Mobile phones play a pivotal role in supporting ubiquitous and unobtrusive sensing of human activities. However, maintaining a highly accurate record of a user's behavior throughout the day imposes significant energy demands on the phone's battery. In this paper, we present the design, implementation, and evaluation of METIS: an adaptive mobile sensing platform that efficiently supports social sensing applications. The platform implements a novel sensor task distribution scheme that dynamically decides whether to perform sensing on the phone or in the infrastructure, considering the energy consumption, accuracy, and mobility patterns of the user. By comparing the sensing distribution scheme with sensing performed solely on the phone or exclusively on the fixed remote sensors, we show, through benchmarks using real traces, that the opportunistic sensing distribution achieves over 60% and 40% energy savings, respectively. This is confirmed through a real world deployment in an office environment for over a month: we developed a social application over our frameworks, that is able to infer the collaborations and meetings of the users. In this setting the system preserves over 35% more battery life over pure phone sensing.
Wristbands have been traditionally designed to track the activities of a single person. However there is an opportunity to utilize the sensing capabilities of wristbands to offer activity tracking services within the domain of team-based sports games. In this paper we demonstrate the design of an activity tracking system capable of detecting the players' activities within a one-to-one basketball game. Relying on the inertial sensors of wristbands and smartphones, the system can capture the shooting attempts of each player and provide statistics about their performance. The system is based on a two-level classification architecture, combining data from both players in the game. We employ a technique for semi-automatic labeling of the ground truth that requires minimum manual input during a training game. Using a single game as a training dataset, and applying the classifier on future games we demonstrate that the system can achieve a good level of accuracy detecting the shooting attempts of both players in the game (precision 91.34%, recall 94.31%).