Wireless Communication Systems Based on Spatial Modulation MIMO
Abstract
Spatial modulation (SM) is a unique single-stream, multiple-input multiple-output (MIMO)
transmission technique. Unlike traditional MIMO schemes, SM sends out signals through a
single active antenna, and achieves multiplexing gains by encoding information bits into the
index of the currently active antenna. In contrast to multi-stream MIMO systems, this particular
characteristic offers great superiority in two main aspects. Firstly, SM completely avoids
inter-channel interference. Secondly, SM requires a single radio-frequency chain, regardless
of the number of antennas used, and therefore exhibits a significant energy saving. However,
the property of a single active antenna challenges the channel estimation process for SM: the
transmit antennas have to be activated sequentially for sending pilot signals. As a result, the
time consumed in pilot transmission is proportional to the number of transmit antennas. However,
this fact has so far been neglected in related research. Also, published research on SM
has focused on point-to-point communications, and few have covered a network perspective. In
this thesis, a comprehensive study is undertaken on SM systems in single-user, multi-user and
multi-cell scenarios.
As a unique three-dimensional modulation scheme, SM enables a trade-off between the size of
the signal constellation diagram and the size of the spatial constellation diagram. In this thesis,
an optimum transmit structure is proposed for SM to employ an adaptive scale of antennas
against channel correlations. Unlike traditional antenna selection methods, this new approach is
not sensitive to fast fading, due to the exploitation of statistical channel state information (CSI)
instead of instant CSI. The proposed transmit structure is demonstrated to have a near-optimal
performance against exhaustive search, while achieving very low computational complexity.
In addition, three novel methods are developed to improve the channel estimation process for
SM. A first method estimates the entire MIMO channel by sending pilot signals through only
one of the transmit antennas, among which the channel correlation is exploited. In a similar
way but focusing on the receiver, a second method can improve the estimation accuracy without
increasing the pilot sequence length. A third method balances the transmission power between
pilot and data to minimise the bit error rate. A framework of combined channel estimation is
also proposed, in which the three methods are jointly applied.
Furthermore, the antenna allocation in multi-user SM is studied, in order to explore multi-user
diversity gains. A method that jointly manages transmit antennas and receive antennas for all
co-channel users is proposed. The aim of this new method is to maximise the channel capacity
for each user, and the fairness among users is taken into account. It is demonstrated that the
proposed method significantly improves the performance of multi-user SM, especially when
serving a large number of users.
Finally, a novel cooperative scheme is proposed for SM in a multi-cell scenario. Based on the
concept of coordinated multi-point transmission (CoMP), this scheme enables the coordinated
users to swap the base station antennas pertaining to them. A three-tier cellular architecture is
further developed to switch between CoMP and the cooperative scheme.