Recommended scene determination method and device, storage medium and electronic equipment

Abstract

The invention discloses a recommendation scene determination method and device, a storage medium and electronic equipment, and relates to the technical field of smart home, and the method comprises the steps: obtaining first metadata of home equipment in a first time period, the first metadata is used for representing the space where the home equipment is located and the use data of the home equipment; inputting the first metadata into a scene recommendation model for processing, and determining a first recommendation scene corresponding to the home equipment according to a first output result of the scene recommendation model; sending the first recommendation scene to the target object, and obtaining a first feedback returned by the target object based on the first recommendation scene; and determining a second recommendation scene according to the first feedback and the first recommendation scene, the second recommendation scene being applied to the home device. By adopting the technical scheme, the problems that in the related technology, automatic control of the intelligent household electrical appliance depends on manual parameter setting of a user, and the intelligent degree of the household electrical appliance is low are solved.

Description

Technical Field

[0001] This application relates to the field of smart home technology, and more specifically, to a method and apparatus for determining recommended scenarios, a storage medium, and an electronic device. Background Technology

[0002] In related technologies, the automation control of most smart home appliances still relies on manual parameter settings by users, such as preset times and operating modes. While this provides users with basic customization capabilities, the lack of in-depth understanding of user behavior and adaptive capabilities means that home appliances cannot adjust accordingly to changes in user habits. For example, users' appliance usage patterns may differ significantly in different seasons and on weekdays versus weekends, but manually set parameters are often fixed and cannot flexibly match these changes, thus reducing the level of intelligence of the appliances and the user experience.

[0003] There is still no effective solution to the problem that the automated control of smart home appliances relies on manual parameter settings by users, resulting in low levels of intelligence in these appliances. Summary of the Invention

[0004] This application provides a method and apparatus for determining recommended scenarios, a storage medium, and an electronic device to at least solve the problem in related technologies that the automated control of smart home appliances relies on users manually setting parameters, resulting in a low level of intelligence in home appliances.

[0005] According to one embodiment of this application, a method for determining a recommended scenario is provided, comprising: acquiring first metadata of a home appliance within a first time period, wherein the first time period is prior to the current time, and the first metadata is used to characterize the space where the home appliance is located and the usage data of the home appliance; inputting the first metadata into a scenario recommendation model for processing, and determining a first recommended scenario corresponding to the home appliance based on a first output result of the scenario recommendation model; sending the first recommended scenario to a target object, and obtaining first feedback returned by the target object based on the first recommended scenario; and determining a second recommended scenario based on the first feedback and the first recommended scenario, wherein the second recommended scenario is applied to the home appliance.

[0006] In an optional embodiment, before determining the first recommended scenario corresponding to the home appliance based on the first output result of the scenario recommendation model, the method further includes: extracting features from the first metadata using a convolutional neural network to obtain multiple data features; determining multiple association relationships between the multiple data features using a recurrent neural network, and inputting the multiple association relationships into the scenario recommendation model for processing to obtain multiple usage habits of the target object for the home appliance and the confidence levels of the multiple usage habits, wherein the multiple association relationships are used to indicate the relationship between any N data features among the multiple data features, the multiple usage habits and the multiple association relationships correspond one-to-one, and N is a positive integer; and determining the first usage habit with the highest confidence level among the multiple usage habits as the first output result.

[0007] In an optional embodiment, determining the first recommended scenario corresponding to the home appliance based on the first output result of the scenario recommendation model includes: determining multiple recommended scenarios matching the first output result in the home appliance function rule base, wherein the multiple recommended scenarios include the first recommended scenario, and the home appliance function rule base contains multiple scenarios corresponding to the home appliance; determining the scores of the multiple recommended scenarios, and determining the recommended scenario with the highest score among the multiple recommended scenarios as the first recommended scenario, wherein the multiple scores are used to characterize the usage frequency of the multiple recommended scenarios.

[0008] In an optional embodiment, determining the second recommended scenario based on the first feedback and the first recommended scenario includes: if the first feedback indicates acceptance of the first recommended scenario, determining the first recommended scenario as the second recommended scenario; if the first feedback indicates modification of the first recommended scenario, modifying the first recommended scenario based on the first feedback, and determining the modified first recommended scenario as the second recommended scenario; if the first feedback indicates rejection of the first recommended scenario, determining a third recommended scenario from among the multiple recommended scenarios as the second recommended scenario, wherein the third recommended scenario is the recommended scenario with the highest rating among the other recommended scenarios, and the other recommended scenarios are scenarios other than the first recommended scenario among the multiple recommended scenarios.

[0009] In an optional embodiment, after determining the second recommended scenario based on the first feedback and the first recommended scenario, the method further includes: obtaining second metadata of the home device within a second time period, wherein the second time period is the time period during which the second recommended scenario is applied to the home device, and the second metadata is of the same type as the first metadata; if a first parameter in the second metadata changes relative to a second parameter in the first metadata, the second metadata is input into the scenario recommendation model for processing, and a fourth recommended scenario of the home device is determined based on the second output result of the scenario recommendation model, wherein the first parameter and the second parameter are of the same type.

[0010] In one optional embodiment, obtaining the first metadata of a home device within a first time period includes: encrypting the fourth metadata of the home device within a third time period to obtain the fifth metadata, wherein the start time of the third time period is before the start time of the first time period, and the end time of the third time period is the current time; storing the fifth metadata on a local server and a cloud server respectively, and determining the encrypted first metadata from the fifth metadata according to the first time period; and decrypting the encrypted first metadata to obtain the first metadata.

