Pilot- and CP-Aided Channel Estimation in MIMO Non-Orthogonal Multi-Carriers

Shahab Ehsanfar, Maximilian Matthé, Marwa Chafii, Gerhard P. Fettweis

Research output: Contribution to journalArticlepeer-review

Abstract

Motivated by 5G application requirements that challenge the use of orthogonal frequency division multiplexing (OFDM), non-orthogonal multi-carriers are being investigated. Unlike OFDM that takes advantage of orthogonal pilot observation, in non-orthogonal waveforms, pilots are contaminated by interference from multiple dimensions, i.e., inter-subsymbol-, inter-carrier-, and inter-antenna-interference, when multiple-input-multiple-output (MIMO) is also part of the transmission. Employing cyclic-prefix (CP) in multi-carrier systems not only protects the signal from inter-symbol-interference but also allows circular interpretations of the channel, which simplifies the estimation and equalization techniques. Nevertheless, the CP information is usually discarded at the receiver side. In this paper, by considering the fact that non-orthogonal waveforms suffer from multiple dimensions of interference, we derive a MIMO linear-minimum-mean-squared-error (LMMSE)-based parallel-interference-cancellation (PIC) method for joint channel estimation and equalization of non-orthogonal waveforms. Unlike the common practice, by properly localizing the pilots in time domain, we also use the pilots' information from CP. We apply our proposed algorithm to a flexible non-orthogonal waveform known as generalized frequency division multiplexing (GFDM). Taking advantage of block-circularity of GFDM, we investigate the complexity aspects for such CP-aided LMMSE-PIC channel estimation. Through simulation results, we show that using CP information of pilots for GFDM gains up to 2.4-dB better frame error rate performance than an OFDM signal.

Original languageEnglish (US)
Article number8573153
Pages (from-to)650-664
Number of pages15
JournalIEEE Transactions on Wireless Communications
Volume18
Issue number1
DOIs
StatePublished - Jan 2019

Keywords

  • Channel estimation
  • GFDM
  • interference cancellation
  • iterative processing
  • MIMO equalization

ASJC Scopus subject areas

  • Computer Science Applications
  • Electrical and Electronic Engineering
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Pilot- and CP-Aided Channel Estimation in MIMO Non-Orthogonal Multi-Carriers'. Together they form a unique fingerprint.

Cite this