Edge intelligence for automated off-grid structural triage in day-zero earthquake response

Abstract

Post-earthquake structural damage assessment traditionally relies on manual visual inspections, creating critical bottlenecks during disaster response. To automate this process, this paper introduces ChilloRa, an end-to-end edge-to-GIS framework for rapid Day-Zero structural triage. Unlike conventional data-logging sensors, ChilloRa implements a physics-informed Finite State Machine (FSM) directly at the edge, autonomously converting raw kinematic measurements into machine-driven structural damage states. To address communication bottlenecks, the framework integrates a burst-resilient dual-receiver LoRa gateway that mitigates collision-induced channel congestion. Under a simulated 2000-node network stress test scenario, the proposed cyber-physical architecture achieved a near-lossless 99.84% Packet Delivery Ratio while maintaining automated telemetry aggregation. The resulting decentralized data stream is automatically processed through a custom ArcGIS toolbox to generate prioritized building-level risk maps. By transforming manual scouting into a zero-touch, data-driven spatial intelligence pipeline, ChilloRa provides a scalable framework for post-earthquake infrastructure management and supports rapid Urban Search and Rescue (USAR) prioritization.