Smart home management method and apparatus

Abstract

The application relates to a smart home management method and device, and relates to the technical field of smart home management. The method comprises the following steps: after a first control instruction indicating change working state information is sent to a first actuator, actuator state information of the first actuator and first actual environment information of a linkage device group to which the first actuator belongs are acquired; in the case that the actuator state information of the first actuator and the first actual environment information are acquired, environment association information of the first actuator is acquired; based on the environment association information of the first actuator and the actuator state information of the first actuator, current theoretical environment information of the first actuator is determined; the first actual environment information and the current theoretical environment information are matched to obtain a matching result; and based on the matching result, a device abnormal state of the linkage device group is determined. Through the above method, automatic abnormal detection and diagnosis of smart home devices can be realized.

Description

Technical Field

[0001] This application relates to the field of smart home management technology, and in particular to a smart home management method and device. Background Technology

[0002] Smart home refers to the automation, intelligence, and networking of home life through the integration of various smart technologies and devices.

[0003] However, the detection of anomalies in smart homes relies heavily on manual inspection. With the rapid increase in the number and types of smart home devices, the efficiency of anomaly inspection in smart home management cannot meet the actual needs of intelligent operation and maintenance, resulting in poor management performance. Summary of the Invention

[0004] This application provides a smart home management method and device, which can realize automated anomaly detection and diagnosis of smart home devices, improve the reliability and intelligent operation and maintenance level of smart home system, thereby improving the overall management effect and ensuring user experience. The technical solution is as follows.

[0005] On the one hand, a smart home management method is provided, the method comprising: After sending a first control command to the first actuator to indicate a change in working status information, the actuator status information of the first actuator and the first actual environmental information of the linkage equipment group to which the first actuator belongs are obtained; the linkage equipment group includes at least one actuator and at least one sensor, and the environmental information collected by the sensor in the linkage equipment group is related to the change in the working status of the actuator; Having obtained the actuator state information and the first actual environment information of the first actuator, the environmental association information of the first actuator is obtained; the environmental association information is used to characterize the theoretical environment information corresponding to each actuator state information of the actuator; Based on the environmental association information and the actuator status information of the first actuator, the current theoretical environmental information of the first actuator is determined; The first actual environment information is matched with the current theoretical environment information to obtain a matching result; Based on the matching results, the abnormal state of the equipment in the linked equipment group is determined.

[0006] On the other hand, a smart home management device is provided, the device comprising: The first information acquisition module is used to acquire the actuator status information of the first actuator and the first actual environmental information of the linkage equipment group to which the first actuator belongs after sending a first control command to the first actuator indicating a change in working status information; the linkage equipment group includes at least one actuator and at least one sensor, and the environmental information collected by the sensor in the linkage equipment group is related to the change in the working status of the actuator. The second information acquisition module is used to acquire environmental association information of the first actuator when the actuator state information and the first actual environment information of the first actuator are acquired; the environmental association information is used to characterize the theoretical environment information corresponding to each actuator state information of the actuator; The environmental information determination module is used to determine the current theoretical environmental information of the first actuator based on the environmental association information of the first actuator and the actuator state information of the first actuator; The matching module is used to match the first actual environment information with the current theoretical environment information to obtain a matching result; The status determination module is used to determine the abnormal status of the equipment in the linked equipment group based on the matching result.

[0007] In one possible implementation, the state determination module is used for, If the matching result indicates that the current theoretical environment information matches the first actual environment information, it is determined that no equipment abnormality has occurred in the linked equipment group; If the matching result indicates that the current theoretical environment information does not match the first actual environment information, a fault judgment is performed on the linkage equipment group to determine the abnormal state of the equipment in the linkage equipment group.

[0008] In one possible implementation, the state determination module includes: The instruction issuing submodule is used to send a second control instruction indicating a change in working status information to the second actuator in the linkage equipment group; The information acquisition submodule is used to acquire the second actual environmental information of the linked equipment group; The fault determination submodule is used to determine that the first actuator has a device fault when the first actual environmental information is inconsistent with the second actual environmental information; and to determine that the sensor in the linkage equipment group has a device fault when the first actual environmental information is consistent with the second actual environmental information.

[0009] In one possible implementation, the device further includes: The strategy execution module is used to execute a device adjustment strategy corresponding to the first actuator when it is determined that the first actuator has a device fault. The device adjustment strategy includes at least one of the following: adjusting the actuator status information of at least one third actuator, or enabling at least one fourth actuator, the fourth actuator having the same function as the first actuator. The device adjustment strategy is generated based on at least one of the following: the function type of the first actuator, the first actual environmental information, and historical adjustment strategies, wherein the device adjustment strategy is used to match the theoretical environmental information with the first actual environmental information.

[0010] In one possible implementation, the state determination module is used for, If the actuator status information of the first actuator is not obtained, it is determined that the communication of the first actuator is abnormal; Without obtaining the first actual environmental information, it is determined that the sensor communication in the linkage equipment group is abnormal.

[0011] In one possible implementation, the device further includes: The operation information generation module is used to generate recommended operation information corresponding to the first actual environment information through a recommendation model when the first actual environment information is obtained. The recommended operation information is used to assist users in smart home management. The information push module is used to push the recommended operation information to the user terminal; The recommendation model is trained based on a set of user behavior samples, which includes historical environment information and corresponding user historical operation information.

[0012] In one possible implementation, the recommendation model includes a first sub-recommendation model and a second sub-recommendation model; The operation information generation module includes: The information generation submodule is used to input the first actual environment information into the first sub-recommendation model and the second sub-recommendation model respectively to obtain the first candidate operation information and the second candidate operation information; The operation information determination submodule is used to determine either the first candidate operation information or the second candidate operation information as the recommended operation information when the first candidate operation information and the second candidate operation information are the same; and to determine both the first candidate operation information and the second candidate operation information as the recommended operation information when the first candidate operation information and the second candidate operation information are different.

[0013] In one possible implementation, the device further includes: The sample update module is used to update the user behavior sample set according to the received user selection operation based on the recommended operation information after pushing the recommended operation information to the user terminal, when the first candidate operation information is different from the second candidate operation information. The model optimization and training module is used to optimize and train the target sub-recommendation model based on the updated user behavior sample set. The target sub-recommendation model is the sub-recommendation model corresponding to the candidate operation information that was not selected by the user in the recommendation operation information.

[0014] In one possible implementation, the model optimization training module is used for, When the target sub-recommendation model is a long short-term memory network recommendation model, and when the proportion of candidate operation information recommended by the target sub-recommendation model that is not selected by the user is greater than a preset proportion threshold, the updated user behavior sample set is grouped by k grouping time nodes to generate k+1 different user behavior sample subsets. Each user behavior sample subset contains user behavior samples from the corresponding grouping time node to the latest time node, where k is a positive integer. The target sub-recommendation model is optimized and trained sequentially using the k+1 user behavior sample subsets in different training rounds. The optimized target sub-recommendation model is then applied sequentially in each training round to generate candidate recommendation operation information until the optimization completion condition is met. The optimization completion condition includes: the candidate operation information output by the optimized target sub-recommendation model is consistent with the candidate operation information output by another sub-recommendation model, or the number of user behavior sample subsets used is equal to k+1.

