> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

1

Abstract—This paper presents a single-feed ultra-wideband

circularly polarized (CP) antenna with high front-to-back ratio.

The antenna is composed of two orthogonally placed elliptical

dipoles which are printed on both sides of a substrate. It has a low

profile, compact size and is demonstrated to be able to operate

over a 3:1 frequency range. A simple but effective method of

impedance matching is employed to broaden the bandwidth of the

antenna. To realize an ultra-wideband CP antenna with high

front-to-back ratio, a novel composite cavity is proposed and

integrated with the presented crossed dipoles. The composite

cavity effectively reduces the side-lobe and back-lobe of the

antenna. Detailed analysis of the working mechanism of this

cavity is provided. To verify the concept, a prototype covering the

GNSS and S frequency bands has been designed and

manufactured. Measured results are in good agreement with

simulated results and both demonstrate an impedance bandwidth

from 0.9 GHz to 2.87GHz (3.19:1=104.5%) and a 3-dB axial ratio

(AR) bandwidth from 1GHz to 2.87GHz (2.87:1=96.6%). The

measured back-lobe level is below -15dB and the front-to-back

ratio (FBR) is over 25dB across the whole GNSS band, showing

the antenna has superior multi-path mitigation performance. The

electric field distributions at different frequency points are given

in order to analyze the ultra-wideband CP characteristics of the

antenna. Compared to other reported single-feed CP antennas,

the antenna has advantages such as a wider CP bandwidth, lower

back-lobe radiation and better multi-path mitigation. It is a

promising candidate as a multi-function antenna covering several

wireless systems.

Index Terms—Crossed dipoles, single-feed antennas, circular

polarization, wideband antenna, GNSS, satellite communication,

back-lobe reduction.

I. INTRODUCTION

IRCULARLY polarized (CP) antennas are widely used in

global navigation satellite systems (GNSS) [1], satellite

Manuscript received March 13, 2015. L. Zhang, S. Gao, Q. Luo and P. R. Young are with the School of Engineeri

ng and Digital Arts, University of Kent, Canterbury CT2 7NT, UK. (emails:

lz76@kent.ac.uk; s.gao@kent.ac.uk; qiluo@ieee.org; p.r.young@kent.ac.uk). Q. Li is with the Department of Electronics and Information Engineering,

Huazhong University of Science and Technology, Wuhan 430074, China.

(email: qingxia_li@hust.edu.cn) Y. L. Geng is with the Institute of Antenna and Microwaves, Hangzhou

Dianzi University, Xiasha, Hangzhou, Zhejiang 310018, China.

R. A. Abd-Alhameed is with the Antennas and Applied Electromagnetics Research Group, University of Bradford, Bradford BD7 1DP, U.K.

communication systems, RFID and wireless power

transmission systems [2] due to their capabilities of reducing

polarization mismatch and suppressing multipath interferences

[3].

GNSS systems have found important applications globally in

both the military arena and commercial and consumer markets

[4]. However, there are several challenges in developing

high-performance GNSS antennas. To increase the positioning

speed, reliability and availability, the antenna should be able to

receive signals from various navigation systems such as global

positioning system (GPS), GLONASS, Galileo, Compass and

even Indian Regional Navigation Satellite System (IRNSS).

Consequently, the operation bandwidth for a high-performance

GNSS antenna should extend from 1.1GHz to 1.62GHz and

even to 2.5GHz (for IRNSS) with good circular polarization

properties.

Designing such a wideband circularly polarized antenna is

always challenging. Spiral antennas exhibit interior broadband

circular polarization characteristics and thus are extensively

studied [5-7]. However, their bi-directional radiation properties

make these conventional spiral antennas unsuitable for GNSS

applications since a directional radiation pattern is needed to

decrease the effects of the reflections from the ground [8].

Several directional spiral antennas including spiral-mode

microstrip antenna (SMM) [9], conducting plane backed spiral

[10], and cavity backed spiral [11] have been proposed. Helical

antennas are another kind of broadband CP antenna, which

have been extensively used in the GNSS scenario as well [12].

Another method to make the antenna fulfill wideband

coverage requirement is to employ multi-band antennas. Novel

multiband antennas, such as an aperture coupled stacked

microstrip antenna [13] have been proposed to cover GPS L1

and L2 bands and a triple band CP antenna for GPS application

is presented in [14].

