PARTIAL ADAPTIVITY FOR MONOPULSE ANGLE ESTIMATION IN JAMMING

Kai-Bor Yu

Shanghai Key Laboratory of Intelligent Sensing and Recognition, Shanghai Jiao Tong University,
Shanghai, China. Email: kbyu77@yahoo.com

ABSTRACT cancellation technique for coping with multiple jammers

based on DBF. In Section 5, we extend the product pattern
Mainlobe cancellation makes use of an additional delta-delta concept to sub-array beamforming where adaptivity and
beam compared with the conventional monopulse system. monopulse processing can be achieved in sub-array or at
The salient feature of this technique is monopulse- array level. In Section 6, we illustrate these techniques for
preserving in jamming. This feature can be described in multiple mainlobe and sidelobe jammers based on DBF for
terms of orthogonal nulling and product pattern. The rectangular aperture. Section 7 is the summary. The MLC
orthogonal nulling can be extended to cope with multiple for non-rectangular aperture and for the 1-D linear array is
jammers based on digital beamforming (DBF) on discussed in a companion paper [7].
rectangular aperture. The product pattern concept can be
extended through sub-array beamforming with array factor II. MONOPULSE PROCESSING
and sub-array pattern. In both schemes, partial adaptivity
can be applied for achieving monopulse preservation in Current surveillance and tracking radars mostly employ
jamming. monopulse processing for the detection and angle estimation
of a target. Conventional monopulse processing involves
Index Terms— Monopulse radar, mainlobe cancellation, digital one beam in transmit and multiple simultaneous beams on
beamforming, partial adaptivity, non-rectangular aperture. receive. Typically a sum beam without any tapering is
employed in the transmit array for full power operation. On
I. INTRODUCTION
receive two or more beams are formed for target detection
and angle estimation, i.e. the sum beam, the delta-azimuth
Mainlobe cancellation technique [1-3] makes use of an
beam and the delta-elevation beam. The sum beam is used
additional delta-delta beam in comparison with the
for surveillance search and target detection. Once a target is
conventional monopulse system. The salient feature of the
detected, the ratio of the delta-azimuth beam over the sum
mainlobe cancellation technique is monopulse preserving
beam is used for the azimuth angle estimation, and the ratio
while canceling a jammer within the main-beam. This
of the delta-elevation beam over the sum beam is used for
technique makes use of an orthogonal nulling concept where
the elevation angle estimation. This approach for angle
jamming is canceled in one domain and monopulse
estimation is computationally efficient as it requires only the
processing is carried out in another domain, or in terms of
computation of the monopulse ratios and a table look-up for
product patterns, the adapted pattern being common to the
the angles. Practical radar received beams are typically
overall sum and delta patterns can be canceled out leaving
tapered for sidelobe control such as Taylor weighting for the
the monopulse ratio un-distorted. This concept can be
sum beam and Bayliss weighting or the difference beam.
extended to cope with multiple jammers using adaptive
digital beamforming (DBF) on one domain and monopulse
The delta-delta beam or the double-difference beam is a
beamforming on another domain [4-6]. In this paper, we
four-lobed far-field antenna response formed by an aperture
extend product pattern concept for sub-array beamforming
illumination that has an anti-symmetrical phase property in
where adaptivity and monopulse beamforming can be
both dimensions of a planar antenna array. The delta-delta
achieved in the complementary sub-array pattern or array
beam can be illustrated by considering a planar array with
factor. It should be noted that this approach for monopulse
the following analog beamforming architecture: row-based
preserving based on sub-array beamforming is not limited to
beamforming followed by the column-based beamforming,
rectangular aperture.
or vice-versa. Row-based beamfoming generates the row-
sum and row-delta beams. The sum and difference of the
In Section 2, we give an overview of monopulse processing
row-sum beams generate the overall sum and elevation-delta
and monopulse beamforming including a description of the
beam. The sum and difference of row delta-beam generates
delta-delta beam. In Section 3, we review the mainlobe
the azimuth-delta beam and the delta-delta beam. This delta
cancellation technique with a discussion of the orthogonal
delta beam is typically available but terminated because
nulling and product patterns. In Section 4, we describe two
conventional monopulse processing uses only the sum
schemes in extending the orthogonal nulling mainlobe

‹,(((

Authorized licensed use limited to: UNIVERSIDAD CARLOS III MADRID. Downloaded on June 13,2020 at 18:44:38 UTC from IEEE Xplore. Restrictions apply.
beam, the azimuth-delta beam and the elevation-delta beam. 30
Sum and Delta Beam

