Control method and device for household equipment and storage medium

Abstract

The invention relates to the technical field of smart home, and discloses a control method and device for home equipment and a storage medium, and the method comprises the steps: carrying out the adaptation of key features between a source domain data set and a target domain data set, the source domain data set comprises sensor data in a training environment, and the target domain data set comprises sensor data in a training environment; the target domain data set comprises sensor data in an actual use environment; performing feature alignment on the source domain data set and the target domain data set; training a home control model based on the source domain data set and the target domain data set; and controlling the home equipment to operate based on the real-time sensor data and the trained home control model. The home control model obtained by the method can better deal with complex and changeable situations, can control the home equipment to operate according to the corresponding mode based on the real-time sensor data under different situations, is beneficial to improving the automation degree of the home equipment, and ensures that the operation mode of the home equipment can meet user requirements.

Description

Technical Field

[0001] The present disclosure relates to the technical field of smart homes, for example, to a control method, device and storage medium for home appliances. Background Art

[0002] With the development of smart home technology, users have an increasingly strong demand for automatic adaptation of home devices. Different users may have very different usage habits and environmental conditions, and users expect home devices to automatically adjust their operating modes based on their usage habits and environmental conditions.

[0003] In the related art, preset rules or templates are usually used to achieve automatic control of home appliances, which is difficult to cope with complex and changing situations, requires a lot of manual intervention and debugging, is costly and cannot truly achieve automation.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] In order to provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. The summary is not an extensive review, nor is it intended to identify key / critical components or delineate the scope of protection of these embodiments, but rather serves as a prelude to the detailed description that follows.

[0006] The embodiments of the present disclosure provide a control method, device and storage medium for home appliances. The home control model thereof can better cope with complex and changeable situations, and can control home appliances to operate in corresponding modes based on real-time sensor data in different situations, thereby helping to improve the degree of automation of home appliances.

[0007] According to a first aspect of the present disclosure, a control method for a household device is provided, comprising:

[0008] Adapting key features between a source domain dataset and a target domain dataset, wherein the source domain dataset includes sensor data in a training environment, and the target domain dataset includes sensor data in an actual use environment;

[0009] Perform feature alignment on the source domain dataset and the target domain dataset;

[0010] Train the home control model based on the source domain dataset and the target domain dataset;

[0011] Control the operation of home devices based on real-time sensor data and trained home control models.

[0012] In some embodiments, adapting key features between a source domain dataset and a target domain dataset includes:

[0013] Calculate the distribution deviation degree corresponding to each feature dimension between the source domain dataset and the target domain dataset;

[0014] Based on the distribution deviation degree corresponding to each feature dimension, the key features between the source domain dataset and the target domain dataset are determined;

[0015] Adapt the key features between the source domain dataset and the target domain dataset.

[0016] In some embodiments, based on the distribution deviation degree corresponding to each feature dimension, determining the key features between the source domain dataset and the target domain dataset includes:

[0017] The feature dimension whose distribution deviation degree ranks before the Nth place is determined as the key feature between the source domain dataset and the target domain dataset, where N is a positive integer; or, the feature dimension whose distribution deviation degree is greater than a preset deviation degree threshold is determined as the key feature between the source domain dataset and the target domain dataset.

[0018] In some embodiments, the degree of distribution shift includes KL divergence or JS divergence.

[0019] In some embodiments, performing feature alignment on a source domain dataset and a target domain dataset includes:

[0020] A feature alignment network is used to perform preliminary feature alignment on the source domain dataset and the target domain dataset, and the feature alignment network is made to learn feature mappings that eliminate the differences between the source domain dataset and the target domain dataset;

[0021] The feature alignment network is used to regularize the source domain dataset and the target domain dataset to enhance the feature alignment effect.

