Bayesian link margins for opportunistic downlink packet systems

Downlink cellular packet systems utilizing opportunistic scheduling and adaptive modulation must deal with outdated channel state information. A fade margin is part of the link budget which is used to mitigate the effect of channel variation. Channel prediction reduces this margin, improving the transmission rate. We derive a fade margin which keeps packet outage constant for any prediction error variance. In our model both the signal and the interferer powers are modeled as autoregressive stochastic processes and predicted. The MMSE power predictor is also derived. Improvement in system throughput due to channel prediction is shown for a proportional fair scheduled downlink packet system, and compared to systems without prediction and opportunistic scheduling.