READ ME File For 'Differential Evolution Algorithm Aided Turbo Channel Estimation and Multi-User Detection for G.Fast Systems in the Presence of FEXT'

Dataset DOI: https://doi.org/10.5258/SOTON/D0550

ReadMe Author: Jiankang Zhang, Southampton Wireless Group, Electronics and Computer Science, University of Southampton

This dataset supports the publication:
Zhang, Jiankang; Chen, Sheng; Zhang, Rong; Anas F. Al Rawi; Hanzo, Lajos.
Differential Evolution Algorithm Aided Turbo Channel Estimation and Multi-User Detection for G.Fast Systems in the Presence of FEXT. 
IEEE Access

This dataset contains the data that are used for generating Fig.1 and Fig5 to Fig.14. These figures are plotted using GLE (Graphics Layout Engine). The scripts of Gle are also included in the folds for each figures. In order to generate these figures, you should install Gle
http://glx.sourceforge.net/

The figures are as follows:

Fig. 1: (a) Measured direct channel and crosstalk channel strengths, and (b) the average noise power amplified by the ZF FEXT canceller.
Fig. 5: NMSE versus the subcarrier index: (a) Eb/N0 = 20 dB, and (b) Eb/N0 = 30 dB.
Fig. 6: Idealized SER based on perfect CSI versus the subcarrier index: (a) Eb/N0 = 20 dB, and (b) Eb/N0 = 30 dB.
Fig. 7: NMSE versus the number of iterations: (a) Eb/N0 = 20 dB, and (b) Eb/N0 = 30 dB.
Fig. 8: Idealized BER based on perfect CSI versus the number of iterations: (a) Eb/N0 = 20 dB, and (b) Eb/N0 = 30 dB.
Fig. 9: (a) Training-based NMSE performance versus Eb/N0, and (b) idealized BER performance versus Eb/N0 relying on perfect CSI, for the system bandwidths of 52.5MHz, 105.0MHz, 157.5MHz and 210.0MHz.
Fig. 10: Achievable performance of the DEA aided turbo CE and MUD: (a) NMSE versus Eb/N0 parametrized by the number of turbo iterations, and (b) BER versus Eb/N0 parametrized by the number of versus turbo iterations.
Fig. 11: Impact of loop length on the achievable detection performance of four idealized MUDs associated with perfect CSI.
Fig. 12: Impact of impulse noise on the achievable detection performance of four idealized MUDs associated with perfect CSI.
Fig. 13: Comparison of the achievable detection performance of four MUDs based on LS-CE acquired by training as well as based on perfect CSI.
Fig. 14: Computational complexity of the DEA-MUD expressed as the ratio of its required CF evaluations over the total CF evaluations imposed by the ML-MUD. The perfect CSI is assumed


Date of data collection: from November 1th, 2017 to , December 1st, 2017

Information about geographic location of data collection: University of Southampton, U.K.

Related projects:
the European Research Council¡¯s Advanced Fellow Grant
the Royal Society Wolfson Research Merit Award
the EPSRC project EP/N004558/1


Date that the file was created: June 2018