Iterative detection of a CFSK system

The MAP consensus decoder using iterative detection for a collision frequency-shift-keyed (CFSK) system is investigated in this paper. This consensus decoder decodes each user's message in a joint and iterative manner. For a specific user, interference from other users is canceled by the iteration process. The metric of the CFSK channel will introduce high complexity to the MAP algorithm when the system is highly loaded. The K-most probable combinations (KMPC) approach and two-most probable combinations (TMPC) approach are applied to simplify the metric. Analysis shows that the TMPC method keeps fairness for all the users, even though some users put high weights on the probabilities of part of the frequencies.