Failure Node Detection and Recovery in Wireless Sensor Networks

Abstract

A Wireless Sensor Network may often consist of hundreds of distributed sensors. Our goal is to formulate wireless sensor networks (WSNs) fault identification problem in terms of pattern classification and to introduce a newly developed algorithm, neighbor node hidden conditional algorithm (NHCA) to determine the unknown path through which packets are transmitted from source to destination. We propose a concept of fault recovery in WSN using clustering. This includes the protocols of Dynamic Delegation based Efficient Broadcast and Neighborhood Hidden Conditional Random Field Algorithm. The history of transmission is classified according to the pattern, sorted and ranked along with the cluster information. The data privacy is maintained with in the cluster during the packet transmission apart from the destination which may present outside the cluster. The leader of the cluster is restricted only to view the transmission path in order to maintain the confidentiality of the data transmitted. Our simulation results strongly enforce fault recovery in quick time and also maintain the confidentiality of data.