A Trust-Aware DRL-Assisted Secure Routing and Anomaly-Resilient Alert Aggregation Framework with Blockchain Validation for Underwater IoT Networks

Abstract

Secure, reliable and energy-efficient communication is crucial for underwater Internet of Things (UIoT) networks, particularly in the context of critical environmental monitoring applications. This paper suggests a trust-aware DRL-aided secure routing and alert aggregation system that validates with the help of blockchain in underwater IoT networks. Deep Reinforcement Learning is used to dynamically optimize routing decisions based on residual energy, latency, trust scores and alert delivery reliability. A light-weight cryptographic mechanism is used to ensure the safe and low-latency delivery of environmental alerts. Anomaly detection at the sink layer is built in to detect and isolate malicious or corrupted packets, thereby providing enhanced robustness of the network. Moreover, blockchain validation offers tamper-proof and transparent gathering of alert information. Experimental results show that the system outperforms the traditional solutions in terms of packet delivery, packet delay, trust accuracy and resilience to security attacks. The proposed framework is an efficient and scalable solution for the secure monitoring system for the underwater environment.