Smart home linkage control system and method oriented to scene self-adaption

Abstract

The invention relates to the technical field of intelligent home linkage control, in particular to an intelligent home linkage control system and method oriented to scene self-adaption. Working data and sensing data of a smart home are standardized to obtain virtual sensing data, the confidence degree of occurrence of various preset home unit events is obtained through the virtual sensing data, scene feature data are obtained, and finally, smart home linkage control is performed according to the scene feature data. Working data generated by the smart home device and limited environment sensing data are ingeniously subjected to standardization processing and converted into virtual sensing information capable of comprehensively reflecting the environment state and user behaviors, the overall sensing coverage range of the system is remarkably expanded, the innovative mechanism of the confidence coefficient of the preset home unit event is matched, and the user experience is improved. Therefore, the system can identify the scene more accurately and make a relatively reliable linkage control decision, and the problem of linkage misalignment caused by insufficient sensing data coverage of the smart home in the prior art is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smart home linkage control, and particularly relates to a smart home linkage control system and method for scene adaptation. BACKGROUND

[0002] In the field of smart home, linkage control technology significantly improves the intelligent level of living environment and user experience by integrating the collaborative operation of multiple types of intelligent devices. Based on environmental parameters and user behavior characteristics, the technology can automatically trigger the linkage response between devices, not only realizing dynamic adaptive adjustment of the home environment, but also greatly reducing the user's manual operation burden, making the home system have stronger flexibility and humanization characteristics.

[0003] However, the smart home linkage system often highly depends on the comprehensive arrangement of multiple types of sensors and intelligent home devices to support accurate scene recognition and linkage decision-making, such as the need to set up location tracking devices or environmental monitoring sensors, such as Bluetooth beacons, infrared sensors, and behavior perception modules, such as cameras and microphones, in users' homes.

[0004] However, the current situation is that most users' smart home systems are not "one-step" construction, but are formed by gradually replacing or supplementing traditional non-intelligent devices with newly added intelligent devices. Currently, only in specific scenarios such as hotels and dormitories, can the sensor layout be planned in advance. This results in a large number of blind spots in the collection of some sensor data when users use smart home, directly affecting the accuracy of the linkage system's judgment. Therefore, people need a new linkage control technology to solve the problem of linkage inaccuracy caused by insufficient sensor data coverage in current smart home. SUMMARY

[0005] Therefore, the present application provides a smart home linkage control system for scene adaptation to solve the problem of linkage inaccuracy caused by insufficient sensor data coverage in current smart home in the prior art.

[0006] The present application provides a smart home linkage control system for scene adaptation, comprising: a virtual sensor module for obtaining working data and sensor data of a smart home, and standardizing the working data and sensor data to obtain virtual sensor data; an event analysis module for obtaining the confidence of occurrence of a plurality of preset home unit events according to the virtual sensor data, wherein the preset home unit event is a factor index used to combine to describe different home scenes of a user; a scene analysis module for obtaining scene feature data according to the confidence of occurrence of a plurality of preset home unit events; A linkage control module is configured to perform intelligent home linkage control according to scene feature data.

[0007] In a preferred implementation: according to the virtual sensor data, a plurality of preset home unit event occurrence confidence levels are obtained, including: Obtain operation data of the mobile phone; According to the operation data and the virtual sensor data, a plurality of preset home unit event occurrence confidence levels are obtained.

[0008] In a preferred implementation: according to the operation data and the virtual sensor data, a plurality of preset home unit event occurrence confidence levels are obtained, including: According to the operation data, a mobile phone operation vector is established; According to the virtual sensor data, a virtual sensor vector is established; Splicing the mobile phone operation vector and the virtual sensor vector to obtain a sensor feature vector; The sensor feature vector is input into a preset neural network model to obtain a confidence level vector output by the neural network model, and each element in the confidence level vector corresponds to a confidence level of occurrence of a preset home unit event.

[0009] In a preferred implementation: one intelligent home generates a group of virtual sensor data at a time, one virtual sensor vector is established according to one group of virtual sensor data, and one sensor feature vector only includes one virtual sensor vector; according to the confidence levels of a plurality of preset home unit events, scene feature data is obtained, including: Obtain the confidence level vector corresponding to each intelligent home; Superimpose and normalize all the confidence level vectors, and take the obtained vector as the scene feature data.

