A Class of Hybrid Coupled Serially Concatenated Codes

Spatially coupled serially concatenated codes (SC-SCCs) outperform the spatially coupled parallel concatenated codes (SC-PCCs) both in the waterfall region and the error-floor region. Although the SC-SCCs with a weak outer convolutional code (CC) and a strong inner CC reveal good waterfall performance, their error-floors are still too high. This paper introduces the recursive encoding structure into SC-SCCs to reduce the error-floor. We introduce the hybrid coupled serially concatenated codes (HC-SCCs), in which spatial coupling is achieved by taking a portion of the encoding outputs of both the outer and the inner encoders as the input of the current inner encoder. As a result, the proposed HC-SCCs are recursive in nature. The exact density evolution equations of the HC-SCCs are derived. Numerical results are provided to show the performance advantages of the proposed HC-SCCs. Particularly, simulation results show that, compared to the original SC-SCCs, about two orders of magnitude error-floor performance improvement are obtained with the proposed HC-SCCs.