[0022] In some embodiments, a feature alignment network is used to perform preliminary feature alignment on a source domain dataset and a target domain dataset, and the feature alignment network is made to learn a feature mapping that eliminates the difference between the source domain dataset and the target domain dataset, including:

[0023] Extract high-dimensional features from source domain datasets and target domain datasets;

[0024] The extracted high-dimensional features are preliminarily aligned using a feature alignment network, and the feature alignment network learns feature mapping that eliminates the differences between the source domain dataset and the target domain dataset.

[0025] In some embodiments, training a home control model based on a source domain dataset and a target domain dataset includes:

[0026] The features of the source domain dataset and the target domain dataset are extracted by the feature extractor, and the extracted features are input into the domain classifier in the feature alignment network, wherein an attention mechanism is used between the feature extractor and the domain classifier;

[0027] The input features are processed by a domain classifier, and the processed features are used to train a home control model, wherein a collaborative training mechanism is adopted between the home control model and the domain classifier.

[0028] According to a second aspect of the present disclosure, there is provided a control device for a household appliance, comprising:

[0029] A feature adaptation module is configured to adapt key features between a source domain dataset and a target domain dataset, wherein the source domain dataset includes sensor data in a training environment, and the target domain dataset includes sensor data in an actual use environment;

[0030] A feature alignment module, configured to perform feature alignment on a source domain dataset and a target domain dataset;

[0031] A model training module, configured to train a home control model based on a source domain dataset and a target domain dataset;

[0032] The device control module is configured to control the operation of home devices based on real-time sensor data and a trained home control model.

[0033] According to a third aspect of the present disclosure, a control device for home appliances is provided, comprising a processor and a memory storing program instructions, wherein the processor is configured to execute the control method for home appliances provided by the first aspect of the present disclosure when running the program instructions.

[0034] According to a fourth aspect of the present disclosure, a storage medium is provided, in which computer program instructions are stored. When the computer program instructions are executed by a processor, the control method for home appliances provided in the first aspect of the present disclosure is executed.

[0035] The control method, device and storage medium for home appliances provided by the embodiments of the present disclosure can achieve the following technical effects:

[0036] The control method for home appliances provided by the disclosed embodiment enables the home control model to learn the ability to control the operation of home appliances based on real-time sensor data by training the home control model based on the source domain data set and the target domain data set. The home control model can better cope with complex and changing situations, and can control the home appliances to operate in corresponding modes based on real-time sensor data in different situations, which helps to improve the automation level of home appliances. At the same time, the process of adapting and aligning key features helps to reduce the difference between the source domain data set and the target domain data set at a lower cost and higher efficiency, thereby improving the generalization ability of the home control model in the target domain (actual use environment), so that the home control model has a higher accuracy in the control ability of home appliances in the target domain, and ensures that the operation mode of the home appliances can meet the needs of users.

[0037] The foregoing general description and the following description are exemplary and explanatory only and are not intended to limit the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] One or more embodiments are exemplarily described by corresponding drawings, which do not limit the embodiments. Elements with the same reference numerals in the drawings are shown as similar elements, and the drawings do not constitute a scale limitation, and wherein:

[0039] Figure 1 is a flow chart of a control method for household appliances provided by an embodiment of the present disclosure;

[0040] Figure 2 is a flow chart of another control method for household appliances provided by an embodiment of the present disclosure;

[0041] Figure 3 is a flow chart of another control method for household appliances provided by an embodiment of the present disclosure;

[0042] Figure 4 is a schematic diagram of a feature alignment network provided by an embodiment of the present disclosure;

[0043] Figure 5 is a schematic diagram of a training process of a home control model provided by an embodiment of the present disclosure;

[0044] Figure 6 is a structural schematic diagram of a control device for household appliances provided by an embodiment of the present disclosure;

[0045] Figure 7 It is a structural schematic diagram of another control device for home appliances provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0046] In order to be able to understand the features and technical contents of the embodiments of the present disclosure in more detail, the implementation of the embodiments of the present disclosure is described in detail below in conjunction with the accompanying drawings. The attached drawings are for reference only and are not used to limit the embodiments of the present disclosure. In the following technical description, for the convenience of explanation, a full understanding of the disclosed embodiments is provided through multiple details. However, one or more embodiments can still be implemented without these details. In other cases, to simplify the drawings, well-known structures and devices can be simplified for display.