Taylor Sum Beam

1.5
Monopulse Ratio and Fit

The delta-delta beam provides an additional degree-of- 20

Bayliss Delta Beam

freedom for mainlobe cancellation, specifically, delta- 10

azimuth beam and delta-elevation beam serve as the 0 0.5

Monopulse Ratio
pattern (dB)
auxiliary beam for cancellation of jamming in the sum -10
0
-20

beam, and the delta-delta beam serves as the auxiliary beam -30 -0.5

for cancellation of jamming in the delta-azimuth and delta- -40

elevation beams. -50

-1

Monopulse Ratio
Linear Fit

-60 -1.5
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 -0.05 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05

III. MAINLOBE CANCELLATION AND

u(sine) u(sine)

ORTHOGONAL NULLING
The set of monopulse beams including the delta-delta beam
is considered for rectangular aperture. Tapering can be
applied such as Taylor coefficients for the sum beam and
Bayliss coefficients for the delta beam. The illumination
functions along the columns and rows are independent
leading to the product antenna patterns. These patterns are
separable in the directional cosines u and v. The antenna
patterns for the four beams can be expressed as the
following:
(1)
Fig. 1 The 1-dimensional sum and delta beam pattern factors (left)
and the corresponding monopulse ratio (right)
nulls at the jammer elevation angle. The adapted monopulse
ratio is given as the ratio of the adapted azimuth delta beam
over the adapted azimuth sum beam. The patterns can be
expressed as product patterns in azimuth and elevation. The
elevation patterns form identical nulls are canceled out in
the ratio thus leaving the azimuth monopulse ratio un-
distorted. A similar procedure applies to the adapted
elevation monopulse processing. This mainlobe cancellation
capability maintains target detection over the entire beam
except at the jammer location and the monopulse angle
estimation accuracy except at the null strips. The following
expressions illustrate the monopulse-preserving processing
just described:

Monopulse ratios can be determined by finding the ratios of

the delta-azimuth beam and the delta-elevation beam to the
sum beam respectively:

(2)

The antenna illumination functions are separable and the (3)

corresponding patterns can be expressed in terms of product
of the pattern factors. The common factor in the orthogonal
dimension is canceled out from both the numerator and
denominator leading to a ratio which is dependent only on
the concerned angle variable. The pattern factors and the
monopulse ratio are illustrated in Fig. 1.
Mainlobe cancellation technique makes use of the delta
beams to cancel jamming in the sum beam, and the delta-
delta beam is used to cancel jamming in the delta-azimuth
beam or the delta-elevation beam [3]. This nulling
approach cancels jamming in one dimension while forming
undistorted monopulse ratio in the other dimension. This
orthogonal nulling concept can be illustrated by examining
the adapted antenna patterns. The adapted azimuth sum
beam and the adapted azimuth delta beam have identical
IV. DBF EXTENSION OF ORTHOGONAL
NULLING ON RECTANGULAR APERTURE
This technique can be extended to cope with multiple
mainlobe and/or sidelobe jammers [4-6]. The extension is
based on digital beamforning (DBF) on rectangular aperture
with separable weighting vectors resulting in product
patterns. The separable beamforming corresponds to row-
based beamforming and column-based beamforming where
adaptive beamforming can be carried out in row (or column)
and monopulse beamforming can be carried out in column
(or row) respectively. DBF provides multiple DOFs for
canceling multiple mainlobe and/or sidelobe jamming. As a
result, the final full array beam patterns can be expressed as
product of two pattern factors, one pattern factor is the

Authorized licensed use limited to: UNIVERSIDAD CARLOS III MADRID. Downloaded on June 13,2020 at 18:44:38 UTC from IEEE Xplore. Restrictions apply.
monopulse sum or delta beam in one domain, and the other
factor is the adapted pattern in the other domain. The Mono
adapted pattern factor being common is canceled out when Monopulse Adapt Ang
forming the monopulse ratio. {ri} RBF CBF Est

Array
Input {ci} Monopulse Adapt Mono
RBF Ang
Adapted Fixed CBF Est
RBF Monopulse
CBF
Ratio

Adapted Monopulse
Fig. 3 Monopulse RBF/CBF followed by adaptive CBF/RBF and
Fixed monopulse angle estimation
CBF RBF Ratio

(5)
Fig. 2 Adaptive RBF/CBF followed by fixed CBF/RBF and
monopulse angle estimation