[0011] In an optional embodiment, after obtaining the first metadata of the home device within a first time period, the method further includes: obtaining the target object as the working mode determined by the scene recommendation model, wherein the working mode includes a local mode and a cloud mode; when the working mode is the local mode, determining the category of the scene recommendation model as a local scene recommendation model; when the working mode is the cloud mode, determining the category of the scene recommendation model as a cloud scene recommendation model.

[0012] According to another aspect of the present invention, a device for determining a recommended scenario is also provided, comprising: a first acquisition module, configured to acquire first metadata of a home appliance within a first time period, wherein the first time period is prior to the current time, and the first metadata is used to characterize the space where the home appliance is located and the usage data of the home appliance; a first determination module, configured to input the first metadata into a scenario recommendation model for processing, and determine a first recommended scenario corresponding to the home appliance based on a first output result of the scenario recommendation model; a second acquisition module, configured to send the first recommended scenario to a target object, and acquire first feedback returned by the target object based on the first recommended scenario; and a second determination module, configured to determine a second recommended scenario based on the first feedback and the first recommended scenario, wherein the second recommended scenario is applied to the home appliance.

[0013] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer program, and the computer program is configured to execute the determination method of the above-mentioned recommended scenario when running.

[0014] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the method for determining the recommended scenario through the computer program.

[0015] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the method for determining the above-mentioned recommended scenario.

[0016] In this embodiment, first metadata of the home appliance within a first time period is obtained, wherein the first time period is prior to the current time, and the first metadata is used to characterize the space where the home appliance is located and the usage data of the home appliance; the first metadata is input into a scene recommendation model for processing, and a first recommended scene corresponding to the home appliance is determined based on the first output result of the scene recommendation model; the first recommended scene is sent to a target object, and a first feedback returned by the target object based on the first recommended scene is obtained; a second recommended scene is determined based on the first feedback and the first recommended scene, wherein the second recommended scene is applied to the home appliance. By adopting the above technical solution, the problem in related technologies that the automated control of smart home appliances relies on manual parameter settings by the user, resulting in a low level of intelligence in the home appliances is solved. Attached Figure Description

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

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the hardware environment for a method of determining a recommended scenario according to an embodiment of this application;

[0020] Figure 2 This is a flowchart of a method for determining recommended scenarios according to embodiments of this application;

[0021] Figure 3 This is a first schematic diagram of a method for determining a recommended scenario according to an optional embodiment of this application;

[0022] Figure 4 This is a second schematic diagram of a method for determining a recommended scenario according to an optional embodiment of this application;

[0023] Figure 5 This is a third schematic diagram illustrating a method for determining a recommended scenario according to an optional embodiment of this application;

[0024] Figure 6 This is a structural block diagram of a recommended scenario determination device according to an embodiment of this application. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] According to one aspect of the embodiments of this application, a method for determining recommended scenarios is provided. This method for determining recommended scenarios is widely applicable to whole-house intelligent digital control application scenarios such as smart homes, smart home ecosystems, smart home device ecosystems, and intelligence house ecosystems. Optionally, in this embodiment, the above-mentioned method for determining recommended scenarios can be applied to, for example... Figure 1 The hardware environment shown consists of home appliances 102 and computer terminal 104. Figure 1As shown, home device 102 is connected to computer terminal 104 via a network and can be used to provide services (such as application services) to home device 102 or clients installed on home device 102. A database can be set up on home device 102 or independently of home device 102 to provide data storage services for home device 102 and computer terminal 104. Cloud computing and / or edge computing services can be configured on home device 102 or independently of home device 102 to provide data processing services for home device 102 and computer terminal 104.

[0028] The aforementioned networks may include, but are not limited to, at least one of the following: wired network, wireless network. The aforementioned wired network may include, but is not limited to, at least one of the following: wide area network, metropolitan area network, local area network. The aforementioned wireless network may include, but is not limited to, at least one of the following: Wi-Fi (Wireless Fidelity), Bluetooth. Home appliances 102 and computer terminals 104 may not be limited to PCs, mobile phones, tablets, smart air conditioners, smart range hoods, smart refrigerators, smart ovens, smart stoves, smart washing machines, smart water heaters, smart washing equipment, smart dishwashers, smart projectors, smart TVs, smart clothes racks, smart curtains, smart audio-visual equipment, smart sockets, smart speakers, smart speakers, smart fresh air systems, smart kitchen and bathroom equipment, smart bathroom equipment, smart robot vacuum cleaners, smart window cleaning robots, smart mopping robots, smart air purifiers, smart steam ovens, smart microwave ovens, smart water heaters, smart air purifiers, smart water dispensers, system platforms, etc.

[0029] This embodiment provides a method for determining recommended scenarios, applied to the aforementioned home appliances. Figure 2 This is a flowchart of a method for determining a recommended scenario according to an embodiment of this application. The process includes the following steps:

[0030] Step S202: Obtain first metadata of the home device within a first time period, wherein the first time period is before the current time, and the first metadata is used to characterize the space where the home device is located and the usage data of the home device;

[0031] Step S204: Input the first metadata into the scene recommendation model for processing, and determine the first recommended scene corresponding to the home device based on the first output result of the scene recommendation model;

[0032] Step S206: Send the first recommended scenario to the target object and obtain the first feedback returned by the target object based on the first recommended scenario;

[0033] Step S208: Determine a second recommended scenario based on the first feedback and the first recommended scenario, wherein the second recommended scenario is applied to the home device.

[0034] In this embodiment, first metadata of the home appliance within a first time period is obtained, wherein the first time period is prior to the current time, and the first metadata is used to characterize the space where the home appliance is located and the usage data of the home appliance; the first metadata is input into a scene recommendation model for processing, and a first recommended scene corresponding to the home appliance is determined based on the first output result of the scene recommendation model; the first recommended scene is sent to a target object, and a first feedback returned by the target object based on the first recommended scene is obtained; a second recommended scene is determined based on the first feedback and the first recommended scene, wherein the second recommended scene is applied to the home appliance. By adopting the above technical solution, the problem in related technologies that the automated control of smart home appliances relies on manual parameter settings by the user, resulting in a low level of intelligence in the home appliances is solved.