[0047] The terms "first", "second", etc. in the specification and claims of the embodiments of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate, so that the embodiments of the embodiments of the present disclosure described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions.

[0048] Unless otherwise stated, the term "plurality" means two or more.

[0049] In the embodiment of the present disclosure, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B indicates: A or B.

[0050] The term "and / or" is a description of the association relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or, A and B.

[0051] The term "correspondence" may refer to an association relationship or a binding relationship. The correspondence between A and B means that there is an association relationship or a binding relationship between A and B.

[0052] The disclosed embodiment provides a control device for home appliances, which can train a home control model based on a source domain data set and a target domain data set, and then use the trained home control model to control the operation of the home appliances. Here, the control device can be integrated into the home appliance. Alternatively, the control device can be an independent device, for example, the control device can be a server, or a local computing device (such as a personal computer, etc.).

[0053] Combination Figure 1 As shown, the embodiment of the present disclosure provides a control method for a home appliance, the execution subject of the method is a control device for the home appliance (hereinafter referred to as the control device), and the control method for the home appliance includes:

[0054] S101, a control device adapts key features between a source domain dataset and a target domain dataset.

[0055] In the embodiment of the present disclosure, the source domain data set includes sensor data in the training environment, and the target domain data set includes sensor data in the actual use environment. The sensor data may include environmental data and / or user data, wherein the environmental data may include temperature, humidity, brightness, etc., and the user data may include the number of indoor users, user trajectory data, and user face information, etc.

[0056] In the disclosed embodiment, by comparing the features of each dimension in the source domain dataset and the target domain dataset, the features with a large degree of difference are queried as key features, and then the key features are adapted. Here, for the identified key features, methods such as feature space calibration and sample resampling can be used to adapt the key features, reduce the distribution difference between the source domain dataset and the target domain dataset, and help improve the generalization performance of the home control model in the target domain.

[0057] In the disclosed embodiment, only the key features between the source domain dataset and the target domain dataset are adapted, which can enable more targeted optimization and improve the adaptation effect and efficiency.

[0058] S102: The control device performs feature alignment on the source domain dataset and the target domain dataset.

[0059] S103: The control device trains a home control model based on the source domain dataset and the target domain dataset.

[0060] S104, the control device controls the operation of the home appliances based on the real-time sensor data and the trained home control model.

[0061] The control method for home appliances provided by the disclosed embodiment enables the home control model to learn the ability to control the operation of home appliances based on real-time sensor data by training the home control model based on the source domain data set and the target domain data set. The home control model can better cope with complex and changing situations, and can control the home appliances to operate in corresponding modes based on real-time sensor data in different situations, which helps to improve the automation level of home appliances. At the same time, the process of adapting and aligning key features helps to reduce the difference between the source domain data set and the target domain data set at a lower cost and higher efficiency, thereby improving the generalization ability of the home control model in the target domain (actual use environment), so that the home control model has a higher accuracy in the control ability of home appliances in the target domain, and ensures that the operation mode of the home appliances can meet the needs of users.

[0062] In some embodiments, key features between a source domain dataset and a target domain dataset are adapted, including: calculating the degree of distribution deviation corresponding to each feature dimension between the source domain dataset and the target domain dataset; determining the key features between the source domain dataset and the target domain dataset based on the degree of distribution deviation corresponding to each feature dimension; and adapting the key features between the source domain dataset and the target domain dataset.

[0063] Combination Figure 2 As shown, the embodiment of the present disclosure provides another control method for a home appliance, and the control method for a home appliance includes:

[0064] S201, the control device calculates the distribution deviation degree corresponding to each feature dimension between the source domain data set and the target domain data set.

