IEEE Aerospace and Electronic Systems Magazine - June 2020 - 65

Credit: Image licensed by Ingram Publishing

element has the capability of dynamically and simultaneously changing its characteristics, i.e., mode of operation, by reconfiguring its radiation pattern, polarization,
operational frequency, or a combination thereof [30]-
[33]. In parasitic layer pixel based MRAs, a reconfigurable
parasitic layer housing electrically small segments, called
pixels, which are interconnected via RF switches, is placed
in the near field of a driven antenna. By activating different
switch configurations, the geometry on the metallic surface
changes and antenna operates in different modes each having different pattern, polarization, and frequency properties. Depending on the number of pixels and switches in the
design, a single MRA may have a large number of different
radiation patterns called modes. It is important to note that
reconfiguration or control of an MRA corresponds to an
actual hardware adaptation, which is accomplished through
management of switch states, i.e., ON or OFF, within the
MRA. For an MRA array (MRAA), each element of an
array can be controlled to generate one of many modes,
which leads to generating many different beampatterns.
For example, an MRAA is shown to provide higher SNR
compared to fixed pattern phased array systems [34]-[36]
where MRA elements have multiple modes with tilted
beam properties. Additional degrees of freedom provided
by MRAs make them very useful in areas such as wireless
communications [34], [35], radar, and direction of arrival
(DoA) estimation [37], [38].
In an initial version of this study presented in [38] and
[39], we have shown that changing the MRA modes leads to
significant variations in terms of metrics such as the CramerRao lower bound (CRLB) for DoA parameters or signal-tonoise ratio (SNR) providing an opportunity to cognitively
select the modes that have potential to result in enhanced
performance of the array in terms of problems involving estimation, detection, or tracking tasks. Optimal selection of
modes for each antenna element in an array is not an easy
task considering that an MRAA of N antennas with each
MRA having L different modes could be in LN total number
of array "states." We have shown that a suboptimal, but
computationally efficient mode selection scheme can be
used to determine the MRA modes that will optimize CRLB
or SINR metrics [38]. To utilize MRAs for cognitive radars,

JUNE 2020

we have adopted the general cognitive radar framework for
target detection and tracking detailed in [14]. The work in
[14] formalizes a general flexible framework on Bayesian
target detection (track initialization/termination) and tracking (state estimation) along with the cognitive feedback
mechanism and the optimization criterion to obtain the next
set of sensor data to optimize the considered task. System
resources, such as observation time on each sensor, are cognitively allocated to optimize tracking performance in [14]
as a demonstration of the general framework. We have
developed a cognitive radar target detection and tracking
system built on this framework utilizing mode selection for
MRAAs for enhanced tracking performance.
The main contributions of this article are as follows.
1) To the best knowledge of the authors, for the first
time, a hardware component of a cognitive radar
system, i.e., MRA array, is physically and cognitively reconfigured leading to enhanced target tracking performance.
2) A Bayesian cognitive target tracking technique is
proposed utilizing the general framework in [14] to
minimize the CRLB of DoA parameters through
adaptive selection of the MRA modes.
3) A computationally efficient but suboptimal iterative
mode selection (IMS) algorithm is provided to find
the modes of each antenna in an array configuration
that will locally minimize the CRLB so that the cognitive controller can reconfigure the transmission
antenna parameters.
4) The developed radar signal model, CRLB of DoA
parameters for MRAA system and MUSIC-based
DoA estimation technique that achieves the CRLB
[38] is used to create cognitive MRA based tracking.
5) Proposed cognitive tracking technique with MRAAs
can provide enhanced tracking performance compared to fixed pattern arrays or MRAA random
mode arrays of the same size and average radiation
power. Analysis with respect to SNR and number of
modes are provided.

IEEE A&E SYSTEMS MAGAZINE

65
IEEE Aerospace and Electronic Systems Magazine - June 2020

Table of Contents for the Digital Edition of IEEE Aerospace and Electronic Systems Magazine - June 2020
