|dc.description.abstract||Smart antenna technology is expected to play an important role in future wireless
communication networks in order to use the spectrum efficiently, improve the
quality of service, reduce the costs of establishing new wireless paradigms and
reduce the energy consumption in wireless networks. Generally, smart antennas
exploit multiple widely spaced active elements, which are connected to separate
radio frequency (RF) chains. Therefore, they are only applicable to base stations
(BSs) and access points, by contrast with modern compact wireless terminals with
constraints on size, power and complexity. This dissertation considers an alternative
smart antenna system the electronically steerable parasitic array radiator
(ESPAR) which uses only a single RF chain, coupled with multiple parasitic elements.
The ESPAR antenna is of significant interest because of its
flexibility in beamforming by tuning a number of easy-to-implement reactance loads connected
to parasitic elements; however, parasitic elements require no expensive RF circuits.
This work concentrates on the study of the ESPAR antenna for compact
transceivers in order to achieve some emerging techniques in wireless communications.
The work begins by presenting the work principle and modeling of the ESPAR
antenna and describes the reactance-domain signal processing that is suited to the
single active antenna array, which are fundamental factors throughout this thesis.
The major contribution in this chapter is the adaptive beamforming method
based on the ESPAR antenna. In order to achieve fast convergent beamforming
for the ESPAR antenna, a modified minimum variance distortionless response
(MVDR) beamfomer is proposed. With reactance-domain signal processing, the
ESPAR array obtains a correlation matrix of receive signals as the input to the
MVDR optimization problem. To design a set of feasible reactance loads for a desired
beampattern, the MVDR optimization problem is reformulated as a convex
optimization problem constraining an optimized weight vector close to a feasible
solution. Finally, the necessary reactance loads are optimized by iterating the convex problem and a simple projector. In addition, the generic algorithm-based
beamforming method has also studied for the ESPAR antenna.
Blind interference alignment (BIA) is a promising technique for providing an optimal
degree of freedom in a multi-user, multiple-inputsingle-output broadcast
channel, without the requirements of channel state information at the transmitters.
Its key is antenna mode switching at the receive antenna. The ESPAR
antenna is able to provide a practical solution to beampattern switching (one
kind of antenna mode switching) for the implementation of BIA. In this chapter,
three beamforming methods are proposed for providing the required number of
beampatterns that are exploited across one super symbol for creating the channel
fluctuation patterns seen by receivers. These manually created channel
patterns are jointly combined with the designed spacetime precoding in order to
align the inter-user interference. Furthermore, the directional beampatterns designed
in the ESPAR antenna are demonstrated to improve the performance of
BIA by alleviating the noise amplification.
The ESPAR antenna is studied as the solution to interference mitigation in small
cell networks. Specifically, ESPARs analog beamforming presented in the previous
chapter is exploited to suppress inter-cell interference for the system scenario,
scheduling only one user to be served by each small BS at a single time. In
addition, the ESPAR-based BIA is employed to mitigate both inter-cell and intracell
interference for the system scenario, scheduling a small number of users to be
simultaneously served by each small BS for a single time.
In the cognitive radio (CR) paradigm, the ESPAR antenna is employed for spatial
spectrum sensing in order to utilize the new angle dimension in the spectrum
space besides the conventional frequency, time and space dimensions. The twostage
spatial spectrum sensing method is proposed based on the ESPAR antenna
being targeted at identifying white spectrum space, including the new angle dimension.
At the first stage, the occupancy of a specific frequency band is detected
by conventional spectrum-sensing methods, including energy detector and
eigenvalue-based methods implemented with the switched-beam ESPAR antenna. With the presence of primary users, their directions are estimated at the second
stage, by high-resolution angle-of-arrival (AoA) estimation algorithms. Specifically, the compressive sensing technology has been studied for AoA detection with
the ESPAR antenna, which is demonstrated to provide high-resolution estimation
results and even to outperform the reactance-domain multiple signal classification.||en_US