An improved A~* decoding is presented for binary block codes that performs maximum-likelihood soft-decision decoding. The decoder uses A~* algorithm to search through a code tree for an optimal path that minimizes a pth metric function. When a candidate codeword is found, it is checked with general threshold. If checked successfully, the candidate codeword is the msot likely(ML) codeword and the search stops. Otherwise the process of search continues until an optimal path is found. Simulation results show that the porposed decoding scheme achieve optimal performance meanwhile significantly reducing decoding coplexity compared with A~* decoding avaliable.
展开▼