Performance Analysis for Generalised Spatial Modulation

In this paper, the performance of generalised spatial modulation (GSM) using large scale multiple-input multiple-output (MIMO) systems is analysed from an information-theoretic point of view. The analysis is extended to account for the average bit error ratio (ABER). We introduce a simple model for the capacity which allows us to derive a closed-form expression for the capacity of GSM. This is achieved by modelling GSM as two independent sources of information, where we show that GSM has the potential to achieve the capacity of MIMO systems. Moreover, using Monte-Carlo simulation for link-level mutual information, it is shown that the theoretical limit can be reached with practical signal constellations in operationally important signal-to-noise-ratio (SNR) regions. Furthermore, the ABER performance of GSM is analysed in this paper, and a tight upper bound for correlated and uncorrelated, Rayleigh and Rician channels is derived. Finally, the performance of GSM is validated through Monte Carlo simulations and compared with the performance of spatial modulation (SM). It is shown that GSM outperforms SM by up to 3.5 dB while the same low complexity of SM is retained. Specifically, the complexity is equal to that of single-input multiple-output (SIMO) systems. Moreover, a further advantage of SM is maintained which is that only a single radio frequency (RF) chain is required. Thus, GSM is a perfect candidate for large scale MIMO systems.