Multi-path mitigation is another big challenge for

high-performance GNSS antenna design. Since non-LOS (line

of sight) multi-path signals could be generated by reflection of

satellite signals, the system performance will be degraded once

these multi-path signals are received by antennas [15]. Choke

ring ground plane [16] is an effective way to mitigate

multi-path signals through suppression of surface wave

propagating along ground planes. Based on choke ring ground

plane, a novel wideband non-cutoff ground plane has been

Single-Feed Ultra-Wideband Circularly

Polarized Antenna with Enhanced Front-to-Back

Ratio

Long Zhang, Steven Gao, Member, IEEE, Qi Luo, Member, IEEE, Paul R. Young, Senior Member,

IEEE, Qingxia Li, Member, IEEE, You-Lin Geng and Raed A. Abd-Alhameed, Senior Member, IEEE

C

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

2

presented in [8]. Choke rings are, however, rather bulky and

expensive. In order to reduce the complexity as well as the cost

of choke ring ground plane, a compact-size cross-plate reflector

ground plane (CPRGP) has been proposed [17]. The bandwidth

of these ground planes is limited and none of them are able to

cover such a wide frequency range as required for a

high-performance GNSS antenna. One of the objectives of this

presented work is to fill this gap.

In this paper, an ultra-wideband CP antenna is presented. It is

able to cover all GNSS frequency bands and a part of S-band.

At first, a crossed elliptical dipole CP antenna is proposed and a

simple but effective impedance matching method is given.

Then, a novel composite cavity with unequal-length crossed

fins is designed to obtain low backward radiation and high FBR.

The proposed ultra-wideband CP antenna can work from 1GHz

to 2.87GHz while maintain very low back-lobe level across the

whole GNSS band.

The paper is organized as follows: section II introduces the

antenna structure and explains the simple but effective

impedance matching method for bandwidth enhancement;

section III presents the novel composite cavity and its operation

mechanism; section IV presents the measurement results and

comparisons with simulation results as well as the analysis of

electric fields above the cavity aperture; the conclusion is given

in section V.

II. ULTRA-WIDEBAND CROSSED ELLIPTICAL DIPOLE ANTENNA

Single-feed crossed dipoles with integrated phase delay lines,

producing circular polarization, have been initially presented in

[18]. However, the input impedance of this antenna is about

122+j122 Ω and thus cannot be directly connected to a 50Ω

feed line. To compensate the impedance mismatching as well

as enhance the axial ratio (AR) bandwidth, parasitic open loops

have been used [19]. Through changing the shape of dipole, this

kind of antenna can be designed for wideband [20, 21] or

multi-band [22, 23] operation.

In this section, the antenna configuration will be given at first

and then the proposed simple but effective impedance matching

method is explained.

A. Antenna Geometry

L1

W1

R1

R3

R2

Arm1

Arm2

Arm3

Arm4

(a)

W2

L2

Arm1Arm3

Arm2

Arm4

(b)

PinTop Layer

Bottom Layer

Substrate

Coaxial Line

h

(c)

Fig. 1. Geometry of proposed antenna: (a) top view, (b) bottom view, (c) side

view.

Fig. 1 shows the configuration of the proposed single-feed

broadband CP antenna. The antenna consists of two pairs of

double-sided elliptical dipoles which are placed

perpendicularly to each other. The dipole arm 1 & 2 and 3 & 4

are both connected by ring-shaped phase delay lines, which

introduce a 90o phase difference between the orthogonally

placed dipole pairs to produce circularly polarized radiation.

The antenna is etched on a 0.817 mm thick Rogers RO4003

substrate with a relative permittivity of 3.55 and a loss tangent

of 0.0027. The four elliptical dipole arms are all characterized

by a major axis length R1 and minor axis length R2 while the

phase delay line is a 3/4 ring with inner radius R3 and line width

W2. A pair of partly overlapped rectangular patches with width

W1 are printed on both sides of the substrate and are used for

placing the coaxial feed line. Furthermore, the patches are

employed to tune the impedance matching of the proposed

antenna, which is discussed in detail in the following text.

Similar to other crossed dipole CP antenna, the proposed

antenna radiates in bi-direction. As shown in Fig. 1, current

phase on arm 1 is ahead of current phase on arm 2 and thus a

RHCP wave is excited along broadside direction while a LHCP

wave propagates along downward direction.

