Proposed Security Models for Node-level and Network-level Aspects of Wireless Sensor Networks Using Machine Learning Techniques

Abstract

     As a result of the pandemic crisis and the shift to digitization, cyber-attacks are at an all-time high in the modern day despite good technological advancement. The use of wireless sensor networks (WSNs) is an indicator of technical advancement in most industries. For the safe transfer of data, security objectives such as confidentiality, integrity, and availability must be maintained. The security features of WSN are split into node level and network level. For the node level, a proactive strategy using deep learning /machine learning techniques is suggested. The primary benefit of this proactive approach is that it foresees the cyber-attack before it is launched, allowing for damage mitigation. A cryptography algorithm is put forth and contrasted with the current algorithms at the network level. Elliptic Curve Cryptography combined with the Koblitz encoding technique produced superior results. By implementing machine learning and deep learning techniques, wireless sensor networks are protected against cyber-attacks, and the suggested encryption approach ensures the confidentiality of data transfer. The estimated encryption and decryption times were evaluated with various file sizes and contrasted with the current systems. The suggested solutions were successful in achieving security at both the node level and network level.