[0010] In a preferred implementation: superimposing and normalizing all the confidence level vectors, and taking the obtained vector as the scene feature data, including: Obtain the distance between the mobile phone and the intelligent home corresponding to the confidence level vector, and establish a superposition weight of the confidence level vector based on the distance; Superimpose and normalize all the confidence level vectors based on the superposition weight, and take the obtained vector as the scene feature data.

[0011] In a preferred implementation: the elements in the virtual sensor vector include a device type field and a plurality of virtual sensor parameter fields; the preset home unit events include behavior unit events and environment unit events, wherein the behavior unit events are used to represent the behavior of the user, and the environment unit events are used to represent the environment state of the indoor.

[0012] In a preferred implementation: according to the scene feature data, intelligent home linkage control is performed, including: Obtain target scene feature data and a preset standard control strategy set, the standard control strategy set including a plurality of preset standard scene feature data and corresponding control strategies; Match a plurality of standard scene feature data closest to the scene feature data from the standard control strategy set; Integrate the control strategies corresponding to the plurality of standard scene feature data closest to the scene feature data to obtain a target control strategy; Based on the target control strategy, perform linkage control on the intelligent home actually existing in the target control strategy.

[0013] The application further provides a scene-adaptive intelligent home linkage control method, comprising: Obtain working data and sensing data of the intelligent home, and standardize the working data and the sensing data to obtain virtual sensing data; According to the virtual sensing data, obtain a plurality of preset home unit event occurrence confidence degrees, wherein the preset home unit event is a factor index used for combining to describe different home scenes of a user; According to the plurality of preset home unit event occurrence confidence degrees, obtain scene feature data; According to the scene feature data, perform intelligent home linkage control.

[0014] The application further provides an electronic device, comprising: A memory and a processor; The memory is used for storing a program, and the processor is used for executing steps in the above scene-adaptive intelligent home linkage control method when executing the program.

[0015] The application further provides a computer readable storage medium used for storing computer readable programs or instructions, which can realize steps in the above scene-adaptive intelligent home linkage control method when executed by a processor.

[0016] The above scheme has the following beneficial effects: The application provides a scene-adaptive smart home linkage control system, which obtains working data and sensing data of a smart home through a virtual sensing module, standardizes the working data and the sensing data to obtain virtual sensing data, obtains confidence levels of occurrence of a plurality of preset home unit events according to the virtual sensing data through an event analysis module, wherein the preset home unit event is a factor index used for combining to describe different home scenes of a user, obtains scene feature data according to the confidence levels of occurrence of the plurality of preset home unit events through a scene analysis module, and performs smart home linkage control according to the scene feature data through a linkage control module. The application breaks through the limitation of traditional dependence on special environmental sensors, and ingeniously standardizes working data generated by smart home equipment itself and limited environmental sensing data to convert the working data and the limited environmental sensing data into virtual sensing information capable of fully reflecting environmental states and user behaviors, so that ordinary smart equipment originally without environmental sensing capability is changed into an indirect provider of environmental information, thereby significantly expanding the overall sensing coverage of the system without increasing additional hardware costs. On this basis, the confidence level of the preset home unit event is introduced to evaluate the innovative mechanism, so that the system does not simply make a control decision directly based on the virtual sensing data, but firstly maps the data to a plurality of basic elements describing home scenes of the user, then analyzes the confidence level of occurrence of each event to effectively quantify data uncertainty, and then comprehensively analyzes the confidence levels of a plurality of events, so that the system can more accurately identify composite scene features, reduce misjudgment caused by virtual data as a sensing data source, and enable the system to make a relatively reliable linkage control decision through multi-source data fusion and probability reasoning. The linkage misalignment problem of the smart home caused by insufficient sensing data coverage in the prior art is solved. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A system architecture diagram of the scene-adaptive smart home linkage control system provided by the application is provided. Figure 2 A method flowchart of the scene-adaptive smart home linkage control method provided by the application is provided. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0019] In combination with Figure 1 a specific embodiment of the application, a scene-adaptive smart home linkage control system is disclosed, which comprises: The virtual sensor module 110 is configured to acquire working data and sensing data of the smart home, and standardize the working data and the sensing data to obtain virtual sensing data. The event analysis module 120 is configured to obtain confidence degrees of occurrence of a plurality of preset home unit events according to the virtual sensing data, wherein the preset home unit event is a factor index used to combine different home scenes of the user. The scene analysis module 130 is configured to obtain scene feature data according to the confidence degrees of occurrence of the plurality of preset home unit events. The linkage control module 140 is configured to perform smart home linkage control according to the scene feature data.