B. Impedance Matching

The impedance matching for a crossed dipole CP antenna is

usually challenging. Methods such as using 75 Ω coaxial line

[18], employing parasitic elements [19] and changing the shape

of dipole [21] have been proposed to achieve good impedance

matching. However, these approaches are either complex or

space-consuming.

A simple but effective impedance matching method has been

utilized in the proposed antenna, which is simpler and more

effective than the aforementioned ones. By using this method, a

3:1 impedance bandwidth can be achieved without increasing

complexity and space of the presented antenna.

A magnified picture of the partly overlapped rectangular

patch is given by Fig. 2. As shown in the picture, the pair of

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

3

rectangular patches is printed on both sides of the substrate with

an overlapped area of W1 L1. This overlapped structure of two

patches results in a pair of parallel-plate capacitors, i.e.

capacitor C1 and C2 shown in Fig. 2. It is found that these two

capacitors can effectively enhance the impedance

characteristics of proposed antenna and thus broaden the

antenna’s bandwidth.

L1 L2

W1

C1 Slot

Pin

C2

Fig. 2. Configuration of overlapped rectangular patch: (a) top view, (b) bottom

view.

To analyze the proposed impedance matching mechanism,

qualitatively, the equivalent circuit model of the antenna is

provided in Fig. 3. The symbol YA denotes the antenna

admittance without the overlapped rectangular patches while

C1 and C2 denote the parallel-plate capacitors and Yin represents

the total input admittance of the proposed antenna.

YA C2C1

Yin

Fig. 3. Equivalent circuit model of proposed antenna.

Using the formula provided in [24], the capacitance of

parallel-plate capacitors C1 and C2 can be calculated by the

following equation when the electric charge density on the

plates is uniform and the fringing fields at the edges can be

neglected.

(1)

The approximation derives from the small slot on bottom

rectangular patch and thus the input admittance Yin of antenna

can be denoted by:

(2)

It is shown in equation (2) that the antenna input admittance

Yin is affected by the rectangular patch width W1 and length L1.

Therefore, the impedance characteristic of the antenna can be

tuned by these two parameters to achieve a good impedance

matching.

In order to demonstrate, intuitively, the effect of changing

parameters W1 and L1, the simulated input impedances under

different pairs of W1 and L1 are studied while keeping the other

antenna geometry parameters fixed, as in TABLE I. TABLE I

ANTENNA PARAMETERS

L2 W2 R1 R2 R3 h

6.2mm 1.5mm 23.4mm 18mm 5.2mm 0.817mm

The simulated input impedance of the proposed antenna for

various values of L1 is shown in Fig. 4. It is worth noting that

the rectangular patch width W1 is kept constant as 7.5mm while

the length L1 is varied. As can be seen from Fig. 4, the input

impedance without the overlapped patch (case L1 = 0 mm) is

large inductively and the majority of the impedance loci falls

outside the VSWR = 2 circle, indicating that the bandwidth at

this state is very narrow. By increasing L1, the impedance loci

begin to move along the admittance circle due to the increasing

capacitance of C1 and C2. This phenomenon can be well

explained by equation (2) as well.

L1 = 0 mm

L1 = 1.5 mm

L1 = 2.5 mm

L1 = 3.5 mm

3 GHz

1 GHz

VSWR = 2

Fig. 4. Simulated input impedance of proposed antenna with different patch

length L1

W1 = 1.5 mm

W1 = 3.5 mm

W1 = 5.5 mm

W1 = 7.5 mm

3 GHz

1 GHz

VSWR = 2

Fig. 5. Simulated input impedance of proposed antenna with different patch

width W1 Fig. 5 depicts the input impedance loci under different patch

widths, W1, when patch length L1 is 2.5 mm. Similar to the

phenomenon observed in Fig. 4, the impedance loci will move

along the admittance circle when W1 becomes larger. This

impedance matching procedure is similar to the method using

lumped LC elements to tune the impedance bandwidth.

However, the proposed method is more advantageous in terms

of antenna efficiency and complexity due to the absence of a

lossy -type or M-type LC circuit.

(a) (b)

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

4

C. Results

The prototype of the proposed antenna is shown in Fig. 6. As

shown in Fig. 6 , the antenna is etched on a 115mm 115mm

Rogers RO4003 substrate and is fed directly by a 50Ω SMA

connector.

Fig. 6. Prototype of proposed antenna: (a) top view, (b) bottom view.

