Unformatted text preview: on Using Fill Bit. When using the SEC decoding
mode (single bit error correction/double bit error detection), a single bit error in reception of a
normal codeword will be corrected, making the codeblock acceptable. The following
discussion only applies when using the SEC decoding mode where such corrections are
possible. The tail sequence is not a codeword but is instead a specially constructed sequence
designed to appear to the decoder as an "uncorrectable pattern". This means that when the tail
sequence is received without change, the decoder will declare that the received sequence is an
uncorrectable pattern, which forces it to leave the DECODE state and enter the SEARCH state.
However, if a single bit of this tail sequence becomes altered in the channel, the received
sequence may no longer be recognized as "uncorrectable", and an improper correction may take
place. Because the expected CODEBLOCK REJECTION did not occur, we say the tail
sequence was "missed" (not recognized). The probability of this happening is given by
Missing a tail sequence has far-reaching effects, since it is possible that the decoder will not be
in SEARCH state for the next CLTU and therefore the entire subsequent CLTU may be
This problem may be significantly reduced and the reliability of identifying a tail sequence
improved by using the fill bit to selectively inhibit the SEC mode. As described in Reference
, each valid codeblock terminates on a fill bit which is always set to "0". A tail sequence
(which consists of an alternating pattern of ones and zeros starting with a zero) will always
contain a "1" in the position normally occupied by the fill bit in a regular codeblock. This
information can be exploited at the receiving end to inhibit the SEC mode whenever the fill bit
is a "one". This process substantially improves the ability to identify a tail sequence with a
negligible effect on the codeword acceptance process.
Issue 6 Page D-20 January 1987 CCSDS REPORT CONCERNING TELECOMMAND: SUMMARY OF CONCEPT AND SERVICE The recommended algorithm at the receiving end is as follows:
(1) Test the received pattern alone (ignoring the fill bit) for errors. (2) If NO errors are detected in the received pattern, ACCEPT the codeword and
IGNORE the fill bit. (3) If ONE error is detected in the received pattern, TEST the fill bit.
(b) (4) If the fill bit = 0, correct the error in the pattern and ACCEPT it.
If the fill bit = 1, declare a CODEBLOCK REJECTION. If TWO errors are detected in the received pattern, declare a CODEBLOCK
REJECTION. IGNORE the fill bit. By selectively testing the fill bit, the probability of missing a tail sequence can be found from
the probability of a single bit being changed in any portion of the tail sequence except the fill
bit, AND that the received fill bit has also been changed. That is, TWO changes have to occur
in the same codeblock, one of which must be the fill bit p...
View Full Document
This document was uploaded on 03/06/2014.
- Spring '14