise level spikes at night", and uses the association between "increased heart rate", "disrupted sleep structure" and "anxiety state" or "increased cardiovascular load" in the knowledge graph, the health status model may determine that there is a user risk scenario of "decreased sleep quality and physiological stress caused by stress or anxiety".

[0075] By utilizing professional models for deep feature association and knowledge reasoning, a more clinically relevant and accurate assessment of users' health status is achieved, thereby identifying potential risks that are easily overlooked by simple rules.

[0076] Step 404: Aggregate the multidimensional features sent by multiple edge nodes to obtain aggregated update information; In this embodiment of the invention, a secure aggregation algorithm under the federated learning framework can be used to aggregate and calculate the encrypted feature gradients uploaded by multiple edge nodes, enabling cross-node collaborative model training without accessing the original user data. Aggregation processing is a technique for collaborative learning, such as federated learning, under the premise of privacy protection. The server can coordinate multiple edge nodes to train the model using local multidimensional feature historical data, but only uploads the updated gradients of the model parameters or the encrypted intermediate calculation results, not the data itself.

[0077] In some examples, taking federated learning as an example, the server can first distribute the current model parameters to the edge nodes participating in this round of training. Each node calculates the health state model loss based on local multidimensional features and generates parameter update gradients. Then, the nodes process the gradients using homomorphic encryption or differential privacy technology and upload them to the server. The server aggregates the encrypted gradients through a secure multi-party computation protocol to eliminate random noise from individual users, and finally decrypts the data to obtain aggregated update information that reflects the distribution of knowledge in the group.

[0078] By employing federated learning to securely aggregate distributed feature data, it is possible to extract common patterns and knowledge from massive user data without obtaining the original data.

[0079] Step 405: Update the health status model based on the aggregated update information; In this embodiment of the invention, the server can use the aggregated update information generated in step 404 to update the health status model, that is, adjust the parameters of the model.

[0080] In some examples, aggregated updates may indicate a stronger correlation between the environmental characteristic of "increased pollen concentration in spring" and "changes in nighttime respiratory characteristics in users with allergic rhinitis" across a large number of users. Once the model incorporates this new knowledge, its assessment of respiratory abnormalities in spring will be more accurate for users with a history of allergies. In this way, the model can continuously learn from broader population data, discover new association patterns, and evolve its performance, while strictly adhering to data privacy principles.

[0081] By leveraging collective knowledge to continuously iterate and optimize professional models, the server's core analytical capabilities have been automatically evolved and continuously enhanced. By adapting to changes in population habits, environmental factors, and new medical discoveries, the health risk assessments and service recommendations provided are up-to-date, maintaining their accuracy and cutting-edge nature.

[0082] Step 406: Generate a proactive care decision based on the user risk scenario; In this embodiment of the invention, after clarifying the user's risk scenario, the process can proceed to the decision-making and planning stage to generate a proactive care decision. The logic for generating the proactive care decision can be based on a query of the policy tree in the server, or it can be based on calculations using an optimization algorithm.

[0083] In some examples, for a high-risk scenario like "detecting a suspected fall in an elderly person living alone," the decision generator might follow a pre-defined emergency strategy tree: First, it attempts to confirm via voice inquiry through a smart speaker; if there is no effective response within 30 seconds, it executes the following decisions: "1. Turn on all lights in the house; 2. Push an alarm to the emergency contact app and dial a pre-set phone number; 3. Unlock the front door to allow emergency personnel to enter." For a long-term health risk scenario like "detecting a user who has been staying up all night and exhibits fatigue characteristics," the decision might be more moderate and planned: "One hour before the user's usual rest time the next day, automatically dim the living room lights, play soothing music, and suggest the user to rest earlier via voice." The generated decisions can ultimately be encapsulated into device-independent, serializable pre-processed strategies and sent to edge nodes to update the edge rule engine.

[0084] Step 407: Send the proactive care decision to the edge node so that the edge node controls the home device to execute the proactive care decision.

[0085] In this embodiment of the invention, the server can send an instruction packet containing the proactive care decision to the target edge node based on the target edge node ID specified in the proactive care decision. The instruction packet may include a sequence of actions to be executed, execution constraints, expiration time, and requirements for reporting the execution results.

[0086] By establishing a traceable command issuance and confirmation mechanism, it is ensured that important control commands can be ultimately executed regardless of the network state of the edge node.

