In this letter, we apply the concept of media-based modulation with radio frequency (RF) mirrors to uncoded spacetime labeling diversity (USTLD). Several configurations are investigated. Results demonstrate a significant improvement in the error performance over USTLD and the recently proposed space-time channel modulation for fast frequency-flat Rayleigh fading channels with and without channel estimation error. Improvement is also shown for the quasi-static channel case. The analytical average bit error probability is formulated for the scheme and is shown to agree well with Monte Carlo simulation results at moderate-to-high signal-to-noise ratios. Furthermore, the detection computational complexity of each configuration is analysed in terms of the number of real operations.

The use of radio frequency (RF) mirrors to improve spectral efficiency is a recently introduced technique. The authors investigate its application to single RF chain and inter‐channel interference‐free communications. The technique is first applied to improve the error performance of a conventional single‐input multiple‐output (SIMO) scheme, in the form of SIMO media‐based modulation (SIMO‐MBM) with single‐mirror activation pattern (MAP). Significant improvement in error performance is demonstrated. The theoretical average bit error probability (ABEP) of SIMO‐MBM with single MAP is formulated and serves to validate the Monte‐Carlo simulation results. The application of RF mirrors to further improve spectral efficiency in quadrature spatial modulation is then proposed, in the form of quadrature spatial media‐based modulation (QSMBM) and is able to achieve very large spectral efficiencies, while maintaining relatively good error performance. The theoretical ABEP of QSMBM is formulated and shown to agree well with Monte‐Carlo simulation results. Finally, MAP selection for QSMBM is investigated. Due to the high computational complexity and memory requirements of optimal MAP selection, low‐complexity suboptimal algorithms for MAP selection are investigated. A significant improvement in the error performance of QSMBM is demonstrated with these algorithms.

Link adaptation in the form of Euclidean distance‐based antenna selection (EDAS) at the transmitter, has demonstrated a significant improvement in the error performance of spatial modulation (SM). However, when a large transmit antenna array is required, the computational complexity imposed by the EDAS scheme is inherently high. Thus, there is great motivation for the development of low‐complexity transmit antenna selection (LCTAS) schemes for SM. In this study, the authors propose several LCTAS schemes for SM. Assuming full knowledge of the channel at the receiver, the proposed schemes select a subset of the candidate transmit antenna elements based on channel amplitude, antenna correlation, Euclidean distance between transmit vectors or a combination of these; demonstrating a significant improvement in error performance compared with conventional SM, while imposing significantly lower computational complexity than EDAS‐SM. Moreover, the proposed schemes employ a novel complexity reducing technique, which allows parallel processing of algorithms and can easily be applied to existing schemes.

Spatial modulation (SM) is an efficient transmission scheme based on multiple‐input multiple‐output systems. SM, by nature, requires coherent detection to achieve optimal error performance. This requires that the receiver has full knowledge of the channel state information (CSI), however, determining the CSI results in an increase in the detection complexity at the receiver. Differential modulation systems are capable of performing detection without knowledge of the CSI at the receiver, however, they suffer a 3 dB error performance penalty as compared to their coherent counterparts. In this study, the authors present a generalised differential scheme for SM (GD‐SM) based on generalised differential modulation, which is capable of reducing the performance penalty incurred through optimal power allocation. Furthermore, they extend the architecture of GD‐SM and optimal power allocation to conventional differential SM. The architecture of the proposed systems, advantageously permits the use of either M ‐ary quadrature amplitude modulation or M ‐ary phase shift keying constellations. Simulation and theoretical results demonstrate that GD‐SM incurs only about 0.5 dB performance loss as compared to coherent SM when the frame length is 400.