[0015] On the other hand, a computer device is provided, the computer device including a processor and a memory, the memory storing at least one computer program, the at least one computer program being loaded and executed by the processor to implement the above-described smart home management method.

[0016] On the other hand, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, the computer program being loaded and executed by a processor to implement the above-described smart home management method.

[0017] On the other hand, a computer program product is provided, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform to implement the smart home management method provided in the above-described various optional implementations.

[0018] The technical solution provided in this application may include the following beneficial effects: The smart home management method provided in this application constructs a group of interconnected devices consisting of actuators and sensors. After sending control commands to the first actuator, it simultaneously acquires the actuator's status information and the first actual environmental information of the interconnected device group. Based on pre-set environmental association information, it determines the current theoretical environmental information, and then matches the actual environmental information with the theoretical environmental information. Based on the matching result, it automatically determines whether there are any device malfunctions in the interconnected device group. This method can achieve automated anomaly detection and diagnosis of smart home devices, effectively solving the problems of reliance on manual inspection, low efficiency, and difficulty in adapting to the rapid growth in the number and types of devices in traditional smart home management. It significantly improves the reliability and intelligent operation and maintenance level of the smart home system, thereby improving the overall management effect and ensuring user experience.

[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0020] 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.

[0021] Figure 1 A schematic diagram of an exemplary embodiment of the present application is shown. Figure 2 This illustration shows an interaction diagram between a smart home management system and a user terminal provided in an exemplary embodiment of this application; Figure 3 A flowchart of an exemplary embodiment of this application is shown; Figure 4 A block diagram of a smart home management device provided in an exemplary embodiment of this application is shown; Figure 5 A structural block diagram of a computer device illustrated in an exemplary embodiment of this application is shown; Figure 6 A structural block diagram of a computer device illustrated in another exemplary embodiment of this application is shown. Detailed Implementation

[0022] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0023] Smart home refers to a series of solutions that integrate various smart technologies and devices to automate, intelligentize, and network home life, thereby improving living comfort, security, convenience, and energy efficiency. With the development of IoT technology, smart home devices are more closely integrated with other smart devices and services, forming a more complete living ecosystem. Different smart home systems can be applied in different scenarios. For example, a smart home system may include: Intelligent security systems, including access control systems, video surveillance, and intrusion alarms, are used to improve home security.

[0024] Environmental control systems, such as intelligent lighting, temperature control, and curtain control, can automatically adjust according to environmental changes or user settings.

[0025] Entertainment systems, including smart speakers, multimedia centers, and home theaters, provide a high-quality entertainment experience.

[0026] Smart home appliances, such as smart refrigerators, smart washing machines, and smart ovens, can be remotely controlled and managed via the Internet.

[0027] Health management system: Monitors the user's health status, such as blood glucose monitoring, blood pressure monitoring, etc.

[0028] Energy management system: intelligently monitors and controls household electricity consumption to achieve energy conservation and emission reduction.

[0029] The aforementioned smart home systems include various actuators and sensors to achieve smart home control; Figure 1 A schematic diagram of an exemplary embodiment of the smart home management system provided in this application is shown, such as... Figure 1 As shown, the smart home management system can consist of a hardware architecture 110, a software architecture 120, and a network communication protocol 130.

[0030] The hardware architecture 110 includes several smart home devices, generally including a smart gateway, sensors, and actuators. The smart gateway acts as a central controller for discovering smart home devices, managing groups, and forwarding control commands. Sensors are used to collect environmental information, such as temperature, humidity, and light intensity. Actuators are used to execute control commands, such as turning lights on and off or adjusting curtains.

[0031] The software architecture 120 can include various components adapted to smart home management functions, such as device discovery components, group management components, scene configuration components, user interface components, device self-test components, adaptive adjustment components, and operation recommendation components.

[0032] The system includes: a device discovery component for automatically scanning and discovering newly connected smart home devices; a group management component for grouping devices according to user-defined rules or system-recommended schemes; a scene configuration component for supporting user-defined scene modes, such as sleep mode and movie-watching mode, i.e., user-defined rules or schemes; a user interface component for providing a graphical user interface for easy device control and status queries; a device self-test component for analyzing whether smart home devices have malfunctioned based on environmental information and actuator operating status; an adaptive adjustment component for adaptively adjusting the smart home device's operating scheme when a malfunction occurs, ensuring that the smart home device's operation meets user needs; and an operation recommendation component for obtaining recommended operations based on collected information and a recommendation model when a user initiates smart home device control, and then pushing these recommendations.

[0033] The network communication protocol 130 may include standard protocols such as Zigbee, Wi-Fi, and Bluetooth to achieve interconnection between smart home devices; among them, Zigbee is suitable for short-range, low-power device communication, Wi-Fi is suitable for remote control and Internet access, and Bluetooth is suitable for direct connection between mobile devices and smart home devices.

[0034] In the smart home management scenario provided in this application embodiment, the software architecture can be deployed in computer devices, such as computer equipment or terminal devices, while the hardware architecture is implemented as various smart home devices. Figure 2 This illustration shows an interaction diagram between a smart home management system and a user terminal provided in an exemplary embodiment of this application, such as... Figure 2 As shown, the smart gateway in smart home device 210 establishes a communication connection with a computer device to obtain device information of the connected devices and upload it to the computer device; it is also used to receive information sent by the computer device and forward it to the corresponding sensors and actuators; the information sent by the computer device includes: group management information, control commands, etc. Sensors are used to collect environmental information, such as temperature, humidity, and light intensity, and upload this information to computer devices via a smart gateway. They are also used to receive information from computer devices forwarded by the smart gateway and perform corresponding tasks.

[0035] The actuator is used to upload the current working status information to the computer device through the smart gateway, and also to receive information sent by the computer device forwarded by the smart gateway and perform corresponding tasks.

[0036] User terminal 230 establishes a communication connection with computer device 220 via a network communication protocol for uploading operation information to the computer device. The user terminal can also connect to a smart gateway for inputting control commands, directly reducing network latency and facilitating timely user control of smart home devices. Furthermore, the user terminal scans the identification codes of smart home devices and sends them to the computer device; these identification codes can be QR codes, product codes, or other unique identifiers of the smart home devices. It also provides a graphical user interface for convenient device control and status queries; specifically, users input operation information through the terminal, which is then uploaded to the computer device.

[0037] Computer device 220 includes: device discovery component, group management component, scene configuration component, user interface component, device self-test component, adaptive adjustment component, and operation recommendation component. In the smart home management scenario, the device discovery component is used to obtain the identification code of the smart home device scanned by the terminal and obtain the device information of the smart home device in order to access new smart home devices.

[0038] The group management component is used to group devices according to user-defined grouping schemes or system-recommended grouping schemes. Grouping schemes include: grouping by smart home device product type, grouping by smart home device location, or customizing any smart home device into a group. System-recommended grouping schemes can be derived from grouping schemes of similar users, set as a default scheme, or generated after analyzing the user's residence map and preset needs. For example, if the user's residence map shows a three-bedroom, one-living room, one-bathroom, and one-kitchen layout, and the user inputs preset needs such as whole-house lighting control and single-room lighting control via a terminal, then all smart lights in the house can be grouped together, and the smart lights in each room can be grouped together. This allows the user to directly turn off a group of lights or perform lighting changes within a group, realizing whole-house lighting control and single-room lighting control.