[0087] This invention, when communicating with a server, receives multidimensional features generated from multimodal data uploaded by edge nodes; acquires a health status model constructed based on historical data and medical knowledge; inputs the multidimensional features into the health status model to determine user risk scenarios; aggregates multidimensional features from multiple edge nodes to update the model; optimizes the health status model using the aggregated update information; generates proactive care decisions based on the determined risk scenarios; and distributes the decisions to edge nodes to control device execution. This invention, by introducing and continuously optimizing a health status model based on medical knowledge, makes the identification of user risk scenarios more scientific and accurate. Simultaneously, by aggregating multi-node data to update the model, it achieves federated evolution of model parameters and knowledge sharing, thereby continuously improving the intelligence and personalization of the entire system's health care decisions while protecting data privacy.

[0088] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0089] Figure 5 This is a structural block diagram of a control device for home appliances provided in an embodiment of the present invention; like Figure 5 As shown in the figure, an embodiment of the present invention provides a control device for home appliances, applied to an edge node. The edge node is communicatively connected to multiple home appliances and a server in a home environment. The device may specifically include the following modules: The data acquisition module 501 is used to acquire multimodal data in the home environment; the multimodal data includes at least one of user physiological data, user behavior data, environmental data, and home appliance operation data. Feature extraction module 502 is used to perform desensitization processing on the multimodal data and extract features to obtain multidimensional features; The device preprocessing module 503 is used to determine a preprocessing strategy that matches the multidimensional features in a preset edge rule engine, and control the corresponding home appliances according to the preprocessing strategy. The feature sending module 504 is used to send the multidimensional features to the server; The decision receiving module 505 is used to receive the proactive care decision generated by the server based on the multidimensional features, and control the corresponding home devices to execute the proactive care decision.

[0090] In this embodiment of the invention, the edge rule engine includes multiple rules, which describe the preprocessing strategy to be executed when a preset multidimensional feature is detected; the device preprocessing module includes a target rule determination submodule, which is used to determine the target rule matching the multidimensional feature among the multiple rules of the preset edge rule engine, and control the corresponding home device according to the preprocessing strategy in the target rule.

[0091] In this embodiment of the invention, the device further includes: The edge rules engine update module is used to receive proactive care decisions sent from the server and update the edge rules engine according to the proactive care decisions.

[0092] This invention simplifies the proactive care decision-making process for home devices by acquiring multimodal data from a home environment; de-identifying and extracting features from the multimodal data to obtain multidimensional features; determining a preprocessing strategy matching the multidimensional features in a locally deployed edge rule engine and executing the preprocessing strategy to control home appliances; sending the multidimensional features to a server so that the server can generate proactive care decisions based on the multidimensional features; and receiving and executing the proactive care decisions issued by the server to control the corresponding home appliances. This invention simplifies the proactive care decision-making process for home appliances by enabling edge nodes to instantly generate and execute local preprocessing strategies after identifying multimodal data matching the edge rule engine, and uploading the de-identified feature data to the server for joint decision-making, thereby improving control efficiency while protecting user privacy.

[0093] Figure 6 This is a structural block diagram of another home appliance control device provided in an embodiment of the present invention.

[0094] like Figure 6 As shown in the embodiment of the present invention, another control device for home appliances is applied to a server. The server is communicatively connected to at least one edge node and communicates with multiple home appliances in the home environment. The device may specifically include the following modules: The feature receiving module 601 is used to receive multidimensional features uploaded by the edge node when communicating with the server; the multidimensional features are generated by the edge node based on multimodal data in the home environment; The scenario determination module 602 is used to determine the user risk scenario based on the multidimensional features; The decision generation module 603 is used to generate proactive care decisions based on the user risk scenario. The decision sending module 604 sends the proactive care decision to the edge node so that the edge node controls the home device to execute the proactive care decision.

[0095] In this embodiment of the invention, the scene determination module includes: The model acquisition module is used to acquire a health status model; the health status model is constructed based on historical multidimensional feature vectors and a medical knowledge graph. The multidimensional features are input into the health status model to determine the user's risk scenarios.

[0096] In this embodiment of the invention, the device further includes: The feature processing module is used to aggregate the multidimensional features sent by multiple edge nodes to obtain aggregated update information. The model update module is used to update the health status model based on the aggregated update information.

[0097] This invention, in its embodiments, receives multi-dimensional features generated from multimodal data uploaded by edge nodes during communication with a server; determines user risk scenarios based on these features; generates proactive care decisions based on the determined risk scenarios; and sends the decisions to the edge nodes to control home appliances to execute them. This invention, by centrally processing the feature-rich data from edge nodes, achieves accurate determination of user risk scenarios and generation of care decisions on the server, effectively leveraging the server's global analysis advantages. It provides the edge side with proactive care decisions that differ from local rules, forming a collaborative architecture between server-side decision-making and edge execution.

