A Lightweight User State Monitoring System on Android Smartphones

Smartphones are widely used around the world, which are also equipped with some sensors that can be used for the awareness of their users' state. These sensors include GPS, accelerometer, and microphone among others. In this paper, we present an empirical way to identify user's state including daily activity like walking, running, accidental threats like falling-down, and emotional status like sadness, joy, and anger. The monitoring should be realized in a non-intrusive way. We realize this idea by the design and implementation of a comprehensive run time user state monitoring system on Android smartphones, as less instructive as possible. The experiments show that it has good performance in terms of both monitored state accuracy and footprint incurred while conduct monitoring. The evaluations also show that the power consumption of the monitoring system is even neglectable which proves the usability of the proposed monitoring system.