[0039] The scene configuration component allows users to customize various scene modes or obtain recommended scene modes, such as sleep mode and movie-watching mode, which are user-defined smart home device operating schemes. Under the corresponding scene mode, the smart home devices will adjust accordingly. For example, in sleep mode, the bedroom air conditioner will be turned on, all lights in the house will be turned off, and the door will be automatically locked. Users can set up according to their actual needs. For example, if there are elderly people or children in the house who need to be monitored, a security monitoring mode can be set up, and the home cameras can be set to track the corresponding people. A corresponding analysis monitoring component can also be set up to analyze the monitoring video to determine if there are any falls and issue timely warnings, that is, push the corresponding warning information to the user terminal.

[0040] The user interface component is used to receive operation information and coordinate with other components to perform corresponding operations.

[0041] The device self-test component is used to analyze whether smart home devices have malfunctioned based on environmental information and actuator operating status information.

[0042] The adaptive adjustment component is used to adaptively adjust the working scheme of smart home devices when they malfunction, ensuring that the smart home devices can work to meet the user's needs.

[0043] The operation recommendation component is used to obtain recommended operation information based on the collected information and the recommendation model when the user starts to control smart home devices in the terminal, and then push it.

[0044] In such Figure 1 and Figure 2 Based on the smart home system shown, this application provides a smart home management method that can improve the smart home management effect, realize automatic anomaly detection in the smart home system, and improve the anomaly detection efficiency. Figure 3 A flowchart illustrating an exemplary embodiment of this application is provided for a smart home management method. This method can be executed by a computer device, which can be implemented as a server or a terminal device. Actuators and sensors in the smart home system can establish a communication connection with the computer device through a smart gateway to achieve smart home management. Figure 3 As shown, the smart home management method may include the following steps.

[0045] Step 310: After sending a first control command to the first actuator to indicate a change in working status information, the actuator status information of the first actuator and the first actual environmental information of the linkage equipment group to which the first actuator belongs are obtained; the linkage equipment group includes at least one actuator and at least one sensor, and the environmental information collected by the sensor in the linkage equipment group is related to the change in the working status of the actuator.

[0046] In this embodiment, the first actuator can be any smart home device in the smart home management system. The computer device can pre-configure a group of interconnected devices, which includes at least one actuator and at least one sensor. Through coordinated operation between the devices, corresponding functions are achieved in specific scenarios. Simultaneously, changes in the operating state of any actuator in the interconnected device group will affect the environmental information collected by the sensors within the group. For example, the computer device can configure a light sensor, an automatic curtain switch, and smart lighting fixtures into the same interconnected device group. The configuration of the interconnected device group can be adjusted according to different application scenarios and needs. This device group can be automatically generated by the computer device based on the user's historical configuration, or it can be manually set and submitted to the computer device by the user according to specific needs; this application does not impose any restrictions on this.

[0047] The first control command sent by the computer device to the first actuator can be generated and issued based on the received user operation, or it can be automatically generated and issued based on changes in environmental information.

[0048] After sending a first control command to the first actuator to indicate a change in its operating status information, the first actuator will normally respond to the command and adjust the corresponding operating status information. For example, adjusting the actuator's operating status information may include adjusting operating parameters, such as adjusting the brightness of smart lights (e.g., from 50% to 80%), adjusting the set temperature of an air conditioner (e.g., from 26℃ to 22℃), or changing the opening / closing ratio of smart curtains, etc. Alternatively, adjusting the actuator's operating status information may also include switching operating modes. Switching operating modes may involve adjusting a set of preset operating parameters, such as switching an air purifier from "silent mode" to "powerful mode," which may simultaneously increase fan speed, adjust noise levels, and extend operating time; or switching smart lighting from "reading mode" to "cinema mode," correspondingly achieving overall changes in color temperature, brightness, and even multi-lamp linkage, etc.

[0049] To verify whether there is any equipment malfunction in the linked equipment group, the computer equipment can obtain the actuator status information of the first actuator and collect the first actual environmental information through the sensors in the linked equipment group, which can be used as the basis for judging whether the equipment is malfunctioning.

[0050] On the one hand, computer equipment can determine the abnormal state of the corresponding equipment based on the information acquisition results. For example, if the actuator status information of the first actuator is not obtained, it can be determined that the communication of the first actuator is abnormal. Without obtaining the first actual environmental information, it was determined that the sensor communication in the linkage equipment group was abnormal.

[0051] The communication anomaly can include two situations: one is equipment failure, such as damaged equipment hardware or firmware error; the other is communication connection anomaly, such as network interruption, signal interference or configuration error, etc. After detecting the communication anomaly, the computer device can provide relevant anomaly information to the user to assist the user in locating and troubleshooting the anomaly.

[0052] On the other hand, after the computer equipment obtains the actuator status information and the actual environmental information, it can determine that the communication connection between the actuator and the sensor is normal. In this case, it can further diagnose whether there is a fault in the equipment. The fault diagnosis process is shown in steps 320 to 350.

[0053] Step 320: After obtaining the actuator state information and the first actual environment information of the first actuator, obtain the environment association information of the first actuator; the environment association information is used to characterize the theoretical environment information corresponding to each actuator state information of the actuator.

[0054] In other words, this environmental correlation information is used to characterize the theoretical environmental information expected by the actuator under different operating states. Specifically, it represents the correspondence between changes in the actuator's state information and changes in environmental parameters during normal operation. This environmental correlation information can be a state-environment theoretical mapping established based on the inherent influence relationship between the actuator's working principle and the physical quantities monitored by sensors. For example, for a smart air conditioner, in "cooling mode," for every 1°C decrease in the set temperature, the theoretical temperature change in the corresponding space should follow a thermodynamic model, allowing the room temperature to approach the set value within a certain time. For a smart humidifier, in "continuous humidification" mode, there is a clear correspondence between its mist output and the theoretical increase in indoor humidity. Alternatively, this environmental correlation information can also be obtained through statistical analysis of historical normal data.

[0055] Step 330: Based on the environmental association information and actuator status information of the first actuator, determine the current theoretical environmental information of the first actuator.

[0056] After obtaining the environmental information, the computer device can deduce the theoretically present environmental parameters, i.e., the current theoretical environmental information, by querying the information and combining it with the current actual working state of the first actuator (i.e., actuator state information).

[0057] Step 340: Match the first actual environmental information with the current theoretical environmental information to obtain the matching result.

[0058] The matching result can include whether the first actual environmental information matches the current theoretical environmental information, or whether they do not match. Since the key environmental parameters on which different actuators adjust their operating states vary—for example, smart air conditioners primarily respond to temperature changes, smart humidifiers primarily regulate humidity, and smart switches are often linked to light intensity—the actual environmental information typically contains multiple types of information, such as natural environmental information (temperature, humidity, light intensity), location information (the location of the device), time information (the time of the operation), and weather information (indoor and outdoor weather conditions). Therefore, when performing the matching judgment, the computer device needs to extract the corresponding actual key parameters from the acquired first actual environmental information based on the device type of the first actuator, match these actual key parameters with the theoretical key parameters in the current theoretical environmental information, and determine the matching result based on the degree of deviation between the two. Illustratively, if the difference between the actual key parameters and the theoretical key parameters is less than a first difference threshold, the matching result is determined to be a match; otherwise, the matching result is determined to be a mismatch.

