The Impact of Mutual Coupling on MIMO Radar
Emissions
Brian Cordill
*
, Justin Metcalf
*
, Sarah A. Seguin
*
, Deb Chatterjee
†
and Shannon D. Blunt
*
Abstract — The effects of mutual coupling be-
tween antenna elements are considered with regard
to the impact upon co-located MIMO radar emis-
sions. Because this sensing scheme intentionally cou-
ples the spatial and fast-time (waveform) domains,
it is shown that MIMO radar is sensitive to any elec-
tromagnetic mutual coupling effects that are not ad-
equately characterized in the transmit array mani-
fold. This sensitivity leads to mismatch that will
degrade the radars sensitivity on receive.
1 INTRODUCTION
For traditional radar emissions where a single wave-
form is transmitted from all array elements the im-
pact of mutual coupling is generally viewed as a
limitation for receive-mode array processing. How-
ever, if different elements in the array emit dif-
ferent waveforms via the co-located multiple-input
multiple-output (MIMO) radar paradigm [1] then
mutual coupling effects must likewise be considered
on transmit. Due to constructive/destructive com-
bining in space (far field) of the different emissions
from the various antenna elements, different tem-
porally modulated signals are emitted in different
spatial directions. This intentional coupling of the
spatial and fast-time (waveform) domains can be
affected by the mutual coupling between proximate
antenna elements. Thus the time-domain signal
propagating in a particular direction may be some-
what different from what is presumed under the
assumption of no mutual coupling. As a result, a
mismatch is induced that will subsequently degrade
receive performance.
To assess the impact that mutual coupling has
upon the intentional space/fast-time coupling of
MIMOemissionsweconsidertheuseofacylindrical
array of dipoles due to the azimuthal invariance of
the spatial beampattern (to control for spatial vari-
ation and thereby focus on temporal distortion).
That said, the task of obtaining a high-fidelity EM
model of the cylindrical array manifold is quite in-
volvedandisatopicofon-goinginvestigation. Here
the current efforts for modeling the array manifold
are discussed followed by preliminary results using
a simplistic mutual coupling model that indicates
the performance degradation that is expected if the
presence of mutual coupling is not addressed.
∗
The University of Kansas
†
The University of Missouri – Kansas City
2 CYLINDRICAL DIPOLE ARRAY
The formulation for calculating the embedded ele-
ment pattern of a single dipole is available in [2]
including the development of the periodic Green’s
function and its subsequent simplification via the
use of the Floquet theorem [3] to reduce the entire
analysis to a single cell. The results thus obtained
apply strictly to infinite arrays, though it is shown
later that the embedded element pattern of the cen-
tral element in a small (e.g. 5 × 5) finite array can
still be adequately computed via this approach. It
is expected, however, that patterns for the edge el-
ements in the finite array would be different.
2.1 Dipole Embedded Element Pattern
The two dimensional cylindrical array consisting of
narrow, strip dipole elements is shown in Figure 1.
When a single dipole is excited it induces currents
on all the other elements, which manifests in the
form of mutual coupling. The total field at any
observation point is the superposition of the fields
radiated from the excited and passive dipole ele-
ments. Salient features of the full-wave MoM for-
mulation, leading to the solution of the electric field
integral equation (EFIE) for the array problem, is
described below from [2].
The currents on the individual dipole elements
are expressed in terms of a finite number Q of the
entire domain (sinusoidal) basis functions
J =ˆ z
Q
q=1
ˆ
i
q
sin
qπ
L
z +
π
2
. (1)
Employing the infinite, periodic, Green’s function
for the cylindrical configuration shown in Figure 1,
the Galerkin’s method is employed to obtain the
MoM matrix equation for solving the (complex)
current coefficients in (1) as
Q
q=1
Z
pq
ˆ
i
q
= V
p
, where, p =1, 2, 3, ··· , Q. (2)
In (2) the impedance ( Z
pq
) and voltage ( V
p
) terms
978-1-61284-978-2/11/$26.00 ©2011 IEEE
644