On the Design of Near-Optimal Variable-Length Error-Correcting Codes for Large Source Alphabets

Yen Ming Chen, Feng Tsang Wu, Chih Peng Li, Pramod K. Varshney

Research output: Contribution to journalArticlepeer-review

Abstract

In this paper, we focus on near-optimal designs for variable-length error-correcting (VLEC) codes, while considering the performance in terms of both the average codeword length (ACL) and the symbol-error rate (SER). The criteria for narrowing the search space and enhancing the SER performance are investigated. An efficient code construction algorithm is then devised based on a random search process. Taking advantage of the significantly reduced search complexity, we are able to construct near-optimal VLEC codes for large source alphabets. Numerical results obtained for various free distance values and alphabet sizes show that constructed VLEC codes have both reduced ACL values and enhanced SER performances. Performance improvement is more notable for large source alphabets.

Original languageEnglish (US)
Article number9197658
Pages (from-to)7896-7910
Number of pages15
JournalIEEE Transactions on Communications
Volume68
Issue number12
DOIs
StatePublished - Dec 2020

Keywords

  • Joint source-channel coding (JSSC)
  • average codeword length (ACL)
  • digital communications
  • variable-length error-correcting (VLEC) codes

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this