Randomized Sleeping for Energy Conservation and Robustness in Dense Sensor Networks

This work presents an approach for exploiting high density of homogeneous nodes in a sensor network to conserve energy by allowing nodes to “sleep”, while providing robustness to changing topologies and satisfying application constraints. Unlike deterministic algorithms that assume static connectivity, the approach proposed here does not require nodes to have awareness of their individual neighbors and uses a randomized algorithm to provide robustness to network connectivity that is time-varying due to fluctuations in channel quality, “death” and “birth” of nodes, and nodes’ powering down. The algorithm allows each node to decide for itself when to “sleep” or be “awake” based solely on local observations, provides for load balancing among all nodes in a region, and achieves energy savings proportional to the density of nodes within the region.