Analytic Mathematical Models of Tactical Military Communications Channels
Quarterly progress rept. no. 6, 1 Oct-31 Dec 1972
ILLINOIS UNIV AT URBANA COORDINATED SCIENCE LAB
Pagination or Media Count:
The Viterbi decoding algorithm yields minimum probability of error when applied to a memoryless channel provided that all input sequences are equally likely. In this report, the algorithm was generalized for application to channels with finite memory and it was shown that the generalized algorithm is also maximum likelihood decoding. It was also shown that the generalized Viterbi algortithm on a simple memory channel performs better than the original Viterbi algorithm with the same decoding complexity. The M-state Markov model was reviewed in this report. The process of identifying the parameters of the M- state model from the coefficients A sub i and A sub i n sub j, n sub j1 of the gap model was determined to be more complicated than was anticipated. An alternative, the simple partitioned Markov model was examined to determine the effect of the second order statistics, namely the interdependence of the gaps, on the error burst distribution. An alternative definition of the burst was adopted to speed up this investigation. The difference or similarity between these two definitions will be determined.
- Non-Radio Communications