An Integrated Manet Approach For Biometric Authentication System: Applications And Comparative Analysis Of Deep Learning Techniques

Abstract

Security is a major concern for providing trusted communications in mobile ad hoc networks (MANETs) in a potentially hostile environment. This concern is mainly due to the peer-to-peer (P2P) architecture in MANETs, system resource constraints, shared wireless medium, and highly dynamic network topology. Authentication is a common prevention-based approach used in MANETs to reduce intrusions. Biometrics provides some possible solutions to authentication used in MANETs since it has the direct connection with user identity and needs little user interruption.The biometric data can't be recreated or hacked, so it makes the distinctive verification more strong.Unimodal biometrics has to face several challenges such as noise in sensed data, intra-class variations, inter-class similarities. Security problems could be resolved by adopting multimodal biometric systems. Multimodal biometric system presents a more reliable authentication method due to the combination of statistically independent biometric traits.In this article, different varieties of bio-metric approaches and the deep learning methods are compared where effective algorithm is identified by the comparison.