[0098] As the apparatus embodiment is basically similar to the method embodiment, it is described in a relatively simple manner. For relevant details, please refer to the description of the method embodiment.

[0099] This invention also provides an electronic device, including a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the various processes of the above-described home appliance control method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0100] It should be noted that the electronic devices in the embodiments of the present invention include the mobile electronic devices and non-mobile electronic devices described above.

[0101] This invention also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described home appliance control method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0102] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0103] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0104] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0105] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0106] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0107] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0108] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0109] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0110] The control method, device, electronic device, and computer-readable storage medium for home appliances provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for controlling home appliances, characterized in that, Applied to edge nodes, which are communicatively connected to multiple home devices and a server in a home environment, the method includes: Acquire multimodal data in the home environment; the multimodal data includes at least one of user physiological data, user behavior data, environmental data, and home appliance operation data; The multimodal data is desensitized and feature extracted to obtain multidimensional features; A preprocessing strategy matching the multidimensional features is determined in a preset edge rule engine, and the corresponding home appliances are controlled according to the preprocessing strategy. The multidimensional features are sent to the server; The system receives proactive care decisions generated by the server based on the multidimensional features and controls the corresponding home appliances to execute these decisions.

2. The control method for home appliances according to claim 1, characterized in that, The edge rule engine includes multiple rules, which describe the preprocessing strategies to be executed when a preset multidimensional feature is detected. The step of determining a preprocessing strategy that matches the multidimensional features in a preset edge rule engine, and controlling the corresponding home appliances according to the preprocessing strategy, includes: Among the multiple rules of the preset edge rule engine, the target rule for multidimensional feature matching is determined, and the corresponding home appliances are controlled according to the preprocessing strategy in the target rule.

3. The control method for home appliances according to claim 1, characterized in that, After sending the multidimensional features to the server, the method includes: Receive proactive care decisions from the server and update the edge rules engine based on the proactive care decisions.

4. A method for controlling home appliances, characterized in that, Applied to a server, the server having a communicative connection with at least one edge node and communicating with multiple home devices in a home environment, the method includes: When communicating with the server, the edge node uploads multidimensional features; these multidimensional features are generated by the edge node based on multimodal data from the home environment. Based on the aforementioned multidimensional features, user risk scenarios are determined; Based on the user risk scenarios, a proactive care decision is generated; The proactive care decision is sent to the edge node so that the edge node controls the home appliances to execute the proactive care decision.

5. The control method for home appliances according to claim 4, characterized in that, The process of determining user risk scenarios based on the multidimensional features includes: A health status model is obtained; the health status model is constructed based on historical multidimensional feature vectors and a medical knowledge graph. The multidimensional features are input into the health status model to determine the user's risk scenarios.

6. The control method for home appliances according to claim 5, characterized in that, Also includes: The aggregated update information is obtained by aggregating the multidimensional features sent by multiple edge nodes. The health status model is updated based on the aggregated update information.

7. A control device for home appliances, characterized in that, An apparatus applied to edge nodes, which are communicatively connected to multiple home devices and a server in a home environment, comprising: The data acquisition module is used to acquire multimodal data in the home environment; the multimodal data includes at least one of user physiological data, user behavior data, environmental data, and home appliance operation data. The feature extraction module is used to desensitize the multimodal data and extract features to obtain multidimensional features; The device preprocessing module is used to determine a preprocessing strategy that matches the multidimensional features in a preset edge rule engine, and control the corresponding home devices according to the preprocessing strategy. A feature sending module is used to send the multidimensional features to the server; The decision receiving module is used to receive the proactive care decision generated by the server based on the multidimensional features, and control the corresponding home devices to execute the proactive care decision.

8. A control device for home appliances, characterized in that, Applied to a server, the server communicating with at least one edge node and multiple home devices in a home environment, the device includes: The feature receiving module is used to receive multidimensional features uploaded by the edge node when communicating with the server; the multidimensional features are generated by the edge node based on multimodal data in the home environment; The scenario determination module is used to determine the user's risk scenario based on the multidimensional features; The decision generation module is used to generate proactive care decisions based on the user's risk scenario. The decision sending module sends the proactive care decision to the edge node so that the edge node controls the home device to execute the proactive care decision.

9. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory, which, when executed by the processor, implement the steps of the control method for the home appliance as described in claims 1-3 or 4-6.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the control method for the home appliance as described in claims 1-3 or 4-6.

Images