02-05-2013 дата публикации
Номер: US20130111305A1
A channel decoding method and decoder are disclosed. The decoding method is based on a Circular Viterbi Algorithm (CVA), rules out impossible initial states one by one through iterations according the received soft information sequence, and finally finds the global optimal tail-biting path. In the present invention, all impossible iterations are ruled out through multiple iterations, and only the initial state having most likelihood with the received sequence survives. The algorithm is finally convergent to an optimal tail-biting path to be output. In addition, the method also updates a metric of a maximum likelihood tail-biting path (MLTBP) or rules out impossible initial states through the obtained surviving tail-biting path, thereby effectively solving the problem that the algorithm is not convergent due to a circular trap, providing a practical optimal decoding algorithm for a tail-biting convolutional code, reducing the complexity of an existing decoding scheme, and saving the storage space. 1. A channel decoding method , comprising:{'sub': path,0', '0', '0', 'MLTBP', 'MLTBP', '0', 'path,L', 'state', 'path,L, 'sup': i', 'O', 'O', '1', '1, 'S101: during the first iteration, i.e. i=1, initializing the metric M(S) of the surviving path that enters state s at position 0 to 0, where s∈S, Sindicates the state space at position 0, and i indicates the number of iterations; setting the metric of the optimal maximum likelihood tail-biting path (MLTBP) Pto 0, that is, letting M=0; performing an MVA, and searching for MLTBP in current iteration; for all s∈S, setting the state metric of state s to be equal to M(s), that is, letting M(s)=M(s);'}{'sub': MLTBP', 'MLTBP', 'MLTBP', 'MLTBP', 'MLTBP', 'MLTBP', 'MLTBP', 'MLTBP', 'MLTBP, 'sup': i', 'i', 'O', 'i', 'O', 'O', 'O', '1', 'i, 'S102: if the net path metric Mof P(s′,s′) found in current iteration is greater than the metric M, that is, M>M, updating (P, M) with (P(s′,s′),M);'}{'sub': L', 'L', 'state', 'MLTBP', 'state', ' ...
Подробнее
