Improved Fault Diagnosis in Wireless Sensor Networks using Deep Learning Technique
G. Kiruthiga1, P. MayilVel Kumar2, K.M.Murugesan3, T. Yawanika4

1Dr. G. Kiruthiga, IES College of Engineering, Chittilappilly (Kerala), India.
2Dr. P. MayilVel Kumar, Karpagam Institute of Technology, Coimbatore (Tamil Nadu), India.
3Dr. K. M. Murugesan, Karpagam Institute of Technology Coimbatore (Tamil Nadu), India.
4T. Yawanika, Karpagam Institute of Technology Coimbatore (Tamil Nadu), India.
Manuscript received on 06 June 2019 | Revised Manuscript received on 30 June 2019 | Manuscript Published on 04 July 2019 | PP: 757-760 | Volume-8 Issue-1S4 June 2019 | Retrieval Number: A11400681S419/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (

Abstract: In recent times, Wireless Sensor Network (WSN) has increased its attention due to its positive impact on surveillance in its surrounding environment. Numerous researches have been report since decades, however, the studies on diagnosing the network fault in critical conditions have received little attention. This could be another area of interest in WSN to increase its overall lifespan and network scalability. However, owing to ad-hoc characteristics of sensor nodes, the scalability of network reduces and this makes the network administrator to poorly observe the conditions of the network. The other major limitations associated with the fault diagnosis in WSN includes: short communication range, limited energy resource, limited processing power, low bandwidth, storage in sensor node, conditionally independent transmission of signals, high power in transmission and signal acquisition and faulty sensory reading under harsh operating condition. The present study considers improving the lifetime and scalability of sensor nodes using passive fault diagnosis using a deep learning approach named Conventional Neural Network. This method effectively classifies the faulty sensor nodes and eliminates it from communicating with other sensor nodes.
Keywords: Recurrent Neural Network, Fault Diagnosis, Network Scalability, WSN.
Scope of the Article: Deep Learning