Noncoherent Detection with Polar Codes

Chaofan Chen, Li Li*, Li Wang, Shuai Wang, Xiangming Li, George K. Karagiannidis

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)

Abstract

Polar codes (PCs) have attracted significant attention in the last decade, especially after their adoption in the forthcoming 5G wireless networks. However, previous studies focused on coherent polar codes, which always rely on the strong assumption of available perfect channel state information. Instead, in this paper, we investigate the use of PCs in noncoherent systems. First, a binary differential phase shift keying (BDPSK) demodulator is concatenated with a polar decoder to form the noncoherent detector, where successive cancellation (SC) is applied. The simulation results demonstrate that the SC-based PCs for noncoherent detection have approximately a discrepancy of only 3 dB compared with the coherent counterpart in noncoherent channels. Furthermore, in order to further decrease this discrepancy, we replace the BDPSK demodulator with a soft-input soft-output (SISO) multiple symbol differential sphere decoding demodulator. Similarly, the SC-based PC decoder is replaced by the SISO belief propagation-based PC decoder, and by using this novel architecture, an iterative noncoherent detector is constructed. Benefiting from further invoking extrinsic information transfer chart tool and the dynamic window-size detection scheme, the performance of the proposed iterative noncoherent detector becomes competitive with its coherent one in practical applications, since the performance degradation is reduced to 1 dB.

Original languageEnglish
Article number8589036
Pages (from-to)6362-6372
Number of pages11
JournalIEEE Access
Volume7
DOIs
Publication statusPublished - 2019

Keywords

  • EXIT chart
  • Polar codes
  • SISO-MSDSD
  • iterative detection
  • noncoherent detection

Fingerprint

Dive into the research topics of 'Noncoherent Detection with Polar Codes'. Together they form a unique fingerprint.

Cite this