A Novel Non-invasive Method to Measure Power Consumption on Smartphones

Abstract

Studies related to smartphone power consumption have intensified in recent years motivated not only by the significant increase in the number of users, but also by the increase in daily usage time by an average user. Therefore, considering that the power supply of mobile devices is limited by the battery capacity, an accurate method to measure power consumption in mobile devices is necessary. However, the current hardware approach used by researchers and industry can cause damage in non-removable battery devices or it may require a long time to measure precisely the power consumption on smartphones. In this paper, we use a new non-invasive method to measure the energy consumed by smartphones. The evaluation results confirm that this method ensures support for all models of smartphones that use the Android operational system based with Linux kernel so far. The switch mode method has an accuracy of 98,18% when compared with the current method used by industry and avoids the high costs of the invasive approaches. The power consumption of important features of a smartphone were measured and divided in five classes. The results show that camera, display, speaker and 5G are the hardware components with highest power consumption compared to the other parts of the smartphone. As future work, the results obtained shall be used to describe a model to estimate and reduce the high power consumption from hardware consumption classes.