Two techniques can be designed to support the orthogonal
nulling approach for the digital beamforming array. One
technique is based on identical adaptive processing on the
rows (or columns) followed by column-based (or row-
based) monopulse beamforming [4,6]. This scheme is
described by Fig. 2. On the top part of the diagram, adaptive
row-based beamforming (RBF) is followed by fixed
monopulse sum and delta column-based beamforming
(CBF). On the lower part of the diagram, adaptive CBF is
followed by fixed monopulse sum and delta RBF. The
resulting beams will have the product patterns. In each case,
the adapted pattern factor is common to the overall sum and
delta beams cancels out leaving the monopulse ratio un-
distorted. The described processing will have monopulse
patterns that can be expressed as the following product
pattern factors:
This orthogonal nulling concept is considered partial
adaptivity because adaptivity is employed only in one
dimension and degrees-of-freedom (DOFs) in the other
dimension is used for monopulse processing. The
computational complexity and sample requirement is
substantially less than for a fully adaptive array. For a
rectangular array of size M x N, full adaptivity would have
MN DOFs and each snapshot is of dimension MN. For the
partial adaptivity considered here where identical row
processing or column processing are employed, each
snapshot for row or column is of dimension M or N,
respectively. Thus an array measurement X would
correspond to M snapshots for row beams and N snapshots
for column beams, i.e.

(6)

The array observation can be stacked up along the rows or

(4) along the columns as following:

The second technique is to generate the sum and delta

monopulse beams for the rows (or columns) and to follow (7)
by adaptive processing on the resulting row- (or column-)
beams [5,6]. This scheme is described by Fig. 3. In a similar
fashion as the first scheme, the overall beams will have
product patterns with a common adaptive factor that will be The adaptive row and column weights can then be computed
canceled out when forming monopulse ratio as shown by the as following:
following expressions:

Authorized licensed use limited to: UNIVERSIDAD CARLOS III MADRID. Downloaded on June 13,2020 at 18:44:38 UTC from IEEE Xplore. Restrictions apply.
 (8)
with the covariance matrices computed as following:
(9)
V. PRODUCT PATTERN USING SUB-ARRAY
BEAMFORMING
The salient feature of monopulse preserving in jamming
described in section 3 makes use of product pattern for
orthogonal nulling. This product pattern concept can be
extended to sub-array beamforming where the overall
pattern is the product of array factor and sub-array beam-
pattern. In this set-up, adaptivity can be applied at sub-array
level or at the array factor level. In each case, the adapted
pattern factor is common to the overall sum and delta beam
patterns which can be canceled out leaving the monopulse
ratio un-distorted. These processing architectures can be
summarized in Figure 4 and Figure 5. The monopulse
preserving in jamming property can be observed in the
following product pattern equations for adaptive sub-array
beamforming followed by monopulse processing (equation
10) and monopulse sub-array beamforming followed by
adaptive processing (equation (11)). In each case, the
adaptive pattern being common to the numerator and
denominator are canceled out, leaving the monopulse ratio
un-distorted:
(10)
(11)
Authorized licensed use limited to: UNIVERSIDAD CARLOS III MADRID. Downloaded on June 13,2020 at 18:44:38 UTC from IEEE Xplore. Restrictions apply.
Fig. 4 Adaptive sub-array beamformig followed by monopulse
sum/delta array beamforming
Fig.5 Monopulse sum/delta beamforming followed by adaptive
array beamforming
VI. SIMULATION EXAMPLE
We illustrate the salient features of the orthogonal nulling
scheme by considering a 32 x 32 digital array. The set of
quiescent monopulse beams are generated which includes
the sum beam, the delta-azimuth beam, the delta-elevation
beam and the delta-delta beam. The sum beam has Taylor
weighting and the delta beam has Bayliss weighting. The
corresponding product patterns can be described to be
product of Taylor-and-Taylor, Bayliss-and-Taylor, Taylor-
and-Bayliss and Bayliss-and-Bayliss, respectively. Fig. 6
shows the monopulse beam patterns. The patterns are
product patterns with the Taylor sum and Bayliss delta beam
pattern factors and the corresponding monopulse ratio
shown in Fig 1. Fig. 7 shows the 2-D monopulse curves for
azimuth and elevation.
The performance with multiple jammers have been carried
out to demonstrate the preservation of monopulse ratios in
multiple jamming. The scenarios selected has jammers at
(u,v) equal to (0.038, 0.036) and (0.02, -0.3) in sine space.
The element jammer-to-noise ratio (JNRe) is selected to be
30 dB. The adapted pattern factors are derived and
monopulse ratios are shown to be preserved except at the
nulls (Fig. 8). Monopulse angle estimation in the clear is
evaluated at element SNR of 0 dB as the baseline
performance. Performance such as root-mean-squares errors
(RMSE) are shown in Fig. 8. Fig. 9 shows the adapted sum-
azimuth beam and adapted delta-azimuth beam. Fig. 10
shows the beam cuts and the adapted azimuth monopulse
which are un-dsitorted. The simulation is repeated for the
following target with element SNRs of 0 dB, 10 dB, 20 dB,
30 dB in jamming. Angle estimation performance RMSE for
u and v are evaluated. Fig. 11 shows that performance
degradation along the nulls and the performance will
improve and the nulls will narrow with increase in SNR.
 Sum Beam Pattern (dB)
-0.5 Delta-Azimuth Beam Pattern (dB)
-0.5
20
-0.4 20

