Decentralized fuzzy controlling for target classification using wireless sensor networks

Target classification is one of the applications of wireless sensor networks that aims to recognize the type of mobile targets that navigate within a sensing field. This paper presents a fuzzy-based controller module using MaxMin and MinMax Distributed K-Nearest Neighbors (DKNN) algorithms for ground vehicle classification in order to achieve efficient energy usage and better classification accuracy. This fuzzy module has embedded in an existing target classification system. The fuzzy-based controller module handles the wireless sensor nodes sensing rate (refresh rate) dynamically. A simulation-based study has carried out to test our approach and the simulation results have compared to well-known MaxMin and MinMax DKNN algorithms from literature. Simulation results show that our proposed approach prolongs the network lifetime and achieves better target classification accuracy.