Multiuser Random Coding Techniques for Mismatched Decoding

Jonathan Scarlett,Alfonso Martinez,Albert Guillén i Fàbregas
DOI: https://doi.org/10.1109/TIT.2016.2555317
2016-04-15
Abstract:This paper studies multiuser random coding techniques for channel coding with a given (possibly suboptimal) decoding rule. For the mismatched discrete memoryless multiple-access channel, an error exponent is obtained that is tight with respect to the ensemble average, and positive within the interior of Lapidoth's achievable rate region. This exponent proves the ensemble tightness of the exponent of Liu and Hughes in the case of maximum-likelihood decoding. An equivalent dual form of Lapidoth's achievable rate region is given, and the latter is shown to extend immediately to channels with infinite and continuous alphabets. In the setting of single-user mismatched decoding, similar analysis techniques are applied to a refined version of superposition coding, which is shown to achieve rates at least as high as standard superposition coding for any set of random-coding parameters.
Information Theory
What problem does this paper attempt to address?