[0020] In the above process, the working data is data representing the running state of the smart home device itself, for example, the mode of the air conditioner, the set temperature, whether the light is on or not, whether the face recognition function of the door lock is on, and the like. The sensing data refers to limited sensing data that can be collected by the smart home device, for example, the temperature collected by the temperature sensor built in the air conditioner, the humidity collected by the humidity sensor built in the air conditioner, the door and window state collected by the magnetic sensor built in the smart door and window, and whether there is an obstruction collected by the infrared sensor of the smart door lock.

[0021] The virtual sensing data is data obtained by standardizing the working data and the sensing data and virtually representing the sensing data. It should be noted that the smart home in the embodiment includes not only the smart home device with a preset related hardware or software algorithm and capable of home linkage, but also any non-smart device capable of being acquired with control data and remotely controlled. For example, some old air conditioners have a temperature sensor built in, but the temperature data is not stored or sent out, but only displayed on the air conditioner screen, so that other devices in the home cannot acquire the temperature data collected by the air conditioner. However, the air conditioner can be remotely controlled by the mobile phone infrared, so that the indoor temperature can be indirectly inferred according to the remote control data of the air conditioner, and the inferred temperature is the virtual sensing data.

[0022] Some specific examples of virtual sensing data are given below: 1. Assuming that the user adjusts the brightness of the TV to 60% (higher than 40% when watching TV regularly), and plays a dark scene movie, it can be inferred that the user actively adjusts the screen brightness because of insufficient ambient light, so the ambient light level can be inferred according to the on-off state, screen brightness, volume, use time, and playing content (movie, dark scene).

[0023] 2、Assume that the refrigerator compressor runs continuously at high load (60% duration), and the user opens the door multiple times in a short period of time (causing cold air loss and frequent compressor startup) - combined with the summer environment, it is determined that the kitchen environment temperature is relatively high, indirectly providing the user with an "kitchen hot" environmental perception (such as starting the kitchen fan). Therefore, the kitchen environment temperature can be inferred according to the running mode of the refrigerator, the internal temperature (the refrigerator's built-in sensor can only measure the internal temperature), the number of door openings, and the compressor cooling duration ratio.

[0024] 3、Assume that the water heater is set to 55°C high temperature (far exceeding the daily hand washing requirement), and the remaining hot water quantity is sufficient to support bathing, combined with the user's "8 pm bathing" habit, it can be determined that the user will enter the bathroom soon, and the bathroom temperature needs to be raised in advance. Therefore, user behavior perception can be performed according to the heating state (continuous heating) of the intelligent electric water heater, the set temperature, the remaining hot water quantity, and the water use period.

[0025] 4、Assume that the user adjusts the mist level from 3 to 5 within 1 hour (indicating that the current mist level is not sufficient to alleviate dryness), and the humidifier runs continuously for 4 hours - combined with the characteristics of dry air in winter, it is determined that the "air dryness is high". Therefore, the air dryness level can be inferred according to the mist level, the running time, the remaining water quantity in the water tank, and other data of the humidifier device.

[0026] 5、Assume that the cleaning robot collides significantly more times in the living room carpet area than in other areas (indicating that there are more obstacles on the ground, such as dust and debris), and the dust box is nearly full within a short period of time - it is determined that the area is highly contaminated (such as the user's APP pushing a "living room carpet needs to be manually cleaned" reminder, or automatically booking the next deep cleaning). Therefore, the indoor dirt level can be inferred according to the cleaning state, the cleaning area, the number of collisions, and the dust box capacity.

[0027] 6、According to the opening record, time, and abnormal interception state of the intelligent door lock, it can be inferred whether the user member is at home.

[0028] 7、According to the opening degree of intelligent doors and windows, intelligent curtains, and other facilities, combined with the current time, the current light intensity can be determined.

