Nowadays multiple antenna wireless systems have gained considerable attention due to their
capability to increase performance. Advances in theory have introduced several new schemes
that rely on multiple antennas and aim to increase data rate, diversity gain, or to provide
multiuser capabilities, beamforming and direction finding (DF) features. In this respect, it
has been shown that a multiple antenna receiver can be potentially used to perform radio
localization by using the direction of arrival (DoA) estimation technique.
In this field, the literature is extensive and gathers the results of almost four decades
of research activities. Among the most cited techniques that have been developed, we find
the so called high-resolution algorithms, such as multiple signal classification (MUSIC), or
estimation of signal parameters via rotational invariance (ESPRIT). Theoretical analysis
as well as simulation results have demonstrated their excellent performance to the point
that they are usually considered as reference for the comparison with other algorithms.
However, such a performance is not necessarily obtained in a real system due to the presence
of non idealities. These can be divided into two categories: the impairments due to the
antenna array, and the impairments due to the multiple radio frequency (RF) and acquisition
front-ends (FEs). The former are strongly influenced by the manufacturing accuracy and,
depending on the required DoA resolution, have to be taken into account. Several works
address these issues in the literature. The multiple FE non idealities, instead, are usually
not considered in the DoA estimation literature, even if they can have a detrimental effect
on the performance. This has motivated the research work in this thesis that addresses the
problem of DoA estimation from a practical implementation perspective, emphasizing the
impact of the hardware impairments on the final performance. This work is substantiated
by measurements done on a state-of-the-art hardware platform that have pointed out the
presence of non idealities such as DC offsets, phase noise (PN), carrier frequency offsets
(CFOs), and phase offsets (POs) among receivers. Particularly, the hardware platform will
be herein described and examined to understand what non idealities can affect the DoA
estimation performance. This analysis will bring to identify which features a DF system
should have to reach certain performance.
Another important issue is the number of antenna elements. In fact, it is usually limited by practical considerations, such as size, costs, and also complexity. However, the most
cited DoA estimation algorithms need a high number of antenna elements, and this does not
yield them suitable to be implemented in a real system. Motivated by this consideration,
the final part of this work will describe a novel DoA estimation algorithm that can be
used when multipath propagation occurs. This algorithm does not need a high number
of antenna elements to be implemented, and it shows good performance despite its low
implementation/computational complexity.
