Design and validation of Novel Transmit beam Shaping Technique for Distributed Phased Array Radars

Download Full-Text PDF Cite this Publication

Text Only Version

Design and validation of Novel Transmit beam Shaping Technique for Distributed Phased Array Radars

Shashi Shekhar 1, Sarath Chandra 2, Harshvardhan Paithane3 D&E-RS-NS2, Bharat Electronics Limited, Defence PSU, Bangalore, India

Abstract3D surveillance radar with distributed phased array architecture requires different Tx and Rx beams to meet the Radar's vertical coverage requirement.

In typical distributed array of MxN elements, elements are arranged into M number of rows with each row having N elements and each row of N elements are fed with a single high power Transmitter module through RF power division network. While it is advantageous to have multiple narrow beams in receive to do elevation coverage, it is better to have a single wide

Height

Broaden Tx Beam

Multiple Tx Beams

Range

beam in transmit due to azimuth scan timing constraints. In such type of applications, the efficient beam forming technique can be used for same antenna to create broader beam on transmit and narrow switchable beams on receive without increasing the complexity and RF distribution losses.

In this paper, the application of one of the phase only techniques "Quadratic phase weighting technique" is validated for a phased array antenna used in L band surveillance radars. This paper brings out the comparison between results of MATLAB modeled algorithm with EM simulation of hardware implemented in CST. This result gives insights of implementation aspects of phase only technique for Tx beam spoiling.

Keywords–Phased Array Antenna, Beam Width, Beam forming, Radar, MATLAB, CST, Gain, Phase, T/R module

  1. INTRODUCTION

    Semi active Phased Array Antenna with distributed array architecture is quite popular for Air/surface surveillance radar applications. Unlike fully active Phased Array Antenna, where in each antenna element is fed by a separate T/R module, group of elements in a semi active distributed array are fed with common RF input/output through power distribution networks. This architecture gives advantage in terms of cost due to reduced Transmit/Receive hardware. But, the challenge lies in designing RF distribution networks with required amplitude and phase weights to form desired beams in Azimuth and Elevation.

    In general, the Azimuth beam width (AZ_BW) of antenna array for rotating radar is fixed constant (Typ: 20-50). Taylor

    Fig.1.Transmit beams (Narrow and Broadened)

    The typical natural Elevation beam width of antenna array is in the range of 20 to 30. In general, 3D radar for air surveillance is expected to scan about 20 to 450 in elevation. For example, to scan200in elevation, narrow beam of 20has to be switched 10 times. For AZ_BW of 20 to 50, the dwell time available at 12 rpm is in the range of 26 to 70 msec. A very little time of 1.3 to 3.5msec is available to switch multiple high power narrow beams to meet the required vertical coverage. Considering the scanning and switching time constraints, it is advantageous to have broadened beam on transmit rather than multiple switchable beams. On receive multiple beams can be formed using either analog or digital beam forming based on system requirement.

  2. BEAMSHAPING APPROACH FOR 16X16 PHASED ARRAY ANTENNA

    Consider a 16X16 Phased Array Antenna arranged in rectangular fashion with 16 elements per row and 16 of such rows placed one over other in vertical direction. Each row has equally spaced 16 elements with each element fed with 1:16 power divider (power combiner on receive) and element amplitude is tapered /weighted as per Taylor distribution (N=16, SLL 30dB, Nbar5).But all the elements are in same phase as other. This gives us a narrow beam in Azimuth with side lobe level of 30dB.

    or Chebyshev distributions are popularly used to give amplitude weights for elements in row to generate required

    A8,Ø

    0

    A7,Ø

    0

    A6,Ø

    0

    A5,Ø

    0

    A4,Ø

    0

    A3,Ø

    0

    A2,Ø

    0

    A1,Ø

    0

    A1,Ø

    0

    A2,Ø

    0

    A3,Ø

    0

    A4,Ø

    0

    A5,Ø

    0

    A6,Ø

    0

    A7,Ø

    0

    A8,Ø

    0

    AZ_BW and Side Lobe Levels (SLL)[1].The value of AZ_BW is kept same irrespective of Transmit/Receive mode.Fig.1 illustrates the two possible beam forming techniques to meet vertical coverage requirement on transmit.

    1:16 POWER DIVIDING/COMBINING NETWORK

    TRANSMIT/RECEIVE MODULE WITH AMPL,PHASE CONTROL

    Fig.2.Single Row of 16 elements with Taylor Amplitude tapering and 1:16 power division

    As shown in Fig.2, each single row of 16 elements is fed with a T/R module which can further control the common amplitude and phase applied to the entire row. As shown in Fig.3,sixteen of such rows with T/R module are stacked vertically to form the complete 16×16 array. Now we have 16 T/R modules whose transmit and receive amplitudes & phases can be controlled to achieve desired beam shape in Elevation.

    Fig.3.16 Rows stacked vertically to form 16×16 array

    There are different amplitude and phase distribution methods which we can use for beam shaping. The most common techniques are amplitude and phase tapering. In the distributed array radar where there is one transmitter per row, the transmitter is always driven near to saturation. Hence, any extra loss in amplitude tapering cannot be compensated at

    pattern in space. Similarly, we can synthesize an antenna radiation pattern with required beam width by calculating individual element phase excitation weights.

