A deep Bi-LSTM based Fault Detection Algorithm for WSNs

Abstract

The demand for the Internet of Things (IoT) has raised in today's era significantly. Wireless Sensor Networks (WSNs) is the main component of IoT. However, the sensor nodes are prone to failures due to their deployment in hostile environments. In WSNs, fault in sensor nodes extensively degrades the network performances. The existing fault detection approaches are unable to diagnosis different types of faults in the sensor. Also, the fault detection accuracy of the existing approaches is very less. In this paper, a deep Bidirectional LSTM model is used to detect the fault in nodes. The proposed methodology can detect various types of faults such as soft permanent, intermittent and transient faults. The results show that the proposed algorithm outperforms the other state-of-the- art fault detection approaches in terms of fault detection accuracy, false alarm rate, and false-positive rate.