Intelligent management message routing in ubiquitous sensor networks

Abstract

This paper first presents an intelligent ubiquitous sensor network implementation for agricultural and livestock farms and designs an efficient cyclic routing scheme for network management messages, aiming at improving productivity and profit in those industries. Instead of multiple point-to-point message transmissions for status indication collection, our management message traverses the specific set of nodes of interest one by one. The management application collects routing information from the neighbor table commonly available in the current sensor protocol to calculate the communication cost between each pair of target nodes. Based on this topology view, a genetic algorithm solver decides the traversal sequence of a cyclic management path. The experiment result discovers that the multithreaded version, in which each thread runs its own initial population, can find a much better solution, efficiently escaping local traps. © 2011 Springer-Verlag Berlin Heidelberg.