[0029] It can be understood that the specific inference method of the virtual sensor data described above can be flexibly set according to actual conditions in combination with experience and tests. It is worth noting that it is inevitable to make errors by simply relying on virtual sensor data to infer, for example, the user pulling the curtain may indicate that the current sunlight is too strong, or it may indicate that the current is night and the user needs to rest. Therefore, on the one hand, the present application will combine a variety of virtual sensor data for unified analysis in subsequent scene analysis, and through cross-validation of multiple data sources, the ambiguity of a single signal can be eliminated, and the reliability of scene judgment can be greatly improved. On the other hand, the present application will further quantify the possibility of occurrence of each basic scene element (such as "need to block light", "night rest", "privacy protection", etc.) by "evaluating the confidence level of the occurrence of the preset home unit event", rather than making a binary judgment. This probabilistic processing mechanism can dynamically adjust the confidence weight of different events according to the reliability of each virtual sensor data, and finally through the comprehensive confidence level of all preset home unit events, the system can more accurately identify the composite scene characteristics, and avoid false triggering caused by single data anomaly or ambiguity. On this basis, the present application allows the inference process of virtual sensor data to have errors, greatly improving the practicality of the present application, so that the smart home system can still make accurate and reliable linkage control decisions under the condition of insufficient sensor coverage.

[0030] In the present application, the preset home unit event refers to a factor index that describes different home scenes of the user. These events are single, but through different combinations, complex actual scenes can be described. For example, the preset home unit event can include any artificially set unit event such as whether the user is at home, whether the door and window are open, whether the gas stove is used, whether the space temperature is too low, etc.

[0031] In a preferred embodiment, the preset home unit event includes a behavior unit event and an environment unit event, wherein the behavior unit event is used to represent the behavior of the user, and the environment unit event is used to represent the environmental state of the indoor. The following are some specific examples of preset home unit events: The behavior unit event includes: whether the user is at home, whether the user is in the kitchen / living room / bedroom, whether the user is working / resting / entertaining / eating, whether the user is resting, etc. broader events, or can be more detailed as whether the washing machine is used, whether the refrigerator is turned on, whether the music is played, etc. Indirectly describe the user's behavior events.

[0032] The environment unit event includes: the temperature level of the indoor, the humidity level, the ventilation level, the light level, the air quality level, the noise level, etc. Any event that can describe the indoor environment.

[0033] The embodiment uses virtual sensor data to comprehensively determine the credibility of the occurrence of the preset home unit event, to indirectly analyze the actual user scene in the room, forming a benign cycle of adaptive system: virtual sensor data makes up for the defects of insufficient number of physical sensors, and expands the range of the environment that the system can perceive; and the confidence evaluation mechanism ensures the reliability of decision-making under the condition of incomplete data by quantifying uncertainty. The combination of the two makes the method particularly suitable for the progressive smart home upgrading scene with incomplete sensor coverage and unreasonable layout, which not only avoids the cost and complexity of users being forced to carry out large-scale sensor reconstruction in pursuit of complete functions, but also significantly improves the environmental perception ability and linkage control precision of the existing smart home system under the condition of limited hardware, providing users with a more intelligent, reliable and economical and practical home experience.

[0034] Further, in a preferred embodiment, the step performed by the event analysis module 120 of obtaining the confidence of the occurrence of a plurality of preset home unit events according to the virtual sensor data specifically includes: obtaining operation data of the mobile phone; obtaining the confidence of the occurrence of a plurality of preset home unit events according to the operation data and the virtual sensor data.

[0035] In the embodiment, the operation data of the mobile phone refers to any data that can be used to describe the current mobile phone usage state of the user, such as location information (such as GPS positioning, Wi-Fi connection), time context (such as schedule, usage habit), interactive behavior (such as application usage record, screen operation mode), mobile phone running mode (standby, shutdown, screen power, charging, etc.

[0036] As one of the most important electronic products at present, the mobile phone can directly represent the user's behavior in most cases, because the mobile phone data often has stronger time continuity and behavior correlation, which can provide key context supplement and confidence calibration for virtual sensor data. This design fully utilizes the reality that users almost never leave their mobile phones, so that the system can obtain reliable data sources without additional user investment. Therefore, the embodiment ingeniously uses the characteristics of the mobile phone as the most direct and comprehensive digital record carrier of user behavior, and combines the operation data of the mobile phone and the virtual sensor data for comprehensive analysis, which not only greatly expands the data dimension that can be used for scene inference, so that the system can make up for the defects of insufficient home sensor coverage (such as determining whether the user is at home through the mobile phone GPS, and inferring the user's behavior intention through the application usage record) through the mobile phone data, but also significantly improves the accuracy and practicality of confidence analysis.

