Energy-synchronized computing for sustainable sensor networks

Abstract

To ensure sustainable operations of wireless sensor systems, environmental energy harvesting has been regarded as one of the most fundamental solutions for long-term applications. In energy-dynamic environments, energy conservation is no longer considered necessarily beneficial, because energy storage units (e.g., batteries or capacitors) are limited in capacity and leakage-prone. In contrast to legacy energy conservation approaches, we aim at energy-synchronized computing for wireless sensor devices. The starting point of this work is TwinStar, which uses ultra-capacitor as the only energy storage unit. To efficiently use the harvested energy, we design and implement leakage-aware feedback control techniques to match the activities of sensor nodes with dynamic energy supply from environments. We conduct system evaluation under both indoor and outdoor typical real-world settings. Results indicate our leakage-aware energy-synchronized control can effectively utilize energy that could otherwise leak away.