Adaptive Payload-Agnostic LoRa Mesh Routing for Underground Environments

Abstract

Underground environments such as tunnels and buried infrastructure present major communication challenges due to severe attenuation, dynamic obstruction, and the limited practicality of extending wired networks as monitored zones evolve. While LoRa offers attractive long-range and low-power operation for such settings, conventional single-hop or staticrouting approaches are often insufficient for maintaining resilient end-to-end telemetry delivery. This paper presents an adaptive payload-agnostic LoRa mesh routing architecture for underground environments, designed to support reliable multi-hop forwarding from deep underground sensors to ground-level servers and monitoring platforms. The proposed system combines a three-tier LoRa mesh of sensor, relay, and edge nodes with LoRaWAN backhaul and a monitoring dashboard. Its key feature is a lightweight adaptive routing mechanism in which parent selection is driven by route rank, received signal strength indicator (RSSI), and queue occupancy. Sensor nodes apply single-parent hysteresis-based selection for route stability, while relay nodes maintain a multi-parent active set for adaptive forwarding under changing network conditions. A hardware prototype using ESP32-based LoRa nodes, WisGate gateways, and a Raspberry Pi dashboard was evaluated under baseline, high-load, stress, and node-failure scenarios. Across these tests, packet delivery ratio ranged from 78.7% to 95.8%, mean latency ranged from 1.9s to 6.2s, and throughput increased from 59 bps to 110 bps as traffic intensity rose. The results demonstrate that the proposed routing architecture supports resilient end-to-end operation under congestion and partial path degradation, while also highlighting remaining challenges in latency, security, and field-scale deployment.