Smart home decision-making method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a smart home decision-making method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring equipment information and user information; establishing a local learning model for local intelligent equipment; the equipment information and the user information serve as input information, a local learning model is adopted for training, and initial training data are obtained; controlling the local intelligent device to send the initial training data to the cloud intelligent decision assistant; the cloud intelligent decision assistant is configured to generate decision training data based on the initial training data and the user information; using the cloud smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; the smart home decision-making result is used for controlling the local smart devices, so that decision-making control for multiple local smart devices is realized, the effectiveness and rationality of smart home decision making are improved, and the user experience is further improved.

Description

Technical Field

[0001] The present invention relates to the field of smart home decision-making technology, and in particular to a smart home decision-making method, a smart home decision-making device, an electronic device and a computer-readable storage medium. Background Art

[0002] When multiple smart devices respond at the same time, how to handle priorities is a difficult problem. Currently, many smart home devices can optimize models based on user habits or other information to make the devices more in line with user habits. However, in a home with multiple devices, other smart devices, especially those that are less used by users, often lack data for optimization models, making it impossible to make effective decisions and controls for multi-smart device scenarios.

[0003] Therefore, how to make control decisions for smart devices is a technical problem that technicians in this field need to overcome. Summary of the invention

[0004] The embodiments of the present invention provide a smart home decision-making method, device, electronic device and computer-readable storage medium to solve the problem of how to make control decisions for smart home devices.

[0005] An embodiment of the present invention discloses a smart home decision-making method, which is applied to a local smart device, wherein the local smart device is configured with a corresponding cloud-based smart decision-making assistant, and the method may include:

[0006] Get device information and user information;

[0007] Build local learning models for local smart devices;

[0008] Taking the device information and the user information as input information, and using the local learning model for training to obtain initial training data;

[0009] Controlling the local smart device to send the initial training data to the cloud-based smart decision assistant;

[0010] The cloud-based intelligent decision assistant is configured as follows:

[0011] generating decision training data based on the initial training data and the user information;

[0012] The cloud-based smart home decision assistant is used to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

[0013] Optionally, the local smart device is a smart lighting device, the user information includes user location information, the device information includes current time information, ambient light intensity information and brightness information, and the step of using the device information and the user information as input information and using the local learning model for training to obtain initial training data may include:

[0014] Determine the initial weights and initial biases;

[0015] Determining initial predicted brightness information based on the user location information, the current time information, the initial weight, the initial bias, and the ambient light intensity information;

[0016] Determining an update weight and an update bias for the intelligent lighting device based on the initial predicted brightness information, the brightness information, the initial weight and the initial bias;

[0017] Determining updated predicted brightness information based on the updated weight and the updated bias;

[0018] The updated predicted brightness information is used as initial training data.

[0019] Optionally, the step of generating decision training data based on the initial training data and the user information may include:

[0020] Establish a cloud-based federated learning model;

[0021] Determine a target smart device from a plurality of local smart devices;

[0022] The initial training data of the plurality of local smart devices are used as input data of the cloud-based joint learning model to generate decision training data for the target smart device.

[0023] Optionally, it may also include:

[0024] The cloud-based intelligent decision assistant is used to update the local learning model, and the updated local learning model is sent to the corresponding local smart device.

[0025] Optionally, the smart home decision result may include adjusting a lighting angle and / or adjusting lighting brightness.

[0026] Optionally, the step of determining the initial predicted brightness information based on the user location information, the current time information, the initial weight, the initial bias and the ambient light intensity information may include:

[0027] The user location information, the current time information, the initial weight, the initial bias and the ambient light intensity information are input into Formula 1, which is:

[0028]

[0029] in, is the initial predicted brightness information, x1 is the user location information, x2 is the current time information, x3 is the ambient light intensity information, w1, w2 and w3 are initial weights, and b is the initial bias.

[0030] Optionally, the step of determining an update weight and an update bias for the intelligent lighting device based on the initial predicted brightness information, the brightness information, the initial weight and the initial bias may include:

[0031] The initial predicted brightness information, the brightness information, the initial weight and the initial bias are input into Formula 2, which is:

[0032]

[0033] Wherein, MSE is the mean square error, n is the total number of information of the initial predicted brightness information, y i is the ith brightness information, is the i-th initial predicted brightness information, w1, w2 and w3 are initial weights, and b is the initial bias.

[0034] Optionally, the step of using the initial training data of the plurality of local smart devices as input data of the cloud-based joint learning model to generate decision training data for the target smart device may include:

[0035] Determine the number of local smart devices;

[0036] The number of devices and the initial training data are input into Formula 3, which is:

[0037]

[0038] Among them, θ new is the decision training data, N is the number of devices, θ i is the initial training data corresponding to the i-th local smart device.

[0039] The embodiment of the present invention also discloses a smart home decision method, which is applied to a cloud-based smart decision assistant, wherein the cloud-based smart decision assistant is configured with a corresponding local smart device, and the local smart device is configured to obtain device information and user information; establish a local learning model for the local smart device; use the device information and the user information as input information, and use the local learning model for training to obtain initial training data; control the local smart device to send the initial training data to the cloud-based smart decision assistant; the method may include:

[0040] generating decision training data based on the initial training data and the user information;

[0041] The cloud-based smart home decision assistant is used to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

[0042] The embodiment of the present invention further discloses a smart home decision-making device, which is applied to a local smart device, and the local smart device is configured with a corresponding cloud-based smart decision-making assistant. The device may include:

[0043] User information acquisition module, used to obtain device information and user information;

[0044] A local learning model building module is used to build a local learning model for local smart devices;

[0045] An initial training data generating module, used to use the device information and the user information as input information, and adopt the local learning model for training to obtain initial training data;

[0046] An initial training data sending module, used to control the local smart device to send the initial training data to the cloud-based smart decision assistant;

[0047] The cloud-based intelligent decision assistant is configured to: generate decision training data based on the initial training data and the user information; use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; and the smart home decision result is used to control local smart devices.