[0065] In some embodiments, the distribution shift degree may include KL divergence (Kullback-Leibler divergence), which is an asymmetric statistical distance metric used to measure the difference between two probability distributions. The KL divergence corresponding to the i-th feature dimension between the source domain dataset and the target domain dataset can be calculated by the following formula:

[0066] D KL (P||Q)=∑ i P(i)log Q(i)P(i).

[0067] In the formula, D KL (P||Q) represents the KL divergence, P and Q represent the probability distribution of the source domain and the target domain respectively, and P(i) and Q(i) reflect the distribution of the source domain and the target domain on the same feature dimension.

[0068] In some embodiments, the distribution shift degree may include JS divergence (Jensen-Shannon divergence), which is a symmetric version of KL divergence and is also a statistical distance metric. The JS divergence corresponding to the i-th feature dimension between the source domain dataset and the target domain dataset can be calculated by the following formula:

[0069]

[0070] D KL (P||M)=∑ i P(i)log M(i)P(i),

[0071] D KL (Q||M)=∑ i Q(i)log M(i)Q(i).

[0072] In the formula, DJS (P||Q) represents the JS divergence, P and Q represent the probability distribution of the source domain and the target domain respectively, and P(i) and Q(i) reflect the distribution of the source domain and the target domain on the same feature dimension.

[0073] S202: The control device determines the key features between the source domain dataset and the target domain dataset based on the distribution deviation degree corresponding to each feature dimension.

[0074] In some embodiments, based on the distribution deviation degree corresponding to each feature dimension, determining the key features between the source domain data set and the target domain data set includes: determining the feature dimension whose distribution deviation degree is ranked before the Nth place as the key feature between the source domain data set and the target domain data set. Wherein, N is a positive integer, and the value of N can be determined according to actual design requirements.

[0075] Specifically, the feature dimensions between the source domain dataset and the target domain dataset may be sorted in descending order of the distribution shift degree, and then the feature dimension with the distribution shift degree ranked before the Nth place is determined as the key feature.

[0076] In some embodiments, based on the distribution deviation degree corresponding to each feature dimension, the key features between the source domain dataset and the target domain dataset are determined, including: determining the feature dimension whose distribution deviation degree is greater than a preset deviation degree threshold as the key feature between the source domain dataset and the target domain dataset.

[0077] Specifically, the embodiment of the present disclosure may pre-set a deviation degree threshold, then compare the distribution deviation degree corresponding to each feature dimension with the deviation degree threshold, and determine the feature dimension whose distribution deviation degree is greater than the preset deviation degree threshold as a key feature.

[0078] S203: The control device adapts key features between the source domain dataset and the target domain dataset.

[0079] S204: The control device performs feature alignment on the source domain dataset and the target domain dataset.

[0080] S205: The control device trains a home control model based on the source domain dataset and the target domain dataset.

[0081] S206, the control device controls the operation of the home appliances based on the real-time sensor data and the trained home control model.

[0082] In some embodiments, feature alignment is performed on a source domain dataset and a target domain dataset, including: performing preliminary feature alignment on the source domain dataset and the target domain dataset using a feature alignment network, and enabling the feature alignment network to learn a feature mapping that eliminates the difference between the source domain dataset and the target domain dataset; and regularizing the source domain dataset and the target domain dataset using the feature alignment network to enhance the feature alignment effect.

[0083] In some embodiments, training a home control model based on a source domain dataset and a target domain dataset includes: extracting features of the source domain dataset and the target domain dataset using a feature extractor, inputting the extracted features into a domain classifier in a feature alignment network, wherein an attention mechanism is used between the feature extractor and the domain classifier; processing the input features by the domain classifier, and training the home control model using the processed features, wherein a collaborative training mechanism is used between the home control model and the domain classifier.

[0084] Combination Figure 3 As shown, the embodiment of the present disclosure provides another control method for a home appliance, and the control method for a home appliance includes:

[0085] S301, the control device adapts key features between a source domain dataset and a target domain dataset.