VII. SUMMARY
-0.4
-0.3 10
10
-0.3
-0.2
0
-0.2 0
-0.1
-10
-0.1
-10

This paper extends monopulse-preserving capability for 0

v(sine)
-20
0
0.1 -20

orthogonal nulling on rectangular aperture and sub-array 0.2

0.3
-30

-40
0.1

0.2
-30

beamforming product pattern. DBF on rectangular aperture 0.4 -50 0.3

-40

-50
0.4

supports multiple jamming cancellation in one domain while 0.5

-0.5 -0.4 -0.3 -0.2 -0.1
u
0 0.1 0.2 0.3 0.4 0.5
-60

0.5
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
-60

monopulse beamforming in another domain. For sub-array Delta-Elevation Beam Pattern (dB) u(sine)Pattern (dB)
Delta-Delta Beam
-0.5 -0.5
20
20
-0.4 -0.4

beamforming, the overall pattern is the product of sub-array -0.3

10
-0.3
10

beam-pattern and array factor pattern. Adaptivity for -0.2

-0.1
0

-10
-0.2

-0.1 -10

jamming cancellation can be carried out at sub-array level or

v(sine)

v(sine)
0 0
-20 -20

0.1 0.1

at array level, and the resulting adapted pattern factor being 0.2
-30
0.2
-30

-40
-40

common to the sum beam and delta beam will be cancelled 0.3

0.4
-50
0.3

0.4
-50

out leaving the monopulse ratio un-distorted. An array 0.5

-0.5 -0.4 -0.3 -0.2 -0.1 0
u(sine)
0.1 0.2 0.3 0.4 0.5
-60 0.5
-0.5 -0.4 -0.3 -0.2 -0.1 0
u(sine)
0.1 0.2 0.3 0.4 0.5
-60

simulation is carried out to demonstrate the monopulse-

preserving capability of the technique.
Fig. 6 Monopulse patterns of the sum beam (top left), the azimuth-
delta beam (top left), the elevation-delta beam (bottom left) and the
delta-delta beam (bottom right).
REFERENCES

[1] K.-B. Yu, “ Advanced Monopulse Processing of Phased Array Azimuth Monopulse Ratio
Elevation Monopulse Ratio
-0.04
Radar,” Proceedings of the 2015 IEEE Radar Conference, -0.04 1
1

0.8
-0.03
Cincinnati, OH, USA, May 2015.
0.8
-0.03

0.6 0.6
-0.02 -0.02

0.4 0.4

[2] K.-B. Yu, “ Enhanced Angle Estimation of 4-Channel Radar -0.01

0.2
-0.01
0.2

v(sine)
v(sine)

System,” Proceedings of the 2016 IET International Radar 0 0

-0.2
0 0

-0.2

Conference, Hangzhou, China, October 2015. 0.01

-0.4
0.01

-0.4
0.02 0.02
-0.6 -0.6

0.03 -0.8 0.03

[3] C.R. Clark, “ Main beam jammer cancellation and target angle 0.04
-1
-0.8

-1
0.04
estimation with a polarization-agile monopulse antenna,” Proc. of -0.04 -0.03 -0.02 -0.01 0
u(sine)
0.01 0.02 0.03 0.04
-0.04 -0.03 -0.02 -0.01 0
u(sine)
0.01 0.02 0.03 0.04

1989 IEEE National Radar Conference, Dallas, TX, pp. 95-100.

[4] K.-B. Yu, “ Adaptive digital sub-array beamforming and
deterministic sum and difference beamforming, with jamming Fig. 7 Monopulse ratio curves for the azimuth angle (left) and for
cancellation and monopulse ratio preservation,” U.S. Patent the elevation angle (right).
6,661,366, issued December 9, 2003.
RMSE for u angle estimate (msine) RMSE for v angle estimate (msine)
[5] K.-B. Yu, “ Adaptive digital beamforming architecture for -0.04
4
-0.04

target detection and angle estimation in multiple mainlobe and -0.03 -0.03 4

3.5

sidelobe jamming, “ U.S. Patent Number: 6,697,009, issued -0.02 -0.02

3.5

February 24, 2004 -0.01

3
-0.01
3
v(sine)

v(sine)