Fig. 7 shows the simulated and measured VSWR of the

proposed broadband circularly polarized antenna. It can be seen

from the measured result that the antenna can achieve a VSWR

< 2 bandwidth from 0.96GHz to 3.02GHz (3.14:1 = 103.5%),

which is much wider than aforementioned crossed dipole

antennas [18-23]. Meanwhile, it is also worth noting that the

antenna size is only 106mm 106mm (0.34λm 0.34λm, λm

denotes the free space wavelength at the minimum working

frequency), which is smaller than the regular crossed dipole

antenna as shown in TABLE II.

Fig. 7. Simulated and measured VSWR of proposed antenna

TABLE II COMPARISON OF CROSSED DIPOLE CP ANTENNAS

Ref. No. Antenna

Size (mm)

Minimum

Working Frequency

fm (GHz)

Antenna Size

(in wavelength)

Impedance

Bandwidth

[18] 120 120 2.01 0.8λm 0.8λm 30.7%

[19] 120 120 1.97 0.79λm 0.79λm 38.2%

[20] 60 60 0.877 0.18λm 0.18λm 11.8%

[21] 55 55 1.99 0.36λm 0.36λm 50.2%

This Paper 106 106 0.96 0.34λm 0.34λm 103.5%

III. NOVEL COMPOSITE CAVITY FOR MULTIPATH MITIGATION

APPLICATION

Multipath interference, caused by the deleterious

superposition of signals received at a user’s antenna via

multiple paths from a satellite, is one key source of error in

relative positioning of GNSS systems [25]. For GNSS

receiving antennas, the ability of multi-path mitigation is

critical for achieving high performance. Theoretically, a good

multipath mitigating GNSS antenna should be able to have a

cross-polarization rejection ratio (or polarization isolation)

dB for signals incoming at any positive elevation angle

and back radiation less than -10 dB to decrease the effects of the

reflections from the ground [8]. In addition, the pattern roll-off

from zenith to horizon should be between 8 and 14 dB, while

phase center variation should be less than 2 mm [17].

The aforementioned requirements are very difficult to

achieve simultaneously and thus certain compromises need to

be made within the antenna design procedure. Moreover,

considering that the proposed antenna in section Ⅱ is a

ultra-wideband antenna, further compromise is expected to be

made. Besides, since the proposed crossed elliptical dipole

antenna is bi-directional, making the antenna radiate in one

direction, i.e. improving front-to-back ratio (FBR), is regarded

as the main goal during the antenna optimization procedure.

A. Design Concept

The proposed novel composite cavity is shown in Fig. 8. It

can be seen from Fig. 8 that the ground plane is composed of a

cylindrical cavity with crossed unequal length fins. The

composite cavity is designed in such a way to reduce the

back-lobe of the antenna maximally.

R4

W3

W4

R5

R6

t2

t1

h1

h2

h3

Fig. 8. Geometry of proposed composite cavity: (a) top view, (b) prospective

view.

E

H

k

EH

k z

x

y

Ⅰ

Ⅱ

E

H

k

EH

kz

x

y

Ⅰ Ⅱ

Fig. 9. Two plane waves on: (a) flat circular ground plane, (b) proposed cavity.

To explain the working mechanism of the proposed

composite cavity, similar analysis methods to [8, 17] are

adopted. It is well known that the surface wave propagating

along the ground plane will reradiate at the ground plane edge

and thus increase the side-lobes as well as the back-lobe of the

antenna. Therefore, suppression of surface waves can reduce

the back-lobe of the antenna.

(a) (b)

(a) (b)

(b) (a)

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

5

Considering the two TEM plane wavesⅠandⅡ shown in Fig.

9, they can be denoted by the following equations respectively.

( ) ( ) (3)

( ) ( ) (4)

The boundary condition of flat conductor ground plane

imposes that no tangential electric field can exist, i.e.

(5)

This equation indicates that plane waveⅠcannot propagate

along the circular ground plane while wave Ⅱ can exist.

Considering the situation in Fig. 9 (b), wave Ⅰ cannot

propagate along the cavity bottom either due to the boundary

condition. In contrast to the circular ground plane, wave Ⅱ

cannot propagate along the cavity either since there is a

mandatory boundary condition imposed by the crossed fins.

Therefore, it can be expected that the proposed composite

cavity can offer better ability of surface wave suppression

compared to regular circular flat ground plane.

As indicated above, the proposed antenna radiates in

bi-direction, RHCP towards the broadside and LHCP backward.