[0086] S302, the control device uses a feature alignment network to perform preliminary feature alignment on the source domain dataset and the target domain dataset, and enables the feature alignment network to learn a feature mapping that eliminates the difference between the source domain dataset and the target domain dataset.

[0087] Combination Figure 4 As shown in the figure, the feature alignment network includes a feature alignment module and a domain classifier. The feature alignment module attempts to learn a feature mapping function to map the features of the source domain dataset and the target domain dataset to a shared latent space. The domain discriminator attempts to distinguish whether the features in the latent space are from the source domain or the target domain. Through adversarial training, the feature alignment module and the domain discriminator compete with each other, perform preliminary feature alignment on the source domain dataset and the target domain dataset, and finally learn a feature mapping that can eliminate the differences between the source domain dataset and the target domain dataset to the greatest extent.

[0088] In some embodiments, a feature alignment network is used to perform preliminary feature alignment on a source domain dataset and a target domain dataset, and the feature alignment network is made to learn a feature map that eliminates the difference between the source domain dataset and the target domain dataset, including: extracting high-dimensional features from the source domain dataset and the target domain dataset; performing preliminary feature alignment on the extracted high-dimensional features using a feature alignment network, and making the feature alignment network learn a feature map that eliminates the difference between the source domain dataset and the target domain dataset.

[0089] Specifically, the embodiments of the present disclosure can use a pre-trained feature extraction model to extract high-dimensional features from the source domain dataset and the target domain dataset. The high-dimensional features contain rich semantic information of the original data. Figure 4 As shown in the figure, the feature alignment network includes a feature alignment module and a domain classifier. The feature alignment module attempts to learn a feature mapping function to map the high-dimensional features of the source domain dataset and the target domain dataset to a shared latent space. The domain discriminator attempts to distinguish whether the high-dimensional features in the latent space are from the source domain or the target domain. Through adversarial training, the feature alignment module and the domain discriminator compete with each other to perform preliminary feature alignment on the source domain dataset and the target domain dataset, and finally learn a feature mapping that can eliminate the differences between the source domain dataset and the target domain dataset to the greatest extent.

[0090] S303, the control device uses a feature alignment network to regularize the source domain dataset and the target domain dataset to enhance the feature alignment effect.

[0091] The disclosed embodiments may introduce some regularization techniques, such as MMD distance minimization (Maximum Mean Discrepancy Minimization) and Adversarial Dropout, etc. The feature alignment network regularizes the source domain dataset and the target domain dataset based on the corresponding regularization techniques to enhance the feature alignment effect, thereby helping the feature alignment network learn more robust and universal feature representations.

[0092] S304, the control device uses a feature extractor to extract features of the source domain data set and the target domain data set, and inputs the extracted features into a domain classifier in a feature alignment network.

[0093] In the disclosed embodiment, an attention mechanism is used between the feature extractor and the domain classifier. By introducing the attention mechanism, the feature extractor can adaptively focus on the important parts of the input data, enhance the ability to extract key features, and thus perform cross-module information fusion in a more targeted manner.

[0094] S305, the control device processes the input features through a domain classifier, and uses the processed features to train a home control model.

[0095] In the disclosed embodiment, a collaborative training mechanism is adopted between the home control model and the domain classifier. The collaborative training mechanism can promote joint optimization between modules, and through collaborative training, each module can learn from each other and improve performance together.

[0096] S306, the control device controls the operation of the home appliances based on the real-time sensor data and the trained home control model.

[0097] Combination Figure 5 As shown, Figure 5 This is a schematic diagram of the training process of a home control model provided by an embodiment of the present disclosure. The features of the source domain data set and the target domain data set are extracted by a feature extractor, and the extracted features are input into the domain classifier in the feature alignment network. At the same time, an attention mechanism is used between the feature extractor and the domain classifier. Here, the graph neural network can be integrated into the feature extractor, and the correlation information between data samples can be used to further improve the expression ability of the features.