[0059] Furthermore, if the actuator's operating state adjustment depends on multiple key environmental parameters, the computer device, when performing a matching judgment, can extract and match multiple actual key parameters with theoretical key parameters. Based on the degree of deviation corresponding to each of the multiple key environmental parameters, the matching result between the first actual environmental information and the current theoretical environmental information is determined. Illustratively, the computer device can calculate the difference for each key environmental parameter. Then, based on the importance of each key environmental parameter to the actuator's operating state adjustment, the deviations are weighted and summed. Finally, the weighted comprehensive deviation value is compared with a preset second difference threshold. If the comprehensive deviation value is less than the second deviation threshold, the matching result is determined to be a match; otherwise, the matching result is determined to be a mismatch. The weights corresponding to different levels of importance can be configured based on actual needs, and this embodiment does not impose any limitations on this.

[0060] Step 350: Based on the matching results, determine the abnormal status of the linked equipment group.

[0061] If the matching result indicates that the current theoretical environment information matches the first actual environment information, it is determined that no equipment abnormality has occurred in the linked equipment group.

[0062] If the matching result indicates a mismatch between the current theoretical environmental information and the first actual environmental information, it is determined that a device malfunction has occurred in the linked equipment group. In this case, the computer can further determine the type of device malfunction, i.e., whether it is a sensor malfunction, an actuator malfunction, or both. Therefore, in one possible implementation, when the matching result indicates a mismatch between the current theoretical environmental information and the first actual environmental information, a fault judgment is performed on the linked equipment group to determine the abnormal state of the linked equipment group.

[0063] In the process of fault diagnosis, the computer equipment can make a preliminary fault diagnosis based on the reasonableness of the current theoretical environmental information and the first actual environmental information. For example, if the first actual environmental information fed back by the sensor exceeds the reasonable physical range, it is determined that the sensor has a fault. For example, the temperature value in the first actual environmental information fed back by the sensor is -100℃. Or, if the current theoretical environmental information corresponding to the actuator deviates from its reasonable working range, it is determined that the actuator has a fault. For example, the theoretical environmental information corresponding to the smart air conditioner in cooling mode is 0℃.

[0064] In another feasible solution, since the operation of smart home devices causes changes in environmental information, sensors and actuators can troubleshoot each other. Therefore, if the linked device group includes actuators other than the first actuator, the computer device can actively adjust the actual state of the environment in which the linked device group is located by changing the operating state of these other actuators that have not changed their operating state. Based on the sensor's collection of environmental information after the adjustment, further fault judgment is performed. In this case, fault judgment is performed on the linked device group to determine the abnormal state of the linked device group, including: Send a second control command indicating a change in operating status information to the second actuator in the linkage equipment group; Obtain the second actual environmental information of the linked equipment group; If the first actual environmental information and the second actual environmental information are inconsistent, it is determined that the first actuator has a device malfunction. If the first actual environmental information and the second actual environmental information are consistent, it is determined that there is a device malfunction in the sensor of the linkage equipment group.

[0065] By changing the operating status information of the second actuator, the actual environmental information of the linked equipment group can be changed. If the actual environmental information fed back by the sensor is consistent before and after the change of the operating status information of the second actuator, it means that the sensor cannot accurately perceive the change of the actual environmental information and the sensor has a device malfunction. If the actual environmental information fed back by the sensor is inconsistent before and after the change of the operating status information of the second actuator, it means that the sensor can accurately perceive the change of the actual environmental information, and the situation of the sensor having a device malfunction can be ruled out. Then it is determined that the first actuator has a device malfunction. Through the above cross-validation mechanism, the computer equipment can distinguish and locate the specific abnormal equipment in the linked equipment group.

[0066] For example, a set of interconnected devices may include a light sensor as the sensor and an automatic curtain switch and a smart light as the actuators. The automatic curtain switch executes the command to open the curtains, changing its reported status from "closed" to "open." The system obtains the current time as 4:00 PM and the weather as cloudy, inferring that the theoretical range of indoor light intensity should be greater than the threshold X. However, the actual light intensity value collected by the light sensor is lower than X, which does not match the expected environment. To diagnose the fault, the system controls the smart light to switch from off to on to actively change the ambient lighting conditions, and simultaneously monitors the changes in the light sensor readings: if the light intensity value changes significantly, it indicates that the sensor is responding normally, and it can be determined that the automatic curtain switch has not actually opened properly, indicating an execution fault; if the light intensity value remains unchanged, it indicates that the sensor has failed to effectively respond to changes in light, and it can be determined that the light sensor has a data acquisition fault.

[0067] Furthermore, after identifying a faulty device, the computer generates corresponding fault information and sends it back to the user for troubleshooting. Through this self-checking process, the computer can determine the abnormal status of smart home devices, effectively improving the reliability and maintainability of the smart home system. In addition, automated anomaly detection can provide early warnings of potential faults, allowing users to intervene promptly. It also reduces the difficulty and time cost of manually troubleshooting device problems, providing a basis for long-term stable operation and continuous optimization of the system.

[0068] In one possible implementation, when a smart home device malfunctions, the computer can adaptively adjust the device's operating scheme to continuously ensure that the user's needs are met. Illustratively, when the device self-test module detects a malfunction in a smart home device, or when a user actively reports a device malfunction through a terminal, the computer will adaptively adjust the current smart home operating scheme. The adjustment objectives mainly include the following two methods: by coordinating the operating states of other normal devices, keeping the difference between the environmental information and the state when the device is not malfunctioning within a preset acceptable range; or, using normal devices with similar or substitutable functions as substitutes for the malfunctioning device, working collaboratively to achieve the original functional requirements as much as possible. Therefore, when it is determined that the first actuator has a device malfunction, the method also includes: Execute a device adjustment strategy corresponding to the first actuator, the device adjustment strategy including at least one of the following: adjust the actuator status information of at least one third actuator, or enable at least one fourth actuator, the fourth actuator having the same function as the first actuator; The equipment adjustment strategy is generated based on at least one of the following: the function type of the first actuator, the first actual environmental information, and the historical adjustment strategy. The equipment adjustment strategy is used to match the theoretical environmental information with the first actual environmental information.

[0069] The third and fourth actuators can be actuators in the same linked equipment group as the first actuator, or they can be actuators outside of that group. For example, if any smart lights in the house malfunction, the computer will adjust its original operating plan, selecting nearby, functional smart lights to turn on, keeping the light intensity changes in the area before and after the adjustment within a preset range, thus ensuring basic lighting needs. Alternatively, if the air conditioner in a room malfunctions, the computer will update its original operating plan, starting the working air conditioner in a connected room and keeping the room door open, maintaining temperature changes within an acceptable range, ensuring basic cooling or insulation effects. Or, if the monitoring equipment in a specific room malfunctions, the computer can reconfigure the monitoring plan, using monitoring equipment in other locations for functional compensation. For instance, in elderly monitoring scenarios, if the monitoring in the elderly person's bedroom fails, the computer can use other cameras (such as a living room camera) to capture the elderly person's movements, or adjust the orientation of a camera with pan-tilt functionality to cover part of the elderly person's room, thus maintaining basic monitoring capabilities of the elderly person's daily life.

