The Modified Dorsch Decoder

The Holy Grail in coding theory is the maximum likelihood decoder. Paradoxically, the best error-correcting codes are not commonly used in practical communication systems. This is because there is no practical decoder available that achieves anywhere near maximum likelihood performance. Inferior codes, such as LDPC or turbo codes, are used that perform much better with a practical decoder such as the iterative decoder. One day, may be in 100 years time, maximum likelihood decoders will be generally available and today’s best codes will then be used in practical systems. In this chapter a modified Dorsch decoder is described, which achieves near maximum likelihood performance for all binary codes that are not too long. It is a practical decoder for half rate codes having a length less than about 180 bits or so using current digital processors. The outstanding performance achieved, using several examples of the best binary codes known, is presented.