Title: Low-complexity LDPC-convolutional codes based on cyclically shifted identity matrices

Authors: Fotios Gioulekas; Constantinos Petrou; Athanasios Vgenis; Michael Birbas

Addresses: University General Hospital of Larissa, Mezourlo 41110, Greece ' DE P&I Sens Leveling, ASML Netherlands BV, De Run 6501, 5504 DR Veldhoven, The Netherlands ' 26225, Patras, Greece ' Department of Electrical Engineering and Computer Technology, University of Patras, 26500, Greece

Abstract: In this study, a construction methodology for low-density parity-check convolutional codes (LDPC-CCs) ensembles based on cyclically shifted identity matrices is proposed. The proposed method directly generates the syndrome former matrices according to the specified code parameters and constraints i.e., code-rate, degree-distribution, constraint length, period and memory, in contrast to the majority of the available approaches that produce relevant error-correcting codes based on either block ones, protographs or spatially-coupled type of codes. Simulation results show that the constructed ensembles demonstrate advanced error-correcting capability of up to 0.2 dB in terms of frame-error and bit-error rates at the convergence region, when compared with the performance of error-correcting schemes adopted by various communication standards, with equivalent hardware complexity even at short codeword-lengths. Specifically, the constructed LDPC-CCs have been assessed against the corresponding error-correcting codes used in WiMAX and G.hn standards for wireless and wireline telecommunications, respectively.

Keywords: FEC; low-density parity-check convolutional codes; LDPC-CCs; complexity; error-correction; WiMAX; G.hn; cyclically shifted identity matrices; LDPC-block codes; schedulable memory; syndrome-former.

DOI: 10.1504/IJICT.2019.097686

International Journal of Information and Communication Technology, 2019 Vol.14 No.2, pp.139 - 158

Received: 26 Aug 2016
Accepted: 27 Oct 2016

Published online: 17 Jan 2019 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article