Electronic device control method and control apparatus

Abstract

This application provides a control method and device for home appliances. By using a ridge regression model established based on user habits, it predicts the setting values ​​of related parameters on the user-operated home appliance or the setting values ​​of parameters on related home appliances when the user operates the appliance. Then, it controls the user-operated home appliance and / or related home appliances based on the predicted setting values. This solution addresses the problem that current smart home devices often struggle to understand and acquire user habits accurately, or the methods used are overly complex.

Description

Technical Field

[0001] This application relates to the field of electronic device control, and more particularly to an electronic device control method and control device. Background Technology

[0002] With home appliances becoming increasingly intelligent, analyzing and predicting users' habits of using different devices has become particularly important. With user habits in mind, it is possible to more intelligently set the status of devices based on the user and / or set up device linkage effects that conform to the user's usage habits.

[0003] However, current smart home technologies often struggle to understand and acquire user habits, resulting in inaccurate or complex methods. While statistical analysis-based prediction methods are simple and easy to implement, they lack accuracy.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] In view of this, this application proposes a control method and device for home appliances. A ridge regression model is established based on the user's usage habits of the home appliances. When the parameters of one home appliance are set, the established model predicts the setting values ​​of related parameters of at least one home appliance, including the stated appliance. Based on the predicted setting values, the home appliances are controlled. This solution solves the problem that current smart home devices face difficulties in understanding and acquiring user habits, or the methods are often complex. It also makes the use of smart home devices more intelligent and convenient, reducing the user's operational burden and saving user time.

[0006] According to one aspect of this application, a method for controlling a home appliance is provided, characterized in that it includes:

[0007] A ridge regression model was established based on users' habits of using home appliances;

[0008] In response to a user setting at least one parameter of a home appliance, the established ridge regression model predicts the set values ​​of associated parameters of at least one home appliance, including the home appliance itself.

[0009] Based on the predicted values ​​of the associated parameters, the associated parameters of the at least one home device are set to control the at least one home device.

[0010] Optionally,

[0011] A ridge regression model is built based on users' usage habits of home appliances, including:

[0012] Based on users' habits of associating at least two parameters of a single home device, a single-device ridge regression model is established, which is used to predict the association settings between at least two parameters of the single home device.

[0013] Optionally,

[0014] Users' habits regarding the association settings of at least two data points for a single home device, including: the settings of at least two parameters for a single home device and the frequency of use of each setting.

[0015] Optionally,

[0016] Based on users' habits of associating at least two parameters of a single home appliance, a single-device ridge regression model is established, including:

[0017] The frequency of use for each of the aforementioned settings is used as the dependent variable of the ridge regression model, and at least two parameters of a single home appliance are used as the independent variables of the ridge regression model.

[0018] Optionally,

[0019] In response to a user setting at least one parameter of a home appliance, the established ridge regression model predicts the setting value of at least one home appliance associated with the home appliance, including:

[0020] In response to a user setting at least one parameter for a single home appliance, the established single-device ridge regression model predicts the setting value of the associated parameter of the at least one parameter of the single home appliance.

[0021] Optionally,

[0022] A ridge regression model is built based on users' usage habits of home appliances, including:

[0023] Based on users' habits of using multiple home appliances in a coordinated manner, a device linkage ridge regression model is established, which is used to predict the linkage settings of the multiple home appliances.

[0024] Optionally,

[0025] The user's habits of using multiple home devices in a coordinated manner include: the switching parameters of multiple home devices based on time, or the coordination of switching parameters and other parameters, and the frequency of use of each coordination.

[0026] Optionally,

[0027] Based on users' habits of using multiple home appliances in a coordinated manner, a ridge regression model for device coordination is established, including:

[0028] The frequency of use of the aforementioned linkage is used as the dependent variable of the ridge regression model, and the switching parameters of multiple home appliances, or the switching parameters and other parameters, are used as the independent variables of the ridge regression model.

[0029] Optionally,

[0030] In response to a user setting at least one piece of data for a home appliance, the established ridge regression model predicts the set values ​​of associated data for at least one home appliance, including the home appliance itself, comprising:

[0031] In response to a user's setting of the on / off parameters of a home appliance, or the on / off parameters and at least one other parameter, the established device linkage ridge regression model predicts the setting values ​​of the on / off parameters of at least one other home appliance, or the associated parameters of each linked home appliance.

[0032] Optionally,

[0033] For scenarios involving multiple interconnected devices, a device interconnection ridge regression model is established to enable or disable multiple interconnected home devices, and a separate single-device ridge regression model is established for each interconnected home device.

