Dissertation committee:
Καθηγητής Νικόλαος Καλουπτσίδης (επιβλέπων), Καθηγητής Σέργιος Θεοδρωρίδης, Καθηγητής Θωμάς Σφηκόπουλος
Summary:
In this thesis, a family of low complexity convolutional codes is constructed,
by modifying appropriately the trellis diagram of punctured convolutional
codes. The goal is to improve performance at the expense of a reasonable low
increase of the trellis complexity. Many new convolutional codes of various
code rates and values of complexity are provided. In many cases, a small
increase in complexity can lead to a great improvement of performance, compared
to punctured convolutional codes. Furthermore, a method is presented for
designing new flexible convolutional codes, by combining the techniques of path
pruning and puncturing. The new codes can vary their rate, as well as the
complexity of their trellis diagram, and hence the computational complexity of
the decoding algorithm, leading to coding schemes that manage more efficiently
the system resources, compared to variable rate convolutional codes.
Finally, the possibility of applying the aforementioned results using recursive
convolutional encoders, which are used as constituent encoders in turbo codes,
is investigated. The goal is to construct flexible turbo coding schemes.
Simulation results indicate that in specific ranges of the SNR, a decrease in
the computational complexity of the decoding procedure can even result to a
decrease in the bit error rate.
Keywords:
Maximum likelihood decoding, Trellis diagram, Minimum trellis complexity, Convolutional codes, Turbo codes