[0035] In an optional embodiment, before determining the first recommended scenario corresponding to the home appliance based on the first output result of the scenario recommendation model, the method further includes: extracting features from the first metadata using a convolutional neural network to obtain multiple data features; determining multiple association relationships between the multiple data features using a recurrent neural network, and inputting the multiple association relationships into the scenario recommendation model for processing to obtain multiple usage habits of the target object for the home appliance and the confidence levels of the multiple usage habits, wherein the multiple association relationships are used to indicate the relationship between any N data features among the multiple data features, the multiple usage habits and the multiple association relationships correspond one-to-one, and N is a positive integer; and determining the first usage habit with the highest confidence level among the multiple usage habits as the first output result.

[0036] Convolutional Neural Networks (CNNs) are typically used to process data with spatial structure or local correlations, such as images and signals. In this embodiment, a CNN is used to process primary metadata, which may include multi-dimensional information such as time, device type, device location, ambient temperature, humidity, and light intensity. The CNN extracts key features from this data through its internal convolutional and pooling layers, such as device usage frequency, preference settings (temperature, volume, etc.), and usage time patterns (equivalent to multiple data features). Recurrent Neural Networks (RNNs) excel at processing sequential data and can capture trends in data over time. In this embodiment, an RNN can analyze the evolution of multiple data features extracted by the CNN over time to determine multiple relationships between them. For example, an RNN can discover the habitual pattern of "Mr. Zhang always turns on the air conditioner first, then the TV, and sets the lights to reading mode on weekday evenings," or the relationship of "Ms. Li habitually turns on the coffee machine in the kitchen and starts the music system in the living room on weekend mornings." Multiple relationships are input into the scenario recommendation model for processing, resulting in multiple usage habits of the target object for home appliances and the confidence levels of these habits. For example, {Scenario type: master bedroom air conditioner automation, trigger condition: winter weekdays 19:00-23:00, operation command: turn on at 21:00, temperature 26℃; turn off at 7:00 the next day, confidence level: 92%}. The habit with the highest confidence level, i.e., the first usage habit, is selected and determined as the first output result.

[0037] In another optional embodiment, if a first confidence level exists among the multiple confidence levels, the first usage habit corresponding to the first confidence level is determined as the first output result, wherein the first confidence level is greater than a preset value. Optionally, a first usage habit with a confidence level higher than 80% is determined from the multiple usage habits and is determined as the first output result.

[0038] In an optional embodiment, determining the first recommended scenario corresponding to the home appliance based on the first output result of the scenario recommendation model includes: determining multiple recommended scenarios matching the first output result in the home appliance function rule base, wherein the multiple recommended scenarios include the first recommended scenario, and the home appliance function rule base contains multiple scenarios corresponding to the home appliance; determining the scores of the multiple recommended scenarios, and determining the recommended scenario with the highest score among the multiple recommended scenarios as the first recommended scenario, wherein the multiple scores are used to characterize the usage frequency of the multiple recommended scenarios.

[0039] The home appliance function rule base is a database storing various possible scenarios and their corresponding appliance operation rules. This rule base contains multiple scenario definitions for various home appliances (such as air conditioners, televisions, and lighting systems). Each scenario details how the devices collaborate to achieve a specific goal or satisfy a particular usage situation, such as "sleep mode," "away mode," and "movie night." These scenarios are based on common usage situations and device function presets, while also allowing users to customize or modify them to better suit individual needs. The initial output result is matched with scenarios in the home appliance function rule base to find scenarios that match the user's current habits. For example, if the initial output result indicates that the user habitually adjusts the air conditioner to a lower temperature while watching television at night, then scenarios containing the operations of "watching television" and "adjusting the air conditioner temperature" are searched. After finding multiple recommended scenarios that match the initial output result, these scenarios are scored. The scoring mechanism assesses the popularity and practicality of a scenario based on its usage frequency, i.e., the number of times the scenario has been activated in historical data. Scenarios with higher usage frequency generally mean they better meet the user's needs and therefore receive higher scores. Then, the scenario with the highest rating among these matching scenarios is selected as the primary recommended scenario. The aim is to recommend scenarios that both align with current user habits and have been validated as the most popular and practical. For example, if the "Movie Night" scenario has been activated 15 times in the past month, while the "Reading Time" scenario has only been activated 3 times, then "Movie Night" will be selected as the primary recommended scenario because it has a higher usage frequency rating, indicating that users are more likely to choose this scenario in similar situations.

[0040] In an optional embodiment, determining the second recommended scenario based on the first feedback and the first recommended scenario includes: if the first feedback indicates acceptance of the first recommended scenario, determining the first recommended scenario as the second recommended scenario; if the first feedback indicates modification of the first recommended scenario, modifying the first recommended scenario based on the first feedback, and determining the modified first recommended scenario as the second recommended scenario; if the first feedback indicates rejection of the first recommended scenario, determining a third recommended scenario from among the multiple recommended scenarios as the second recommended scenario, wherein the third recommended scenario is the recommended scenario with the highest rating among the other recommended scenarios, and the other recommended scenarios are scenarios other than the first recommended scenario among the multiple recommended scenarios.