[0034] Specifically, in response to a user setting at least one parameter of a home appliance, the established ridge regression model predicts the setting value of at least one home appliance associated with the home appliance, including:

[0035] In response to a user turning a home appliance on or off, the established device linkage ridge regression model predicts whether at least one other home appliance needs to be turned on or off, and the established single-device ridge regression model predicts the setting values ​​of related parameters (excluding switch parameters) of at least two home appliances, including the aforementioned home appliance.

[0036] Optionally, the prediction is made with the principle of not reducing the value of the dependent variable.

[0037] According to another aspect of this application, a control device is provided, including a memory and a processor, characterized in that the processor performs the method described above.

[0038] According to the control method and device for home appliances proposed in this application, a ridge regression model established based on user habits is used to predict the setting values ​​of related parameters on the user-operated home appliance or the setting values ​​of parameters on related home appliances when the user operates the home appliance. Then, the user-operated home appliance and / or related home appliances are controlled based on the predicted setting values. This solution solves the problem that current smart home devices face difficulties in understanding and acquiring user habits, or the methods are often complex. It also makes the use of smart home devices more intelligent and convenient, reduces the user's operational burden, and saves user time.

[0039] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, the preferred embodiments of this application are described in detail below with reference to the accompanying drawings. Attached Figure Description

[0040] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0041] Figure 1 A schematic diagram of a control method for a home appliance according to this application is shown;

[0042] Figure 2 A schematic diagram of an embodiment of a control method for a home appliance according to this application is shown. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0044] Figure 1 A schematic diagram of a control method for a home appliance according to this application is shown.

[0045] As one implementation method, home appliances can include all electrically controllable devices, including but not limited to smart home devices, such as refrigerators, air conditioners, fans, controllable lights, electric curtains, televisions, robot vacuums, etc., and can also include mobile phones, tablets, computers, etc.

[0046] Step S1: Establish a ridge regression model based on users' usage habits of home appliances.

[0047] By learning user habits and establishing mathematical models based on these habits, the control of home appliances can be made more intelligent, convenient, and user-friendly. This application employs a ridge regression model, a type of linear regression model. Unlike ordinary regression models, ridge regression addresses the problem of excessively strong correlations between independent variables, ensuring that each independent variable is relatively independent. It can predict the linear relationship between the dependent variable and multiple independent variables. Based on this relationship, we can understand how the correlation between multiple independent variables should change to keep the dependent variable constant or above a certain value. This helps us better understand the working principles of the devices and user habits, thereby optimizing device performance and user experience.

[0048] As one implementation approach, for the scenario of a single device being used independently, a single-device ridge regression model can be established based on the user's habits regarding the association settings of at least two parameters of a single home appliance. This model is used to predict the association settings between at least two parameters of the single home appliance. When a single device is used independently, only the association parameters of that device need to be predicted based on user habits.

[0049] Optionally, a user's habits regarding the associated settings of at least two data points for a single home appliance can include the settings of at least two parameters of the single home appliance and the frequency of use for each setting. The frequency of use serves as a representation of user satisfaction. For example, for air conditioners, the combination of settings for parameters such as temperature, humidity, fan speed, mode, and swing, reflecting user usage habits, along with the frequency of use for each combination, can be used as observation / training samples to create a ridge regression model. As an example, the frequency of use for each setting can be used as the dependent variable of the ridge regression model, and the at least two parameters of the single home appliance can be used as the independent variables. This yields a linear relationship between the dependent variable and multiple independent variables. Furthermore, when a user sets one independent variable, the ridge regression model can predict the settings of other independent variables to maintain the dependent variable at a certain level.

[0050] As one implementation method, once a single-device ridge regression model is established, it can predict the setting values ​​of associated parameters of a single home appliance based on the user's setting of at least one parameter of that single home appliance. Taking an air conditioner as an example, after the single-device ridge regression model is established, when the user sets the air conditioner's temperature parameter again, for example, to 27 degrees Celsius, the single-device ridge regression model will predict the optimal setting values ​​of other parameters such as humidity, fan speed, and air swing based on the current temperature setting.

[0051] As one implementation method, for scenarios involving the coordinated use of multiple devices, a device-linked ridge regression model can be established based on users' habits of using multiple home appliances in conjunction with each other. This model is used to predict the coordinated settings of the multiple home appliances. Specifically, for scenarios involving the coordinated use of multiple devices, a ridge regression model can be established to predict whether other home appliances will simultaneously turn on or off when one appliance is turned on or off. For example, in the past month since the onset of summer, every evening when a user returns home, they habitually turn on the air conditioner, fan, and stereo simultaneously, and turn off the camera. Therefore, by learning the user's habits of using multiple home appliances in conjunction with each other, a ridge regression model can be established. Thus, when the user returns home again and turns on the air conditioner, the established ridge regression model can predict that the fan and stereo should also be turned on, while the camera should be turned off. Of course, the device-linked ridge regression model can be used simultaneously with the single-device ridge regression model described earlier.

