Activity management through Bernoulli scheduling in 802.15.4 sensor clusters

The activity of individual nodes in a sensor network has to be managed in order to meet the simultaneous goals of maximizing the network lifetime and achieving the desired information throughput at the network sink. In this work we investigate the interaction of activity management with the CSMA-CA-based MAC layer in a beacon enabled IEEE 802.15.4-compliant network cluster. Activity management is performed through Bernoulli scheduling of service and sleep periods, which gives better results than the traditional, exhaustive and I-limited scheduling of the active and sleep periods. We derive the event reliability at the cluster coordinator as the function of physical layer parameters, data link layer parameters, Bernoulli scheduling parameter, and the probability distribution of sleep period. Based on the centralized analytical model, we derive distributed algorithm for calculating sleep parameters and demonstrate that the proposed distributed activity management policy is capable of achieving and maintaining the desired network reliability while maximizing the network lifetime.