06-10-2012, 04:20 PM
Reconfigurable MEMS Antennas and Coupling Matrix Estimation
Reconfigurable MEMS.pdf (Size: 4.23 MB / Downloads: 76)
Abstract
One of the demands for future wireless communication systems is higher data
rates. New applications demand higher data rates and higher data rates give
the service providers the possibility to offer new services. To achieve higher
data rates the concept of MIMO (Multiple-Input Multiple-Output) systems
has emerged. The basic principle behind MIMO is to use multiple antennas
in contrast to the currently deployed systems mostly based on single antenna
systems. The handheld devices need to be small and at the same time ver-
satile due to the mobility of the user. To improve the overall performance
following the MIMO paradigm, several antenna elements may be introduced
on each handheld device. Requiring one feed chain per antenna element, this
would result in a considerable increase in space, cost, and complexity and
makes the implementation of large MIMO systems a difficult task. One way
to overcome the setbacks is the use of reconfigurable antennas. For a fixed
number of antenna elements in an antenna array, the choice of reconfigurable
elements will increase the number of possibilities. The reconfigurability is
preferably achieved by integrating switches with the antenna to save space.
RF-MEMS (Radio Frequency Microelectromechanical Systems) switches be-
long to a relatively new concept with advantages such as having low loss,
better bandwidth properties, and demanding low actuation voltage.
INTRODUCTION
The wireless link between a transmit antenna system and a receive antenna
system can be described in a generalized way according to the drawing in Figure 1.1.
Instead of having a single antenna on each side, so-called antenna arrays may be
used. The antenna array consists of several antenna elements resulting in a more
versatile and less static antenna system. Different parameters may be adjusted to
fit the specific purpose of the antenna system. The choice of antenna elements, the
distance between them, and the relative feeding of them could be such parameters.
One of the demands for future wireless communication systems is higher data
rates. The trend is roughly the same as in the world of Internet. New applications
demand higher data rates and higher data rates give the service providers the
possibility to offer new services. To achieve higher data rates the concept of MIMO
(Multiple-Input Multiple-Output) systems has emerged.
Background – Reconfigurable Antennas
A schematic overview representing wireless communication between a base station
and several handheld devices is found in Figure 1.2. In this scenario, the handheld
devices need to be small and at the same time versatile due to the mobility of the
user. To improve the overall performance following the MIMO paradigm, several
antenna elements may be introduced on each handheld device. This corresponds to
nT > 1 and nR > 1 in Figure 1.1.
Previous Work – Reconfigurable Antennas
Different antenna structures for wireless communication have been proposed in
the literature, including a meandered slot antenna [WSEL00] and a dual frequency
PIFA [LHW97]. These are motivated by their low profile, low cost, and easy fabrication.
Also, monopole antennas have been used in wireless communications because
of their wide-band characteristics and design simplicity [KL07]. The concept of
MEMS switches has been discussed [Bro98] and manufactured examples of such
have been reported [OLS03]. Combining antennas and switches yields the concept
of reconfigurable antennas as discussed in [SLS+00]. With the reconfigurability
antenna parameters such as frequency, polarization, and radiation pattern may
be altered. Polarization reconfigurability is presented in [SJK04], while [HFZB03]
present a combination of radiation pattern reconfigurability and frequency reconfigurability.
The latter was implemented using copper tape instead of actual switches.
The challenge lies in combining the antenna structure and the MEMS switch package
into a working reconfigurable antenna, which was recently suggested in [RRS08].
An example of a process for monolithic integration of printed circuit boards and
MEMS switches was presented in [CQC+03]. In addition to combining the antenna
with MEMS switches, one likes to evaluate the impact of the reconfigurable antenna
on the overall wireless channel. With multiple antenna elements on both
the transmitter and receiver, the communication over a Multiple-Input-Multiple-
Output (MIMO) channel [GSS+03] can increase the capacity of a wireless channel
considerably [WFGV98] without the requirement of additional bandwidth or transmit
power [CHIO03] [KS05]. Recently, a reconfigurable array with a 2-element
monopole array and 6 parasitic elements was designed and measured with dummy
switches in an indoor environment [MPS06]. The proposed solution was shown to
improve the performance of the wireless link. Also, a 2-element array using printed
dipoles combined with 2 parasitic elements was designed and measured using PIN
diodes [PKF+08]. The array was evaluated by measurements in terms of indoor
channel capacity and a notable improvement in capacity with respect to a system
without reconfiguration was reported.
Background – Estimation of Coupling Matrix
Wireless communications includes many different aspects ranging from hardware
to software issues. One of them is the notion of adaptive antenna arrays. The
base station in Figure 1.2 could consist of an adaptive antenna array. The presence
of noise in a system and the increase of users leading to interference benefits from
such an antenna system. With adaptive antenna arrays, the radiation patterns may
change continuously over an infinite number of options.
Compared to switched beam systems (with multiple fixed beam patterns to
choose from) and single antenna systems, the range of a adaptive antenna array
system can be increased. In adaptive antenna arrays, signal processing is used
as a basis for decisions. An accurate estimate of the incoming signal is therefore
of importance. As part of that, the modeling of the antenna array is crucial.
Otherwise, the estimated signal could be biased and the decision made based on
that will deviate from the optimal choice. Assuming ideal behavior by the antenna
array when estimating the incoming signal is typically something that leads to
results that deviate from the optimum and reduces the performance of the adaptive
antenna array. [FW94], [Lin05a]
Previous Work – Estimation of Coupling Matrix
The effect of mutual coupling between the antenna elements based on array performance
has been reported for antenna arrays and found to be significant [GK83]
[LLSB98]. Different methods to compensate for the mutual coupling has been
proposed in the literature. Except for the method of matrix multiplication compensation
[SH90] used here, the introduction of dummy elements [Lun96] and minimization
of the inter-element coupling [LLSB98] has been reported. To find the
coupling matrix, methods based on far-field calculations [SH90], and lest-squares
estimation [SDL00] has been reported.
The work presented in Part II of this thesis is based on previous work done by
Dr. Björn Lindmark [Lin05a]. Presented there is the problem of compensating for
the mutual coupling while considering the phase shift and element factor. Presented
here are two modified approaches to the same problem with the introduction of the
Cramér-Rao Lower Bound (CRB) as a theoretical tool for analyzing the estimator.
Thesis Outline
This thesis is divided into two parts related to the left and right side of Figure 1.2.
When communicating with the base station, knowledge about the environment
is helpful for optimizing the different parameters in hand. One such parameter to
change could be the physical structure of each antenna element that constitute the
antenna array of the handheld device. Part I deals with reconfigurable antennas
with three different antennas presented and discussed.
At the same time, the antenna system at the base station needs to estimate
the incoming signals. Typically, the signal is estimated based on measured data at
the terminals of the antenna system. The aim is to find out what was transmitted
and pass on the information to some post-processing unit. Part II deals with the
estimation of the coupling matrix given data from a calibration measurement of
an antenna array along with a discussion including the Cramé-Rao Lower Bound
(CRB).