The Decoupling Methods for Increasing the Isolation
between Two Antennas
Min Li and Lijun Jiang
Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong
Abstract—This paper presented the parasitic structure
decoupling method and other related approaches to increase the
isolation between two closely coupled antennas. The design
parameters of the resultant decoupling structures based on the
proposed methods can be precisely derived instead of fitted to
achieve high antenna isolations. A two-element monopole array is
employed as a benchmark to demonstrate the effectiveness of the
proposed methods. The simulation results show that good
impedance matching and high port isolation could be achieved
simultaneously based on the proposed methods.
Keywords—Isolation, decoupling methods, MIMO antenna,
network parameters, impedance matching
I.
I
NTRODUCTION
For a compact MIMO system, the mutual coupling (MC)
among antennas could degrade the system performance.
Various decoupling techniques (DTs) including the decoupling
network (DN) [1], artificial metamaterial (MM) [2],
neutralization line (NL) [3], defected ground structure (DGS)
[4], parasitic structure (PS) [5], etc., have been developed to
reduce the MC. However, most DNs suffer from narrow
isolation bandwidths. The decoupling designs employing MMs
usually require large antenna separations. Complex
optimization processes were adopted for the NL, DGS and PS
to realize high antenna isolations.
In this paper, we present derivation based methods to increase
the isolation between two antennas. These methods are realized
by adopting different decoupling structures, i.e., the PS, DN, NL
and shorting structure (SS). The design parameters of the
resultant decoupling structure can be directly derived to
achieve high antenna isolations. These decoupling structures
have different formations but similar design procedure:
utilizing the reactive component to manipulate and eliminate
the MC between antennas.
II. M
ETHODOLOGY
In the following discussion, a theoretical network model is
developed first by using the parasitic structure (PS). The high
antenna isolation is achieved by deriving the network
parameters and manipulating its condition for the minimum
coupling. Notably, this network model can also be applied to
the other methods. But certain modifications are required.
A. Isolation Enhancement by the Parasitic Strucure (PS)
The network model of the proposed method for using the
parasitic structure assisted by reactive component is shown in
Fig. 1. To decouple the coupling between Ant 1 and Ant 2, a
parasitic structure (PS) is inserted in between them. Two
additional transmission lines (TLs) with the same characteristic
impedance Z
0
(normally equal to 50 Ω) and the electrical
lengths
θ
1
and
θ
2
are added to the antennas’ feed lines. The PS
is connected with a TL (Z
0
,
θ
3
) and terminated by a reactive
load Z
L1
. Hence, the problem is changed to find appropriate
parameter values of
θ
1
,
θ
2
,
θ
3
and Z
L1
to achieve the high
antenna isolation.
For a two-element antenna system with one PS in the middle,
as shown in Fig. 1, the 3 by 3 matrix [S
t
1
] at the reference plane
t
1
describes the relationship between two antennas and the
effect of the added PS. After adding the additional TLs with the
same characteristic impedances Z
0
and different electrical
lengths
, [1,3]
i
i
θ
∈∈θ
to their feed lines, a new scattering
matrix [S
t
2
] is obtained at the reference plane t
2
.
The voltage-current relations of the new 3-port network can
be expressed by its impedance matrix [Z
t
2
] as
= ⋅
222
ttt
VZI
(1)
where [Z
t
2
] can be obtained by applying an S-to-Z
transformation to [S
t
2
]. If the PS is terminated by a reactive
component with an impedance Z
L1
, the voltage-current
relationship across this reactive load can be describe by
22
313
tt
L
VZI=− ⋅
. (2)
Substituting (2) into (1) yields a new 2×2 impedance matrix for
the input ports of the two objective antennas only, i.e.,
= ⋅
333
ttt
VZI
(3)
For a perfect isolation, the following decoupling condition shall
be satisfied in (3):
3
21
0
t
Z =
. (4)
Solving (4) yields Z
L1
as
1
(
L
f= θ)
(5)
where
(
θ)
is a complex function with the input phase-delay
vector θ=[
θ
1
,
θ
2
,
θ
3
]. For MIMO antennas, higher radiation
efficiency is always preferred. Hence, the Z
L1
must be purely
imaginary to avoid any unwanted ohmic loss, i.e.,
1
Re{ } Re{ ( } 0
L
Ζ f==θ)
. (6)
Hence, the phase-delay vector θ is derived from (6) directly.
Then they are used to compute the purely imaginary Z
L1
using
(5). The reactance Z
L1
can be consequently realized by
appropriate inductor or capacitor.
In addition, implanting the resultant Z
L1
into [Z
t
3
] in (3) yields
the input impedances of both decoupled antennas. They are
used to design the matching circuits with little impact to the
resultant high isolation.
978-1-7281-0716-5/19/$31.00 ©2019 IEEE
Authorized licensed use limited to: Missouri University of Science and Technology. Downloaded on February 23,2024 at 17:31:47 UTC from IEEE Xplore. Restrictions apply.