[0070] The aforementioned equipment adjustment strategies can be generated by a computer device based on the function type of the first actuator, the first actual environmental information, and historical adjustment strategies. Specifically, the computer device can search for available devices with similar or alternative functions in the linked equipment group or whole-house equipment based on the function type of the first actuator. For example, if the faulty device is a lighting device, other smart lights in the same area are prioritized as the adjustment target. After determining the adjustment target, the target adjustment amount required to maintain environmental parameters within an acceptable range is calculated based on the currently collected first actual environmental information, and the working state parameters of the substitute device are determined accordingly. Furthermore, the computer device can retrieve effective adjustment strategies executed in similar fault scenarios in the past and make adaptive adjustments based on the current environment and equipment status. If multiple successful alternatives exist in the historical records, the computer device can prioritize the strategy with the best overall effect. Alternatively, the computer device can train a strategy generation model based on historical adjustment strategies, and through continuous iteration and optimization, obtain a trained strategy generation model. This model, when input into the current function type of the first actuator and the first actual environmental information, outputs an equipment adjustment strategy that conforms to the current situation.

[0071] Furthermore, to facilitate smart home operation for users, the smart home management method provided in this application also involves generating recommended smart home operations. When a user initiates control of a smart home device on their terminal, recommended operation information is pushed to the user based on the collected actual environmental information. This process can be implemented as follows: Once the first actual environmental information is obtained, a recommendation model is used to generate recommended operation information corresponding to the first actual environmental information. This recommended operation information is used to assist users in managing smart homes. Push recommended operation information to the user's terminal; The recommendation model is trained based on a set of user behavior samples, which includes historical environmental information and corresponding user historical operation information.

[0072] The aforementioned user behavior sample set can be a sample set established by a computer device after receiving operation information uploaded by a user through a terminal, and associating and storing the operation information with the corresponding environmental information.

[0073] During the model training phase, the system uses historical environmental information from the user behavior sample set as input features and corresponding user historical operation information as training labels to iteratively optimize the recommendation model. The training objective is to make the recommended operation information output by the model as consistent as possible with the user's actual historical operations, thereby gradually improving the accuracy of the model's personalized recommendations. Through continuous iteration and learning, a well-trained recommendation model with user behavior prediction capabilities is finally obtained.

[0074] During the model application phase, after inputting the first actual environment information into the recommendation model, the model outputs recommended operation information corresponding to the first actual environment information. Furthermore, the computer device can push this recommended operation information to the user terminal for the user to select settings or refer to. For example, the computer device can push the information in the form of a pop-up window or a floating window.

[0075] In one possible implementation, in order to improve the accuracy of the recommendation operation information of the model, the recommendation model in this embodiment can be composed of two models with different architectures. Schematic, the recommendation model includes a first sub-recommendation model and a second sub-recommendation model. The recommendation model generates recommended operation information corresponding to the first actual environment information, including: The first actual environment information is input into the first sub-recommendation model and the second sub-recommendation model respectively to obtain the first candidate operation information and the second candidate operation information; If the first candidate operation information and the second candidate operation information are the same, the first candidate operation information or the second candidate operation information shall be determined as the recommended operation information. If the first candidate operation information and the second candidate operation information are different, both the first candidate operation information and the second candidate operation information will be determined as recommended operation information.

[0076] In other words, if the candidate operation information generated by the two recommendation sub-models is consistent, the computer device can integrate it into the same candidate operation information and recommend it to the user terminal as recommended operation information; if the candidate operation information generated by the two recommendation sub-models is inconsistent, the computer device can use the candidate operation information obtained by each sub-model as two different recommended operation information and recommend them to the user terminal for the user to choose or refer to.

[0077] In one possible implementation, the two recommendation sub-models can be a neural network recommendation model and a long short-term memory network recommendation model, respectively. For the neural network recommendation model, to predict user actions based on actual environmental information, this embodiment constructs a recommendation sub-model based on a backpropagation (BP) neural network. This model has a three-layer feedforward network structure, including an input layer, a hidden layer, and an output layer. The input layer takes the current actual environmental information as input. The dimension of the actual environmental information is determined by the number of sensors actually deployed. If p sensors are deployed, the number of nodes in the input layer is p+3. The output layer outputs the predicted operation information, corresponding to one node. The number of nodes in the hidden layer can be determined by the following formula: Where l is the number of nodes in the hidden layer, n is the number of nodes in the input layer, m is the number of nodes in the output layer, and a is an adjustment constant, which can be a value between 1 and 10. Taking a=6 as an example, the number of nodes in the hidden layer is determined to be 9 after calculation.

[0078] The neural network recommendation model incorporates activation functions. The hidden layer neurons use the sigmoid tangent function (tansig) as their activation function due to its good differentiability and nonlinear mapping ability. The output layer neurons also use the tansig function to accommodate the normalized representation of the output values.

[0079] After constructing the network structure, a set of user behavior samples is used as training samples. The model is iteratively trained using the backpropagation algorithm, continuously adjusting the network weights and biases until the model converges. The trained model can output recommended operation information that is highly consistent with the user's actual operational tendencies based on the actual input environment information.

[0080] For Long Short-Term Memory (LSTM) network recommendation models, to construct a time-series-based operation prediction model, the computer device first constructs a time series from the set of user behavior samples. For example, the time series can be represented as [(t1, d1), (t2, d2), ..., (t...]. n , d n [ ], where t represents a time point and d represents the operation information at the corresponding time point. Afterwards, the computer equipment normalizes the time series data to improve model training performance. For example, MinMaxScaler can be used to scale the data range to the [0,1] interval. After preprocessing, the model uses historical operation information from several consecutive time points as input to predict operation information at subsequent time points. By constructing a sequence prediction model containing a multi-layer LSTM (Long Short-Term Memory) network structure and a fully connected output layer, and using mean squared error as the loss function for training, the model can learn temporal dependencies and achieve effective prediction of future operation information.

[0081] In this embodiment of the application, for the Long Short-Term Memory Network recommendation model, the input can be set as actual environmental information and the output can be set as operation information. For example, the environmental information sequence of multiple consecutive time steps can be used as input to predict the operation information at the current or next moment. After training the Long Short-Term Memory Network recommendation model with a set of user behavior samples, a sub-recommendation model is obtained.

[0082] After a user makes a selection based on the recommended operation information from the computer device, the user terminal can upload the selection to the computer device. Upon receiving the user's selection, the computer device associates and stores this operation information with the current actual environment information, forming a new user behavior sample. This user behavior sample can then be used for incremental learning or periodic optimization training of the recommendation model (including each sub-recommendation model), thereby continuously improving the model's prediction accuracy and personalized adaptation capabilities in real-world usage scenarios.

[0083] Furthermore, when the first candidate operation information differs from the second candidate operation information, the computer device can also perform collaborative training based on different sub-recommendation models; after pushing the recommended operation information to the user terminal, the method further includes: Update the user behavior sample set based on the received user selection actions based on recommendation action information; The target sub-recommendation model is optimized and trained based on the updated user behavior sample set. This target sub-recommendation model is the sub-recommendation model corresponding to the candidate operation information that has not been selected by the user in the recommendation operation information.

