CN-122018351-A - Smart home control method, device, equipment and medium

Abstract

The application relates to the technical field of intelligent home, and discloses an intelligent home control method, device, equipment and medium. The method comprises the steps of obtaining user behavior data and equipment state data, conducting self-attention mechanism processing on the user behavior data to capture time sequence dependency characteristics among user behaviors to obtain first user characteristics, conducting association processing on the user behavior data and the equipment state data to capture space association characteristics among user behaviors and controlled equipment to obtain second user characteristics, generating dynamic user intention characteristics based on the first user characteristics and the second user characteristics, and conducting dynamic control on the controlled equipment based on the dynamic user intention characteristics. The embodiment of the application can improve the control precision of intelligent home control.

Inventors

• YANG QICHANG
• SHI WEIFENG

Assignees

• 深圳市华曦达科技股份有限公司

Dates

Publication Date

20260512

Application Date

20260210

Claims (10)

1. 1. The intelligent home control method is characterized by comprising the following steps of: Acquiring user behavior data and equipment state data; Performing self-attention mechanism processing on the user behavior data to capture time sequence dependency characteristics among user behaviors and obtain first user characteristics; performing association processing on the user behavior data and the equipment state data to capture spatial association characteristics between the user behavior and the controlled equipment so as to obtain second user characteristics; Generating a dynamic user intent feature based on the first user feature and the second user feature; and dynamically controlling the controlled equipment based on the dynamic user intention characteristic.
2. 2. The smart home control method of claim 1, wherein the self-attention mechanism processing of the user behavior data comprises: Performing time sequence coding processing on the user behavior data to obtain user behavior coding characteristics; Performing matrix transformation processing on the user behavior coding features to obtain a self-attention query vector, a self-attention key vector and a self-attention value vector; And carrying out self-attention mechanism processing based on the self-attention query vector, the self-attention key vector and the self-attention value vector to obtain the first user characteristic.
3. 3. The smart home control method according to claim 1, wherein the associating the user behavior data and the device state data includes: constructing a dynamic relationship graph based on the user behavior data and the equipment state data; based on the dynamic relationship graph, carrying out association analysis processing on the user behavior and the controlled equipment to obtain spatial association characteristics; And adjusting the dynamic relation graph based on the spatial association characteristic, and generating the second user characteristic based on the adjusted dynamic relation graph.
4. 4. The smart home control method according to claim 3, wherein the performing association analysis processing on the user behavior and the controlled device based on the dynamic relationship map includes: Generating an equipment node embedding vector based on the dynamic relation graph; Extracting characteristics of the embedded vectors of the equipment nodes to capture association characteristics between the user behaviors and the controlled equipment, so as to obtain equipment node state characteristics; Based on the dynamic relation graph and the equipment node state characteristics, carrying out association strength calculation on the user behavior and the controlled equipment to obtain an association strength value between equipment nodes; The spatial correlation feature is generated based on correlation strength values between the device nodes.
5. 5. The smart home control method of claim 1, wherein the generating dynamic user behavior features based on the first user features and the second user features comprises: Performing fusion processing on the first user characteristics and the second user characteristics to obtain fusion characteristics; based on the statistical significance index of the fusion feature, matching the fusion feature with a preset user intention template feature to inquire corresponding user intention reference feature; and carrying out intention level abstract reasoning on the fusion features and the user intention reference features based on a preset user intention level system to obtain the dynamic user intention features.
6. 6. The smart home control method of claim 5, wherein the fusing the first user characteristic and the second user characteristic comprises: determining the first user characteristic as a cross-attention query vector; performing transformation processing on the second user characteristic to obtain a cross attention key vector and a cross attention value vector; and performing cross attention processing based on the cross attention query vector, the cross attention key vector and the cross attention value vector to obtain the fusion characteristic.
7. 7. The smart home control method of claim 5, wherein performing intent level abstract reasoning on the fused features and the user intent reference features comprises: based on the fusion characteristics, carrying out characteristic embedding processing on the user intention reference characteristics to obtain user intention embedding characteristics; And carrying out feature mapping processing on the user intention embedded features to enable the user intention embedded features to be mapped layer by layer along a hierarchical path of the user intention hierarchical system, so as to obtain the dynamic user intention features.
8. 8. An intelligent home control device, which is characterized by comprising: the first module is used for acquiring user behavior data and equipment state data; The second module is used for carrying out self-attention mechanism processing on the user behavior data so as to capture time sequence dependency characteristics among the user behaviors and obtain first user characteristics; The third module is used for carrying out association processing on the user behavior data and the equipment state data so as to capture the space association characteristic between the user behavior and the controlled equipment and obtain a second user characteristic; a fourth module for generating a dynamic user intent feature based on the first user feature and the second user feature; And a fifth module, configured to dynamically control the controlled device based on the dynamic user intention feature.
9. 9. An electronic device comprising a memory storing a computer program and a processor implementing the smart home control method of any one of claims 1 to 7 when the computer program is executed by the processor.
10. 10. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the smart home control method of any one of claims 1 to 7.

