A Survey of Methods and Architectures for Multi-Attack Detection using Forensics and Coherent Integrated Photonic Neural Networks-Based Prevention for Secure IoT-MANETs

Abstract

The rapid growth of the Internet of Things (IoT) and Mobile Ad Hoc Networks (MANETs) has enabled decentralized and scalable communication, but their distributed nature and resource constraints make them highly vulnerable to cyber-attacks such as DDoS, black hole, wormhole, and spoofing, which degrade performance and compromise data integrity. Traditional intrusion detection systems (IDS) often fail to address these complex threats due to limited scalability and adaptability. To overcome these challenges, recent approaches leverage deep learning, forensic analytics, and photonic neural networks for efficient multi-attack detection. Deep learning models, including CNN, LSTM, and hybrid architectures, have shown high accuracy in identifying anomalous traffic patterns, while forensic techniques enhance post-attack analysis and support proactive defense strategies. Additionally, photonic neural networks provide ultra-fast processing with reduced latency and energy consumption, making them suitable for real-time IoT-MANET environments. Comparative studies indicate that hybrid deep learning models combined with optimization techniques outperform conventional methods in terms of detection accuracy and computational efficiency. However, challenges such as dataset imbalance, real-time deployment, and energy efficiency still persist. This study highlights key advancements, identifies research gaps, and provides insights into future directions for developing robust and scalable intrusion detection mechanisms to secure IoT-MANET systems against multi-attack scenarios.