Machine learning approach for secure communication in wireless video sensor networks against denial-of-service attacks

Mobile ad hoc networks (MANETs) own a flexible framework with the absence of a server, where conventional security components fail to compensate the level of MANET security conditions since it is confined to a particular environment, its data transfer potential, and battery and memory constrains. MANET provides a well-grounded path and an efficiency in communication, but the confidentiality of the trust parameters remains a great challenge since it may be overheard by the impostor. This demands the need of exchanging the encrypted mathematical values. The proposed machine learning security paradigm provides firm and trustworthy network in spite of establishment over additional network platform. The QoS is improved through support vector machine for denial-of-service attack. The node has to be clustered to accomplish its respective task. The clustering is done with the help of LEACH protocol, where cluster head and Cluster member are fixed to transfer the data in the network. Low Energy adaptive clustering heirarchy (LEACH) propagates energy to abstain from draining of battery and malignant network. A secure framework is built along with encryption and decoding to protect from denial-of-service attack. Acknowledgement-based flooding attack has been focused with the help of support vector machine algorithm. The messages are encoded in from the source node and coded again during transmission phase to obtain the original message. Defending the traditional methodologies, the proposed work provides excellent QoS when compared and tested with other protocols. The results obtained ensure its efficiency when support vector machine technique is combined with encryption scheme.