On the capacity of the relay channel with known state at the source

Abstract

We investigate the general relay channel with state that is noncausally known at the source. We first present a multiletter characterization of the capacity using the Gelfand-Pinsker encoding scheme. We then tailor the result to the general Gaussian relay channel with additive interference noncausally known at the source and compute a closed-form lower bound to the capacity using the joint dirty paper coding (DPC) and multidimensional linear relaying. We finally demonstrate that a potentially suboptimal evaluation of the proposed lower bound can simultaneously outperform known state-of-art coding strategies including DPC with signal description or state description along with compress-and-forward, noisy network coding and decode-and-forward.