The Dispersion of Universal Joint Source-Channel Coding for Arbitrary Sources and Additive Channels
share
We consider a universal joint source channel coding (JSCC) scheme to transmit an arbitrary memoryless source over an arbitrary additive channel. We adopt an architecture that consists of Gaussian codebooks for both the source reproduction sequences and channel codewords. The natural minimum Euclidean distance encoder and decoder, however, need to be judiciously modified to ensure universality as well as to obtain the best (highest) possible communication rates. In particular, we consider the analogue of an unequal error (or message) protection scheme in which all sources are partitioned into disjoint power type classes. We also regularize the nearest neighbor decoder so an appropriate measure of the size of each power type class is taken into account in the decoding strategy. For such an architecture, we derive ensemble tight second-order and moderate deviations results. Our first-order result generalizes seminal results by Lapidoth (1996, 1997). The dispersion of our JSCC scheme is a linear combination of the mismatched dispersions for the channel coding saddle-point problem by Scarlett, Tan and Durisi (2017) and the rate-distortion saddle-point problem by the present authors, thus also generalizing these results.
READ FULL TEXT