[0041] When a user accepts (the first feedback) the first recommended scenario (e.g., a "movie night" scenario, including dimming lights and turning on the sound system) through an interactive interface (such as a mobile app or voice assistant), the first recommended scenario (the scenario confirmed by the user) is designated as the second recommended scenario for application to the home appliances. If the user requests modifications to the first recommended scenario (e.g., wanting to lower the sound volume), the current first recommended scenario is adjusted accordingly based on the user's first feedback (specific modification instructions), including but not limited to changing device operating parameters or adding / removing devices from the scenario. The modified first recommended scenario is designated as the second recommended scenario. When the user explicitly rejects the first recommended scenario, the third recommended scenario, which has the highest rating among the previously matched recommended scenarios (excluding the first recommended scenario), is selected as the second recommended scenario. The rating mechanism plays a role here again, ensuring that even if the user rejects the original recommended scenario, another more likely popular scenario can be recommended based on historical usage frequency and user preferences.

[0042] In an optional embodiment, after determining the second recommended scenario based on the first feedback and the first recommended scenario, the method further includes: obtaining second metadata of the home device within a second time period, wherein the second time period is the time period during which the second recommended scenario is applied to the home device, and the second metadata is of the same type as the first metadata; if a first parameter in the second metadata changes relative to a second parameter in the first metadata, the second metadata is input into the scenario recommendation model for processing, and a fourth recommended scenario of the home device is determined based on the second output result of the scenario recommendation model, wherein the first parameter and the second parameter are of the same type.

[0043] The second time period refers to the time during which the second recommended scenario (i.e., the recommended scenario last confirmed or modified by the user) is applied to home appliances. During this time period, operational data of the home appliances is continuously collected; this data is called second metadata. The second metadata is consistent in type with the previously collected first metadata, including information such as device operation time, device status, environmental parameters (e.g., temperature, humidity), and user operation patterns. Comparing the second metadata with the first metadata, the first and second parameters refer to the values ​​of the same type of parameter in different time periods (the first time period and the second time period). For example, the first parameter could be the set temperature of an air conditioner in the second time period, while the second parameter is the set temperature in the first time period. If the first parameter changes during the second time period (e.g., the set temperature is adjusted from 26℃ to 24℃), it is considered a signal of user preference or habit adjustment. When such parameter changes are detected, the second metadata is input into the scenario recommendation model for processing. The scenario recommendation model generates a new output result (the second output result) based on the analysis of the second metadata. This result includes the updated user usage habits and recommended actions related to the changed parameters. A new recommended scenario, the fourth recommended scenario, is determined based on the second output result to reflect the user's latest preferences or environmental changes. For example, if the scenario recommendation model predicts that users prefer to watch movies in a cool environment, then the fourth recommended scenario might suggest setting the air conditioning temperature to 24°C when watching movies at night.

[0044] In one optional embodiment, obtaining the first metadata of a home device within a first time period includes: encrypting the fourth metadata of the home device within a third time period to obtain the fifth metadata, wherein the start time of the third time period is before the start time of the first time period, and the end time of the third time period is the current time; storing the fifth metadata on a local server and a cloud server respectively, and determining the encrypted first metadata from the fifth metadata according to the first time period; and decrypting the encrypted first metadata to obtain the first metadata.

[0045] The system collects fourth-generation metadata from the time home devices are first used until the current time. To protect user data security, this fourth-generation metadata is encrypted to generate encrypted fifth-generation metadata. Encryption technologies can include AES-256, RSA, etc. The fifth-generation metadata is stored in two locations: a local server (e.g., the built-in encrypted storage unit of a home smart gateway) and a cloud server (e.g., a secure data center provided by the brand). This dual-storage strategy aims to balance data security and computing resource utilization. Local server storage allows for fast access, reduces reliance on the internet, and is more suitable for real-time scenario recommendations. Cloud server storage leverages its powerful computing capabilities for large-scale data analysis and AI model training, improving the accuracy of recommended scenarios. When making scenario recommendations based on user habits, the encrypted data for the first time period (i.e., the encrypted first-generation metadata) is determined from the fifth-generation metadata. Then, using a pre-agreed decryption algorithm and key, this encrypted data is decrypted to recover the original first-generation metadata. The decrypted data is then used for model analysis to generate personalized scenario recommendations.

[0046] In an optional embodiment, after obtaining the first metadata of the home device within a first time period, the method further includes: obtaining the target object as the working mode determined by the scene recommendation model, wherein the working mode includes a local mode and a cloud mode; when the working mode is the local mode, determining the category of the scene recommendation model as a local scene recommendation model; when the working mode is the cloud mode, determining the category of the scene recommendation model as a cloud scene recommendation model.

[0047] The system retrieves the user's chosen operating mode for the scene recommendation model, which can be either local or cloud-based. Users select this mode via an interactive interface (such as a mobile app), choosing whether the data processing environment is a local device or a cloud server. If the user selects local mode, the scene recommendation model is categorized as a local scene recommendation model. This means that user operation data (such as device usage time and preference settings) will be processed and analyzed locally, typically within the computing unit built into a smart home gateway or smart appliance. The local scene recommendation model can be deployed using a lightweight AI model (such as a trimmed version of the Llama 2 model) to adapt to the hardware limitations of local devices. This approach ensures data security and privacy while reducing reliance on the internet, facilitating service delivery offline or in situations with unstable networks. Conversely, if the target user selects cloud mode, the scene recommendation model is categorized as a cloud-based scene recommendation model. Cloud mode utilizes large-scale pre-trained AI models (such as an optimized version of the BERT model based on the Transformer architecture), deployed on a cloud GPU server cluster, providing powerful multi-dimensional data correlation analysis capabilities.

[0048] To better understand the process of determining the above-mentioned recommended scenarios, the implementation flow of the above-mentioned recommended scenarios will be described below in conjunction with optional embodiments, but this is not intended to limit the technical solutions of the embodiments of this application.

