Phone Call Detection Based on Smartphone Sensor Data

Smartphones are now equipped with as many as 30 embedded sensors, which have been widely used in human activity recognition, context monitoring, and localization. In this paper, we propose a phone call detection scheme using smartphone sensor data. We design Android applications to record, upload and display smartphone sensor data. We show how proximity and orientation sensors together can be used to accurately predict phone calls. Furthermore, the activity state during a phone call can be classified into three categories: sitting/standing, lying down, and walking. Features are extracted from proximity and orientation sensors to determine the range of values satisfying each state. Our system achieves an overall accuracy of 85 %.