An Efficient Outlier Detection Technique in Wireless Sensor Networks

摘要

An outlier in wireless sensor networks is a sensor reading generated by malfunctioned sensor nodes or external environmental effect. Since an outlier is an important indicator of environment change or malfunctioned sensors, many techniques to detect the outliers have been proposed. However, the previous outlier detection techniques mainly depend on the size of data clusters and consume much energy to obtain the outliers. In this paper, we propose an efficient outlier detection technique based on data clustering. In order to construct data clusters of the sensor readings, we adapt the Pigeonhole Principle, and then extract the outliers based on the clusters. In our experiment, we demonstrate the efficiency of our proposed technique compared to other outlier detection techniques.