WSJT-X/lib/ldpc/dep-H.html
Steven Franke 5ac886855d Add ldpc sandbox folder.
git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@6437 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
2016-01-25 00:04:21 +00:00

86 lines
3.0 KiB
HTML
Executable File

<HTML><HEAD>
<TITLE> Linear Dependence in Parity Check Matrices </TITLE>
</HEAD><BODY>
<H1> Linear Dependence in Parity Check Matrices </H1>
<P>If a code is specified by means of a <I>M</I> by <I>N</I> parity
check matrix, <B>H</B>, in which some rows are linearly dependent - a
situation that is usually avoided - it would be possible to map more
than the usual <I>K=N-M</I> message bits into a codeword, since one or
more rows of <B>H</B> could have been deleted without affecting which
bit vectors are codewords.
<P>However, this software does not increase the number of message bits
in this case, but instead produces a generator matrix in which some
rows are all zero, which will cause some bits of the codeword to
always be zero, regardless of the source message. Referring to the <A
HREF="encoding.html#gen-rep">description of generator matrix
representations</A>, this is accomplished by partially computing
what would normally be <B>A</B><SUP><SMALL>-1</SMALL></SUP> (for a
dense or mixed representations) or the <B>L</B> and <B>U</B> matrices
(for a sparse representation), even though singularity prevents this
computation from being carried out fully.
<P><B>Example:</B> The parity check matrix created below is redundant,
since the 10100 row is equal to the sum of the 11000 and 01100 rows.
<UL><PRE>
<LI>make-pchk dep.pchk 4 5 0:0 0:1 1:1 1:2 2:0 2:2 3:3 3:4
<LI>print-pchk -d dep.pchk
Parity check matrix in dep.pchk (dense format):
1 1 0 0 0
0 1 1 0 0
1 0 1 0 0
0 0 0 1 1
<LI>make-gen dep.pchk dep.gen dense
Note: Parity check matrix has 1 redundant checks
Number of 1s per check in Inv(A) X B is 0.2
<LI>print-gen dep.gen
Generator matrix (dense representation):
Column order (message bits at end):
0 1 3 2 4
Inv(A) X B:
0
0
1
0
</PRE></UL>
The generator matrix above can be used to encode message blocks containing
one bit. This message bit is copied unchanged to the last bit (numbered 4)
of the codeword, and the first four bits of the codeword are set by multiplying
this message bit (seen as a vector of length one) by
<B>A</B><SUP><SMALL>-1</SMALL></SUP><B>B</B>, shown above, and then
storing the results in positions given by the column ordering.
The result is that bit 3 of the codeword produced is
also set to the message bit, and bits 0, 1, and 2 are set to zero.
<P>Which bits are used for message bits, and which bits are fixed at
zero, depends on arbitrary choices in the algorithm, which may differ
from one encoding method to another. No attempt is made to make the
best choice.
<P>Note that codeword bits that are always zero can arise even when <B>H</B>
does not have linearly dependent rows. For example, if a row of <B>H</B>
has just one 1 in it, the codeword bit at that position must be zero in any
codeword. The way the software handles parity check matrices with less
than <I>M</I> independent rows is equivalent to adding additional rows
to <B>H</B> in which only one bit is 1, in order to produce <I>M</I>
independent checks.
<HR>
<A HREF="index.html">Back to index for LDPC software</A>
</BODY></HTML>