On the Capacity of Compound State-Dependent Channels with States Known at the Transmitter

International audienceThis paper investigates the capacity of compound state-dependent channels with non-causal state information available at only the transmitter. A new lower bound on the capacity of this class of channels is derived. This bound is shown to be tight for the special case of compound channels with stochastic degraded components, yielding the full characterization of the capacity. Specific results are derived for the compound Gaussian Dirty-Paper (GDP) channel. This model consists of an additive white Gaussian noise (AWGN) channel corrupted by an additive Gaussian interfering signal, known at the transmitter only, where the input and the state signals are affected by fading coefficients whose realizations are unknown at the transmitter. Our bounds are shown to be tight for specific cases. Applications of these results arise in a variety of wireless scenarios as multicast channels, cognitive radio and problems with interference cancellation