A biologically inspired networking model for wireless sensor networks

Wireless sensor networks have emerged in strategic applications such as target detection, localization, and tracking, where the large scale renders centralized control prohibitive. In addition, the finite batteries of the nodes demand energy aware network control. In this article we propose an energy-efficient topology management model inspired by biological intercell signaling schemes, which allows sensor nodes to form clusters around imminent targets in a purely distributed and autonomous fashion. In particular, nodes in the target vicinity collaborate to form a cluster according to their relative observation quality values, based on a bioinspired lateral induction process. Subsequently, the clustered nodes compete according to their energy levels until some of them gain active status while the rest remain idle, based on a bio-inspired lateral inhibition process. A final phase of the model has the active cluster members compete until one of them becomes the cluster head, again based on the lateral inhibition process. We examine the behavior of such a network control flow in both finite-size and infinite-size networks. Specifically, we show that the proposed model is inherently stable and discuss its convergence for networks of finite size. Furthermore, we discuss the asymptotic behavior when the number of nodes goes to infinity, where we study the average number of active cluster members.