[0049] Figure 3 This is a first schematic diagram of a method for determining a recommended scenario according to an optional embodiment of the present invention, as shown below. Figure 3 As shown, it specifically includes the following:

[0050] The data acquisition module is responsible for capturing real-time operational data from smart home appliances, such as power-on time, temperature settings, and user identity. This data is then transmitted to the data storage and processing mode selection module. Specifically, the data acquisition module is equipped with a series of communication interfaces, including Universal Serial Bus (USB), Bluetooth 5.0, Wi-Fi 6, and ZigBee 3.0. These interfaces are used to collect operational data from smart home appliances, including operational behavior, time, environmental conditions, user identity, and appliance location. Through the GPIO interface of the appliance control chip, the module sensitively captures the start and stop signals of the device, such as the high level when the air conditioner is turned on and the low level when it is turned off. The addition of the RTC module enables precise recording of operation time, ensuring that the timestamp is accurate to the second, which is crucial for analyzing user habits. The linkage between the temperature sensor and the calendar algorithm can automatically identify seasonal changes and provide more suitable usage scenario analysis for environmentally sensitive devices such as air conditioners. The application of user account login information or fingerprint recognition technology allows the system to customize personalized scenario recommendations based on different user identities. The recording of appliance location data further enriches the scenario settings, making the recommendations more granular to specific areas in the home. The collected raw data is processed and uniformly converted into JSON format. For example, a record may be described as: {"Appliance type":"wall-mounted air conditioner","Operation time":"2025-12-07 21:00:00","Operation season":"Winter","Operator":"User A","Appliance location":"Master bedroom","Operation behavior":"Turns on, temperature 26℃"}.

[0051] In the data storage and processing mode selection module, users can choose between local mode and cloud mode based on their preferences and privacy and security needs. This means storing data in encrypted form on a local storage device or uploading it to a cloud server for processing. Specifically, users can flexibly switch data processing modes through the mode selection interface of the mobile terminal application, choosing either "cloud mode" or "local mode" to meet their personal data security preferences. The user's selection is sent to the smart gateway's control module in the form of encrypted commands. High and low level signals are used to instantly switch the data path, ensuring the secure execution of user commands. The collected home device operation data is encrypted using the AES-256 symmetric encryption algorithm. The encryption key is generated by combining the user's account password with unique identification information such as the smart gateway's MAC address, ensuring that each user's data is independent and highly encrypted. The encrypted data is stored with a timestamp, effectively preventing data replay attacks and ensuring the security of data transmission and storage. In cloud mode, encrypted data is uploaded to the cloud via the TLS 1.3 security protocol and stored in a physically isolated distributed database. In local mode, data is stored in the smart gateway's built-in SQLCipher encrypted database, accessible only within the local area network, avoiding the risk of exposure to the internet. This embodiment also supports seamless switching between modes. When a user switches from cloud mode to local mode, or vice versa, the data is synchronized and encrypted through the TLS 1.3 protocol to ensure data consistency. At the same time, the system will automatically clean up redundant data in the original mode to prevent data residue and protect user privacy and security.

[0052] Next, the AI ​​analysis module, based on the user-selected data processing mode (local or cloud), invokes a lightweight local AI model or a large-scale cloud-based AI model to perform in-depth analysis of the encrypted data, uncovering user habits and preferences. The AI ​​analysis results are then transmitted to the scene recommendation module. Specifically, in cloud mode, the system deploys an optimized version of the BERT model based on the Transformer architecture. This large-scale pre-trained model leverages the powerful computing capabilities of cloud GPU server clusters to process and analyze high-dimensional user data, discovering complex relationships hidden within massive amounts of data and improving the intelligence level of scene recommendations. In local mode, the system uses a trimmed and optimized version of the Llama 2 open-source model. Through model quantization technology, computationally intensive operations are transformed into low-complexity operations, enabling it to run efficiently on the ARM architecture processor of the smart gateway. Even offline or with unstable networks, it can provide real-time scene analysis and recommendation services. The AI ​​analysis module's workflow consists of three stages: feature extraction, pattern discovery, and model update. First, a convolutional neural network (CNN) is used to extract feature vectors from the collected multi-dimensional data, converting information such as operation time, season, and operator into a machine-understandable form. Then, a recurrent neural network (RNN) is used to perform deep analysis on these feature vectors, uncovering the inherent relationships and patterns between the data. For example, it identifies the preference behavior of specific users turning on the air conditioner on winter weekday evenings. Based on these relationships, user habits (equivalent to AI analysis results) are determined, including four key dimensions: "scenario type," "trigger condition," "operation command," and "confidence level." For example, in the "Master Bedroom Air Conditioner Automation" scenario, the trigger condition is "winter weekday 19:00-23:00," the operation command is "turn on at 21:00, set the temperature to 26℃, turn off at 7:00 the next day," and the confidence level is as high as 92%.

[0053] The scenario recommendation module generates a series of personalized scenario recommendations (first recommended scenario) based on AI analysis results. Specifically, it matches the AI ​​analysis results with multiple recommended scenarios in the built-in home appliance function rule library. The home appliance function rule library stores executable operation information for various home appliances, such as temperature adjustment and mode switching for air conditioners, and heating time settings for water heaters. Based on usage habits, matching operation instructions are selected from the rule library to generate a series of automated scenarios (first recommended scenario) that both conform to user habits and are suitable for home appliance functions. Each recommended scenario includes four parts: "Scenario Name," "Specific Operation," "Expected Effect," and "Modification Entry." For example: "[Weekday Air Conditioner Automation] Specific Operation: Automatically turn on the master bedroom air conditioner at 20:50 every night, set the temperature to 26℃, and automatically turn it off at 7:00 the next day; Expected Effect: Preheating in advance to meet bedtime needs, automatic shutdown saves energy; Modification Entry: Click to adjust the start time / temperature," making it easy for users to understand and modify. These recommendations are presented to users through interactive feedback units (such as mobile apps and voice assistants). After receiving a recommendation, users can confirm, modify, or reject it. These feedback instructions are sent to smart home appliances via the device control unit, triggering the devices to execute the corresponding scenario. Simultaneously, the instruction information is fed back to the AI ​​analysis module, serving as a crucial basis for model optimization and promoting the continuous improvement of the accuracy of automated scenario recommendations. Optionally, if a user confirms a certain scenario three times consecutively, that scenario is set as a "frequently used scenario" and prioritized for recommendation.

