SmartFieldMesh: A Self-Healing IoT Mesh Network for Resilient Communication in Harsh Environments

IoT mesh networks deployed in harsh environments face frequent node dropouts, volatile links, and limited access to centralized control. Under these conditions, widely used mesh protocols such as Zigbee and Thread often recover slowly and incur higher control overhead, which reduces reliability and drains constrained devices. This paper introduces SmartFieldMesh, a fully decentralized self-healing mesh framework that combines real-time anomaly cues, trust-aware neighbor selection, and federated reinforcement learning. Each node monitors local communication symptoms to flag abnormal behavior, maintains a short-horizon trust score for its neighbors, and updates its routing policy from local experience while periodically exchanging compact model updates with nearby peers. Simulation results show that SmartFieldMesh improves packet delivery and recovery latency, and extends network lifetime relative to the evaluated baselines. The design remains lightweight and scalable, making it practical for deployments such as disaster response, remote sensing, and tactical IoT applications.