Outlier Detection Approach Using Bayes Classifiers in Wireless Sensor Networks

Abstract

Wireless sensor networks (WSN) have become a new information collection and monitoring solution for a variety of applications. Sensor nodes may occasionally produce incorrect measurements due to battery depletion, damage of device and other causes. Those measurements that significantly deviate from the normal pattern of sensed data are considered as outliers. To address the problem of outlier detection in WSN, we propose in this paper a two-level sensor fusion-based outlier detection technique for WSN. The first level of outlier is conducted locally inside the sensor nodes, while the second level is carried out in a level higher (e.g., in a cluster head or gateway). The proposed approach functionality was tested by simulation using a real sensed data obtained from Intel Berkeley Research Lab. The experiment results show that the approach achieved a high-level of detection accuracy and a low percentage of false alarm rate.