Conventionally, a large ground plane or a cavity is required to

reduce the back-lobe of such kinds of antenna by reflecting the

downward waves into the upward direction [19, 22]. The

proposed composite cavity placed underneath the antenna can

afford better back-lobe reduction ability by minimizing

diffracted waves; the mechanism can be explained as follows.

zx

y

Incident

Wave

Reflected

Wave

Diffracted

Wave

(a)

z

x

yIncident

Wave

Reflected

Wave

Diffracted

Wave

(b)

Ey

Hxk

z

x

y

Ex

Hy

k

Ⅲ

Ⅳ

(c)

Fig. 10. Downward wave above: (a) flat circular ground plane, (b) cylindrical

cavity, (c) proposed composite cavity.

Fig. 10 depicts the propagation of downward LHCP wave

excited by the antenna under different situations. When the

proposed antenna is placed above a large flat ground plane, the

majority of the LHCP wave will be reflected by the ground

plane. However, considerable downward waves can still

propagate across the ground plane due to the diffraction of

these waves and result in large back-lobe. This diffraction

effect can be decreased by replacing the flat ground plane with

a cylindrical cavity shown in Fig. 10 (b) as the peripheral

vertical wall can stop part of the diffracted waves.

A LHCP wave depicted in Fig. 10 can be expressed by the

following equation.

{ ⃗ ( ) ⃗⃗⃗⃗ ( ⁄ ) ⃗⃗ ⃗⃗ ( )

| ⃗⃗⃗⃗ | | ⃗⃗ ⃗⃗ | (6)

As a circularly polarized wave, the electric field vector ⃗ will rotate on xy plane. To analyze the downward wave around

the peripheral wall of the proposed composite cavity, two

transient plane waves Ⅲ and Ⅳ whose electric field are along

the +x and +y direction, respectively, have been plotted in Fig.

10 (c). These two waves cannot propagate along the vertical

fins due to the boundary condition given by equation (5).

Actually all the 16 vertical fins can be used to block the

diffracted waves when the direction of the transient E vector is

parallel to these fins. Therefore, the proposed composited

cavity is more advantageous for back-lobe reduction than the

regular cylindrical cavity.

B. Comparison of Back-lobe Reduction

It has been shown in the last section that the proposed

composite cavity can offer better back-lobe reduction ability

than a flat ground plane and regular cylindrical cavity. This

performance enhancement is mainly from the suppression of

surface waves and the blocking of diffracted waves around the

edge of the ground plane. To verify this analysis, a comparison

between a flat plane ground, regular cylindrical cavity and the

proposed composite cavity, in terms of the back-lobe reduction,

is given.

The geometry dimensions of proposed composite cavity

have been given in TABLE III. TABLE III

COMPOSITE CAVITY DIMENSIONS (MM)

W3 W4 R4 R5 R6 h1 h2 h3 t1 t2

35 25 100 95 110 50 52 55 5 3

For the flat ground plane and regular cylindrical cavity, the

most critical dimensions for back-lobe reduction are the ground

plane size as well as the height from antenna to ground plane.

To make an unprejudiced and reasonable comparison, the

proposed antenna in section Ⅱ is placed at a height of h1 = 50

mm to the ground plane for all the three cases. Meanwhile, the

radius of the ground plane is kept at 110mm.

The simulated back-lobe levels for the three ground plane

reflectors are shown in Fig. 11. As can be seen from this figure,

the back-lobe is very high when a flat ground plane is placed

underneath the broadband CP antenna. By using the regular

cylindrical cavity, back-lobes smaller than -12 dB can be

achieved across the whole GNSS band (1.1 - 1.62 GHz).

Further back-lobe reduction has been observed through

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

6

utilizing the proposed composite cavity. It can be noted that a

minimum 8 dB and 2dB enhancement in back-lobe reduction

can be achieved compared to flat ground plane and regular

cylindrical cavity respectively when the proposed composite

cavity is used as a reflector of ultra-wideband CP crossed

elliptical dipole.

Fig. 11. Comparison of back-lobe level using different ground plane reflectors.

C. Cavity Optimization

The proposed cavity is composed of regular cylindrical

cavity and unequal length fins which are perpendicularly

inserted into the vertical wall. The unequal length fins are

chosen in order to enhance the back-lobe performance across

the whole GNSS frequency band.