[0037] In addition, the embodiment can also utilize the mobile phone to perceive the direct interaction between the user and the home device through Bluetooth, NFC and other technologies, thereby providing a "data analysis base station" with high frequency, high precision and high coverage for the system. Even the system can be directly run in the mobile phone, which not only ensures the universality of the smart home linkage system in various home environments, but also greatly improves the practicality, user acceptance and actual deployment feasibility of the scheme by taking the mobile phone as the most familiar device for the user as the system platform.

[0038] It can be understood that, in the above process, the specific implementation mode of judging the confidence degree of each preset home unit event can also adopt any existing technology, such as mapping table, conditional judgment model or complex mathematical calculation model, which is established through experience or experiment, to combine and map different virtual sensing data into the probability of occurrence of different scenes. The present application provides some more preferred confidence degree analysis modes: Specifically, in the above process, the confidence degree of the occurrence of each preset home unit event is obtained according to the operation data and the virtual sensing data, which specifically includes: establishing a mobile phone operation vector according to the operation data; establishing a virtual sensing vector according to the virtual sensing data; splicing the mobile phone operation vector and the virtual sensing vector to obtain a sensing feature vector; inputting the sensing feature vector into a preset neural network model to obtain a confidence degree vector output by the neural network model, each element in the confidence degree vector corresponding to the confidence degree of the occurrence of one preset home unit event.

[0039] The embodiment converts the high-frequency and high-precision operation data generated by the mobile phone, which is the most commonly used device for the user, into a structured mobile phone operation vector, intelligently splices the mobile phone operation vector with the virtual sensing vector collected by the limited sensors and the running state of the home device, forms a sensing feature vector that comprehensively reflects the user's home situation, and then utilizes the powerful nonlinear mapping capability of the neural network to automatically learn and extract the complex potential correlation between these multi-source heterogeneous data, and accurately outputs the confidence degree of the occurrence of each preset home unit event. Through the end-to-end learning of the neural network, the complex correlation between the mobile phone operation and the home environment data, which is difficult for humans to explicitly define, can be automatically mined (such as the specific application usage mode combined with the change of the curtain state, which may indicate that the user is preparing to rest), avoiding the subjectivity and limitations of the manually designed rules in the traditional method. Moreover, the neural network model can be continuously optimized with the evolution of the user's usage habits, and through data-driven adaptive learning, the contribution weights of each data source are constantly adjusted, so that the system becomes more and more accurate.

[0040] In this embodiment, the input data to the neural network consists of sensor feature vectors including mobile phone operation data and virtual sensor data, while the output data is the confidence level of various preset home unit events. Clearly, obtaining both types of data samples is very convenient when training the neural network. Besides experiments, the input data can be directly obtained from user usage records, and actual sensor data can be used as virtual sensor data. Simultaneously, the output data can be directly represented using one-hot encoding based on whether an event has occurred, making the neural network model in this invention highly feasible and practical. Furthermore, the neural network model can be implemented using any existing neural network technology, such as a feedforward neural network model, an LSTM model capable of analyzing time-series data, or a graph neural network model capable of representing home spatial relationship information. Specifically, if a graph neural network is used as the neural network model in this invention, each room can be considered as a node, and the graph neural network can then be used to comprehensively analyze the sensor feature vectors of multiple rooms.

[0041] Based on the use of neural networks, in a preferred embodiment, the elements in the virtual sensing vector include a device type field and various virtual sensing parameter fields. The virtual sensing parameter fields are data fields directly obtained from the virtual sensing device, such as temperature values ​​or temperature levels, illuminance, etc. If a home device cannot obtain certain data, for example, a refrigerator cannot determine illuminance data, then this field can be left empty or represented by a specific value. In this embodiment, the elements in the virtual sensing vector include a device type field to represent a specific type. This embodiment, through a clearly defined device type field, enables the neural network to accurately identify the nature and functional boundaries of each data source device, understand the inherent data acquisition limitations of different devices, and allow the virtual sensing parameter field to be dynamically filled according to the actual capabilities of the device. Data fields not possessed by the device are set to null values ​​or specific markers, ensuring the integrity and consistency of the vector structure. In summary, this structured design of the virtual sensing vector in this embodiment allows the neural network to learn the inherent logic of what data different types of devices should provide, automatically adapting to differences in device configuration. Even in complex real-world environments with a mix of old and new devices, incomplete sensor coverage, and some devices lacking functionality, the system can still effectively utilize available data. By guiding the neural network to focus on the effective parameters of relevant devices through the device type field, and ignoring irrelevant or missing data fields, the model's generalization ability and robustness in heterogeneous device environments are significantly improved, providing crucial support for building a truly practical and implementable smart home linkage control system.