[0084] For example, if recommended operation information A (corresponding to candidate operation information recommended by the neural network recommendation model) is different from recommended operation information B (corresponding to candidate operation information recommended by the long short-term memory network recommendation model), and the user ultimately chooses recommended operation information B, the computer device can optimize and train the neural network recommendation model based on the updated set of user behavior samples; if the user ultimately chooses recommended operation information A, the computer device can optimize and train the long short-term memory network recommendation model based on the updated set of user behavior samples. Through collaborative training, each sub-recommendation model can be continuously and dynamically adjusted based on user feedback, achieving co-evolution and performance improvement of multiple models in collaboration.

[0085] In another possible scenario, if no user action is received after the recommended action information is pushed to the user terminal, it means that the current recommended action information fails to meet the user's needs. The computer device can update the user behavior sample set after receiving the user's actual action information, and optimize and train the recommendation model based on the updated user behavior sample set.

[0086] In one possible implementation, when the target sub-recommendation model is a Long Short-Term Memory (LSTM) network recommendation model, the target sub-recommendation model is optimized and trained based on the updated user behavior sample set, including: When the proportion of candidate operation information recommended by the target sub-recommendation model that is not selected by the user is greater than the preset proportion threshold, the updated user behavior sample set is grouped by k grouping time nodes to generate k+1 different user behavior sample subsets. Each user behavior sample subset contains user behavior samples from the corresponding grouping time node to the latest time node, where k is a positive integer. Using k+1 subsets of user behavior samples sequentially, the target sub-recommendation model is optimized and trained in different training rounds. The optimized target sub-recommendation model is then applied sequentially in each training round to generate candidate recommendation operation information until the optimization completion condition is met. The optimization completion condition includes: the candidate operation information output by the optimized target sub-recommendation model is consistent with the candidate operation information output by another sub-recommendation model, or the number of user behavior sample subsets used is equal to k+1.

[0087] In other words, to improve the training performance of the Long Short-Term Memory (LSTM) network recommendation model, the computer device can incorporate a random perturbation mechanism into the model: ; in This represents the number of segments in the time series. The percentage of times the user selects push operation information B out of the total number of pushes (if candidate operation information A and candidate operation information B are the same, and the user selects push operation information, then operation information B is considered to have been selected). This is the disturbance threshold (i.e., the preset percentage threshold). To randomly generate an integer between the minimum number min and the maximum number max, where min is not less than 1 and max is not greater than n.

[0088] That is, when the number of times a user selects the pushed operation information B accounts for a proportion of the total number of pushes greater than or equal to the proportion threshold set according to the needs, the user behavior sample set will not be grouped, and the complete user behavior sample will be used directly for model training.

[0089] When the percentage of times a user selects push notification message B in the total number of push notifications is less than the perturbation threshold set according to requirements... Then, the user behavior sample set will be grouped into data groups, and the grouping method is as follows: The user behavior sample set is constructed into a time series format, such as an operation time series. The data format should be: [(t1, d1), (t2, d2), ..., (t n , d n[(t1, d1, e1), (t2, d2, e2), …, (t]], where t represents a time point and d is the operation information at the corresponding time point; or, operation-environment time sequence, the data format should be: [(t1, d1, e1), (t2, d2, e2), …, (t]] n , d n ,e n ]], t is the time point, d is the operation information at the corresponding time point, and e is the actual environment information at that time; then, MinMaxScaler is used to normalize the historical operation data to ensure that the values ​​are between 0 and 1; the normalized data is then divided into input (X) and output (y) pairs.

[0090] choose Group time nodes T1, T2, ..., T k The time nodes can be selected randomly or sequentially at preset time intervals. Based on these nodes, the normalized input-output pairs are divided into k+1 subsets of user behavior samples: the first set contains samples from t1 to t... n All input / output pairs, the (i+1)th set ( ): Take time node T i After that (if the time points before and after are related to T) i If the distances are equal, then all input-output pairs at the next available time point are selected. These k+1 subsets of user behavior samples together constitute the input-output pair population.

[0091] Subsequently, in each training round, the computer device sequentially selects a subset of user behavior samples from the population from the input and output for model training. The selection order can be sorted from longest to shortest time span corresponding to each subset of user behavior samples, or sorted from most to least number of samples corresponding to each subset of user behavior samples, or it can be randomly sorted.

[0092] During iterative training, follow these steps: Initialization and first-round prediction: Initialize the iteration coefficients s=1; use the s-th input and output to train the model on the set (i.e. the s-th subset of user behavior samples). After training, use the current LSTM model to predict the operation information at several future time points, and perform inverse normalization on the output to obtain the actual usable operation information, i.e., obtain the operation information B output by LSTM.

[0093] Initial consistency judgment: Determine whether operation information A and operation information B are the same. If they are the same, push operation information A or operation information B to the user terminal in a pop-up window and end the current iteration.

[0094] Iteration round determination: If the two are different, then determine whether s is greater than or equal to k+1: If so, it means that everything has been completed. The training process involves a subset of user behavior samples. At this point, operation information A and operation information B are simultaneously pushed to the user terminal in the form of a pop-up window, and the user's choice is obtained. If the user chooses operation information A, the user behavior sample set is updated, and the neural network recommendation model is optimized and trained. If the user chooses operation information B, the user behavior sample set is updated, and the long short-term memory network recommendation model is optimized and trained.

[0095] If not, update s=s+1, and perform training and prediction again using the s-th sample subset to obtain the updated operation information B, and repeat the consistency judgment. If, in subsequent iterations, the updated operation information B is the same as operation information A, one of them will be pushed to the user terminal in the form of a pop-up window, along with the LSTM model currently used for training and its corresponding time segment nodes. These serve as the initial model and the starting time point, respectively, for subsequent optimization.

[0096] If, after training all k+1 sample subsets, the operational information B is still inconsistent with operational information A, then one or more pieces of operational information B generated in this iteration can be selected and pushed to the user terminal along with operational information A. If the user selects a piece of operational information B, then the LSTM model corresponding to operational information B and the corresponding time segment node will be sent. These serve as the initial model and the starting time point, respectively, for subsequent model optimization. This mechanism effectively addresses the impact of sudden changes in usage habits on the recommendation system, thereby improving the robustness and accuracy of predictions.

[0097] In summary, the smart home management method provided in this application constructs a linked device group consisting of actuators and sensors. After sending control commands to the first actuator, it simultaneously acquires the actuator's status information and the first actual environmental information of the linked device group. Based on pre-set environmental association information, it determines the current theoretical environmental information, and then matches the actual environmental information with the theoretical environmental information. Based on the matching results, it automatically determines whether there are any device malfunctions in the linked device group. This method can achieve automated anomaly detection and diagnosis of smart home devices, effectively solving the problems of reliance on manual inspection, low efficiency, and difficulty in adapting to the rapid increase in the number and types of devices in traditional smart home management. It significantly improves the reliability and intelligent operation and maintenance level of the smart home system, thereby improving the overall management effect and ensuring user experience.

[0098] Furthermore, after acquiring current environmental information, the system automatically generates suitable recommended operation information through a recommendation model and pushes it to the user's terminal. This effectively reduces the operational complexity and decision-making burden for users managing smart homes, while personalized recommendations improve control accuracy and scene adaptability, enhancing the usability and user interaction experience of the smart home system.

[0099] Figure 4 This application shows a block diagram of a smart home management device provided in an exemplary embodiment, which can be used to perform functions such as... Figure 3 All or part of the steps in the illustrated embodiments, such as Figure 4 As shown, the smart home management device may include the following modules.

