Emergency communication is vital for search and rescue operations following natural disasters. Unmanned Aerial Vehicles (UAVs) can significantly assist emergency communication by agile positioning, maintaining connectivity during rapid motion, and relaying critical disaster-related information to Ground Control Stations (GCS). Designing effective routing protocols for relaying crucial data in UAV networks is challenging due to dynamic topology, rapid mobility, and limited UAV resources. This paper presents a novel energy-constrained routing mechanism that ensures connectivity, inter-UAV collision avoidance, and network restoration post-UAV fragmentation while adapting without a predefined UAV path. The proposed method employs improved Q learning to optimize the next-hop node selection. Considering these factors, the paper proposes a novel, Improved Q-learning-based Multi-hop Routing (IQMR) protocol. Simulation results validate IQMRs adaptability to changing system conditions and superiority over QMR, QTAR, and QFANET in energy efficiency and data throughput. IQMR achieves energy consumption efficiency improvements of 32.27%, 36.35%, and 36.35% over QMR, Q-FANET, and QTAR, along with significantly higher data throughput enhancements of 53.3%, 80.35%, and 93.36% over Q-FANET, QMR, and QTAR.

翻译：应急通信对自然灾害后的搜救行动至关重要。无人机凭借其灵活部署能力、快速运动中的连接维持能力，以及向地面控制站中继关键灾难信息的能力，可显著增强应急通信效能。由于动态拓扑结构、高速移动性及无人机资源受限等因素，设计高效路由协议以中继无人机网络中的关键数据极具挑战性。本文提出一种新型能量受限路由机制，该机制在无需预设无人机路径的自适应条件下，可实现连接保障、无人机间避碰及网络碎片化后的拓扑恢复。所提方法采用改进型Q学习算法优化下一跳节点选择。综合上述因素，本文提出名为改进Q学习多跳路由（IQMR）的新型协议。仿真结果验证了IQMR对系统动态变化的适应性，以及在能效与数据吞吐量方面相对于QMR、QTAR与QFANET的优越性。与QMR、Q-FANET及QTAR相比，IQMR分别实现32.27%、36.35%与36.35%的能耗效率提升，同时数据吞吐量较Q-FANET、QMR及QTAR分别显著提升53.3%、80.35%与93.36%。