[0048] The embodiment of the present invention further discloses a smart home decision-making device, which is applied to a cloud-based smart decision-making assistant, wherein the cloud-based smart decision-making assistant is configured with a corresponding local smart device, and the local smart device is configured to obtain device information and user information; establish a local learning model for the local smart device; use the device information and the user information as input information, use the local learning model for training, and obtain initial training data; control the local smart device to send the initial training data to the cloud-based smart decision-making assistant; the device may include:

[0049] A decision training data generating module, used to generate decision training data based on the initial training data and the user information;

[0050] A smart home decision result generation module is used to use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

[0051] The embodiment of the present invention further discloses a smart home decision system, which may include a local smart device and a cloud-based smart decision assistant;

[0052] The local smart device is configured to obtain device information and user information;

[0053] Build local learning models for local smart devices;

[0054] Taking the device information and the user information as input information, and using the local learning model for training to obtain initial training data;

[0055] Controlling the local smart device to send the initial training data to the cloud-based smart decision assistant;

[0056] The cloud-based intelligent decision assistant is configured as follows:

[0057] generating decision training data based on the initial training data and the user information;

[0058] The cloud-based smart home decision assistant is used to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

[0059] The embodiment of the present invention further discloses an electronic device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0060] The memory is used to store computer programs;

[0061] The processor is used to implement the method described in the embodiment of the present invention when executing the program stored in the memory.

[0062] The embodiment of the present invention further discloses a computer-readable storage medium on which instructions are stored. When executed by one or more processors, the processors are enabled to execute the method described in the embodiment of the present invention.

[0063] The embodiments of the present invention include the following advantages:

[0064] In an embodiment of the present invention, device information and user information are obtained; a local learning model for a local smart device is established; the device information and the user information are used as input information, and the local learning model is used for training to obtain initial training data; the local smart device is controlled to send the initial training data to the cloud-based smart decision assistant; the cloud-based smart decision assistant is configured to: generate decision training data based on the initial training data and the user information; use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control the local smart device, thereby realizing decision control for multiple local smart devices, improving the effectiveness and rationality of smart home decisions, and further improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 is a flowchart of a smart home decision-making method provided in Embodiment 1 of the present invention;

[0066] Figure 2 This is a system architecture diagram provided in Embodiment 1 of the present invention;

[0067] Figure 3 is a flowchart of another smart home decision-making method provided in Embodiment 1 of the present invention;

[0068] Figure 4 is a flowchart of the steps of a smart home decision-making method provided in Embodiment 2 of the present invention;

[0069] Figure 5 is a structural block diagram of a smart home decision-making device provided in Embodiment 3 of the present invention;

[0070] Figure 6 is a structural block diagram of a smart home decision-making device provided in Embodiment 4 of the present invention;

[0071] Figure 7 It is a hardware structure block diagram of an electronic device provided in each embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0073] In practical applications, when multiple smart home devices need to be controlled simultaneously in a scenario where multiple smart home devices work together, how to reasonably allocate priorities is a complex and challenging problem. At present, many smart home devices have been able to build personalized user models based on user historical interaction data, environmental information, etc. through machine learning and other technologies, so as to achieve accurate prediction of user habits and optimization of device behavior. However, in a home environment where multiple devices coexist, there are significant differences in the frequency of use of different devices by users. For those devices with low usage frequency, due to the lack of sufficient interaction data, the construction and optimization of their personalized models often face problems such as data sparsity and insufficient model generalization ability, which puts the device at a disadvantage when competing for response resources and makes it difficult to provide a satisfactory service experience for users. The embodiment of the present invention provides a smart home decision-making method, which combines user information and device information to make decisions and control smart devices, so as to improve the effectiveness and rationality of smart home decision control and further enhance the user experience.

[0074] Embodiment 1

[0075] Reference Figure 1 , shows a flowchart of a smart home decision-making method provided in Embodiment 1 of the present invention, which may specifically include the following steps:

[0076] Step 101, obtaining device information and user information;

[0077] Step 102, establishing a local learning model for local smart devices;

[0078] Step 103, using the device information and the user information as input information, and using the local learning model for training to obtain initial training data;

[0079] Step 104, controlling the local smart device to send the initial training data to the cloud-based intelligent decision assistant; the cloud-based intelligent decision assistant is configured to: generate decision training data based on the initial training data and the user information; use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control the local smart device.

[0080] In practical applications, the embodiments of the present invention can be applied to local smart devices. For example, the local smart devices can be smart home equipment configured locally on the user's site and connected via a network to achieve remote control, automation, and personalized services, such as smart speakers, smart light bulbs, smart door locks, smart air conditioners, etc.

[0081] In a specific implementation, the embodiment of the present invention can obtain device information and user information; establish a local learning model for local smart devices; use the device information and user information as input information, use the local learning model for training, and obtain initial training data; control the local smart device to send the initial training data to the cloud-based smart decision assistant; the cloud-based smart decision assistant can be configured to generate decision training data based on the initial training data and user information; use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and user information; the smart home decision result is used to control the local smart device. For example, device information and user information can be obtained, and user information can be obtained by setting a smart sensor device. The user information may include user location information and user behavior information. The smart sensor device can collect various data in the user's life through various sensors, thereby indirectly or directly obtaining user information. The smart sensor device can have a device that senses environmental data, human body data, and device interaction data. It can directly obtain the user's physiological characteristics through biometric recognition, such as fingerprint, facial recognition, iris recognition, etc., and can also be through voice recognition, for example, through voice interaction to obtain the user's instructions and information, and can also be through location information, for example, through GPS (Global Positioning System, Global Positioning System) or Wi-Fi (WIreless Fidelity, mobile hotspot) positioning to obtain user location information, the smart sensor device can be a wearable device, a local smart device with a sensor function, etc., the device information can be the environmental data of the local smart device, the setting parameters of the local smart device itself, etc.

[0082] Then, a local learning model targeting local smart devices can be built.

[0083] Preferably, a suitable learning model can be selected for each local smart device. When a suitable learning model is selected for each local smart device, machine learning models can be customized for different devices to achieve more accurate and efficient intelligent functions. This involves comprehensive consideration of device hardware, data volume, task complexity, and model performance requirements. The selection can be based on the hardware limitations of the local smart device, such as the computing power, memory, and storage space of the local smart device; the data volume, data quality, task complexity, and model performance requirements of the local smart device. For example, for image classification and sentiment analysis, models such as support vector machines, random forests, and convolutional neural networks can be selected. For prediction and fitting, models such as linear regression, decision tree regression, and support vector regression can be selected.

[0084] Of course, the above is only an example, and those skilled in the art may adopt other models, including but not limited to recurrent neural networks, long short-term memory networks and other models as local learning models of local smart devices, and the embodiments of the present invention do not impose any restrictions on this.

