Anytime properties of protograph-based repeat-accumulate codes

Abstract

In this paper, a highly efficient protograph-based repeat-accumulate (P-RA) anytime code is proposed, which is derived from the combination of protograph-based low-density parity-check convolutional (P-LDPCC) and spatially coupled low-density parity-check convolutional (SC-LDPCC) codes. Density evolution technique is applied to demonstrate the anytime properties of the P-RA codes over the binary erasure channel (BEC). The simulation results show that P-RA codes perform better than SC-LDPCC and P-LDPCC codes. More importantly, the finite-length performance of P-RA codes is only dependent on the decoding delay rather than the decoding message position.