[0052] Optionally, user habits regarding the coordinated use of multiple home appliances include: the time-based coordination of multiple home appliances and the frequency of use for each coordination. The frequency of use serves as a representation of user satisfaction. As an example, the coordination includes: coordination of switching parameters or switching parameters and other parameters. The frequency of use for these coordination can then be used as the dependent variable in a ridge regression model, and the switching parameters of the multiple home appliances, or switching parameters and other parameters, can be used as the independent variables in the ridge regression model. For instance, in the past week, around 6 PM each day, the frequency with which a user returns home and simultaneously turns on the air conditioner and stereo is A; the frequency with which they simultaneously turn on the air conditioner, fan, and stereo is B; and the frequency with which they simultaneously turn on the air conditioner, stereo, fan, and camera and turn off the camera is C. Therefore, the switching status of the air conditioner, stereo, fan, and camera, and the frequency of their switching combinations, or further, the switching status of the air conditioner, stereo, fan, camera, and other parameters, and the frequency of their combinations, will be collected as observation / learning samples of the user's habits regarding the coordinated use of multiple home appliances.

[0053] In other words, in the case of multiple devices working together, a ridge regression model can be established solely for the on / off states of the devices, or various parameters (including on / off parameters) of multiple devices can be combined to create a single ridge regression model. Thus, when one device, such as an air conditioner, is turned on or its temperature is set, the ridge regression model can not only predict the on / off states of other devices, but also how to set other parameters of each device besides on / off parameters, such as the temperature and humidity of the air conditioner and the fan speed. In other words, the settings of various parameters for multiple devices can also be considered part of the overall system setup.

[0054] Step S2, in response to the user's setting of at least one parameter of a home appliance, predict the setting value of the associated parameter of at least one home appliance, including the home appliance, through the established ridge regression model.

[0055] As one implementation, after establishing a device linkage ridge regression model, in response to a user's setting of the on / off parameters of a home appliance, or the setting of the on / off parameters and at least one other parameter, the established device linkage ridge regression model predicts the on / off parameters of at least one other home appliance, or the setting values ​​of the associated parameters of each linked home appliance. Again, using the example above as an illustration, after establishing the ridge regression model, if around 6 PM it is detected that the user turns on the air conditioner or turns it on and sets the temperature to 26 degrees Celsius, then the ridge regression model predicts whether the audio system, fan, and camera will simultaneously turn on or off, or predicts the setting values ​​of each associated parameter (any associated parameter other than the air conditioner's own on / off parameters) of the air conditioner, audio system, fan, and camera.

[0056] As one implementation approach, in the case of multi-device linkage, a device linkage ridge regression model can be established to enable or disable multiple home appliances in a coordinated manner, and a separate single-device ridge regression model can be established for each linked home appliance. In response to a user enabling or disabling a home appliance, the device linkage ridge regression model predicts whether at least one other home appliance needs to be enabled or disabled, and the single-device ridge regression models predict the settings of related parameters for at least two home appliances, including the first home appliance, excluding the on / off parameters. Because the device linkage ridge regression model has already predicted the on / off parameters, the single-device model only needs to predict the parameter values ​​other than the on / off parameters. In other words, after predicting the enabling or disabling of the linked devices, the model further predicts the settings of other related parameters for each linked device.

[0057] The user has been taking a shower almost every night at 8 PM for the past week, simultaneously turning on the water heater, stereo, washing machine, and other appliances, and setting their preferred settings for each device.

[0058] As an example, for the past two weeks, users have been returning home around 6 PM almost every day and habitually turn on the air conditioner, fan, and stereo simultaneously. They have also set their preferred settings for each device. For instance, they might set the air conditioner to 27 degrees Celsius, fan speed to automatic, fan to the lowest setting, and the stereo to play the day's news at a frequency of D. Alternatively, they might set the air conditioner to 28 degrees Celsius, fan speed to medium, fan to the second lowest setting, and the stereo to play the day's news at a frequency of F. Other combinations are possible, but will not be elaborated upon. By learning this user's usage habits, we can establish a device-linked ridge regression model for multi-device interaction and separate single-device ridge regression models for the air conditioner, fan, and stereo. Therefore, when the user returns home around a similar time and turns on the air conditioner, the device-linked ridge regression model can predict whether the fan and stereo need to be turned on simultaneously. The single-device ridge regression model for the air conditioner can predict the appropriate temperature and fan speed adjustments, the single-device ridge regression model for the fan can predict the appropriate fan speed setting, and the single-device ridge regression model for the stereo can predict the appropriate content to play. Similarly, if a user returns home at a similar time, turns on the air conditioner, and sets the temperature to 27 degrees Celsius, the prediction process will be similar to the above, except that the initial condition will have the air conditioner temperature parameter set to 37 degrees Celsius.