[0085] Subsequent data processing can be performed locally by establishing a local learning model suitable for local devices.

[0086] The device information and the user information can be used as input information, and the local learning model can be used for training to obtain initial training data. Preferably, before training the local learning model, the model parameters corresponding to each local smart device can be initialized so that each local smart device can obtain the initial user information. Then, the device information including various information of the device (such as device model, operating status, sensor data, etc.) and user information related to the user (such as user preferences, usage habits, personal information, etc.) can be used as input, and the local learning model can be used for training on the local device, so that each device can achieve independent model training. Furthermore, different local smart devices can be based on their corresponding fields. For example, a smart air conditioner can optimize parameters such as temperature and humidity according to the user's location, and a smart light bulb can optimize parameters such as lighting area, light intensity and angle according to the user's location, orientation and behavior, etc. These optimized parameters are used as initial training data, and then the local smart device can be controlled to send the initial training data to the cloud smart Decision assistant. The cloud-based intelligent decision assistant can be configured in the cloud. It is an intelligent tool based on cloud computing technology. It can perform large amounts of data analysis and provide decision suggestions. It has functions such as data collection and integration, data analysis, prediction and simulation, and decision suggestions. The cloud-based intelligent decision assistant can be configured to generate decision training data based on initial training data and user information, and generate smart home decision results based on decision training data and user information, and control local smart devices according to the smart home decision results. For example, the cloud-based intelligent decision assistant can receive initial training data uploaded by multiple local smart devices. When the initial training data uploaded by the smart curtain device includes data such as the sunlight angle, brightness, ultraviolet intensity, and glass temperature outside the window, and the smart human body radar sensor obtains data such as the user's current position coordinates and orientation as user information, cloud computing can be performed based on the above initial training data and user information. The training model obtains the lighting angle and brightness for the local smart bulb as the smart home decision result, and controls the smart bulb to adjust the lighting angle and brightness according to the smart home decision result.

[0087] Preferably, after controlling the local smart device according to the smart home decision result once, continuous decision control can continue according to the updated user information.

[0088] Preferably, different local smart devices can send initial training data to the cloud-based intelligent decision-making assistant for continuous updating. The intelligent decision-making assistant integrates the initial training data and can synchronize all connected local smart devices, thereby updating the models and parameters between each local smart device while protecting user privacy and not sharing original data.

[0089] In an embodiment of the present invention, device information and user information are obtained; a local learning model for a local smart device is established; the device information and the user information are used as input information, and the local learning model is used for training to obtain initial training data; the local smart device is controlled to send the initial training data to the cloud-based smart decision assistant; the cloud-based smart decision assistant is configured to: generate decision training data based on the initial training data and the user information; use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control the local smart device, thereby realizing decision control for multiple local smart devices, improving the effectiveness and rationality of smart home decisions, and further improving the user experience.

[0090] Based on the above embodiment, a variant embodiment of the above embodiment is proposed. It should be noted that in order to make the description concise, only the differences from the above embodiment are described in the variant embodiment.

[0091] In an optional embodiment of the present invention, the step of using the device information and the user information as input information and using the local learning model for training to obtain initial training data includes:

[0092] Determine the initial weights and initial biases;

[0093] Determining initial predicted brightness information based on the user location information, the current time information, the initial weight, the initial bias, and the ambient light intensity information;

[0094] Determining an update weight and an update bias for the intelligent lighting device based on the initial predicted brightness information, the brightness information, the initial weight and the initial bias;

[0095] Determining updated predicted brightness information based on the updated weight and the updated bias;

[0096] The updated predicted brightness information is used as initial training data.

[0097] In actual applications, the local smart device in the embodiment of the present invention is a smart lighting device, the user information includes user location information, and the device information includes current time information, ambient light intensity information and brightness information. By way of example, the smart lighting device can be a lighting device that combines traditional lighting with Internet of Things technology and realizes remote control, automation, personalization and other functions through network connection. Compared with traditional lighting devices, smart lighting devices are more interactive and intelligent, and can provide a more comfortable and convenient lighting experience, such as smart bulbs, smart light strips and smart desk lamps. User information may include user location information, for example, the user's position in the local smart lighting device layout scene, and the device information may include current time information, ambient light intensity information and brightness information, wherein the current time information may be the current time, which may be accurate to seconds or minutes, and the brightness information may be the brightness of the smart lighting device itself, which may be expressed in lumens (the international unit of luminous flux).

[0098] In a specific implementation, the embodiment of the present invention can determine an initial weight and an initial bias; determine initial predicted brightness information based on user location information, current time information, and ambient light intensity information; determine an updated weight and an updated bias for the smart lighting device based on the initial predicted brightness information, brightness information, initial weight, and initial bias; determine updated predicted brightness information based on the updated weight and updated bias; use the updated predicted brightness information as initial training data. Exemplarily, when the local smart device is a smart lighting device, device information and user information can be obtained through sensors equipped by the smart lighting device. Preferably, the initial training data of other local smart devices shared by the cloud intelligent decision assistant can also be used to extract relevant data as user information and device information. The user information may include user location information. The device information may include current time information, ambient light intensity information and brightness information. Then, the initial weight and initial bias may be determined, wherein the bias information may refer to an adjustment value relative to a default setting, and may be an adjustment value of corresponding parameters such as brightness, color temperature, and hue. The ambient light intensity information may be the light intensity of the surrounding environment, and lux (1ux, which is the unit of illumination) may be used as a unit. The initial weight, initial bias, current time information, and ambient light intensity information may be input into a local learning model of the smart lighting device for training. For example, a linear regression model may be selected as a local learning model for generating initial predicted brightness information, thereby determining the initial predicted brightness information. Then, the initial predicted brightness information, brightness information, initial weight, and initial bias may be used as input data. For example, MSE (Mean Score) may be used. Squared Error, mean square error) loss function model is used as a local learning model for intelligent lighting devices to determine the update weights and update biases. When the model converges, the newly obtained weight values ​​and bias values ​​can be used as the updated weights and updated biases. Further, the updated weights and updated biases are re-input into the local learning model that generates the initial predicted brightness information, and the obtained values ​​are used as the updated predicted brightness information. The updated predicted brightness information is closer to the actual value and can be used as initial training data for subsequent data processing processes.

[0099] Preferably, the user information may also include user preferences, such as the brightness and light color that the user likes, and may also include the type of user activity, such as reading, resting or working information. Of course, the above is only an example, and those skilled in the art may use other types of data as user information, including but not limited to user age, user behavior habits, etc., and the embodiments of the present invention do not impose any limitations on this.

