On the Maximum Likelihood Performance of Linear Product Codes(HISC2006)

概要

論文の詳細を見る
In this paper we analyze the maximum likelihood performance of linear product codes. Since the weight enumerators of most linear product codes remain unknown, their performance analysis relied on the truncated union bounds using knowledge about the number of codewords with very low weight. Such bounds are very loose at low and medium signal to noise ratios. Through concatenated representations of product codes, we derive averaged weight enumerators of linear product codes over any finite field. The split weight enumerators of popular component codes such as Hamming codes and extended Hamming codes are studied and used in the analysis. Numerical results show that our technique results in bounds on the maximum likelihood performance of product codes which are tight at both high and low signal-to-noise ratios.
社団法人電子情報通信学会の論文
2006-05-18