Abstract
To save battery, consumer devices, e.g., laptops, smartphones, etc., enter low-power states based on user inactivity. Typically, a device enters a low-power state after a fixed time has elapsed since the last user activity, e.g., keyboard or touchscreen activity. However, a fixed timeout does not work evenly for all users; for example, it is found that a substantial fraction of users reactivate the device immediately after the device screen is dimmed or the device has entered sleep state. This disclosure describes machine learning techniques to predict the transition to a low-power state based on user activity patterns and the state of the device. The techniques result in improved user experience due to better prediction of the start of user inactivity and increased battery life due to accurate power management.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Napper, Jon; Amarilio, Omri; Erat, Dan; and Meng, Jia, "User Activity Prediction for Device Screen and Power Management", Technical Disclosure Commons, (March 19, 2020)
https://www.tdcommons.org/dpubs_series/3032