[0059] Based on the preceding description, user frequency is used as the dependent variable. Therefore, when using a ridge regression model for prediction, the independent variables can be dynamically adjusted to maintain the value of the dependent variable without degrading it, thus avoiding a decrease in user experience. However, this is not the only option; the baseline could also be maintaining the value of the dependent variable within a certain range or above or below a specific value.

[0060] The preceding description uses learning or observing user habits over a recent period as an example, but it is not limited to this. It can also learn over longer or shorter periods, and after the model is built, it can continue to learn and update the model based on user actions.

[0061] Step S3: Based on the predicted setting value of the association parameter, set the association parameter of the at least one home device to control the at least one home device.

[0062] As one implementation, this application can use a control device to implement the above solution. This control device includes a memory and a processor, the processor being used to execute the above method steps. This control device can be included in home appliances or can be independent of home appliances, such as in a central control system or in the cloud.

[0063] Figure 2 A schematic diagram of an embodiment of a control method for a home appliance according to this application is shown.

[0064] Step S21: Determine the independent and dependent variables in the analysis process. The independent variables are the various parameters of a certain device; the dependent variable is represented by the user's satisfaction with each parameter.

[0065] Step S22: Collect a certain amount of dataset. Collect a dataset by means of historical data of the device, background data of the APP, etc. Each row in the dataset represents an observation sample, and each column represents a variable.

[0066] The necessary data for collection and analysis can be obtained through questionnaires, app backend tracking data, etc., to acquire users' common device settings. Device data can be used as independent variables, such as various device parameter data (e.g., air conditioner: temperature, fan speed, function, mode). The values ​​of these data serve as independent variables, while user satisfaction with these parameters (represented by the frequency of user settings) serves as the dependent variable. The collected independent and dependent variables are compiled into a dataset, where each row represents an observation sample and each column represents a variable, as shown in the table below.

[0067]

[0068] Device data can be dynamically added or deleted, and the data can vary depending on the device. For example, devices with vertical or horizontal airflow can have airflow settings, and devices with dehumidification functions can have dehumidification settings.

[0069] Step S23: Establish a ridge regression model and perform regression analysis on the established dataset.

[0070] Using the principle of least squares, a norm is added to the regression coefficients of the regression model. (The norm is used to make the regression coefficients approach zero, which can reduce the complexity and variance of the regression model, reduce the model's dispersion, and improve the accuracy of the regression model.) The norm can be expressed as: λ(β1^2+β2^2+…+βk^2).

[0071] The temperature, humidity, fan speed, mode, and function of the air conditioner can be used as independent variables x1, x2, x3, x4, x5, and user satisfaction can be used as the dependent variable y.

[0072] Establish a ridge regression model: y = β0 + β1x1 + β2x2 + β3x3 + β4x4 + β5x5 + ε, where the regularization term ε (which is the norm mentioned earlier) is λ(β1^2 + β2^2 + β3^2 + β4^2 + β5^2), and λ is a positive constant.

[0073] Step S24: Use gradient descent to estimate the regression coefficients and solve for the value of the regression coefficient (β) based on the dataset.

[0074] Step S25: Determine the ridge regression parameter (λ) through cross-validation or graphical analysis.

[0075] Step S26: Evaluate the regression model and analyze the prediction of the dependent variable based on the values ​​of the independent variables.

[0076] For example, r-squared and mean squared error can be used to adjust the goodness of fit and predictive ability of a regression model, enabling analysis and prediction of the dataset and completion of user satisfaction analysis using these parameters. Alternatively, the dataset can be divided into training and validation sets. Different values ​​of β and λ can be calculated in the training set, and then different values ​​of λ can be validated in the validation set to select the most suitable λ.

