FAULT DIAGNOSIS IN WIRELESS SENSOR NETWORKS USING IMPROVED FIREFLY ALGORITHM AND SUPPORT VECTOR MACHINE

Abstract

Wireless Sensor Network (WSN) consists of a large number of sensor nodes deployed in target areas for specific applications. Reliable transmission of data from the cluster head to the base station for processing is one of the most important challenges to ensure the proper performance of applications in wireless sensor networks. The distinction between good and bad data should be properly realized. Classification of applications in fault detection in wireless sensor network is one of the significant solutions in this field. A new model for detecting the accuracy of collected data based on improved support vector machine (SVM) classification for clustering data into cluster heads in a hierarchical wireless sensor network is proposed in this research. Optimum parameter setting in support vector machine to achieve accurate classification is done through improved firefly algorithm. Environmental information such as temperature, humidity, etc. is transmitted from the sensor nodes to the cluster heads so that the error data is categorized and transmitted to the base station. The simulation results of the proposed method show that the proposed approach provides an effective way to send correct data for WSN applications. This system achieves more than 99% accuracy throughout the data learning process. Data test results proved that the performance of the proposed method is more accurate compared to the base paper method (with accuracy equal to 0.9864).