Optimal Packet Scheduling using Adaptive M-QAM and Orthogonal STBC in MIMO Nakagami-m Fading Channels

We study a cross-layer optimization problem of a rate-adaptive MQAM system that maximizes throughput, and minimizes packet error rate, delay and overflow using the framework of Markov decision process. We consider finite size buffer and random incoming traffic arrivals. The fading channel is assumed to be Nakagami-m and modeled with a finite state Markov channel, and transmit diversity is achieved with orthogonal space-time block code. To schedule packets, the transmitter depends on both buffer state and channel state information, and selective-repeat ARQ is used at the data link layer for error detection. Simulation results show that the throughput maximization depends not only on the received SNR but also on the number of actions and incoming traffic statistics.