[0100] The first information acquisition module 410 is used to acquire the actuator status information of the first actuator and the first actual environmental information of the linkage equipment group to which the first actuator belongs after sending a first control command to the first actuator indicating a change in working status information; the linkage equipment group includes at least one actuator and at least one sensor, and the environmental information collected by the sensor in the linkage equipment group is related to the change in the working status of the actuator. The second information acquisition module 420 is used to acquire environmental association information of the first actuator when the actuator state information and the first actual environment information of the first actuator are acquired; the environmental association information is used to characterize the theoretical environment information corresponding to each actuator state information of the actuator; The environmental information determination module 430 is used to determine the current theoretical environmental information of the first actuator based on the environmental association information of the first actuator and the actuator state information of the first actuator; The matching module 440 is used to match the first actual environment information with the current theoretical environment information to obtain a matching result; The status determination module 450 is used to determine the abnormal status of the equipment in the linkage equipment group based on the matching result.

[0101] In one possible implementation, the state determination module 450 is used for, If the matching result indicates that the current theoretical environment information matches the first actual environment information, it is determined that no equipment abnormality has occurred in the linked equipment group; If the matching result indicates that the current theoretical environment information does not match the first actual environment information, a fault judgment is performed on the linkage equipment group to determine the abnormal state of the equipment in the linkage equipment group.

[0102] In one possible implementation, the state determination module 450 includes: The instruction issuing submodule is used to send a second control instruction indicating a change in working status information to the second actuator in the linkage equipment group; The information acquisition submodule is used to acquire the second actual environmental information of the linked equipment group; The fault determination submodule is used to determine that the first actuator has a device fault when the first actual environmental information is inconsistent with the second actual environmental information; and to determine that the sensor in the linkage equipment group has a device fault when the first actual environmental information is consistent with the second actual environmental information.

[0103] In one possible implementation, the device further includes: The strategy execution module is used to execute a device adjustment strategy corresponding to the first actuator when it is determined that the first actuator has a device fault. The device adjustment strategy includes at least one of the following: adjusting the actuator status information of at least one third actuator, or enabling at least one fourth actuator, the fourth actuator having the same function as the first actuator. The device adjustment strategy is generated based on at least one of the following: the function type of the first actuator, the first actual environmental information, and historical adjustment strategies, wherein the device adjustment strategy is used to match the theoretical environmental information with the first actual environmental information.

[0104] In one possible implementation, the state determination module 450 is used for, If the actuator status information of the first actuator is not obtained, it is determined that the communication of the first actuator is abnormal; Without obtaining the first actual environmental information, it is determined that the sensor communication in the linkage equipment group is abnormal.

[0105] In one possible implementation, the device further includes: The operation information generation module is used to generate recommended operation information corresponding to the first actual environment information through a recommendation model when the first actual environment information is obtained. The recommended operation information is used to assist users in smart home management. The information push module is used to push the recommended operation information to the user terminal; The recommendation model is trained based on a set of user behavior samples, which includes historical environment information and corresponding user historical operation information.

[0106] In one possible implementation, the recommendation model includes a first sub-recommendation model and a second sub-recommendation model; The operation information generation module includes: The information generation submodule is used to input the first actual environment information into the first sub-recommendation model and the second sub-recommendation model respectively to obtain the first candidate operation information and the second candidate operation information; The operation information determination submodule is used to determine either the first candidate operation information or the second candidate operation information as the recommended operation information when the first candidate operation information and the second candidate operation information are the same; and to determine both the first candidate operation information and the second candidate operation information as the recommended operation information when the first candidate operation information and the second candidate operation information are different.

[0107] In one possible implementation, the device further includes: The sample update module is used to update the user behavior sample set according to the received user selection operation based on the recommended operation information after pushing the recommended operation information to the user terminal, when the first candidate operation information is different from the second candidate operation information. The model optimization and training module is used to optimize and train the target sub-recommendation model based on the updated user behavior sample set. The target sub-recommendation model is the sub-recommendation model corresponding to the candidate operation information that was not selected by the user in the recommendation operation information.

[0108] In one possible implementation, the model optimization training module is used for, When the target sub-recommendation model is a long short-term memory network recommendation model, and when the proportion of candidate operation information recommended by the target sub-recommendation model that is not selected by the user is greater than a preset proportion threshold, the updated user behavior sample set is grouped by k grouping time nodes to generate k+1 different user behavior sample subsets. Each user behavior sample subset contains user behavior samples from the corresponding grouping time node to the latest time node, where k is a positive integer. The target sub-recommendation model is optimized and trained sequentially using the k+1 user behavior sample subsets in different training rounds. The optimized target sub-recommendation model is then applied sequentially in each training round to generate candidate recommendation operation information until the optimization completion condition is met. The optimization completion condition includes: the candidate operation information output by the optimized target sub-recommendation model is consistent with the candidate operation information output by another sub-recommendation model, or the number of user behavior sample subsets used is equal to k+1.

[0109] In summary, the smart home management device provided in this application constructs a linked device group consisting of actuators and sensors. After sending control commands to the first actuator, it simultaneously acquires the actuator's status information and the first actual environmental information of the linked device group. Based on pre-set environmental association information, it determines the current theoretical environmental information, and then matches the actual environmental information with the theoretical environmental information. Based on the matching result, it automatically determines whether there are any device malfunctions in the linked device group. This device enables automated anomaly detection and diagnosis of smart home devices, effectively solving the problems of reliance on manual troubleshooting, low efficiency, and difficulty in adapting to the rapid increase in the number and types of devices in traditional smart home management. It significantly improves the reliability and intelligent operation and maintenance level of the smart home system, thereby improving the overall management effect and ensuring user experience.

[0110] Figure 5 A structural block diagram of a computer device 500 illustrated in an exemplary embodiment of this application is shown. This computer device can be implemented as a server as described above in this application. The computer device 500 includes a Central Processing Unit (CPU) 501, a system memory 504 including Random Access Memory (RAM) 502 and Read-Only Memory (ROM) 503, and a system bus 505 connecting the system memory 504 and the CPU 501. The computer device 500 also includes a mass storage device 506 for storing an operating system 509, application programs 510, and other program modules 511.

[0111] Without loss of generality, the computer-readable medium may include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable media implemented using any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer storage media include RAM, ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other solid-state storage technologies, CD-ROM, digital versatile disc (DVD) or other optical storage, magnetic tape cassettes, magnetic tape, disk storage, or other magnetic storage devices. Of course, those skilled in the art will recognize that the computer storage media are not limited to the above-mentioned types. The system memory 504 and mass storage device 506 described above can be collectively referred to as memory.

[0112] According to various embodiments of this disclosure, the computer device 500 can also be connected to a remote computer on a network, such as the Internet. That is, the computer device 500 can be connected to a network 508 via a network interface unit 507 connected to the system bus 505, or it can use the network interface unit 507 to connect to other types of networks or remote computer systems (not shown).

[0113] The memory also includes at least one instruction, at least one program, code set, or instruction set, which are stored in the memory. The central processing unit 501 executes the at least one instruction, at least one program, code set, or instruction set to implement all or part of the steps in the smart home management method shown in the above embodiments.

