Authors: Ilić, Velimir 
Dupraz, Elsa
Vasić, Bane
Title: Generic architectures for uniformly reweighted APP decoders
Journal: 2017 13th International Conference on Advanced Technologies, Systems and Services in Telecommunications, TELSIKS 2017 - Proceeding
Volume: 2017-October
First page: 308
Last page: 316
Conference: 13th International Conference on Advanced Technologies, Systems and Services in Telecommunications, TELSIKS 2017; Nis; Serbia; 18 October 2017 through 20 October 2017
Issue Date: 28-Jun-2017
ISBN: 978-1-538-61800-4
DOI: 10.1109/TELSKS.2017.8246286
We propose generic architectures for memory-efficient Uniformly reweighted a posteriori probability (URAPP) Decoders of Low Density Parity Check (LDPC) codes, with the bit error rate close to Belief propagation decoding. The architectures enables a trade off between the throughput, computational and hardware resources. As a special case, we derive previously proposed flooding schedule URAPP decoder and propose new type of the URAPP decoder which uses shuffled scheduling. We conduct a rigorous complexity analysis and provide new formulas for the URAPP decoder throughput in terms of hardware resources and memory requirements. The analysis confirms that the shuffled decoder represents an optimal solution if the hardware resources are limited, which makes it attractive in practical applications. The results are verified by simulations for the case of finite geometry LDPC codes.
Keywords: APP decoding | architecture | complexity | LDPC | memory | throughput
Publisher: IEEE
Project: Development of new information and communication technologies, based on advanced mathematical methods, with applications in medicine, telecommunications, power systems, protection of national heritage and education 
Representations of logical structures and formal languages and their application in computing 
NSF, Grant ECCS-1500170
Networked Center for Data Storage Research, Grant JC-16-2014-US

Show full item record

Page view(s)

checked on May 9, 2024

Google ScholarTM




Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.