[0100] The embodiments of the present invention determine the initial weight and the initial bias; determine the initial predicted brightness information based on the user location information, the current time information, the initial weight, the initial bias and the ambient light intensity information; determine the updated weight and the updated bias for the smart lighting device based on the initial predicted brightness information, the brightness information, the initial weight and the initial bias; determine the updated predicted brightness information based on the updated weight and the updated bias; and use the updated predicted brightness information as initial training data, thereby realizing local user data processing, protecting user privacy, and improving the accuracy of smart home decision-making.

[0101] In an optional embodiment of the present invention, the step of generating decision training data based on the initial training data and the user information includes:

[0102] Establish a cloud-based federated learning model;

[0103] Determine a target smart device from a plurality of local smart devices;

[0104] The initial training data of the plurality of local smart devices are used as input data of the cloud-based joint learning model to generate decision training data for the target smart device.

[0105] In practical applications, the cloud-based federated learning model can be a distributed machine learning technology, or a dynamic federated learning model, which allows multiple organizations or individuals to jointly train a machine learning model without sharing the original data, thereby improving the protection of data privacy.

[0106] In a specific implementation, an embodiment of the present invention can establish a cloud-based joint learning model; determine a target smart device from multiple local smart devices; use the initial training data of multiple local smart devices as input data of the cloud-based joint learning model to generate decision training data for the target smart device. By way of example, a cloud-based joint learning model can be established to determine a target smart device from multiple local smart devices. For example, when the local smart devices include smart curtains and smart light bulbs, the smart light bulbs can be used as target smart devices. When the cloud-based smart home decision assistant receives the initial training data sent by the local smart device, the initial training data can be used as input data of the cloud-based joint learning model for training to generate decision training data for the smart light bulbs, which is used to control the target smart device to adjust parameters such as brightness and lighting angle. For example, the smart light bulbs can be preferentially controlled to lower the brightness based on the initial training data generated by the smart curtains according to sunlight intensity and user position.

[0107] The embodiment of the present invention establishes a cloud-based joint learning model; determines a target smart device from a plurality of local smart devices; uses the initial training data of the plurality of local smart devices as input data of the cloud-based joint learning model to generate decision training data for the target smart device, thereby aggregating the parameters uploaded by the plurality of local smart devices for decision control of a specific smart device, thereby further improving the accuracy of smart home decisions for the plurality of smart devices.

[0108] In an optional embodiment of the present invention, it also includes:

[0109] The cloud-based intelligent decision assistant is used to update the local learning model, and the updated local learning model is sent to the corresponding local smart device.

[0110] In practical applications, local model training has the advantage of strong real-time performance, while cloud-based model computing has the advantages of strong computing power and sufficient storage space. Therefore, complex calculations can be completed in the cloud, while the local model can be continuously updated to improve the accuracy and robustness of specific tasks.

[0111] In a specific implementation, in an embodiment of the present invention, a cloud-based intelligent decision assistant can be used to update a local learning model, and the updated local learning model can be sent to a corresponding local smart device. For example, the cloud-based intelligent decision assistant can collect feedback data, environmental data, etc. of each local smart device for model training and evaluation, and use a machine learning algorithm to train a new model based on the collected data so that it can better complete a specific task, and evaluate the trained model to ensure that its performance meets the requirements. Preferably, the trained model can be optimized, such as pruning, quantization, etc., to reduce the model size and improve the reasoning efficiency, and a compatibility check can be performed to ensure that the optimized model is compatible with the hardware and software of the local device. A unique version number is assigned to each model for easy management and backtracking, and the training parameters, evaluation results, and other information of each model are recorded. The optimized model is then packaged into a deployable package, which contains model files, configuration files, etc., and is transmitted through a secure and reliable communication protocol (such as HTTPS, i.e., Hypertext Transfer Protocol). ProtocolSecure, Hypertext Transfer Protocol) sends the model package to the device to ensure that the model package is sent to the correct device. The local smart device can connect to the cloud through the network, receive the model update instruction, download the new model package, replace the old model with the downloaded model, update the device configuration file to adapt to the new model, and the local smart device loads the new model into the memory. Preferably, a small amount of local data can be used to verify the new model to ensure that it works properly. If the verification is passed, the device starts to use the new model for training. Furthermore, the local smart device can feed back the performance data after using the new model to the cloud-based intelligent decision-making assistant. The cloud further optimizes the model based on the feedback data to form a closed-loop iterative process.

[0112] In the embodiment of the present invention, the local learning model is updated by the cloud-based intelligent decision assistant, and the updated local learning model is sent to the corresponding local smart device, so that the cloud-based intelligent decision assistant can continuously update the local learning model, thereby improving the intelligence level of the local smart device, further improving the accuracy of the results, and enhancing the adaptability, robustness and security of the system.

[0113] In an optional embodiment of the present invention, the smart home decision result includes adjusting the lighting angle and / or adjusting the lighting brightness.

[0114] In a specific implementation, the smart home decision result in the embodiment of the present invention includes adjusting the lighting angle and / or adjusting the lighting brightness. For example, when the local smart device is a smart lighting device, the user information may include user location information, and the device information may include current time information, ambient light intensity information and brightness information. Then, the initial weight and initial bias may be determined, wherein the bias information may refer to an adjustment value relative to a default setting, and may be an adjustment value of corresponding parameters such as brightness, color temperature, and hue. The ambient light intensity information may be the light intensity of the surrounding environment. The initial weight, initial bias, current time information, and ambient light intensity information are input into a local learning model of the smart lighting device for training. For example, when the current time is seven o'clock in the afternoon, the user is in the living room area, the ambient light intensity information is 50lx, and the brightness information is 0lx, a smart home decision result may be generated to adjust the lighting brightness of the living room smart lighting device to 70lx and the lighting angle to 60°.

[0115] The embodiment of the present invention enables effective control of smart lighting devices by making the smart home decision result include adjusting the lighting angle and / or adjusting the lighting brightness, thereby improving the flexibility and rationality of local smart device decision control.

[0116] In an optional embodiment of the present invention, the step of determining the initial predicted brightness information based on the user location information, the current time information, the initial weight, the initial bias and the ambient light intensity information includes:

[0117] The user location information, the current time information, the initial weight, the initial bias and the ambient light intensity information are input into Formula 1, which is:

[0118]

[0119] in, is the initial predicted brightness information, x1 is the user location information, x2 is the current time information, x3 is the ambient light intensity information, w1, w2 and w3 are initial weights, and b is the initial bias.

