| Summary: | 博士 === 國立臺灣大學 === 電機工程學系 === 86 === In this thesis, we propose new multilevel coding schemes, for
which large free distances and low decoding complexities can
be easily achieved. In Chapter 2, a multilevel coding scheme
is proposed for which the encoding is implemented by using a
binary encoder, followed by a delay processor and a signal
mapper. This proposed coding scheme is a generalization of
the trellis coding scheme proposed by Hellstern in 1993.
With the proposed coding scheme, we have the flexibility of
controlling the decoding delay.
The coding technique proposed in Chapter 2 can be iteratively
applied to construct a more generalized coding scheme which is
described in Chapter 3. The encoding is implemented by using a
binary encoder, followed by multiple pairs of delay processors
and signal mappers. In general, a code constructed by this
method will have larger distance properties than by using
a single delay processor and a signal mapper. We derive
distance bounds for this proposed coding scheme. Simulation
results are provided to examine the system performances.
In Chapter 4, we show that the second coding scheme proposed in
Chapter 3 can be constructed from a diferent approach. With this,
we propose the third coding scheme. The encoding is implemented
by using a binary encoder, followed by a convolutional processor
and a signal mapper. By directly designing a proper convolutional
processor without resorting to the equivalent multiple pairs of
multilevel delay processors and signal mappers, it is possible to
obtain better error performances than those of the coding schemes
constructed from the second scheme.
Each of the proposed coding schemes can be suboptimally decoded
by using the trellis of the component binary code C with some
feedback informations.
|
|---|