[0098] For domain classifiers, in order to improve the robustness and generalization of domain classifiers, techniques such as adversarial training and meta-learning can be used. Adversarial training introduces adversarial samples to force the domain classifier to learn more robust feature representations, thereby improving performance in the face of interference or noise. Meta-learning allows the domain classifier to quickly adapt to the tasks of the target domain and enhance its generalization ability. Through transfer learning, the domain classifier can learn valuable features and knowledge from similar tasks or data, thereby achieving better classification results in the target domain. Optionally, the domain classifier can be a hierarchical classifier or a graph-based classifier. This hierarchical or graph-structured classifier can better capture the complex relationships between data, thereby making more accurate predictions.

[0099] After the extracted features are input into the domain classifier in the feature alignment network, the input features are processed by the domain classifier, and the processed features are used to train the home control model. A collaborative training mechanism is adopted between the home control model and the domain classifier.

[0100] For the home control model, meta-learning can be used to enable the home control model to quickly learn new tasks and reduce dependence on a large amount of training data. In addition, transfer learning can be used to utilize the knowledge of related tasks to enhance the performance of the home control model. Furthermore, the neural architecture search technology is introduced to automatically optimize the structure of the home control model. Through the neural architecture search algorithm, better network topology and hyperparameter configuration can be found, thereby improving the performance of the home control model. In order to design a more intelligent task model optimization strategy, reinforcement learning mechanism can be further combined. Reinforcement learning allows the home control model to learn more effective optimization methods through interaction with the environment.

[0101] Combination Figure 6 As shown, an embodiment of the present disclosure provides a control device 600 for home appliances, and the control device 600 includes a feature adaptation module 601, a feature alignment module 602, a model training module 603 and a device control module 604.

[0102] The feature adaptation module 601 is configured to adapt key features between a source domain dataset and a target domain dataset, wherein the source domain dataset includes sensor data in a training environment, and the target domain dataset includes sensor data in an actual use environment.

[0103] The feature alignment module 602 is configured to perform feature alignment on the source domain dataset and the target domain dataset.

[0104] The model training module 603 is configured to train a home control model based on a source domain dataset and a target domain dataset.

[0105] The device control module 604 is configured to control the operation of home devices based on real-time sensor data and a trained home control model.

[0106] The control device for home appliances provided by the disclosed embodiment enables the home control model to learn the ability to control the operation of home appliances based on real-time sensor data by training the home control model based on the source domain data set and the target domain data set. The home control model can better cope with complex and changing situations, and can control the home appliances to operate in corresponding modes based on real-time sensor data in different situations, which helps to improve the degree of automation of home appliances. At the same time, the process of adapting and aligning key features helps to reduce the difference between the source domain data set and the target domain data set at a lower cost and higher efficiency, thereby improving the generalization ability of the home control model in the target domain (actual use environment), so that the home control model has a higher accuracy in the control ability of home appliances in the target domain, and ensures that the operation mode of the home appliances can meet the needs of users.

[0107] In some embodiments, the feature adaptation module 601 is configured to:

[0108] Calculate the distribution deviation degree corresponding to each feature dimension between the source domain dataset and the target domain dataset;

[0109] Based on the distribution deviation degree corresponding to each feature dimension, the key features between the source domain dataset and the target domain dataset are determined;

[0110] Adapt the key features between the source domain dataset and the target domain dataset.

[0111] In some embodiments, the feature adaptation module 601 is configured to: determine the feature dimension whose distribution deviation degree is ranked before the Nth place as the key feature between the source domain dataset and the target domain dataset, where N is a positive integer; or determine the feature dimension whose distribution deviation degree is greater than a preset deviation degree threshold as the key feature between the source domain dataset and the target domain dataset.

[0112] In some embodiments, the degree of distribution shift includes KL divergence or JS divergence.

[0113] In some embodiments, the feature alignment module 602 is configured to:

[0114] A feature alignment network is used to perform preliminary feature alignment on the source domain dataset and the target domain dataset, and the feature alignment network is made to learn feature mappings that eliminate the differences between the source domain dataset and the target domain dataset;

[0115] The feature alignment network is used to regularize the source domain dataset and the target domain dataset to enhance the feature alignment effect.