[0054] It should be clarified that this embodiment supports three interaction methods: mobile terminal APP, home appliance control panel, and smart speaker. The APP, as the core interaction terminal, has full functionality; the home appliance control panel integrates a touch screen and supports simple operations such as scene confirmation / cancellation; the smart speaker receives user commands through voice recognition to achieve voice control.

[0055] The corresponding interaction flow is as follows Figure 4 As shown, it specifically includes the following:

[0056] Data collection initialization: Users register and log in to the mobile APP, add the master bedroom wall-mounted air conditioner and the bathroom storage-type electric water heater, and enter the device type and location information; select "local mode" through the APP, and the system automatically deploys the trimmed Llama 2 lightweight AI model in the smart gateway. The smart gateway and home appliances are paired through the ZigBee protocol, and the initialization is complete.

[0057] User operation data acquisition: The data acquisition module captures user operations in real time through the control chips of the air conditioner and water heater. For example, it records that "User A turns on the air conditioner at 21:00 on a winter weekday, with the temperature set to 26℃; turns it off at 7:00 the next day; turns on the water heater at 9:00 on weekends, heats for 30 minutes, and then turns it off", and converts the data into JSON format and transmits it to the mode selection module.

[0058] Data storage and transmission: The mode selection module uses the AES-256 algorithm to encrypt the data. The key is "user account password + gateway MAC address". The encrypted data is stored in the SQLCipher local database of the smart gateway and is not transmitted to the cloud.

[0059] User habit AI analysis: The localized lightweight AI model reads and decrypts encrypted data from the local database, extracts feature vectors such as "winter, weekday, 21:00" through CNN, and mines the pattern of "air conditioning use time from 21:00 to 7:00 the next day on winter weekdays, with a preferred temperature of 26℃" through RNN, generating user habits with a confidence level of 93%.

[0060] Automated Scene Recommendation Generation: The scene recommendation module matches user habits with multiple recommended scenes in the home appliance function rule library to generate the first recommended scene: "[Weekday Air Conditioner Automation] Automatically turn on the master bedroom air conditioner at 20:50 every night, set the temperature to 26℃, and automatically turn it off at 7:00 the next day; [Weekend Water Heater Automation] Automatically turn on the water heater at 8:50 on Saturdays and Sundays, and turn it off at 9:20", and pushes it to the user through the APP.

[0061] User Interaction and Scene Execution: The user confirms the air conditioning scene and changes the water heater's start time to 9:00 AM. The app transmits the modified command to the smart gateway. The gateway sends control commands to the air conditioner and water heater via the MQTT protocol. The control commands use the format "Device ID + Operation Code + Parameter" to ensure accurate command delivery. For example, "Device ID: AC-001, Operation Code: Power On, Parameter: Temperature 26℃". After the device executes the command, it displays a "Ready" status, and the app displays the feedback information in real time.

[0062] Data Update and Model Optimization: A week later, the system collected new data on users occasionally using the water heater at 8:50 am on weekend mornings. The AI ​​model used this data as a training sample and updated the model parameters through the gradient descent algorithm, optimizing the water heater's morning start time to 8:50 am. New recommendation scenarios were generated and pushed, achieving dynamic optimization of the model and recommendations.

[0063] Obviously, the embodiments described above are only some embodiments of this application, and not all embodiments. To better understand the method for determining the recommended scenarios described above, the following description, in conjunction with embodiments, illustrates the process, but is not intended to limit the technical solutions of the embodiments of this application. Specifically:

[0064] Figure 5 This is a third schematic diagram of a method for determining a recommended scenario according to an optional embodiment of the present invention, as shown below. Figure 5 As shown, it specifically includes the following:

[0065] In smart home systems, the smart gateway plays a central hub role, integrating functions such as data relay, local data storage, and lightweight AI analysis. It coordinates and manages data communication between various hardware devices throughout the system, ensuring the orderly flow and processing of information. The smart gateway achieves effective connectivity with various devices through diverse connection methods: smart appliances such as air conditioners and water heaters connect to the gateway using ZigBee or Bluetooth 5.0 wireless technology, adapting to the reality of dispersed appliance layouts and ensuring smooth communication between devices and the gateway; users' mobile phones connect wirelessly via Wi-Fi to establish a remote interactive channel with the gateway, allowing users to control home devices or receive personalized scene recommendations anytime, anywhere; cloud servers connect to the gateway via encrypted network cables, ensuring high data transmission security; and the built-in local storage chip and lightweight AI chip connect directly to the gateway's core module, responsible for performing local data storage and preliminary analysis.

[0066] Smart home appliances act as execution terminals, running corresponding scene modes based on received instructions; the user's mobile phone serves as an interaction terminal, capable of issuing control commands and displaying recommended information; local storage chips and cloud servers together constitute the storage terminal, responsible for local data storage and cloud data backup, respectively; a lightweight AI chip acts as an analysis terminal, performing preliminary data analysis locally to assist in rapid response. Through the overall coordination and scheduling of the home smart gateway, these terminals form a complementary hardware system, capable of rapidly processing data locally and further optimizing analysis results using the powerful computing capabilities of the cloud, thereby providing users with a highly personalized and secure smart home experience.