[0114] Figure 6 A structural block diagram of a computer device 600 illustrating another exemplary embodiment of this application is shown. This computer device 600 can be implemented as the aforementioned terminal device, such as a smartphone, tablet computer, laptop computer, desktop computer, etc. The computer device 600 may also be referred to as user equipment, portable terminal, laptop terminal, desktop terminal, or other names.

[0115] Typically, computer device 600 includes a processor 601 and a memory 602.

[0116] In some embodiments, the computer device 600 may optionally include a peripheral device interface 603 and at least one peripheral device. The processor 601, memory 602, and peripheral device interface 603 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 603 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of the following: a radio frequency circuit 604, a display screen 605, a camera assembly 606, an audio circuit 607, and a power supply 608.

[0117] In some embodiments, the computer device 600 further includes one or more sensors 609. The one or more sensors 609 include, but are not limited to, an accelerometer 610, a gyroscope 611, a pressure sensor 612, an optical sensor 613, and a proximity sensor 614.

[0118] Those skilled in the art will understand that Figure 6 The structure shown does not constitute a limitation on the computer device 600, and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0119] In one exemplary embodiment, a computer-readable storage medium is also provided, which stores at least one computer program that is loaded and executed by a processor to implement all or part of the steps in the smart home management method described above. For example, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, or optical data storage device, etc.

[0120] In one exemplary embodiment, a computer program product is also provided, comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions that, when executed by a computer, cause the computer to perform the above-described actions. Figure 3 All or part of the steps of the smart home management method shown in the embodiments.

[0121] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the claims.

[0122] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A smart home management method, characterized in that, The method includes: After sending a first control command to the first actuator to indicate a change in working status information, the actuator status information of the first actuator and the first actual environmental information of the linkage equipment group to which the first actuator belongs are obtained; the linkage equipment group includes at least one actuator and at least one sensor, and the environmental information collected by the sensor in the linkage equipment group is related to the change in the working status of the actuator; Having obtained the actuator state information and the first actual environment information of the first actuator, the environmental association information of the first actuator is obtained; the environmental association information is used to characterize the theoretical environment information corresponding to each actuator state information of the actuator; Based on the environmental association information and the actuator status information of the first actuator, the current theoretical environmental information of the first actuator is determined; The first actual environment information is matched with the current theoretical environment information to obtain a matching result; Based on the matching results, the abnormal state of the equipment in the linked equipment group is determined.

2. The method according to claim 1, characterized in that, Determining the abnormal state of the linked equipment group based on the matching result includes: If the matching result indicates that the current theoretical environment information matches the first actual environment information, it is determined that no equipment abnormality has occurred in the linked equipment group; If the matching result indicates that the current theoretical environment information does not match the first actual environment information, a fault judgment is performed on the linkage equipment group to determine the abnormal state of the equipment in the linkage equipment group.

3. The method according to claim 2, characterized in that, The step of fault diagnosis of the linked equipment group, determining the abnormal state of the equipment in the linked equipment group, includes: Send a second control command indicating a change in working status information to the second actuator in the linkage equipment group; Obtain the second actual environmental information of the linked equipment group; If the first actual environmental information is inconsistent with the second actual environmental information, it is determined that the first actuator has a device malfunction. If the first actual environmental information is consistent with the second actual environmental information, it is determined that the sensor in the linkage equipment group has a device malfunction.

4. The method according to any one of claims 1 to 3, characterized in that, If it is determined that the first actuator has a device malfunction, the method further includes: Execute a device adjustment strategy corresponding to the first actuator, the device adjustment strategy including at least one of the following: adjust the actuator status information of at least one third actuator, or enable at least one fourth actuator, the fourth actuator having the same function as the first actuator; The device adjustment strategy is generated based on at least one of the following: the function type of the first actuator, the first actual environmental information, and historical adjustment strategies, wherein the device adjustment strategy is used to match the theoretical environmental information with the first actual environmental information.

5. The method according to claim 1, characterized in that, The method further includes: If the actuator status information of the first actuator is not obtained, it is determined that the communication of the first actuator is abnormal; Without obtaining the first actual environmental information, it is determined that the sensor communication in the linkage equipment group is abnormal.

6. The method according to claim 1, characterized in that, The method further includes: Upon obtaining the first actual environment information, a recommendation model is used to generate recommended operation information corresponding to the first actual environment information. The recommended operation information is used to assist users in managing smart homes. The recommended operation information is pushed to the user's terminal; The recommendation model is trained based on a set of user behavior samples, which includes historical environment information and corresponding user historical operation information.

7. The method according to claim 6, characterized in that, The recommendation model includes a first sub-recommendation model and a second sub-recommendation model; The step of generating recommended operation information corresponding to the first actual environment information through a recommendation model includes: The first actual environment information is input into the first sub-recommendation model and the second sub-recommendation model respectively to obtain the first candidate operation information and the second candidate operation information; If the first candidate operation information and the second candidate operation information are the same, the first candidate operation information or the second candidate operation information shall be determined as the recommended operation information; If the first candidate operation information and the second candidate operation information are different, both the first candidate operation information and the second candidate operation information are determined as the recommended operation information.

8. The method according to claim 7, characterized in that, If the first candidate operation information differs from the second candidate operation information, after pushing the recommended operation information to the user terminal, the method further includes: Update the user behavior sample set based on the received user selection operation based on the recommended operation information; The target sub-recommendation model is optimized and trained based on the updated user behavior sample set. The target sub-recommendation model is the sub-recommendation model corresponding to the candidate operation information that was not selected by the user in the recommendation operation information.

9. The method according to claim 8, characterized in that, When the target sub-recommendation model is a long short-term memory network recommendation model, the step of optimizing and training the target sub-recommendation model based on the updated user behavior sample set includes: When the proportion of candidate operation information recommended by the target sub-recommendation model that is not selected by the user is greater than the preset proportion threshold, the updated user behavior sample set is grouped by k grouping time nodes to generate k+1 different user behavior sample subsets. Each user behavior sample subset contains user behavior samples from the corresponding grouping time node to the latest time node, where k is a positive integer. The target sub-recommendation model is optimized and trained sequentially using the k+1 user behavior sample subsets in different training rounds. The optimized target sub-recommendation model is then applied sequentially in each training round to generate candidate recommendation operation information until the optimization completion condition is met. The optimization completion condition includes: the candidate operation information output by the optimized target sub-recommendation model is consistent with the candidate operation information output by another sub-recommendation model, or the number of user behavior sample subsets used is equal to k+1.

10. A smart home management device, characterized in that, The device includes: The first information acquisition module is used to acquire the actuator status information of the first actuator and the first actual environmental information of the linkage equipment group to which the first actuator belongs after sending a first control command to the first actuator indicating a change in working status information; the linkage equipment group includes at least one actuator and at least one sensor, and the environmental information collected by the sensor in the linkage equipment group is related to the change in the working status of the actuator. The second information acquisition module is used to acquire environmental association information of the first actuator when the actuator state information and the first actual environment information of the first actuator are acquired; the environmental association information is used to characterize the theoretical environment information corresponding to each actuator state information of the actuator; The environmental information determination module is used to determine the current theoretical environmental information of the first actuator based on the environmental association information of the first actuator and the actuator state information of the first actuator; The matching module is used to match the first actual environment information with the current theoretical environment information to obtain a matching result; The status determination module is used to determine the abnormal status of the equipment in the linked equipment group based on the matching result.

Images