Optimal replay-based channel simulation via dithering methods

Abstract

We investigate replay-based channel simulation methods that reuse data collected from expensive field experiments through the use of dither within the standard communication pipeline. While traditional playback simulations rely on accurate channel models and their estimates, the method proposed here reduces the effect of channel estimation errors and unmodeled effects by exploiting similarities between different input signals. A notion of optimality of the proposed scheme will be discussed from the perspective of a related estimation problem. Finally, the proposed scheme is tested both numerically and experimentally with field data collected from MACE 10 experiments.