[0067] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, operating terminal, or network device, etc.) to execute the methods of the various embodiments of this application.

[0068] Figure 6 This is a structural block diagram of a recommended scenario determination device according to an embodiment of this application; as shown below. Figure 6 As shown, it includes:

[0069] The first acquisition module 62 is used to acquire first metadata of the home device within a first time period, wherein the first time period is before the current time, and the first metadata is used to characterize the space where the home device is located and the usage data of the home device.

[0070] The first determining module 64 is used to input the first metadata into the scene recommendation model for processing, and determine the first recommended scene corresponding to the home device based on the first output result of the scene recommendation model.

[0071] The second acquisition module 66 is used to send the first recommended scenario to the target object and acquire the first feedback returned by the target object based on the first recommended scenario;

[0072] The second determining module 68 is used to determine a second recommended scenario based on the first feedback and the first recommended scenario, wherein the second recommended scenario is applied to the home device.

[0073] In this embodiment, first metadata of the home appliance within a first time period is obtained, wherein the first time period is prior to the current time, and the first metadata is used to characterize the space where the home appliance is located and the usage data of the home appliance; the first metadata is input into a scene recommendation model for processing, and a first recommended scene corresponding to the home appliance is determined based on the first output result of the scene recommendation model; the first recommended scene is sent to a target object, and a first feedback returned by the target object based on the first recommended scene is obtained; a second recommended scene is determined based on the first feedback and the first recommended scene, wherein the second recommended scene is applied to the home appliance. By adopting the above technical solution, the problem in related technologies that the automated control of smart home appliances relies on manual parameter settings by the user, resulting in a low level of intelligence in the home appliances is solved.

[0074] In an optional embodiment, the first determining module 64 is further configured to extract features from the first metadata using a convolutional neural network to obtain multiple data features; determine multiple associations between the multiple data features using a recurrent neural network, and input the multiple associations into the scene recommendation model for processing to obtain multiple usage habits of the target object for the home appliances and the confidence levels of the multiple usage habits, wherein the multiple associations are used to indicate the relationship between any N data features among the multiple data features, the multiple usage habits and the multiple associations correspond one-to-one, and N is a positive integer; and determine the first usage habit with the highest confidence level among the multiple usage habits as the first output result.

[0075] In an optional embodiment, the first determining module 64 is further configured to determine multiple recommended scenarios matching the first output result in the home appliance function rule base, wherein the multiple recommended scenarios include the first recommended scenario, and the home appliance function rule base contains multiple scenarios corresponding to the home appliance; determine the scores of the multiple recommended scenarios, and determine the recommended scenario with the highest score among the multiple recommended scenarios as the first recommended scenario, wherein the multiple scores are used to characterize the usage frequency of the multiple recommended scenarios.

[0076] In an optional embodiment, the second determining module 68 is further configured to: determine the first recommended scenario as the second recommended scenario if the first feedback indicates acceptance of the first recommended scenario; modify the first recommended scenario according to the first feedback if the first feedback indicates modification of the first recommended scenario, and determine the modified first recommended scenario as the second recommended scenario if the first feedback indicates rejection of the first recommended scenario; and determine the third recommended scenario among the multiple recommended scenarios as the second recommended scenario if the first feedback indicates rejection of the first recommended scenario, wherein the third recommended scenario is the recommended scenario with the highest rating among the other recommended scenarios, and the other recommended scenarios are scenarios other than the first recommended scenario among the multiple recommended scenarios.

[0077] In an optional embodiment, the second determining module 68 is further configured to obtain second metadata of the home device within a second time period, wherein the second time period is the time period during which the second recommended scenario is applied to the home device, and the second metadata is of the same type as the first metadata; if a first parameter in the second metadata changes relative to the second parameter in the first metadata, the second metadata is input into the scenario recommendation model for processing, and a fourth recommended scenario for the home device is determined based on the second output result of the scenario recommendation model, wherein the first parameter and the second parameter are of the same type.

[0078] In an optional embodiment, the first acquisition module 62 is further configured to encrypt the fourth metadata of the home device within a third time period to obtain the fifth metadata, wherein the start time of the third time period is before the start time of the first time period, and the end time of the third time period is the current time; store the fifth metadata in a local server and a cloud server respectively, and determine the encrypted first metadata from the fifth metadata according to the first time period; and decrypt the encrypted first metadata to obtain the first metadata.

[0079] In an optional embodiment, the first acquisition module 62 is further configured to acquire the target object as the working mode determined by the scene recommendation model, wherein the working mode includes a local mode and a cloud mode; when the working mode is the local mode, the category of the scene recommendation model is determined to be a local scene recommendation model; when the working mode is the cloud mode, the category of the scene recommendation model is determined to be a cloud scene recommendation model.

[0080] Embodiments of this application also provide a storage medium including a stored program, wherein the program executes any of the methods described above when it is run.

[0081] Optionally, in this embodiment, the storage medium may be configured to store program code for performing the following steps:

[0082] S1, Obtain first metadata of home devices within a first time period, wherein the first time period is before the current time, and the first metadata is used to characterize the space where the home devices are located and the usage data of the home devices;

[0083] S2, input the first metadata into the scene recommendation model for processing, and determine the first recommended scene corresponding to the home device based on the first output result of the scene recommendation model;

[0084] S3, send the first recommended scenario to the target object, and obtain the first feedback returned by the target object based on the first recommended scenario;

[0085] S4, determine a second recommended scenario based on the first feedback and the first recommended scenario, wherein the second recommended scenario is applied to the home device.

