Sensor Health Monitoring in Wireless Sensor Networks

Wireless sensor network (WSN) has been widely used in information collection and monitoring applications. At the center of attention in these applications are data, which come from the sensor part of each WSN node. Sensor faults are common due to the sensor device itself and the harsh deployment environment. These faults may degrade the data quality from the very beginning and thus have a possible large influence on the final success of specific WSN application. Our work presented in this paper is to monitor the health status of sensors in WSN. Inspired by the latest development of time series data mining technology, a flexible approach is proposed for unspecific fault monitoring. Each deployed WSN node learns the frequent patterns periodically. Any sensor fault follows a pattern that is different from the nodepsilas learned frequent patterns. The proposed method can find fault patterns efficiently. We show how this method works well in Castalia simulation environment.