[0120] In practical applications, since numerical features can perform various mathematical operations, such as addition, subtraction, multiplication, division, averaging, distance calculation, etc., feature extraction can be used to convert text parameters into numerical values ​​to facilitate subsequent calculations. For example, user location information, current time information, ambient light intensity information, etc. can be converted into numerical values.

[0121] Exemplarily, when the user location information is x1, the current time information is x2, the ambient light intensity information is x3, the initial weights are w1, w2 and w3, and the initial bias is b, if the user location information x1=2, the current time information x2=18, and the ambient light intensity information x3=300, then x2, the ambient light intensity information is x3, the initial weights are w1=0, w2=0, w3=0, and the initial bias is b=0, if the user location information x1=2, the current time information x2=18, and the ambient light intensity information x3=300 are substituted into Formula 1, that is, Therefore, it can be seen that the initial predicted brightness information is 0.

[0122] In the embodiment of the present invention, the user location information, the current time information, the initial weight, the initial bias and the ambient light intensity information are input into Formula 1, where Formula 1 is: in, is the initial predicted brightness information, x1 is the user location information, x2 is the current time information, x3 is the ambient light intensity information, wl, w2 and w3 are initial weights, and b is the initial bias, so as to achieve local training to obtain the initial predicted brightness information, protect user privacy, and provide a basis for subsequent data processing.

[0123] In an optional embodiment of the present invention, the step of determining an update weight and an update bias for the intelligent lighting device based on the initial predicted brightness information, the brightness information, the initial weight and the initial bias comprises:

[0124] The initial predicted brightness information, the brightness information, the initial weight and the initial bias are input into Formula 2, which is:

[0125]

[0126] Wherein, MSE is mean square error, n is the total number of information of the initial predicted brightness information, yi is the i-th brightness information, is the i-th initial predicted brightness information, w1, w2 and w3 are initial weights, and b is the initial bias.

[0127] In practical applications, in order to improve the accuracy of training results, it is necessary to optimize the weight and bias parameters. A loss function, such as the MSE function, can be defined by using the loss function to calculate the gradient descent to measure the difference between the model prediction value and the true value. The gradient represents the direction in which the loss function changes fastest at the current parameter point. The model parameters, such as weights and biases, are updated in the opposite direction of the gradient, so that the value of the loss function gradually decreases, thereby improving the accuracy of the prediction results.

[0128] Exemplarily, when the loss function is mean square error MSE, the total number of information of the initial predicted brightness information is n, the i-th brightness information is yi, and the i-th initial predicted brightness information is When w1, w2 and w3 are initial weights and b is initial bias, they can be calculated by formula 2 as follows:

[0129] MSE = (500-0)2 = 250000

[0130]

[0131]

[0132] The updated weight w1=0-0.01·2000=-20, the updated weight w2=0-0.01·18000=-180, the updated weight w3=0-0.01·300000=-3000, and the updated bias b=0-0.01·1000=-10.

[0133] Preferably, multiple iterations of training can be performed to converge the model and obtain the final updated weights and updated biases.

[0134] In the embodiment of the present invention, the initial predicted brightness information, the brightness information, the initial weight and the initial bias are input into Formula 2, where Formula 2 is:

[0135]

[0136] Wherein, MSE is mean square error, n is the total number of information of the initial predicted brightness information, yi is the i-th brightness information, is the i-th initial predicted brightness information, w1, w2 and w3 are initial weights, and b is initial bias, thereby improving the accuracy of weight parameters and bias parameters, further reducing the difference between the model prediction results and the true values, and improving the accuracy of the final smart home decision results.

[0137] In an optional embodiment of the present invention, the step of using the initial training data of the plurality of local smart devices as input data of the cloud-based joint learning model to generate decision training data for the target smart device includes:

[0138] Determine the number of local smart devices;

[0139] The number of devices and the initial training data are input into Formula 3, which is:

[0140]

[0141] Among them, θnew is the decision training data, N is the number of devices, θ i is the initial training data corresponding to the i-th local smart device.

[0142] In practical applications, the Federated Averaging Method (FedAvg) is one of the most classic algorithms in federated learning. The parameter aggregation method has significant advantages. It allows each participant to share only the updated model parameters instead of the original data, effectively protecting data privacy, avoiding the centralized storage of sensitive data, and reducing the possibility of data leakage. Federated learning can make full use of the massive data distributed on different devices to train a more powerful and accurate model. By aggregating data from different devices, the model can learn richer features, thereby improving the generalization ability of the model. Compared with traditional centralized machine learning, federated learning only needs to transmit model parameters instead of original data, which greatly reduces the communication cost. It can process data from different devices and different distributions, and improve the robustness of the model on non-independent and identically distributed data. Each device can independently decide whether to participate in training based on its own computing resources and communication conditions, which improves the flexibility of the system. The implementation of the algorithm is relatively simple, easy to understand and deploy, and is applicable to a variety of machine learning models, such as linear regression, logistic regression, neural networks, etc. For example, the cloud-based federated learning model can use the Federated Averaging Method for parameter aggregation.

[0143] Exemplarily, the number of local smart devices may be determined using Formula 3:

[0144]

[0145] Among them, θ new is the decision training data, N is the number of devices, θ i is the initial training data corresponding to the i-th local smart device.

[0146] After the cloud-based intelligent decision-making assistant uses the federated learning method to aggregate parameters, the parameter model can be sent to each local smart device to complete the sharing among multiple local smart devices, which is regarded as the completion of a joint learning.

[0147] In the embodiment of the present invention, the number of local smart devices is determined;

[0148] The number of devices and the initial training data are input into Formula 3, which is:

[0149]

[0150] Among them, θ newis the decision training data, N is the number of devices, and θi is the initial training data corresponding to the ith local smart device, thereby enabling the local smart devices to jointly update the global model without sharing the original data, thereby protecting user privacy and further improving the accuracy of the decision results.

[0151] In order to enable those skilled in the art to better understand the embodiments of the present invention, the embodiments of the present invention are described below using a complete example.

