Comparison of iterative MAP and SOVA receivers for the frequency non-selective fading channel

Thumbnail Image
Li, Yim Cheung
Journal Title
Journal ISSN
Volume Title
The Rayleigh frequency non-selective fading channel is a typical wireless communication channel model. When signals are transmitted over this kind of channel, the received signals experience a time-varying multiplicative gain plus the usual additive white Gaussian noise. The multiplicative gain can be modelled as a correlated Gaussian random variable. For interleaved coded signals transmitted over fading channels, a suboptimal yet practical two stage receiver has been proposed by Vitetta and Taylor. The first stage is a channel demodulator to combat the fading distortion and the second stage is a decoder. Later research by Gertsman and Lodge has shown that error performance can be greatly enhanced when the entire detection scheme is carried out iteratively because channel equalization can benefit from the coding structure. Both demodulator and decoder should be able to deliver soft-outputs for a better error correction capability. In this thesis, two receivers, which use Bahl 'et al''s MAP algorithm and Hagenauer and Hoeher's soft-output Viterbi algorithm (SOVA), are implemented and their error performances are compared. (Abstract shortened by UMI.)