AI- DRIVEN INTRUSION DETECTION AND ERROR CONTROL MECHANISMS FOR NEXT - GENERATION IOT NETWORKS
DOI:
https://doi.org/10.7492/z76h6f17Keywords:
AI-based cybersecurity, Machine Learning (ML) intrusion detection, Error-Correction Codes in network security, Internet of Things (IoT) network protection, security using Convolutional Neural Networks (CNNs), Reed-Solomon and Turbo Codes, Real-time threat detection, secure data transmission in IoTAbstract
The rapid expansion of Internet of Things (IoT) networks has brought forth critical cybersecurity concerns, necessitating the
development of advanced strategies to ensure secure and dependable communicationsTraditional security measures frequently fail to mitigate
new cyber threats such as DDoS attacks, malware attacks and data integrity attacks. This study presented an AI augmented network security
framework designed to involve ML, AI and error-correcting codes (ECCs) to help mitigate attacks through new intrusion detection and data
security in the IoT environment. Using deep-learning models including convolutional neural networks (CNNs) and adaptive particle swarm
optimization (APSO), the authors sought to facilitate rapid threat identification in real time. In addition, the use of ECCs like Reed-Solomon
and Turbo Codes were used to minimize errors in transmission and maintain integrity in wireless environments.
The graduated experimental evaluations on standard IoT datasets, NSL-KDD and CICIDS-2017, resulted in improved detection rates +7.64%,
decreased false positive rates -15%, and improved transmission reliability +30% compared with conventional security approaches.
