Two-level-search minimum distance decoding for binary block codes

A new reduced-search minimum distance decoding algorithm for binary block codes is proposed. The time taken in searching for the nearest code word is reduced when all code words are grouped into classes and then each class into subclasses. The first classification is performed using the code weight distribution whilst the second (or subclass) level uses the weight distribution in some N-coordinates. The order of search is determined from statistical data based on the weight of the received vector and channel statistics.