Low-complexity finite precision decoders for low-density parity-check codes

Shiva Kumar Planjery, Bane Vasic, David Declercq, Michael W. Marcellin

Research output: Contribution to journalConference article

Abstract

We present a new class of finite-precision decoders for low-density parity-check (LDPC) codes. These decoders are much lower in complexity compared to conventional floating-point decoders such as the belief propagation (BP) decoder, but they have the potential to outperform BP. The messages utilized by the decoders assume values (or levels) from a finite discrete set. We discuss the implementation aspects as well as describe the underlying philosophy in designing these decoders. We also provide results to show that in some cases, only 3 bits are required in the proposed decoders to outperform floating-point BP.

Original languageEnglish (US)
JournalProceedings of the International Telemetering Conference
Volume46
StatePublished - Dec 1 2010
Event46th Annual International Telemetering Conference and Technical Exhibition - Overcoming Telemetry Obstacles with Technology, ITC/USA 2010 - San Diego, CA, United States
Duration: Oct 25 2010Oct 28 2010

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Instrumentation
  • Computer Networks and Communications
  • Signal Processing

Fingerprint Dive into the research topics of 'Low-complexity finite precision decoders for low-density parity-check codes'. Together they form a unique fingerprint.

  • Cite this