On sensor selection in mobile devices based on energy, application accuracy, and context metrics

Abstract

Continuously collecting sensor data on mobile devices helps in capturing important contextual information related to the user location, activity, and surrounding environment. This contextual information is required by different context aware applications. However, continuously capturing this information incurs an unacceptable energy cost that leads to draining the limited phone's battery very quickly. In this paper, we propose a sensor selection procedure which decides what sensors to use according to energy, accuracy, and required context metrics. We present an accuracy-energy based algorithm that reduces energy consumption while conserving the required application accuracy. We use this algorithm to build a sensor selection model to choose among location sensors. Experimental evaluation of this model shows energy savings of up to 80% in selected scenarios.