Markov Chain-Based Joint Queue-and-Channel-Aware QoS Assurance for Emergency Networks

Abstract

Emergency communication networks require reliable delivery of delay-sensitive traffic over time-varying wireless links, often under limited buffer resources and rapidly changing channel conditions. In such scenarios, physical-layer adaptation alone may be power efficient but queue blind, whereas queuedriven transmission control may protect reliability at excessive power cost. To address this issue, this paper proposes a Markov chain-based joint queue-and-channel-aware QoS assurance framework for finite-buffer emergency networks. A twodimensional discrete-time Markov chain (DTMC) is established to jointly characterize queue evolution and quantized fadingchannel dynamics. Based on this model, a cross-layer adaptive transmission policy is developed through normalized queuechannel weighting, zero-overflow feasible-set construction, and position-mapped probabilistic mode selection. The feasible set provides explicit buffer-overflow protection, while the probabilistic mapping exploits both queue state information (QSI) and channel state information (CSI) to avoid unnecessary highpower transmissions. Three benchmark policies, namely MACOnly, Water-Filling, and Channel Inversion, are considered for comparison. Numerical results show that, under the considered finite-buffer, constant-arrival, and truncated operational channel setting, the proposed policy achieves zero buffer overflow and improves the energy-delay-reliability tradeoff. At the reference SNR of 15 dB, it reduces the average transmit power by about 9.4% and 13.7% compared with MAC-Only and Channel Inversion, respectively, while providing substantially stronger delaytail control than the queue-independent Water-Filling benchmark.