[0077] The ridge regression model yields a linear relationship between user satisfaction with device settings and various settings: y = β0 + β1x1 + β2x2 + β3x3 + β4x4 + β5x5 + ε. Based on this linear relationship, when a user sets one or more settings, the system can then adjust the settings accordingly. y Value maximization is used to predict the corresponding values ​​of other settings, thereby dynamically adjusting the remaining settings. For example:

[0078] User usage frequency = 0.0035 - 0.0208 × air conditioner temperature + 0.0017 × air conditioner fan speed + 0.0049 × air conditioner mode 制冷 +0.0060× Air Conditioning Mode 制热 +0.0062×Air Conditioning Mode 送风 +0.0055× Air Conditioning Mode 除湿 -0.0019×Air Conditioner Humidity +∈

[0079] The ridge regression model predicts a negative correlation between the dependent and independent variables. Temperature and frequency are negatively correlated; the higher the temperature, the greater the decrease in usage frequency, and the same applies to other variables. When the temperature value is increased, because temperature is negatively correlated with frequency, other independent variables need to be dynamically adjusted to approximately maintain or even increase the frequency, thus preventing a decrease in the frequency (dependent variable) and avoiding a decline in user experience. Adjusting other independent variables can be achieved using mathematical methods such as multiple linear programming.

[0080] Those skilled in the art will understand that embodiments of the present invention can be provided in various forms, such as methods, apparatus, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0081] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0082] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0083] It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of this invention can be combined with each other. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this invention shall still fall within the scope of the technical solution of this invention.

Claims

1. A method for controlling home appliances, characterized in that, include: A ridge regression model is built based on users' usage habits of home appliances, including: Based on users' habits of associating at least two parameters of a single home device, a single-device ridge regression model is established, which is used to predict the association settings between at least two parameters of the single home device. Based on users' habits of using multiple home appliances in a coordinated manner, a device coordination ridge regression model is established. The device coordination ridge regression model is used to predict the coordination settings of the multiple home appliances. The user's habits of using multiple home appliances in a coordinated manner include: the time-based switching parameters of multiple home appliances, or the coordination of switching parameters and other parameters, and the usage frequency of each coordination situation; the other parameters include: the temperature, humidity, fan speed, mode or function of the air conditioner. Based on users' habits of using multiple home appliances in a coordinated manner, a ridge regression model for appliance coordination is established, including: using the frequency of use of the coordinated situation as the dependent variable of the ridge regression model, and using the switching parameters of multiple home appliances, or the switching parameters and other parameters as the independent variables of the ridge regression model. In response to a user setting at least one parameter of a home appliance, the established ridge regression model predicts the set values ​​of associated parameters of at least one home appliance, including the home appliance itself. Based on the predicted values ​​of the associated parameters, the associated parameters of the at least one home device are set to control the at least one home device.

2. The method as described in claim 1, characterized in that: Users' habits regarding the associated settings of at least two parameters for a single home device, including: the settings of at least two parameters for a single home device and the frequency of use of each setting.

3. The method as described in claim 2, characterized in that: Based on users' habits of associating at least two parameters of a single home appliance, a single-device ridge regression model is established, including: The frequency of use for each of the aforementioned settings is used as the dependent variable of the ridge regression model, and at least two parameters of a single home appliance are used as the independent variables of the ridge regression model.

4. The method according to any one of claims 1-3, characterized in that: In response to a user setting at least one parameter of a home appliance, the established ridge regression model predicts the setting value of at least one home appliance associated with the home appliance, including: In response to a user setting at least one parameter for a single home appliance, the established single-device ridge regression model predicts the setting value of the associated parameter of the at least one parameter of the single home appliance.

5. The method as described in claim 1, characterized in that: In response to a user setting at least one parameter of a home appliance, the established ridge regression model predicts the setting value of at least one home appliance associated with the home appliance, including: In response to a user's setting of the on / off parameters of a home appliance, or the on / off parameters and at least one other parameter, the established device linkage ridge regression model predicts the setting values ​​of the on / off parameters of at least one other home appliance, or the associated parameters of each linked home appliance.

6. The method as described in claim 1, characterized in that: For multi-device linkage scenarios, a device linkage ridge regression model is established based on whether multiple home devices are linked on or off, and a separate single-device ridge regression model is established for each linked home device. Specifically, in response to a user setting at least one parameter of a home appliance, the established ridge regression model predicts the setting value of at least one home appliance associated with the home appliance, including: In response to a user turning a home appliance on or off, the established device linkage ridge regression model predicts whether at least one other home appliance needs to be turned on or off, and the established single-device ridge regression model predicts the setting values ​​of related parameters (excluding switch parameters) of at least two home appliances, including the aforementioned home appliance.

7. The method as described in claim 1 or 3, characterized in that: The predictions are based on the principle of not reducing the value of the dependent variable.

8. A control device, comprising a memory and a processor, characterized in that: The processor performs the method as described in any one of claims 1 to 7.

Images