Sovereign Routing: An AI-Driven Framework for Disaster-Resilient Communication Networks
NOEL G, MRUDHUL NR, AKILESHWARAKRISHNAN A
Department of Artificial Intelligence and Data Science, Nehru Institute of Engineering and Technology
Coimbatore, Tamil Nadu 641105, India
Mentor: Kirubhakaran Marisamy, Department of Artificial Intelligence and Data Science
ABSTRACT
Natural disasters such as earthquakes, floods, cyclones, and landslides frequently cause severe damage to conventional communication infrastructure, leading to network outages that disrupt coordination between emergency responders, rescue teams, and affected communities. Maintaining reliable communication during such situations is critical for effective disaster management and timely rescue operations. Traditional routing protocols are typically designed for stable network environments and often fail to adapt when network nodes or communication links become unavailable. Research in disaster communication systems highlights the importance of decentralized, adaptive, and self-organizing network architectures that can sustain connectivity even when existing infrastructure is partially or completely damaged. This paper proposes Sovereign Routing, an intelligent and resilient communication framework designed to maintain connectivity in disaster-affected environments. The proposed system integrates artificial intelligence techniques, including machine learning and reinforcement learning, to monitor network conditions, analyze link reliability, and dynamically determine optimal routing paths for data transmission. By enabling autonomous decision-making within network nodes, the framework reduces dependence on centralized control and enhances the ability of the network to self-organize and recover from failures. The architecture supports distributed communication using lightweight ESP32 hardware nodes which can form temporary mesh networks for emergency deployment. The framework is evaluated using a simulation environment developed with Python, NetworkX, and NS-3 to model network disruptions and node failures realistically. Experimental results indicate that the proposed routing strategy achieves a Packet Delivery Ratio of 93%, reduces average end-to-end latency by 56%, cuts network recovery time by 66%, and nearly triples throughput compared with conventional routing methods. These findings demonstrate that AI-driven routing mechanisms can significantly improve communication resilience during disasters and provide a scalable, cost-effective solution for maintaining connectivity in highly disrupted environments.
Keywords: Artificial Intelligence, Disaster Communication Networks, Reinforcement Learning, Adaptive Routing, Network Resilience, ESP32, IoT Communication, Emergency Communication Systems
