|
AbstractA class of powerful error-correcting codes called
parallel concatenated codes, or turbo codes, has been proposed
recently. Such a code consists of two or more parallel recursive
systematic convolutional codes, each "subcode" is used to encode
the same but independently interleaved data sequence. Together
with a maximum a posteriori probability (MAP) iterative decoding
rule, turbo codes have been shown to achieve bit error rate
performance close to the Shannon's AWGN limit. The performance
and behavior of turbo codes in practical multipath fading
channels, however, remains to be determined. The fading effect
usually incurs burst errors that greatly hinder a channel code's
error correcting capability. Modification of the decoder
structure is needed to account for the influence of multipath
fading. The purpose of this thesis is to investigate the impacts
of channel conditions and various system design alternatives on
the overall system performance. Channel models used are:
independent and correlated Rayleigh fading channels and those
based on Jakes' model. Given the channel model we then study how
the channel state information is derived and such information
can be used to help decoding. Computer simulation is employed to
quantify various system behavior and to provide numerical
examples that give us system solutions.
|