Description

Smart home control method, device, equipment and medium Technical Field The application relates to the technical field of intelligent home, in particular to an intelligent home control method, device, equipment and medium. Background Along with the rapid development of intelligent home, the individuation and automation business volume brought along with the development of intelligent home are also rapidly promoted. The demand for home scene automation is also increasing, and with the development of the artificial intelligence model capability, the business is realized and processed through the artificial intelligence model, so that the learning and operation cost of a user can be reduced, and the intelligent experience is improved. However, the intelligent home control scheme in the related art has the defect of static modeling of user behavior, only relies on fixed time or single equipment interaction data, and is difficult to capture the behavior of dynamic change of a user and multi-equipment linkage habit, so that the personalized service accuracy is insufficient. Disclosure of Invention The application aims to provide an intelligent home control method, device, equipment and medium, which take dynamic user intention characteristics as scene control basis and can improve the control precision of intelligent home control. The embodiment of the application provides an intelligent home control method, which comprises the following steps: Acquiring user behavior data and equipment state data; Performing self-attention mechanism processing on the user behavior data to capture time sequence dependency characteristics among user behaviors and obtain first user characteristics; performing association processing on the user behavior data and the equipment state data to capture spatial association characteristics between the user behavior and the controlled equipment so as to obtain second user characteristics; Generating a dynamic user intent feature based on the first user feature and the second user feature; and dynamically controlling the controlled equipment based on the dynamic user intention characteristic. In some embodiments, the self-attention mechanism processing of the user behavior data includes: Performing time sequence coding processing on the user behavior data to obtain user behavior coding characteristics; Performing matrix transformation processing on the user behavior coding features to obtain a self-attention query vector, a self-attention key vector and a self-attention value vector; And carrying out self-attention mechanism processing based on the self-attention query vector, the self-attention key vector and the self-attention value vector to obtain the first user characteristic. In some embodiments, the associating the user behavior data and the device state data includes: constructing a dynamic relationship graph based on the user behavior data and the equipment state data; based on the dynamic relationship graph, carrying out association analysis processing on the user behavior and the controlled equipment to obtain spatial association characteristics; And adjusting the dynamic relation graph based on the spatial association characteristic, and generating the second user characteristic based on the adjusted dynamic relation graph. In some embodiments, the performing, based on the dynamic relationship graph, association analysis processing on the user behavior and the controlled device includes: Generating an equipment node embedding vector based on the dynamic relation graph; Extracting characteristics of the embedded vectors of the equipment nodes to capture association characteristics between the user behaviors and the controlled equipment, so as to obtain equipment node state characteristics; Based on the dynamic relation graph and the equipment node state characteristics, carrying out association strength calculation on the user behavior and the controlled equipment to obtain an association strength value between equipment nodes; The spatial correlation feature is generated based on correlation strength values between the device nodes. In some embodiments, the generating dynamic user behavior features based on the first user feature and the second user feature comprises: Performing fusion processing on the first user characteristics and the second user characteristics to obtain fusion characteristics; based on the statistical significance index of the fusion feature, matching the fusion feature with a preset user intention template feature to inquire corresponding user intention reference feature; and carrying out intention level abstract reasoning on the fusion features and the user intention reference features based on a preset user intention level system to obtain the dynamic user intention features. In some embodiments, the fusing the first user characteristic and the second user characteristic includes: determining the first user characteristic as a cross-attention query vector