Analysis of Relay Channels and Improved Iterative Decoders for Turbo-Coded Decode-and-Forward Relay Channels
Three-node relay channels which consist of a source, a relay and a destination
are considered in this thesis. In the first part, an uncoded relay channel
is studied. The maximum likelihood (ML) detection rule for the detectand-
forward relaying in additive white Gaussian noise (AWGN) channels
which takes into account the detecting errors at the relay is formulated.
Based on this rule, the upper bound on the probability of error at the
destination’s receiver is analyzed. The source-relay-destination channel is
assumed to be equivalent to an AWGN channel with an equivalent signalto-
noise ratio (SNR) γeq. For the first time, the γeq is analyzed and its
exact analytical expression is derived.
In the second part, a more sophisticated relay channel called the turbocoded
decode-and-forward relay channel is considered. Two recursive systematic
convolutional (RSC) codes are implemented at the source and the
relay which form a turbo-like encoding scheme. An improved maximum
a posteriori decoder (IMAPD) which takes into account the decoding errors
at the relay is analyzed. This decoder is implemented in an iterative
manner similar to the traditional iterative decoder used for turbo codes.
The performance of this IMAPD is compared with the performance of an
iterative decoder commonly used in the literature which does not take into
account the decoding errors at the relay, by simulation. The comparison
shows that although the proposed IMAPD provides a better performance,
especially when many decoding errors occur at the relay, the improvement
is not significant. Then, another heuristical modification for the iterative
decoder is proposed. In spite of the lack of theoretical analysis, the numerical
results show that the proposed heuristically modified iterative decoder
(HMID) gives significantly better performance than the traditional one.
maximum a posteriori