Localized MAC duty-cycling adaptations for global energy-efficiency in Wireless Sensor Networks

Due to their physical constraints and deployment conditions, power management is a key issue in Wireless Sensor Networks. The radio component being the most responsible for energy wastage, controlling medium access is of prime importance in order to obtain desired network lifetime. Several existing mechanisms realize energy gains by alternating active and passive periods at the radio scale and most real WSN deployments implement static and homogeneous duty-cycling (i.e. invariant and identical for each node in the network). Although preventing any node isolation, such method fails to address the dynamics of the network efficiently. We propose a strategy to enable heterogeneous MAC duty-cycle configurations among nodes in the network. We aim at granting each node a specific sleep-depth, according to criteria specific to the deployment (e.g. applicative criteria, location in the routing structure). To implement this idea, sensor nodes are divided into disjoint subsets, each of them standing for a given duty-cycle configuration and leading to a network performance managed at its best (e.g. energy consumption, loss rate). We detail to what extent our approach preserves network connectivity with coherent heterogeneous duty-cycling, thus reaching a compromise between energy consumption and reactivity. The presented experimental campaign was led over the SensLAB testbed. It demonstrates that our solutions provide up to 61% energy saving, yet preserving the loss-rate below 10%, and guaranteeing the connectivity of the network.