An Adaptive Mountain Clustering based Anomaly Detection for Distributed Wireless Sensor Networks

Abstract

Data inconsistency is an important issue in Wireless sensor networks (WSN). Due to the environmental change, the sensor nodes which are deployed in unattended area are damaged by various factors. Additionally, some of the sensor nodes may be hacked for stealing and altering its original sensed data, which affects the consistency of data sensed by faulty or hacked sensor nodes. The reason for this data inconsistency is the improper identification of anomalous data in the network. We propose an adaptive mountain clustering technique with new density measure for detecting anomalous data that avoids the unwanted data transmission to the end user and improves the data accuracy for the same. The proposed method successfully detects the anomalies with low false alarm rate.