Fig. 12 depicts the back-lobe level under different pairs of

W3 and W4. As indicated in this figure, the back-lobe

performance degrades when the fins length increase to 35mm

under equal length fins. The back-lobe reduction under unequal

length fins state is similar to the equal 25mm fins situation

across the GNSS band. However, the latter case features a sharp

increase in back-lobe level after exceeding the GNSS

frequency band and thus an un-equal length fins composite

cavity is finally chosen to minimize the backward radiation.

Fig. 12. Back-lobe level under different W3 and W4.

IV. RESULTS AND ANALYSIS

In this section, the simulated and measured results of the

proposed composite cavity backed ultra-wideband CP antenna

are given to verify the overall radiation performance. The

electric field distribution in the aperture is provided to

demonstrate its circularly polarized characteristics intuitively.

A. VSWR

The prototype of the proposed ultra-wideband composite

cavity backed CP antenna is shown in Fig. 13. The rim of this

cavity is made by gluing two semi-circle shaped conductors

together. Therefore, the rim is not in a perfect circular shape,

which is expected to result in some differences between

simulated and measured results.

Y

XZ

Fig. 13. The prototype of ultra-wideband composite cavity backed CP antenna.

The simulated and measured VSWRs of the proposed

ultra-wideband CP antenna are shown in Fig. 14. As can be

seen from this figure, the measured VSWR is in good

agreement with the simulated one. The ripple around 1.8 GHz

may stem from measurement errors as well as the manufacture

accuracy. The simulated and measured results both indicate a

0.9 GHz to 2.87 GHz (3.19:1=104.5%) impedance bandwidth

can be achieved by the proposed antenna.

Fig. 14. Simulated and measured VSWR of proposed ultra-wideband CP

antenna.

B. Axial Ratio

Axial ratio (AR) is an important index to evaluate a

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

7

circularly polarized antenna. The frequency bandwidth within

which the broadside AR is smaller than 3dB is named as the AR

bandwidth of an antenna. The AR bandwidth is generally

narrower than the impedance bandwidth for a CP antenna and

thus the bandwidth of a CP antenna is usually referred as the

AR bandwidth.

To evaluate the CP performance of the proposed antenna, the

AR characteristics versus frequency relationship is given in Fig.

15. It can be seen from this figure that the measured 3dB AR

bandwidth is from 1 GHz to 2.87 GHz (2.87:1=96.6%), which

can cover the whole GNSS bands (including higher band of

IRNSS) and S band as well.

Fig. 15. Simulated and measured AR of proposed antenna.

C. Radiation Pattern

Since the proposed antenna is measured using a linearly

polarized standard horn antenna, only the two linear

components Eθ and Eφ have be obtained.

Fig. 16 shows the radiation patterns of the proposed

ultra-wideband CP antenna in two main planes (XoZ and YoZ

planes) at 1.15 GHz, 1.4 GHz, 1.7 GHz, 2.1G Hz and 2. 4GHz.

As shown in this figure, the proposed antenna can maintain

smaller than -15dB back-lobe and larger than 25dB FBR across

the whole GNSS frequency band.

XOZ-plane (a) YOZ-plane

XOZ-plane (b) YOZ-plane

XOZ-plane (c) YOZ-plane

XOZ-plane (d) YOZ-plane

XOZ-plane (e) YOZ-plane

Fig. 16. Simulated and measured radiation patterns: (a) 1.15GHz, (b) 1.4GHz,

(c) 1.7GHz, (d) 2.1GHz, (e) 2.4GHz.

D. Analysis of the CP Characteristic

As stated in [26], the radiation of a cavity backed antenna is

determined more directly by the electric field distribution in the

aperture than by the current on the exciter. To investigate the

ultra-wideband CP characteristic of the proposed antenna, the

electric field distributions at 10 mm height above the aperture at

1 GHz, 1.8 GHz and 2.6 GHz are given in Fig. 17. 0 deg 45 deg

90 deg 135 deg

(a)

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

8

0 deg 45 deg

90 deg 135 deg

(b)

0 deg 45 deg

90 deg 135 deg

(c)

Fig. 17. Electric field distribution in the cavity aperture at different frequency: (a) 1GHz, (b) 1.8GHz, (c) 2.6GHz.

As can be seen from the figure, the distribution of electric

field is varied at different frequencies. From Fig. 17 (a), it can

be noticed that the maximum electric field at phase angle 0o and

90o generate at the end of elliptical dipole arm. This indicates

that the proposed antenna works at a half-wavelength dipole

mode at 1 GHz. When the frequency increases to 1.8 GHz, the

strongest electric field occurs at the side edge of the elliptical

dipole arm (see phase angle 45o and 135

o in Fig. 17 (b)).

