Towards distributed wireless intelligent sensor networks

Abstract

Over the past few years, approaches including artificial intelligence models into Wireless Sensor Networks have gained much attention within the research community. However, little attention has been paid to integrate Fuzzy Rule Based-Systems into Wireless Sensor Network. Since traditional approaches of Fuzzy Rule Based-Systems cannot be applied, due to limited resources that the sensors nodes have, the principal aim of this work is to design a distributed knowledge based Wireless Sensor Network, where each node of the network executes an adapted Fuzzy Rule Based-System. This aim has been divided into the followings: 1) to design a Fuzzy Rule Based-System adapted to a sensor node, 2) to design an application protocol which allows transmitting knowledge bases to sensors nodes and 3) to simulate a particular scenario with network simulator ns-2. The performance of a Sun SPOT sensor has been evaluated and a comparison to other devices is reported. The results show that a Sun SPOT sensor can execute properly a Fuzzy Rule Based-System adapted to it, and that it is possible to create a useful distributed knowledge based Wireless Sensor Network.