[0116] In some embodiments, the feature alignment module 602 is configured to:

[0117] Extract high-dimensional features from source domain datasets and target domain datasets;

[0118] The extracted high-dimensional features are preliminarily aligned using a feature alignment network, and the feature alignment network learns feature mapping that eliminates the differences between the source domain dataset and the target domain dataset.

[0119] In some embodiments, the model training module 603 is configured to:

[0120] The features of the source domain dataset and the target domain dataset are extracted by the feature extractor, and the extracted features are input into the domain classifier in the feature alignment network, wherein an attention mechanism is used between the feature extractor and the domain classifier;

[0121] The input features are processed by a domain classifier, and the processed features are used to train a home control model, wherein a collaborative training mechanism is adopted between the home control model and the domain classifier.

[0122] Combination Figure 7 As shown, the embodiment of the present disclosure provides another control device 700 for home appliances, and the control device 700 for home appliances includes a processor (processor) 701 and a memory (memory) 702. Optionally, the control device 700 for home appliances may also include a communication interface (Communication Interface) 703 and a bus 704. Among them, the processor 701, the communication interface 703, and the memory 702 can communicate with each other through the bus 704. The communication interface 703 can be used for information transmission. The processor 701 can call the logic instructions in the memory 702 to execute the control method for home appliances of the above embodiment.

[0123] In addition, the logic instructions in the memory 702 described above may be implemented in the form of software functional units and when sold or used as independent products, may be stored in a computer-readable storage medium.

[0124] The memory 702 is a computer-readable storage medium that can be used to store software programs and computer executable programs, such as program instructions / modules corresponding to the method in the embodiment of the present disclosure. The processor 701 executes the function application and data processing by running the program instructions / modules stored in the memory 702, that is, the control method for the home appliance in the above embodiment is implemented.

[0125] The memory 702 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory 702 may include a high-speed random access memory and may also include a non-volatile memory.

[0126] An embodiment of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured to execute the above-mentioned control method for home appliances.

[0127] The technical solution of the embodiment of the present disclosure can be embodied in the form of a software product, which is stored in a storage medium and includes one or more instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiment of the present disclosure. The aforementioned storage medium may be a non-transient storage medium, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, and other media that can store program codes.

[0128] The above description and the accompanying drawings fully illustrate the embodiments of the present disclosure so that those skilled in the art can practice them. Other embodiments may include structural, logical, electrical, process and other changes. The embodiments represent only possible changes. Unless explicitly required, separate components and functions are optional, and the order of operation may vary. The parts and features of some embodiments may be included in or replace the parts and features of other embodiments. Moreover, the words used in this application are only used to describe the embodiments and are not used to limit the claims. As used in the description of the embodiments and the claims, unless the context clearly indicates, the singular forms of "a", "an" and "the" are intended to include plural forms as well. Similarly, the term "and / or" as used in this application refers to any and all possible combinations of listings containing one or more associated ones. In addition, when used in the present application, the term "comprise" and its variants "comprises" and / or comprising refer to the presence of stated features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or groups thereof. In the absence of further restrictions, the elements defined by the sentence "comprising a ..." do not exclude the presence of other identical elements in the process, method or device comprising the elements. In this article, each embodiment may focus on the differences from other embodiments, and the same and similar parts between the various embodiments may refer to each other. For the methods, products, etc. disclosed in the embodiments, if they correspond to the method part disclosed in the embodiments, then the relevant parts can refer to the description of the method part.

[0129] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software may depend on the specific application and design constraints of the technical solution. The technicians may use different methods for each specific application to implement the described functions, but such implementations should not be considered to exceed the scope of the embodiments of the present disclosure. The technicians may clearly understand that, for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above may refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0130] In the embodiments disclosed herein, the disclosed methods and products (including but not limited to devices, equipment, etc.) can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units can be 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. In addition, the coupling or direct coupling or communication connection between each other 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. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to implement this embodiment. In addition, each functional unit in the embodiment of the present disclosure may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0131] The flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to the embodiment of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowchart and the block diagram in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in a different order from the order disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.