[0152] refer to Figure 2 , showing a system architecture diagram provided in Embodiment 1 of the present invention;

[0153] refer to Figure 3 , showing a flowchart of the steps of another smart home decision-making method provided in Embodiment 1 of the present invention;

[0154] In multiple smart home environments, handling response priorities is a complex issue. There are currently many different approaches. This distributed intelligent decision-making system combines a dynamic joint learning model between multiple devices to make multi-device response decisions more intelligent. The specific process is as follows:

[0155] First, the user's location and behavior can be obtained through smart sensor devices, and the information processing can be passed to the cloud-based smart decision assistant. The communication between smart devices is carried out through the cloud server. The appropriate learning model is selected for each smart device, and the model parameters of each device are initialized so that each device can obtain the initial user information in the cloud-based smart decision assistant. Each device is trained independently according to the field required by the device. For example, smart air conditioners optimize the required temperature and humidity parameters according to the user's location. Smart lights optimize the lighting area, light intensity, angle and other parameters according to the user's location, orientation, and behavior. Different devices will send their own training data to the cloud-based smart decision assistant and continuously update it. The cloud-based smart decision assistant will also integrate this information and synchronize it to smart devices throughout the house. In this way, smart devices can regularly exchange their own learning model parameters and other information, and the cloud-based smart decision assistant completes parameter aggregation;

[0156] Parameter aggregation can be performed using the federated average method. The aggregation formula is shown in the above formula 3, where N is the number of smart devices, θ i is the uploaded model parameter of the i-th device.

[0157] The cloud-based intelligent decision assistant sends the aggregated parameter model to each device to complete the sharing between devices. At this time, a joint learning is completed. Through this method, each device can upload the model parameters obtained by local training without sharing the original data, and jointly update the global model, thereby achieving the purpose of joint learning. This method can protect user privacy, while making full use of the data of each device to optimize the global model.

[0158] After completing a single joint learning, the cloud-based intelligent decision assistant makes decisions based on the information uploaded by each device, and makes specific decisions based on the model's prediction results and the system's strategy. For example, based on the predicted light intensity and user preferences and behaviors, decide whether to adjust the light brightness, color, angle, etc. For example, when the living room smart lamp recognizes that the user has turned on the TV, it will adjust the light angle and brightness, and pass the parameters to the cloud-based intelligent decision assistant. The cloud-based intelligent decision assistant can dynamically adjust the volume of the audio, the brightness of the TV, or adjust the curtain switch, etc. based on the information, and complete the response decision of multiple devices.

[0159] Preferably, the learning model can be stored in a combination of local and cloud storage. The trained model can be stored locally for tasks with high real-time requirements; the cloud can be used for model training and updating, taking advantage of the cloud's strong computing power and ample storage space.

[0160] Preferably, in a smart home environment, when multiple rooms are equipped with smart lighting systems, these lighting systems can automatically adjust the direction and brightness of the lights according to the user's location and activities to improve comfort and energy saving. Data preparation The smart lighting system in each room collects the following data: User location (e.g., coordinates in the room) Ambient light intensity at the current time User activity type (e.g., reading, resting, working) User preferences (e.g., favorite brightness and light color) Model initialization Each smart lighting system initializes a model to predict the optimal light direction and brightness. Suppose we use a simple linear regression model: y=w1·x1+w2·x2+w3·x3+b, where: y is brightness information, x1 is user location information, x2 is current time information, x3 is ambient light intensity information, w1, w2, w3 are initial weights, and b is initial bias;

[0161] Input the user location, current time and ambient light intensity data to calculate the predicted light brightness. For example, you can use formula 1 to calculate. For user location x1=2, current time x2=18, ambient light intensity x3=300, the initial prediction can be

[0162] Then, the updated weights and updated biases can be calculated according to formula 2:

[0163] MSE = (500-0)2 = 250000

[0164]

[0165] The updated weight w1=0-0.01·2000=-20, the updated weight w2=0-0.01·18000=-180, the updated weight w3=0-0.01·300000=-3000, and the updated bias b=0-0.01·1000=-10.

[0166] After multiple iterations of training, the model will gradually converge, and the weights and biases will be adjusted to appropriate values, making the predicted brightness closer to the actual value.

[0167] Preferably, when the local smart devices are smart curtain devices and smart lighting devices, the smart curtain devices recognize the angle, brightness, ultraviolet intensity, glass temperature, and other data of the current sunlight outside the window; the human body radar sensor device can obtain the user's current position coordinates, direction and other data; these two devices send all this data to the cloud-based smart decision assistant, and the cloud-based smart decision assistant can adjust different devices based on all the data. For example, the cloud-based smart decision assistant adjusts lamps and can select what is needed based on all the received data, such as the intensity and angle of sunlight, the user's direction, the current time, etc., and use the training model to calculate the appropriate light angle, brightness and other data, and send it to the smart lighting device for control.

[0168] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0169] Embodiment 2

[0170] Reference Figure 4 , shows a flowchart of a smart home decision-making method provided in Embodiment 2 of the present invention, which may specifically include the following steps:

[0171] Step 401, generating decision training data based on the initial training data and the user information;

[0172] Step 402: Using the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

[0173] In a specific implementation, an embodiment of the present invention can be applied to a cloud-based intelligent decision-making assistant, which is configured with a corresponding local smart device, and the local smart device is configured to obtain device information and user information; establish a local learning model for the local smart device; use the device information and the user information as input information, use the local learning model for training, and obtain initial training data; control the local smart device to send the initial training data to the cloud-based intelligent decision-making assistant.

[0174] As for the second embodiment, since it is basically similar to the first embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0175] Embodiment 3

[0176] Reference Figure 5 , shows a structural block diagram of a smart home decision-making device provided in Embodiment 3 of the present invention, which may specifically include the following modules:

[0177] User information acquisition module 501, used to acquire device information and user information;

[0178] A local learning model building module 502 is used to build a local learning model for a local smart device;

[0179] An initial training data generating module 503 is used to use the device information and the user information as input information, and adopt the local learning model for training to obtain initial training data;

[0180] An initial training data sending module 504 is used to control the local smart device to send the initial training data to the cloud-based smart decision assistant;

[0181] The cloud-based intelligent decision assistant is configured to: generate decision training data based on the initial training data and the user information; use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; and the smart home decision result is used to control local smart devices.

[0182] As for the third embodiment, since it is basically similar to the first embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0183] Embodiment 4

[0184] Reference Figure 6 , shows a structural block diagram of a smart home decision-making device provided in Embodiment 4 of the present invention, which may specifically include the following modules:

[0185] A decision training data generating module 601 is used to generate decision training data based on the initial training data and the user information;

[0186] The smart home decision result generating module 602 is used to use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control the local smart device.

