Skip to main navigation Skip to search Skip to main content

Artificial-noise-aided nonlinear secure transmission for multiuser multi-antenna systems with finite-rate feedback

  • Liang Sun*
  • , Rui Wang
  • , Zhiyao Tang
  • , Victor C.M. Leung
  • *Corresponding author for this work
  • Southern University of Science and Technology
  • Beihang University
  • University of British Columbia

Research output: Contribution to journalArticlepeer-review

Abstract

We consider a low-complexity transceiver design for secure communications of a multiuser multi-antenna system with an external multi-antenna eavesdropper. We propose to employ Tomlinson-Harashima precoder (THP) to simultaneously transmit message-bearing signals and artificial noise (AN) according to the quantized channel state information (CSI). Based on random vector quantization of the channel vectors, we obtain analytical approximations of the ergodic secrecy rate (ESR) of each legitimate receiver (LR) and the ergodic secrecy sum rate of the system with arbitrary system parameters. Based on the obtained analytical results, the near-optimal power allocation to the information signals and AN can be obtained using a numerical method. We show using an information-theoretical method that, besides the advantage in the ESR over the linear precoding scheme, THP can reduce the supported rate of the eavesdropper's channel by preventing the eavesdropper from obtaining the legitimate channels' (quantized) CSI. The ESR loss of each LR compared with the perfect CSI case will increase without bound for a fixed number of feedback bits. We also derive a feedback bit scaling law to solve this problem. Finally, numerical results are provided to verify our analytical results and also the advantage of the proposed secure nonlinear transceiver over the corresponding linear scheme.

Original languageEnglish
Article number8543247
Pages (from-to)2274-2293
Number of pages20
JournalIEEE Transactions on Communications
Volume67
Issue number3
DOIs
StatePublished - Mar 2019

Keywords

  • Physical layer security
  • artificial noise
  • limited feedback
  • nonlinear precoding
  • performance analysis
  • secrecy (sum) rate

Fingerprint

Dive into the research topics of 'Artificial-noise-aided nonlinear secure transmission for multiuser multi-antenna systems with finite-rate feedback'. Together they form a unique fingerprint.

Cite this