Layered Attractor Selection for Clustering and Data Gathering in Wireless Sensor Networks

In this paper we propose a clustering and data gathering scheme for wireless sensor networks, based on a biologically-inspired approach that uses the concept of adaptive response by attractor selection as observed in a gene network. The approach uses a layered concept of clustering and routing that work independently, yet, resulting in an interdependent interaction via a common activity factor. This approach is aimed at establishing resilience and robustness in a two-layered protocol, where the network would perform a distributed clustering and routing of data to a sink. In this scheme, cluster heads are chosen according to their relative residual energy and routing is performed upon the relative cached data sizes and energy recharging rates of the next hop gateway nodes.