[0187] As for the fourth embodiment, since it is basically similar to the first embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0188] Embodiment 5

[0189] The smart home decision-making system includes local smart devices and a cloud-based smart decision-making assistant;

[0190] The local smart device is configured to obtain device information and user information;

[0191] Build local learning models for local smart devices;

[0192] Taking the device information and the user information as input information, and using the local learning model for training to obtain initial training data;

[0193] Controlling the local smart device to send the initial training data to the cloud-based smart decision assistant;

[0194] The cloud-based intelligent decision assistant is configured as follows:

[0195] generating decision training data based on the initial training data and the user information;

[0196] The cloud-based smart home decision assistant is used to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

[0197] As for the fifth embodiment, since it is basically similar to the first embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0198] In addition, an embodiment of the present invention further provides an electronic device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the various processes of the above-mentioned smart home decision-making method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0199] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, each process of the above-mentioned smart home decision-making method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0200] Figure 7 A schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0201] The electronic device 700 includes but is not limited to: a radio frequency unit 701, a network module 702, an audio output unit 703, an input unit 704, a sensor 705, a display unit 706, a user input unit 707, an interface unit 708, a memory 709, a processor 710, and a power supply 711. Those skilled in the art will appreciate that Figure 7 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently. In the embodiments of the present invention, the electronic device includes but is not limited to a mobile phone, a tablet computer, a laptop computer, a PDA, a vehicle-mounted terminal, a wearable device, and a pedometer.

[0202] It should be understood that in the embodiment of the present invention, the radio frequency unit 701 can be used for receiving and sending signals during information transmission or communication. Specifically, after receiving downlink data from the base station, it is sent to the processor 710 for processing; in addition, uplink data is sent to the base station. Generally, the radio frequency unit 701 includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc. In addition, the radio frequency unit 701 can also communicate with the network and other devices through a wireless communication system.

[0203] The electronic device provides users with wireless broadband Internet access through the network module 702, such as helping users to send and receive emails, browse web pages, and access streaming media.

[0204] The audio output unit 703 can convert the audio data received by the RF unit 701 or the network module 702 or stored in the memory 709 into an audio signal and output it as sound. Moreover, the audio output unit 703 can also provide audio output related to a specific function performed by the electronic device 700 (for example, a call signal reception sound, a message reception sound, etc.). The audio output unit 703 includes a speaker, a buzzer, a receiver, etc.

[0205] The input unit 704 is used to receive audio or video signals. The input unit 704 may include a graphics processor (GPU) 7041 and a microphone 7042, and the graphics processor 7041 processes the image data of a static picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The processed image frame can be displayed on the display unit 706. The image frame processed by the graphics processor 7041 can be stored in the memory 709 (or other storage medium) or sent via the radio frequency unit 701 or the network module 702. The microphone 7042 can receive sound and can process such sound into audio data. The processed audio data can be converted into a format output that can be sent to a mobile communication base station via the radio frequency unit 701 in the case of a telephone call mode.

[0206] The electronic device 700 also includes at least one sensor 705, such as a light sensor, a motion sensor, and other sensors. Specifically, the light sensor includes an ambient light sensor and a proximity sensor, wherein the ambient light sensor can adjust the brightness of the display panel 7061 according to the brightness of the ambient light, and the proximity sensor can turn off the display panel 7061 and / or the backlight when the electronic device 700 is moved to the ear. As a kind of motion sensor, the accelerometer sensor can detect the magnitude of acceleration in each direction (generally three axes), and can detect the magnitude and direction of gravity when stationary, which can be used to identify the posture of the electronic device (such as horizontal and vertical screen switching, related games, magnetometer posture calibration), vibration recognition related functions (such as pedometer, tapping), etc.; the sensor 705 can also include a fingerprint sensor, a pressure sensor, an iris sensor, a molecular sensor, a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, etc., which will not be repeated here.

[0207] The display unit 706 is used to display information input by the user or information provided to the user. The display unit 706 may include a display panel 7061, which may be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), or the like.

[0208] The user input unit 707 can be used to receive input digital or character information, and to generate key signal input related to user settings and function control of the electronic device. Specifically, the user input unit 707 includes a touch panel 7071 and other input devices 7072. The touch panel 7071, also known as a touch screen, can collect the user's touch operation on or near it (such as the user's operation on the touch panel 7071 or near the touch panel 7071 using any suitable object or accessory such as a finger, stylus, etc.). The touch panel 7071 may include two parts: a touch detection device and a touch controller. Among them, the touch detection device detects the user's touch orientation, detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into the contact point coordinates, and then sends it to the processor 710, receives the command sent by the processor 710 and executes it. In addition, the touch panel 7071 can be implemented using multiple types such as resistive, capacitive, infrared and surface acoustic waves. In addition to the touch panel 7071, the user input unit 707 may also include other input devices 7072. Specifically, other input devices 7072 may include but are not limited to a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, and a joystick, which are not described in detail here.

[0209] Furthermore, the touch panel 7071 may be overlaid on the display panel 7061. When the touch panel 7071 detects a touch operation on or near it, it transmits the information to the processor 710 to determine the type of the touch event. Then, the processor 710 provides a corresponding visual output on the display panel 7061 according to the type of the touch event. Figure 7 In the figure, the touch panel 7071 and the display panel 7061 are used as two independent components to realize the input and output functions of the electronic device. However, in some embodiments, the touch panel 7071 and the display panel 7061 can be integrated to realize the input and output functions of the electronic device, which is not limited here.

[0210] The interface unit 708 is an interface for connecting an external device to the electronic device 700. For example, the external device may include a wired or wireless headset port, an external power supply (or battery charger) port, a wired or wireless data port, a memory card port, a port for connecting a device with an identification module, an audio input / output (I / O) port, a video I / O port, a headphone port, etc. The interface unit 708 may be used to receive input (e.g., data information, power, etc.) from an external device and transmit the received input to one or more elements within the electronic device 700 or may be used to transmit data between the electronic device 700 and an external device.

[0211] The memory 709 can be used to store software programs and various data. The memory 709 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory 709 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0212] The processor 710 is the control center of the electronic device. It uses various interfaces and lines to connect various parts of the entire electronic device. It executes various functions of the electronic device and processes data by running or executing software programs and / or modules stored in the memory 709, and calling data stored in the memory 709, so as to monitor the electronic device as a whole. The processor 710 may include one or more processing units; preferably, the processor 710 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 710.