Therefore, it can be concluded that the resonance length of

proposed antenna is decreased, which makes the antenna work

at higher frequency band. At 2.6 GHz, the strong electric field

no longer appears just around the antenna (see phase angle 90o

in Fig. 17 (c)) and there are evident strong fields between the

antenna and the cavity. Comparing this field distribution with

those at lower frequencies, it is found that a higher mode is

excited, which results in the deterioration of radiation pattern

shown in Fig. 16 (e).

From the different electric field distributions at phase angle

0o, 45

o, 90

o and 135

o in Fig. 17, it can be seen that a right-hand

rotated electric field has been produced at all given frequencies.

This indicates that the antenna can work as a circularly

polarized antenna in an ultra-wideband frequency range.

In addition, the electric fields are almost all bounded by the

cavity and little fields can propagate over the cavity as shown in

Fig. 17 (a) and (b). This phenomenon also verifies the

back-lobe reduction ability of proposed antenna, which has

been analyzed in section Ⅲ. On the contrary, it is shown in Fig.

17 (c) that the electric fields can propagate over the cavity at 2.6

GHz and thus the back-lobe and side-lobe reduction ability of

the cavity is limited. This conclusion is in good agreement with

the simulated and measured radiation patterns at higher

frequencies.

V. CONCLUSION

A single-feed ultra-wideband circularly polarized antenna

has been proposed and investigated in this paper. A simple but

effective impedance matching method is employed to enlarge

the impedance bandwidth and a composite cavity is presented

to reduce the back-lobe of the antenna. Good agreement

between simulation results and measurement results proves that

this antenna can work in a nearly 100% bandwidth region as a

CP antenna. Analysis of the electric field above the cavity, as

given in the paper, has explained the ultra-wideband CP

operation of the proposed antenna. With a lower than -15dB

back-lobe level and a larger than 25dB FBR character, this

antenna is very promising for high-performance GNSS

application as well as S band applications.

ACKNOWLEDGMENT

The authors would like to thank for the funding from EC FP7

GaNSat project under the contract number 606981. The authors

also thank Mr. Simon Jakes and Mr. Clive Birch for their help

in antenna fabrication.

REFERENCES

[1] M. Maqsood, S. Gao, T. W. Brown, M. Unwin, R. De Vos Van Steenwijk,

J. Xu, et al., "Low-cost dual-band circularly polarized switched-beam array for global navigation satellite system," IEEE Trans. Antennas

Propag, vol. 62, pp. 1975-1982, 2014.

[2] S. Gao, Q. Luo and F. Zhu., "Circularly Polarized Antennas", IEEE Press & Wiley, 2014, UK.

[3] J.-Y. Sze and W.-H. Chen, "Axial-ratio-bandwidth enhancement of a

microstrip-line-fed circularly polarized annular-ring slot antenna,"

Antennas and Propagation, IEEE Transactions on, vol. 59, pp. 2450-2456,

2011.

[4] J. J. Wang, "Antennas for global navigation satellite system (GNSS)," Proceedings of the IEEE, vol. 100, pp. 2349-2355, 2012.

[5] J. Dyson, "The equiangular spiral antenna," Antennas and Propagation, IRE Transactions on, vol. 7, pp. 181-187, 1959.

[6] J. Kaiser, "The Archimedean two-wire spiral antenna," Antennas and

Propagation, IRE Transactions on, vol. 8, pp. 312-323, 1960. [7] M. Nurnberger and J. Volakis, "A new planar feed for slot spiral

antennas," Antennas and Propagation, IEEE Transactions on, vol. 44, pp.

130-131, 1996. [8] F. Scire-Scappuzzo and S. N. Makarov, "A low-multipath wideband GPS

antenna with cutoff or non-cutoff corrugated ground plane," Antennas and

Propagation, IEEE Transactions on, vol. 57, pp. 33-46, 2009. [9] J. J. Wang and V. K. Tripp, "Design of multioctave spiral-mode

microstrip antennas," Antennas and Propagation, IEEE Transactions on,

vol. 39, pp. 332-335, 1991. [10] H. Nakano, K. Nogami, S. Arai, H. Mimaki, and J. Yamauchi, "A spiral

antenna backed by a conducting plane reflector," Antennas and

Propagation, IEEE Transactions on, vol. 34, pp. 791-796, 1986. [11] H. Nakano, H. Oyanagi, and J. Yamauchi, "A wideband circularly

polarized conical beam from a two-arm spiral antenna excited in phase,"

Antennas and Propagation, IEEE Transactions on, vol. 59, pp. 3518-3525, 2011.

> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) <

9

[12] M. Caillet, M. Clénet, A. Sharaiha, and Y. M. Antar, "A broadband folded

printed quadrifilar helical antenna employing a novel compact planar

feeding circuit," Antennas and Propagation, IEEE Transactions on, vol.

58, pp. 2203-2209, 2010.

[13] D. M. Pozar and S. M. Duffy, "A dual-band circularly polarized aperture-coupled stacked microstrip antenna for global positioning

satellite," Antennas and Propagation, IEEE Transactions on, vol. 45, pp.

1618-1625, 1997. [14] O. P. Falade, M. U. Rehman, Y. Gao, X. Chen, and C. G. Parini, "Single

feed stacked patch circular polarized antenna for triple band GPS

receivers," Antennas and Propagation, IEEE Transactions on, vol. 60, pp. 4479-4484, 2012.

[15] L. I. Basilio, R. L. Chen, J. T. Williams, and D. R. Jackson, "A new planar

dual-band GPS antenna designed for reduced susceptibility to low-angle multipath," Antennas and Propagation, IEEE Transactions on, vol. 55, pp.

2358-2366, 2007.

[16] J. Tranquilla, J. Carr, and H. M. Al-Rizzo, "Analysis of a choke ring groundplane for multipath control in global positioning system (GPS)

applications," Antennas and Propagation, IEEE Transactions on, vol. 42,

pp. 905-911, 1994. [17] M. Maqsood, S. Gao, T. Brown, M. Unwin, R. de vos Van Steenwijk, and

J. Xu, "A Compact Multipath Mitigating Ground Plane for Multiband

GNSS Antennas," IEEE Transactions on Antennas and Propagation, vol. 61, pp. 2775-2782, 2013.

[18] J.-W. Baik, K.-J. Lee, W.-S. Yoon, T.-H. Lee, and Y.-S. Kim, "Circularly

polarised printed crossed dipole antennas with broadband axial ratio," Electronics Letters, vol. 44, pp. 785-786, 2008.

[19] J.-W. Baik, T.-H. Lee, S. Pyo, S.-M. Han, J. Jeong, and Y.-S. Kim, "Broadband circularly polarized crossed dipole with parasitic loop

resonators and its arrays," Antennas and Propagation, IEEE Transactions

on, vol. 59, pp. 80-88, 2011. [20] Y.-F. Lin, Y.-K. Wang, H.-M. Chen, and Z.-Z. Yang, "Circularly

polarized crossed dipole antenna with phase delay lines for RFID

handheld reader," Antennas and Propagation, IEEE Transactions on, vol. 60, pp. 1221-1227, 2012.

[21] Y. He, W. He, and H. Wong, "A wideband circularly polarized

cross-dipole antenna," Antennas and Wireless Propagation Letters, IEEE vol. 13, pp. 67-70, 2014.

[22] S. Ta, H. Choo, I. Park, and R. Ziolkowski, "Multi-band, wide-beam,

circularly polarized, crossed, asymmetrically barbed dipole antennas for GPS applications," Antennas and Propagation, IEEE Transactions on vol.

61, pp. 5771-5775, 2013.

[23] K. Saurav, D. Sarkar, and K. Srivastava, "Dual Band Circularly Polarized Cavity Backed Crossed Dipole Antennas," Antennas and Wireless

Propagation Letters, IEEE, vol. 14, pp. 52-55.

[24] H. Nishiyama and M. Nakamura, "Form and capacitance of parallel-plate capacitors," Components, Packaging, and Manufacturing Technology,

Part A, IEEE Transactions on, vol. 17, pp. 477-484, 1994.

[25] C. C. Counselman, "Multipath-rejecting GPS antennas," Proceedings of the IEEE, vol. 87, pp. 86-91, 1999.

[26] S.-W. Qu, C. H. Chan, and Q. Xue, "Wideband and high-gain composite

cavity-backed crossed triangular bowtie dipoles for circularly polarized radiation," Antennas and Propagation, IEEE Transactions on, vol. 58, pp.

3157-3164, 2010.