Abstract:
In the last decade, a lot of work has been done on developing new ways
of enhancing spectral efficiency in multiple-input multiple-output (MIMO)
systems with lower implementation complexity. Spatial modulation (SM) is
one of the techniques developed. Many variants of spatial modulation (SM)
aiming to enhance data rate and bit error rate (BER) performance have been
proposed, however, their performance has not been compared for large-scale
MIMO systems. The first part of this thesis looks at various such schemes,
namely conventional SM, generalized spatial modulation (GSM), quadra
ture spatial modulation (QSM) and enhanced spatial modulation (ESM) to
study their performance for large-scale MIMO systems under generalized
fading conditions. Our results indicate that for the same spectral efficiency
and transmit power, QSM and ESM perform better than GSM and con
ventional SM schemes. For lower order modulation, QSM outperforms all
other schemes under various fading conditions, whereas ESM takes over for
higher order modulation. However, QSM uses one RF chain while ESM uses
two, making QSM a preferred scheme for simpler design and higher energy
efficiency.
In the second part of the thesis, a channel estimation scheme for SM
systems is developed that exploits the correlation between transmit antennas
to estimate channels of inactive antennas using the pilot-based estimate of
viii
the active antenna. It is observed that in a high signal-to-noise ratio (SNR)
regime, the proposed scheme provides about 2 dB and 5 dB gains compared to
the conventional channel estimation (CCE) method for moderately correlated
and highly correlated antennas, respectively. Analytical results correlate with
simulation results, which validates our analysis.
The third part of the thesis looks at the integration of SM with Full
Duplex (FD) radios, and a novel FD-QSM scheme is developed that exploits
multiple antennas to achieve antenna cancellation at the receiving side, to
mitigate the self-interference (SI) signal. It is observed that FD-QSM is
capable of providing about 40% capacity gain over half-duplex (HD)-QSM
and HD-MIMO in the presence of strong and moderate residual SI (RSI),
respectively, in a point-to-point communication scenario. When applied to
the downlink of a cellular network, FD-QSM provides 2dB gain over FD-SM
and 5dB gain over FD-MIMO, operating at the same spectral efficiency, while
huge gains are observed when FD-QSM is used in a non-orthogonal multiple
access (NOMA)-aided FD relay network.