Claims

1. A control method for household equipment, characterized in that: include: Adapting key features between a source domain dataset and a target domain dataset, wherein the source domain dataset includes sensor data in a training environment, and the target domain dataset includes sensor data in an actual use environment; Perform feature alignment on the source domain dataset and the target domain dataset; Train the home control model based on the source domain dataset and the target domain dataset; Control the operation of home devices based on real-time sensor data and trained home control models.

2. The control method for household appliances according to claim 1, characterized in that: Adapt the key features between the source domain dataset and the target domain dataset, including: Calculate the distribution deviation degree corresponding to each feature dimension between the source domain dataset and the target domain dataset; Based on the distribution deviation degree corresponding to each feature dimension, the key features between the source domain dataset and the target domain dataset are determined; Adapt the key features between the source domain dataset and the target domain dataset.

3. The control method for household appliances according to claim 2, characterized in that: Based on the distribution shift degree corresponding to each feature dimension, the key features between the source domain dataset and the target domain dataset are determined, including: The feature dimension whose distribution shift degree ranks before the Nth place is determined as the key feature between the source domain dataset and the target domain dataset, where N is a positive integer; Alternatively, a feature dimension whose distribution deviation is greater than a preset deviation threshold is determined as a key feature between the source domain dataset and the target domain dataset.

4. The control method for household appliances according to claim 2, characterized in that: The degree of distribution shift includes KL divergence or JS divergence.

5. The control method for household appliances according to any one of claims 1 to 4, characterized in that: Perform feature alignment on the source domain dataset and the target domain dataset, including: A feature alignment network is used to perform preliminary feature alignment on the source domain dataset and the target domain dataset, and the feature alignment network is made to learn feature mappings that eliminate the differences between the source domain dataset and the target domain dataset; The feature alignment network is used to regularize the source domain dataset and the target domain dataset to enhance the feature alignment effect.

6. The control method for household appliances according to claim 5, characterized in that: The feature alignment network is used to perform preliminary feature alignment on the source domain dataset and the target domain dataset, and the feature alignment network is made to learn feature mapping that eliminates the differences between the source domain dataset and the target domain dataset, including: Extract high-dimensional features from source domain datasets and target domain datasets; The extracted high-dimensional features are preliminarily aligned using a feature alignment network, and the feature alignment network learns feature mapping that eliminates the differences between the source domain dataset and the target domain dataset.

7. The control method for household appliances according to claim 5, characterized in that: The home control model is trained based on the source domain dataset and the target domain dataset, including: The features of the source domain dataset and the target domain dataset are extracted by the feature extractor, and the extracted features are input into the domain classifier in the feature alignment network, wherein an attention mechanism is used between the feature extractor and the domain classifier; The input features are processed by a domain classifier, and the processed features are used to train a home control model, wherein a collaborative training mechanism is adopted between the home control model and the domain classifier.

8. A control device for household equipment, characterized in that: include: A feature adaptation module is configured to adapt key features between a source domain dataset and a target domain dataset, wherein the source domain dataset includes sensor data in a training environment, and the target domain dataset includes sensor data in an actual use environment; A feature alignment module, configured to perform feature alignment on a source domain dataset and a target domain dataset; A model training module, configured to train a home control model based on a source domain dataset and a target domain dataset; The device control module is configured to control the operation of home devices based on real-time sensor data and a trained home control model.

9. A control device for household appliances, comprising a processor and a memory storing program instructions, characterized in that: The processor is configured to execute the control method for a household device according to any one of claims 1 to 7 when running the program instructions.

10. A storage medium, characterized in that: The storage medium stores computer program instructions, and when the computer program instructions are executed by the processor, the control method for home appliances according to any one of claims 1 to 7 is executed.

Images