ResQ: Edge AI Mesh Networking for Disaster Triage

Mrs.Suma Rajesh1, S Avinash2, S Harsha3, Shamith Gowda4, Vishal Vass5

Assistant Professor, Dept of CSE, KSIT, Karnataka, India1
Student, Dept of CSE, KSIT, Karnataka, India2-5

ABSTRACT
Centralized communication infrastructures, including cellular backhauls and Internet Service Providers (ISPs), repre-sent a "silent but deadly" single point of failure during cata-strophic events such as floods, earthquakes, and cyclones. The resulting "information vacuum" often leads to a surge in mor-tality rates due to the collapse of situational awareness among first responders. This paper introduces ResQ, a decentralized communication framework designed to transform heterogene-ous consumer smartphones into an autonomous, self-healing mesh fabric.

The ResQ architecture provides a threefold contribution:

(1)  a self-forming peer-to-peer (P2P) mesh networking layer utilizing Wi-Fi Direct and Bluetooth Low Energy (BLE) with asynchronous data muling capabilities;

(2)  an edge-based Natural Language Processing (NLP) intent classification system for automated triage; and

(3)   an AI-optimized routing protocol that utilizes ma-chine learning to select relay nodes based on real-time device telemetry. By deploying quantized inference models—specifi-cally Decision Trees (DT), Random Forest (RF), Support Vec-tor Machines (SVM), and XGBoost—directly on the edge, ResQ ensures that high-priority distress signals are routed even in zero-connectivity environments. This framework represents a necessary evolution beyond traditional network paradigms, offering a scalable and resilient infrastructure for modern search and rescue (SAR) operations.

Keywords: Wireless Mesh Networks, Edge AI, Disaster Management, Multi-hop Routing, Triage, Ad-Hoc Net-working, Natural Language Processing, Distributed Sys-tems.