0 2.5
0

2.5

[6] K.-B. Yu, “Mainlobe cancellation, orthogonal nulling and 0.01

2
0.01

product pattern,” Proc. IEEE International Symposium on Phased 0.02 0.02

2

Array Systems and Technology, Boston, MA, October 2016. 0.03

1.5
0.03 1.5

0.04 1 0.04
-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04
[7] K.-B. Yu, “ An Overview of Mainlobe Cancellation with Some u(sine) u(sine)

Extensions,” 2017 IEEE International Conference on Signal

Processing, Communication and Computing,” Xiamen, China, Fig. 8 RMSE angle estimation performance in the clear
October 22-25, 2017.
(SNRe =0 dB)

Authorized licensed use limited to: UNIVERSIDAD CARLOS III MADRID. Downloaded on June 13,2020 at 18:44:38 UTC from IEEE Xplore. Restrictions apply.
 Sum Beam Pattern (dB) Delta-Azimuth Beam Pattern (dB) RMSE for u angle estimate (msine) RMSE for v angle estimate (msine)
-0.5 -0.5 10
-0.04 -0.04 10

20
-0.4 20 -0.4 9 9
-0.03 -0.03

-0.3 10 8
-0.3 8
10
-0.02 -0.02

SLJ -0.2 7
-0.2 0 7
0
-0.01
6 -0.01
-0.1 -0.1 6
-10
-10

v(sine)
v(sine)

v(sine)
v(sine)

0 5
0 0 5
0
-20
-20
4
0.1 0.01 4
0.1 0.01
-30
-30 3
0.2 3
0.2 0.02

0.3 MLJ
-40 0.3

0.4
-40

-50
0.03
2

1
0.02

0.03
2

1
-50
0.4
0.04 0 0.04 0
0.5 -60 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04
-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04
0.5
-0.5 -0.4 -0.3 -0.2 -0.1 0
u(sine)
0.1 0.2 0.3 0.4 0.5
-60 -0.5 -0.4 -0.3 -0.2 -0.1 0
u(sine)
0.1 0.2 0.3 0.4 0.5 u(sine)
(a) u(sine)

RMSE for u angle estimate (msine) RMSE for v angle estimate (msine)
10 10
-0.04 -0.04

9 9
-0.03 -0.03

8 8

-0.02 -0.02
7 7

Fig. 9 Adapted sum beam (left) and adapted delta-azimuth beam -0.01
6
-0.01
6

(right)

v(sine)

v(sine)
0 5 0 5

4 4
0.01 0.01

3 3
0.02 0.02
2 2
Quiescent and Adapted Sum Beam Pattern Azimuth Monopulse Ratio
-0.04 1 0.03 0.03
Quiescent Pattern 1 1
20 Adapted Pattern
0.8 0.04 0.04
-0.03 0 0
-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04

0
-0.02
0.6
u(sine)
(b) u(sine)

0.4 RMSE for u angle estimate (msine) RMSE for v angle estimate (msine)
10 10
-0.04 -0.04
adapted pattern (dB)

-20 -0.01
0.2 9 9
-0.03 -0.03
v(sine)

0 0 8 8
-40
-0.02 -0.02
-0.2 7 7
0.01
-0.01 -0.01
-60 -0.4 6 6

0.02
v(sine)

v(sine)
-0.6 0 5 0 5

-80
0.03 -0.8 4 4
0.01 0.01

-1 3 3
-100 0.04
0.02 0.02
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04
v(sine) u(sine) 2 2

0.03 0.03
1 1

0.04 0 0.04 0
-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04
u(sine)

RMSE for u angle estimate (msine)

(c) u(sine)

10 RMSE for v angle estimate (msine)

-0.04 10
-0.04

Fig. 10 Quiescent and adapted sum beam (left) and azimuth -0.03
9
-0.03
9

monopulse ratio (right) -0.02

8

-0.02
8

7 7

-0.01 -0.01
6 6
v(sine)

v(sine)

0 5 0 5

4 4
0.01 0.01

3 3
0.02 0.02

2 2

0.03 0.03
1 1

0.04 0.04 0
0
-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04
u(sine) u(sine)

(d)

Fig. 11 RMSE angle estimation performance in jamming for

various SNRs: (a) 0 dB, (b) 10 dB, (c) 20 dB and (d) 30 dB. Plots
on left side are for u angle estimate and plots on right side are for v
angle estimate.

Authorized licensed use limited to: UNIVERSIDAD CARLOS III MADRID. Downloaded on June 13,2020 at 18:44:38 UTC from IEEE Xplore. Restrictions apply.