Intelligent electric blanket temperature control method based on geometric smoothing momentum federated learning

By implementing geometric sliding momentum federated learning on smart electric blankets, the issues of privacy and security and communication overhead in the temperature control of smart electric blankets are solved, achieving efficient and stable personalized temperature control and adapting to the communication capability limitations of low-power devices.

CN122111138APending Publication Date: 2026-05-29SHENZHEN TENSOR BOX TECHNOLOGY CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SHENZHEN TENSOR BOX TECHNOLOGY CO LTD
Filing Date
2026-01-28
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing smart electric blankets suffer from data privacy risks, high communication overhead, and short battery life in adaptive temperature control. In particular, machine learning-based solutions cannot balance privacy and energy efficiency when transmitting physiological signal data.

Method used

A temperature control method based on geometric translational sliding momentum federated learning is adopted. By processing physiological signal data on local devices, a lightweight federated learning framework is used for model training. Communication with the aggregation server is only once. Combined with the geometric translational sliding momentum mechanism, the model update is smoothed, ensuring that the data does not leave the local device and reducing the communication burden.

Benefits of technology

It enables efficient and stable personalized temperature control on low-power IoT devices, ensuring user privacy and security, reducing communication overhead, and improving device battery life and temperature control accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122111138A_ABST
    Figure CN122111138A_ABST
Patent Text Reader

Abstract

The application discloses an intelligent electric blanket temperature control method based on geometric smooth momentum federated learning, which comprises the following steps: a global temperature control model parameter is initialized by an aggregation server, and the maximum aggregation round is set; an electric blanket is selected to participate in each aggregation round to obtain the current global temperature control model parameter; the electric blanket takes the current global temperature control model parameter as an initial value, is trained on a physiological signal feature dataset, and local loss is minimized to obtain a local temperature control model parameter; the aggregation server aggregates the local temperature control model parameter to obtain an updated global temperature control model parameter; the final global temperature control model parameter is repeatedly iterated and output; a physiological signal feature data is input into the trained global temperature control model to obtain a predicted temperature value; and the heating power is adjusted based on the difference between the predicted temperature value and the current blanket surface actual temperature to realize temperature control. The application greatly reduces the communication burden and realizes privacy protection.
Need to check novelty before this filing date? Find Prior Art