Improved Tomlinson-Harashima precoding with interference optimization

C. Masouros*, M. Sellathurai, T. Ratnarajah

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

This paper investigates a novel strategy for reducing the power loss in Tomlinson-Harashima precoding (THP) based on optimising the interference to be canceled. The proposed scheme is motivated by the fact that, since both the desired and interfering signals originate from the base station (BS) of the downlink system, the resulting interference can be influenced to reduce the transmission power required to cancel it, without altering the information content of the downlink message. This optimization aims at bringing the interference closer to the replica of the desired symbol in the modulo-extended constellation. By doing so, the quantized distance between the useful signal and interference is reduced and therefore the power required to pre-subtract interference is decreased. Here, the above concept is applied to a practical THP design for multiple input multiple output systems (MIMO-THP). The results show a considerable transmit power reduction compared to conventional THP.

Original languageEnglish
Title of host publication2011 17th International Conference on Digital Signal Processing (DSP)
PublisherIEEE
ISBN (Electronic)9781457702747
ISBN (Print)9781457702730
DOIs
Publication statusPublished - 30 Aug 2011
Event17th International Conference on Digital Signal Processing 2011 - Corfu, Greece
Duration: 6 Jul 20118 Jul 2011

Conference

Conference17th International Conference on Digital Signal Processing 2011
Abbreviated titleDSP 2011
Country/TerritoryGreece
CityCorfu
Period6/07/118/07/11

Keywords

  • known interference
  • multi-user MIMO
  • non-linear precoding
  • Tomlinson-Harashima precoding

ASJC Scopus subject areas

  • Signal Processing

Fingerprint

Dive into the research topics of 'Improved Tomlinson-Harashima precoding with interference optimization'. Together they form a unique fingerprint.

Cite this