Advanced information of parity bits for the A * DecodingAlgorithm

• Main Authors: Chia-Fu Chang, 張家輔
• Other Authors: 林茂昭
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/82746070042650356953

博士 === 國立臺灣大學 === 電信工程學研究所 === 102 === The A* algorithm for decoding linear block codes is known to be an efficient tree-search method for decoding short linear block codes. The A* decoding algorithm and its modified methods can be applied to either maximum-likelihood (ML) soft-decision decoding or stack limited cases of linear block codes. In this thesis, several modified methods of A* decoding algorithm are proposed. The basic concept of this thesis is to employ the early acquired information of some parity bits or some combined parity bits to reduce the needed search tree edges. Analysis for the number of required information bits for obtaining the parity bits is provided. In addition, three versions of the modified A* algorithm which take advantage of the advanced information of parity bits and combined parity bits are proposed. The computer simulation results of several quadratic residue codes and BCH codes are provided. Simulation results show that there is significant advantage in reducing the number of search tree edges in each proposed version as compared with two existing A* algorithms which do not consider the early information of parity bits or combined parity bits.