New Code Construction Method and High-Speed VLSI Codec Architecture for Repeat-Accumulate Codes
Kaibin Zhang,Liuguo Yin,Jianhua Lu
2009-01-01
Abstract:Repeat-Accumulate code has been widely used in various practical communication systems due to its special structure that can be encoded with a low complexity turbo-like encoder and decoded with a high-speed parallel algorithm. However, the construction of RA codes with large girth is still a challenging problem. In this paper, the girth upper bound of RA codes is analyzed and a novel RA code construction scheme is proposed. This scheme generates the parity-check matrix with superimposed structured interleavers, guaranteeing the resulting matrix to be globally optimized. Compared with the RA codes constructed with traditional structured interleavers such as π-rotation interleavers, pseudo-random interleavers, combinatorial interleavers, and modified L-type interleavers, the codes constructed with the proposed method exhibit better error-correcting performance. Moreover, a parallel VLSI codec architecture is proposed for the RA codes generated with the proposed method in this paper. Modifying the conventional RA encoder architecture, the codec for RA codes from our proposed method may have the same hardware implementation complexity as that for traditional RA codes, while the data throughput of our codec may achieve up to 450Mbps, which may be of great value for practical applications.