Abstract

Wireless sensor network is an emerging technology and gaining importance because of its nature of sensing and diverse applications in which it is deployed. Just as there is an increase in applications, likewise erroneous data transmission in Wireless Sensor Network is inevitable. It is because sensor nodes may be deployed in harsh unattended environments. It may be disturbed by environmental factors or malicious nodes. During transmission of data, due to noise or external intervention, the data may turn into erroneous data. Such deviated data is an outlier. As the sensor nodes are used in crucial applications like industrial monitoring system or a patient monitoring systems, an erroneous data should be avoided. In this paper we propose to detect outliers based on Dempster-Shafer Theory. The results obtained are compared using NS2 and JTOSSIM simulators. By detecting outlier nodes, the resilience against node capture can be greatly enhanced as well. Comparative studies of challenges of sensor network and outlier detection methods are provided so that readers can have a rudimentary analysis of the security issues.

Key words: Termsoutlier, erroneous data, wireless sensor networks, Dempster-Shafer theory.