[0042] Further, in a preferred embodiment, one smart home generates a set of virtual sensor data at a time, one virtual sensor vector is established according to one set of virtual sensor data, and one virtual sensor vector is included in one sensor feature vector. In this technology, the steps performed by the scene analysis module 130 described above include: obtaining a confidence vector corresponding to each smart home; superimposing and normalizing all confidence vectors, and taking the obtained vector as scene feature data.

[0043] The design maintains the independence and originality of the data source by defining a one-to-one correspondence between smart homes, virtual sensor data, virtual vectors, and scene feature data, and by converting the virtual sensor data independently generated by each smart home device into a corresponding confidence vector, avoiding mutual interference between different devices in terms of data dimension and magnitude. On this basis, by scientifically superimposing and normalizing the confidence vectors generated by all smart home devices, the system can intelligently integrate multi-source heterogeneous information from different locations and different functional devices in the whole house, forming unified scene feature data reflecting the overall home environment and user state, which not only retains the unique contribution of each device, but also eliminates the bias caused by the difference in the number of devices through normalization, ensuring that even in the case of dense devices in some areas or sparse devices in some areas, the system can still objectively evaluate the comprehensive occurrence probability of each preset home unit event.

[0044] Compared with the way of first aggregating all smart home data and then establishing a scene feature vector for one-time analysis, this analysis method of divide and conquer and then aggregation respects the professionalism and limitations of individual devices and achieves a collaborative understanding of the whole house scene, enabling the system to accurately identify data that contains both local details and overall context, providing a detailed and balanced decision basis for subsequent smart home linkage control, and significantly improving the accuracy and practicality of scene judgment in complex home environments.

[0045] When there is only one mobile phone in the room, the mobile phone can become a platform carrier running the above system. When there are multiple mobile phones in the room, each mobile phone will obtain individual scene feature data based on its own operation data and obtain different control strategies, which may cause conflicts. Therefore, the present application also provides a preferred embodiment for avoiding command conflicts when there are multiple mobile phones in the room.

[0046] Specifically, in the present embodiment, the step of superimposing and normalizing all confidence vectors to obtain the vector as scene feature data includes: acquire the distance between the mobile phone and the smart home corresponding to the confidence vector, and establish the superposition weight of the confidence vector based on the distance; superpose and normalize all the confidence vectors based on the superposition weight, and obtain the vector as the scene feature data.

[0047] The embodiment can run the above system through a mobile phone or through a specified separate computing device (such as a router, a computer, etc.). By introducing the physical distance between the mobile phone and the smart home as a key weight factor, the system can intelligently identify and evaluate the spatial correlation between different users and the current home device, so that the confidence vector provided by the mobile phone closer to the target device obtains a higher weight (for example, the judgment weight of the user's mobile phone in the living room on the lighting control scene in the living room is higher), and the contribution of the mobile phone far away is correspondingly reduced (for example, the influence weight of the user's mobile phone in the bedroom on the living room scene is smaller), thereby effectively avoiding misjudgment caused by multiple users operating at the same time or different location user demand conflicts. At the same time, the dynamic weight distribution mechanism based on distance respects the natural correlation between the actual spatial position and the behavior intention of the user, so that the system can automatically adapt to the differentiated needs of family members in different areas. Even in the case of multiple users at home at the same time and each carrying a mobile phone, the system can still accurately identify the complex situation of different areas requiring different needs. In addition, through the normalization processing after superposition, it is ensured that the scene feature data under different weight contributions still maintains a unified numerical range and comparability, providing a reliable decision basis for subsequent smart home linkage control that considers multiple user needs and avoids command conflicts, significantly improving the accuracy, user satisfaction and actual availability of the smart home system response in a multi-mobile phone environment.

[0048] Further, in a preferred embodiment, the steps performed by the linkage control module 140 described above include: acquiring target scene feature data and a set of preset standard control strategies, the set of standard control strategies including a plurality of preset standard scene feature data and corresponding control strategies; matching a plurality of standard scene feature data closest to the scene feature data from the set of standard control strategies; comprehensively obtaining a target control strategy by combining the control strategies corresponding to the plurality of standard scene feature data closest to the scene feature data; based on the target control strategy, performing linkage control on the smart home actually existing in the target control strategy.

