New trellis codes for Rayleigh fading channels

Abstract

There has been a considerable number of trellis codes published for Rayleigh fading channels. Here we challenge the usual assumption that in the construction of good trellis codes the number of constellation points should only be doubled. Removing this assumption leads to codes which significantly outperform previously published codes without increasing the decoding complexity. To aid in the selection of the codes we also present a modified method of selecting good codes.