4.7 Article

A Novel Low-Complexity GFDM Relay Communication System: Relay Selection and Filter-and-Forward

Journal

IEEE TRANSACTIONS ON SIGNAL PROCESSING
Volume 69, Issue -, Pages 5147-5158

Publisher

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TSP.2021.3108679

Keywords

Finite impulse response filters; Eigenvalues and eigenfunctions; Complexity theory; Prototypes; Relay networks (telecommunication); OFDM; Covariance matrices; Circulant matrix; eigenvalue decomposition; filter-and-forward; GFDM; relay network; relay selection

Funding

  1. University of Torbat Heydarieh, Iran [UTH: 1399/04/2177]

Ask authors/readers for more resources

This paper presents a low-complexity cooperative relay network based on GFDM system, proposing three methods for designing relay nodes. Simulation results demonstrate the performance of the methods in terms of BER and show that the proposed methods can be implemented with low complexity due to the use of circular convolution.
In this paper, we develop a low-complexity cooperative relay network based on generalized frequency division multiplexing (GFDM) system. In this network, multiple relay nodes (RNs) are used to relay GFDM symbols from a source node (SN) toward a destination node (DN). We propose three methods to design the RNs based on the minimization of the total power of the noise after symbol estimation at the DN. In the first method, only one node is selected among all RNs to forward the signal from the SN to the DN. In the second and third methods, a filter-and-forward (FF) scheme is used, such that at each RN, the received GFDM signal from the SN is circularly convolved with a finite impulse response (FIR) filter and then, forwarded to the DN. The FIR filter's coefficients are represented as a circulant matrix. Since a circulant matrix is characterized by its eigenvalues, we obtain the eigenvalues of the FIR coefficient matrices for all RNs. The difference between the second and third methods is that in the former, the eigenvalues of the filter matrices are restricted to be one or zero, however, in the later, the eigenvalues are determined with no restrictions. Simulation results are provided to demonstrate the performance of the proposed methods in terms of bit error rate (BER). Moreover, we show that because of using circular convolution, the proposed methods can be implemented with a low complexity.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available