[0049] The standard control strategy set can be obtained through experiments or experience, and the control strategies in the strategy set can include preset scheme data such as adjustment amounts of various devices. It can be understood that after the confidence vector is obtained, the linkage control strategy of the smart home can also be obtained through a neural network or the like. The embodiment introduces fuzzy matching and strategy fusion technology, so that the system can intelligently identify multiple standard scene features that are most similar to the current complex home situation, rather than demanding a completely consistent match, thereby effectively solving the common problem that scene features in an actual home environment are often between multiple preset standards. By synthesizing the control strategies corresponding to multiple similar standard scenes, the system can absorb the advantage elements of each strategy to generate an optimal target control strategy that takes into account multiple possibilities, which not only maintains the rationality of the control logic, but also has flexibility in dealing with scene ambiguity. In addition, the final execution stage only performs linkage control on the smart home devices actually existing in the user's home, automatically ignores device-related instructions included in the standard strategy but not deployed by the user, and avoids control failure or incorrect operation due to device configuration differences. The embodiment enables the smart home system to provide a reliable control basis based on the standardized strategy library, and to flexibly adapt to different actual home environments and personalized needs, significantly improving the accuracy of control decisions, device compatibility and user satisfaction, and providing a smart and practical linkage control solution for complex and variable home scenes.

[0050] The above is a computer program product provided by the present application, and the present application also provides a scene-adaptive smart home linkage control method, as shown in the accompanying drawings. Figure 2 The present application also provides a scene-adaptive smart home linkage control method, comprising: S201, obtaining working data and sensing data of a smart home, and standardizing the working data and the sensing data to obtain virtual sensing data; S202, obtaining confidence degrees of occurrence of a plurality of preset home unit events according to the virtual sensing data, wherein the preset home unit event is a factor index used to describe different home scenes of a user; S203, obtaining scene feature data according to the confidence degrees of occurrence of the plurality of preset home unit events; S204, performing linkage control of the smart home according to the scene feature data.

[0051] It can be understood that the implementation principle and advantages of the above method are the same as those of the scene-adaptive smart home linkage control system, which can be understood by those skilled in the art, and therefore will not be described in detail herein.

[0052] The present application also provides an electronic device, comprising: a memory and a processor; The memory is configured to store a program, and the processor is configured to execute the steps of the scene-oriented adaptive smart home linkage control method when executing the program.

[0053] The application further provides a computer readable storage medium configured to store a computer readable program or instruction, which can realize the steps of the scene-oriented adaptive smart home linkage control method when executed by a processor.

[0054] The application provides a scene-oriented adaptive smart home linkage control system, which obtains working data and sensing data of a smart home through a virtual sensing module, and standardizes the working data and the sensing data to obtain virtual sensing data, obtains confidence levels of occurrence of a plurality of preset home unit events according to the virtual sensing data through an event analysis module, wherein the preset home unit event is a factor index used for combining description of different home scenes of a user, obtains scene feature data according to the confidence levels of occurrence of the plurality of preset home unit events through a scene analysis module, and performs smart home linkage control according to the scene feature data through a linkage control module. The application breaks through the limitation of traditional dependence on special environmental sensors, and ingeniously standardizes working data generated by smart home devices themselves and limited environmental sensing data to convert the working data and the environmental sensing data into virtual sensing information capable of fully reflecting environmental states and user behaviors, so that ordinary smart devices originally without environmental sensing capability are changed into indirect providers of environmental information, thereby significantly expanding the overall sensing coverage of the system without increasing additional hardware costs. On this basis, the confidence level evaluation of the preset home unit event is introduced to make the system not simply make a control decision directly based on the virtual sensing data, but first map the data to a plurality of basic elements describing home scenes of a user, then analyze the confidence levels of occurrence of each event to effectively quantify data uncertainty, and then integrate the confidence levels of a plurality of events, so that the system can more accurately identify composite scene features, reduce misjudgment caused by virtual data as a sensing data source, and make a relatively reliable linkage control decision through multi-source data fusion and probability reasoning. The linkage misalignment problem of the smart home caused by insufficient sensing data coverage in the prior art is solved.

[0055] It should be noted that each embodiment in the present specification is described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the same and similar parts of each embodiment can be referred to each other.