[0086] Embodiments of this application also provide an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0087] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0088] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0089] S1, Obtain first metadata of home devices within a first time period, wherein the first time period is before the current time, and the first metadata is used to characterize the space where the home devices are located and the usage data of the home devices;

[0090] S2, input the first metadata into the scene recommendation model for processing, and determine the first recommended scene corresponding to the home device based on the first output result of the scene recommendation model;

[0091] S3, send the first recommended scenario to the target object, and obtain the first feedback returned by the target object based on the first recommended scenario;

[0092] S4, determine a second recommended scenario based on the first feedback and the first recommended scenario, wherein the second recommended scenario is applied to the home device.

[0093] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0094] Embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0095] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0096] The embodiments described herein also provide a computer program that includes computer instructions stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in any of the above method embodiments.

[0097] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0098] Obviously, those skilled in the art should understand that the modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular combination of hardware and software.

[0099] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for determining a recommendation scenario, characterized in that, include: Obtain first metadata of home devices within a first time period, wherein the first time period is prior to the current time, and the first metadata is used to characterize the space where the home devices are located and the usage data of the home devices; The first metadata is input into the scene recommendation model for processing, and the first recommended scene corresponding to the home device is determined based on the first output result of the scene recommendation model. Send the first recommended scenario to the target object and obtain the first feedback returned by the target object based on the first recommended scenario; A second recommended scenario is determined based on the first feedback and the first recommended scenario, wherein the second recommended scenario is applied to the home appliance.

2. The method for determining the recommended scenario according to claim 1, characterized in that, Before determining the first recommended scenario corresponding to the home appliance based on the first output result of the scenario recommendation model, the method further includes: Multiple data features are obtained by extracting features from the first metadata using a convolutional neural network; Multiple relationships between the multiple data features are determined by a recurrent neural network, and the multiple relationships are input into the scene recommendation model for processing to obtain multiple usage habits of the target object for the home appliances and the confidence level of the multiple usage habits. The multiple relationships are used to indicate the relationship between any N data features among the multiple data features. The multiple usage habits and the multiple relationships correspond one-to-one, and N is a positive integer. The first usage habit with the highest confidence among the multiple usage habits is determined as the first output result.

3. The method for determining the recommended scenario according to claim 1, characterized in that, The first recommended scenario corresponding to the home appliance is determined based on the first output result of the scenario recommendation model, including: In the home appliance function rule base, multiple recommended scenarios that match the first output result are determined, wherein the multiple recommended scenarios include the first recommended scenario, and the home appliance function rule base contains multiple scenarios corresponding to the home appliances; The ratings of the plurality of recommended scenarios are determined, and the recommended scenario with the highest rating among the plurality of recommended scenarios is determined as the first recommended scenario, wherein the plurality of ratings are used to characterize the frequency of use of the plurality of recommended scenarios.

4. The method for determining the recommended scenario according to claim 3, characterized in that, Based on the first feedback and the first recommended scenario, a second recommended scenario is determined, including: If the first feedback indicates acceptance of the first recommended scenario, the first recommended scenario will be determined as the second recommended scenario; If the first feedback indicates that the first recommended scenario should be modified, the first recommended scenario is modified according to the first feedback, and the modified first recommended scenario is determined as the second recommended scenario. If the first feedback indicates rejection of the first recommended scenario, the third recommended scenario among the multiple recommended scenarios is determined as the second recommended scenario, wherein the third recommended scenario is the recommended scenario with the highest score among the other recommended scenarios, and the other recommended scenarios are scenarios other than the first recommended scenario among the multiple recommended scenarios.

5. The method for determining the recommended scenario according to claim 1, characterized in that, After determining the second recommended scenario based on the first feedback and the first recommended scenario, the method further includes: Obtain the second metadata of the home device within a second time period, wherein the second time period is the time period during which the second recommended scenario is applied to the home device, and the second metadata is of the same type as the first metadata; If the first parameter in the second metadata changes relative to the second parameter in the first metadata, the second metadata is input into the scene recommendation model for processing. The fourth recommended scene for the home device is determined based on the second output result of the scene recommendation model, wherein the first parameter and the second parameter are of the same type.

6. The method for determining the recommended scenario according to claim 1, characterized in that, Obtain the first metadata of home devices within the first time period, including: The fourth metadata of the home device in the third time period is encrypted to obtain the fifth metadata, wherein the start time of the third time period is before the start time of the first time period, and the end time of the third time period is the current time. The fifth metadata is stored on a local server and a cloud server respectively, and the encrypted first metadata is determined from the fifth metadata according to the first time period; The encrypted first metadata is decrypted to obtain the first metadata.

7. The method for determining the recommended scenario according to claim 1, characterized in that, After acquiring the first metadata of the home devices within the first time period, the method further includes: The target object is obtained as the working mode determined by the scene recommendation model, wherein the working mode includes local mode and cloud mode; When the working mode is the local mode, the category of the scene recommendation model is determined to be the local scene recommendation model; When the working mode is the cloud mode, the category of the scene recommendation model is determined to be the cloud scene recommendation model.

8. A device for determining a recommended scenario, characterized in that, include: The first acquisition module is used to acquire first metadata of the home device within a first time period, wherein the first time period is before the current time, and the first metadata is used to characterize the space where the home device is located and the usage data of the home device. The first determining module is used to input the first metadata into the scene recommendation model for processing, and determine the first recommended scene corresponding to the home device based on the first output result of the scene recommendation model. The second acquisition module is used to send the first recommended scenario to the target object and acquire the first feedback returned by the target object based on the first recommended scenario; The second determining module is used to determine a second recommended scenario based on the first feedback and the first recommended scenario, wherein the second recommended scenario is applied to the home device.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method described in any one of claims 1 to 7.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method described in any one of claims 1 to 7 through the computer program.

Images