Context-aware energy-efficient wireless sensor architecture for body activity recognition

In this paper, we present EEWAA, a new Energy-Efficient Wireless sensor Architecture for Activity recognition Applications. Contrary to existing approaches, our solution aims at optimizing the lifetime of every battery-operated devices that may contribute to the activity recognition, including sensors but also smart phones or tablets. Depending on the environment opportunities, our approach balance the load among multiple base stations (mobile phone, tablet, laptop) and when beneficial, enables the energy-limited base stations and/or sensors to offload a part of their tasks to back-end servers. In order to do this, we propose a new decision metric to select the best execution configuration, considering the latency and accuracy requirements, as well as the energy of every device involved in the computation.