[0056] The foregoing description of the disclosed embodiments enables a person skilled in the art to make or use the application. Modifications of these embodiments will occur to persons of skill in the art, and that the appended claims are intended to cover all such modifications that do not depart from the true spirit and scope of the application. Therefore, the application is not limited to the embodiments shown but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A scene-oriented adaptive smart home linkage control system, characterized in that, The method comprises the following steps: a virtual sensor module is used to obtain working data and sensing data of the smart home, and the working data and the sensing data are standardized to obtain virtual sensing data; an event analysis module is used to obtain confidence degrees of occurrence of a plurality of preset home unit events according to the virtual sensing data, wherein the preset home unit events are factor indexes used to combine to describe different home scenes of the user; a scene analysis module is used to obtain scene feature data according to the confidence degrees of occurrence of the plurality of preset home unit events; a linkage control module is used to perform linkage control of the smart home according to the scene feature data. 2.The scene-oriented adaptive smart home linkage control system according to claim 1, wherein, The confidence degrees of occurrence of the plurality of preset home unit events are obtained according to the virtual sensing data, comprising: operation data of a mobile phone is obtained; the confidence degrees of occurrence of the plurality of preset home unit events are obtained according to the operation data and the virtual sensing data. 3.The scene-oriented adaptive smart home linkage control system according to claim 2, wherein, The confidence degrees of occurrence of the plurality of preset home unit events are obtained according to the operation data and the virtual sensing data, comprising: a mobile phone operation vector is established according to the operation data; a virtual sensor vector is established according to the virtual sensing data; a sensing feature vector is obtained by splicing the mobile phone operation vector and the virtual sensor vector; the sensing feature vector is input into a preset neural network model to obtain a confidence degree vector output by the neural network model, and each element in the confidence degree vector corresponds to a confidence degree of occurrence of a preset home unit event. 4.The scene-oriented adaptive smart home linkage control system according to claim 3, wherein, One virtual sensing data is generated by one smart home, one virtual sensor vector is established according to one virtual sensing data, and one sensing feature vector only includes one virtual sensor vector; The scene feature data is obtained according to the confidence degrees of occurrence of the plurality of preset home unit events, comprising: confidence degree vectors corresponding to each smart home are obtained; all the confidence degree vectors are superimposed and normalized, and the obtained vector is taken as the scene feature data. 5.The scene-oriented adaptive smart home linkage control system according to claim 4, wherein, The scene feature data is obtained by superimposing and normalizing all the confidence degree vectors, comprising: a distance between a mobile phone corresponding to the confidence degree vector and a smart home is obtained, and a superposition weight of the confidence degree vector is established based on the distance; all the confidence degree vectors are superimposed and normalized based on the superposition weight, and the obtained vector is taken as the scene feature data. 6.The scene-oriented adaptive smart home linkage control system according to claim 3, wherein, Elements in the virtual sensor vector include a device type field and a plurality of virtual sensing parameter fields; the preset home unit events include behavior unit events and environment unit events, wherein the behavior unit events are used to represent behaviors of the user, and the environment unit events are used to represent environmental states of the indoor. 7.The scene-oriented adaptive smart home linkage control system according to claim 1, wherein, The smart home linkage control is performed according to the scene feature data, comprising: target scene feature data and a preset standard control strategy set are obtained, the standard control strategy set includes a plurality of preset standard scene feature data and corresponding control strategies; a plurality of standard scene feature data closest to the scene feature data are matched from the standard control strategy set; a target control strategy is obtained by comprehensively combining the control strategies corresponding to the plurality of standard scene feature data closest to the scene feature data; the smart home actually existing in the target control strategy is linkage controlled based on the target control strategy. 8.A scene-oriented adaptive smart home linkage control method, characterized in that, The method comprises the following steps: Acquire working data and sensing data of the smart home, standardize the working data and the sensing data, and obtain virtual sensing data; According to the virtual sensing data, obtain the confidence degree of occurrence of a plurality of preset home unit events, wherein the preset home unit event is a factor index used to combine different home scenes of a user; According to the confidence degree of occurrence of the plurality of preset home unit events, obtain scene feature data; According to the scene feature data, perform smart home linkage control.

9. An electronic device, comprising: Comprise: A memory and a processor; The memory is used to store a program, and the processor is used to execute the steps in the scene-adaptive smart home linkage control method of claim 8 when executing the program.

10. A computer-readable storage medium, characterized in that, A computer readable program or instruction for storing, which can realize the steps in the scene-adaptive smart home linkage control method of claim 8 when executed by a processor.

Images