* This is the dataset of the accepted paper (Oct 7, 2017):  R. Rajashekar, K.V.S. Hari, and L. Hanzo, "Transmit Antenna Subset Selection in Spatial Modulation Relying on a Realistic Error-Infested Feedback Channel"

* Paper Abstract:
In this paper, we study the performance of spatial modulation (SM) employing Euclidean distance based antenna selection (EDAS) operating in a realistic error-infested feedback channel, which has hitherto only been studied under ideal feedback channel conditions. Specifically, we model the feedback channel by a bit-flip probability $\delta$ and study its impact on the forward link employing EDAS. We show that the erroneous feedback channel severely degrades the performance of EDAS-aided SM (EDAS-SM) system by imposing an error floor in the forward link. Furthermore, we quantify the error floors associated both with the spatial and with the conventional symbols with the aid of asymptotic symbol error rate analysis. The expressions derived for the error floors in the forward link are utilised for optimizing the feedback signalling, which are shown to help reduce the error floor levels. Furthermore, a pilot-aided selection verification (PSV) algorithm is proposed for mitigating the effects of antenna-set mismatch between the transmitter and the receiver, which eliminates the error floor in the forward link. Simulations are conducted in order to validate the theoretical results presented in the paper. Furthermore, the bit-error ratio (BER) performance of the EDAS-SM is compared to that of the conventional antenna selection (C-AS) both in the PSV as well as in the no selection verification scenarios. It is observed that EDAS-SM outperforms C-AS in both the scenarios considered. Specifically, at a BER of $10^{-5}$, EDAS-SM is observed to give a 3dB signal-to-noise ratio gain compared to the C-AS, when operating at a spectral efficiency of 7 bits per channel use in the face of a feedback BER of $\delta=0.05$.

* Project:
The financial support of the EPSRC projects EP/Noo4558/1 and EP/L018659/1, as well as of the European Research Council’s Advanced Fellow Grant under the Beam-Me-Up project and of the Royal Society’s Wolfson Research Merit Award is gratefully acknowledged.

* This DOI contains the datasets of Figures 2 to 11 of the aforementioned paper. Each Fig#.fig file corresponds to the same numbered figure in the paper. Each .fig file has all the information required to generate the plot. To regenerate the results, just open the Fig#.fig file in Matlab. Exact values of each of the curves can be read from the property editor.