INFLE: An Improved Neighbor-Based Fuzzy Logic Event Detecting Algorithm for Wireless Sensor Networks

Abstract

At present, event detection technologies have become an important part in building efficient wireless sensor networks. One of the popular and Excellent event detecting algorithms is the Neighbor-based Fuzzy Logic algorithm, namely NFLE, which belongs to machine learning technology. However, traditional fuzzy logic algorithm cannot work well in high-precision fire detecting networks. In this paper, we propose an improved NFLE (INFLE) which can dramatically increase the precision of fire detection. In NFLE, the final fire confidence of one node is partly determined by the average readings of its neighbors in fuzzy logic system, which may lead to inaccurate fire detection when event occurred in an area that covers only part of its neighbors. In our proposed INFLE, we select some of neighbors, by specific rules, to determine node’s final state. The simulation results validate that our proposed INFLE outperforms traditional NFLE in event detecting precision.