[0213] The electronic device 700 may also include a power supply 711 (such as a battery) for supplying power to each component. Preferably, the power supply 711 may be logically connected to the processor 710 through a power management system, thereby implementing functions such as charging, discharging, and power consumption management through the power management system.

[0214] In addition, the electronic device 700 includes some functional modules not shown, which will not be described in detail here.

[0215] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0216] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0217] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

[0218] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the embodiments of the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0219] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0220] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0221] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0222] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0223] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0224] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A smart home decision-making method, characterized in that: The method is applied to a local smart device, and the local smart device is configured with a corresponding cloud-based smart decision assistant. The method includes: Get device information and user information; Build local learning models for local smart devices; Taking the device information and the user information as input information, and using the local learning model for training to obtain initial training data; Controlling the local smart device to send the initial training data to the cloud-based smart decision assistant; The cloud-based intelligent decision assistant is configured as follows: generating decision training data based on the initial training data and the user information; The cloud-based smart home decision assistant is used to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

2. The method according to claim 1, characterized in that The local smart device is a smart lighting device, the user information includes user location information, the device information includes current time information, ambient light intensity information and brightness information, and the step of using the device information and the user information as input information and using the local learning model for training to obtain initial training data includes: Determine the initial weights and initial biases; Determining initial predicted brightness information based on the user location information, the current time information, the initial weight, the initial bias, and the ambient light intensity information; Determining an update weight and an update bias for the intelligent lighting device based on the initial predicted brightness information, the brightness information, the initial weight and the initial bias; Determining updated predicted brightness information based on the updated weight and the updated bias; The updated predicted brightness information is used as initial training data.

3. The method according to claim 1, characterized in that: The step of generating decision training data based on the initial training data and the user information comprises: Establish a cloud-based federated learning model; Determine a target smart device from a plurality of local smart devices; The initial training data of the plurality of local smart devices are used as input data of the cloud-based joint learning model to generate decision training data for the target smart device.

4. The method according to claim 1, characterized in that: Also includes: The cloud-based intelligent decision assistant is used to update the local learning model, and the updated local learning model is sent to the corresponding local smart device.

5. The method according to claim 2, characterized in that: The smart home decision result includes adjusting the lighting angle and / or adjusting the lighting brightness.

6. The method according to claim 2, characterized in that The step of determining the initial predicted brightness information based on the user position information, the current time information, the initial weight, the initial bias and the ambient light intensity information comprises: The user location information, the current time information, the initial weight, the initial bias and the ambient light intensity information are input into Formula 1, which is: in, is the initial predicted brightness information, x1 is the user location information, x2 is the current time information, x3 is the ambient light intensity information, w1, w2 and w3 are initial weights, and b is the initial bias.

7. The method according to claim 6, characterized in that The step of determining an update weight and an update bias for the intelligent lighting device based on the initial predicted brightness information, the brightness information, the initial weight and the initial bias comprises: The initial predicted brightness information, the brightness information, the initial weight and the initial bias are input into Formula 2, which is: Wherein, MSE is the mean square error, n is the total number of information of the initial predicted brightness information, y i is the ith brightness information, is the i-th initial predicted brightness information, w1, w2 and w3 are initial weights, and b is the initial bias.

8. The method according to claim 3, characterized in that The step of using the initial training data of the plurality of local smart devices as input data of the cloud-based joint learning model to generate decision training data for the target smart device includes: Determine the number of local smart devices; The number of devices and the initial training data are input into Formula 3, which is: Among them, θ new is the decision training data, N is the number of devices, θ i is the initial training data corresponding to the i-th local smart device.

9. A smart home decision-making method, characterized in that: The method is applied to a cloud-based intelligent decision assistant, wherein the cloud-based intelligent decision assistant is configured with a corresponding local intelligent device, and the local intelligent device is configured to obtain device information and user information; establish a local learning model for the local intelligent device; use the device information and the user information as input information, and use the local learning model for training to obtain initial training data; Controlling the local smart device to send the initial training data to the cloud-based smart decision assistant; the method includes: generating decision training data based on the initial training data and the user information; The cloud-based smart home decision assistant is used to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

10. A smart home decision-making device, characterized in that: The device is applied to a local smart device, and the local smart device is configured with a corresponding cloud-based smart decision assistant. The device includes: User information acquisition module, used to obtain device information and user information; A local learning model building module is used to build a local learning model for local smart devices; An initial training data generating module, used to use the device information and the user information as input information, and adopt the local learning model for training to obtain initial training data; An initial training data sending module, used to control the local smart device to send the initial training data to the cloud-based smart decision assistant; The cloud-based intelligent decision assistant is configured to: generate decision training data based on the initial training data and the user information; use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; and the smart home decision result is used to control local smart devices.

11. A smart home decision-making device, characterized in that: The device is applied to a cloud-based intelligent decision assistant, the cloud-based intelligent decision assistant is configured with a corresponding local intelligent device, and the local intelligent device is configured to obtain device information and user information; establish a local learning model for the local intelligent device; use the device information and the user information as input information, and use the local learning model for training to obtain initial training data; Controlling the local smart device to send the initial training data to the cloud-based smart decision assistant; the device comprises: A decision training data generating module, used to generate decision training data based on the initial training data and the user information; A smart home decision result generation module is used to use the cloud-based smart home decision assistant to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

12. A smart home decision-making system, characterized in that: The smart home decision-making system includes local smart devices and a cloud-based smart decision-making assistant; The local smart device is configured to obtain device information and user information; Build local learning models for local smart devices; Taking the device information and the user information as input information, and using the local learning model for training to obtain initial training data; Controlling the local smart device to send the initial training data to the cloud-based smart decision assistant; The cloud-based intelligent decision assistant is configured as follows: generating decision training data based on the initial training data and the user information; The cloud-based smart home decision assistant is used to generate a smart home decision result based on the decision training data and the user information; the smart home decision result is used to control local smart devices.

13. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; The memory is used to store computer programs; The processor is used to implement the method according to any one of claims 1-8 or 9 when executing the program stored in the memory.

14. A computer-readable storage medium having instructions stored thereon, which, when executed by one or more processors, cause the processors to perform the method according to any one of claims 1 to 8 or 9.

Images