    1. APPLICATION OF QUADRATIC PHASE DISTRIBUTION IN ELEVATION FOR 16X16 ARRAY

      There are many phase only beam shaping techniques available in literature. In this paper, "Quadratic phase weighting method is considered and applied to the 16×16 array configuration to generate the phase weights for each row of 16 elements. The theory and detailed analysis of the method used is elaborated else ware [3]

      Let each row of 16 elements is separated from other row by distance 'd', which is fixed based on frequency of operation, maximum array size limitation and Gain-Beamwidth requirements. Length of the array (L) along the elevation

      direction is given by = ( 1).

      For a symmetric array of even number of elements, the center

      two elements have same amplitude and phasor weights.

      Let is the maximum phase applied at the two extreme end elements of array (Element 1 & 16), Then the phase

      excitation at each element of array () can be calculated as

      2

      low power. Even, high power attenuators at power amplifier

      = 4 ( )

      (3)

      output are not a viable solution. So, Phase only beam shaping

      techniques will solve the problem of beam broadening in distributed array. However, selection of right beam shaping technique depends on transmission line technology used (waveguide/ stripline/ Microstrip/ etc.), ease of design, implementation cost, and complexity constraints.

      III. EFFECT OF PHASE-ONLY EXCITATION WEIGHTING ON ARRAY FACTOR

      For a linear array of 'N' elements spaced apart by a distance d, Array factor (AF) is given as

      Where = Coordinate of nth element along the antenna

      array length 'L'.

      These phase excitation are also explained as phase errors. [4].Fig.4 shows the quadratic curves of N=16 element array

      for different max values with each sample value

      representing particular element excitation phase

      n (1)

      AF = a

      =1

      2

      (1)

      Where = + , =

      is a wavelength

      is the wavenumber and

      an= Amplitude excitation of nth element

      =Phase excitation of nth element

      For phase only weighting, an can be made equal to unity. The

      revised Array factor can be written as

      Fig.4: Element phase excitation for N=16 elements for various values

    2. MATLAB SIMULATION OF 16 x 16 PHASED ARRAY

ANTENNA WITH AZ AND EL WEIGHTS

1 = (1)

=1

(2)

The phase of each element depends on two factors. First one

is ,the phase excitation applied to each element and the second factor is the natural phase difference seen by each

element with respect to other element due to path length and

looking angle difference quantified by .[2].

The phase excitation weights to each element can control the

vector phasor sum resulting in constructive or destructive

MATLAB simulation is done to generate Azimuth and

Elevation radiation patterns for 16×16 antenna array with isotropic elements [5]. Simulation is done at L-band with N=16 elements. Taylor distribution (N=16, SLL -30dB and nbar=5) is used to amplitude weigh elements in azimuth.

Values (1, 2 . 16) generated as per Quadratic phase weighting method for various (00,900,1350,180&2250)

are given as phasor weights for 16 rows in elevation. Table-I

shows the phase excitation weights calculated for

=1800case as per equation(3).

TABLE-I

Row No

Phase excitation weight in degrees

1

180

2

135.2

3

96.8

4

64.8

5

39.2

6

20

7

7.2

8

0.8

9

0.8

10

7.2

11

20

12

39.2

13

64.8

14

96.8

15

135.2

16

180

Table.1.Phase excitation weights calculated as per Quadratic phase method to widen beam in elevation

Fig.5shows that Azimuth radiation pattern of 16X16 array for

various values. It is observed that maximum directivity

of array for =00 is around 28.14 dBi. with Azimuth

Fig.6.Elevation pattern of 16×16 array simulated in MATLAB with Taylor weights in azimuth and quadratic phase weights in elevation

for different values

Fig.6 is of interest which shows the effects of on elevation pattern. As increases, directivity decreases

and the beamwidth flares up and increases. The energy in

main beam direction is distributed to side lobes causing side lobes to increase slightly. Directivity has reduced from

28.14dBi for =00to 22.83dBi in =1800 case.

Interestingly, the beamwidth has increased from 4.50 to

19.20.It can also be found that pattern ripples are observed for

=1800and = 2250.Typically , pattern variation of

±0.3dB is observed for =1800 case .Generally, pattern

variation of ±1dB is accounted in Radar's link budget .So,

selection of optimal for beam broadening depends on

Gain-BW requirements and tolerance to ripple.

VI.VALIDATION OF MATLAB RESULTS WITH CST SIMULATION RESULTS

For validation of the proposed method, EM simulation has been carried out in CST [6] for 16×16 array as configured in

MATLAB for two cases ( =00and =1800).Design

inputs for CST simulation are same as given for MATLAB

including Taylor weights and phase weights except a Microstrip based folded dipole (Gain of 4.8dBi, BWAZ x

0 0

beamwidth of 4.50.=00 case corresponding to the state

where all the elements of array are in phase and corresponds

to maximum directivity. For =1800, directivity at peak

is found to be 22.83dBi.It is to be noted that the directivity is

decreasing as is increasing towards 2250 from 00.But,

Azimuth beamwidth remains constant at 4.50 as phase

weighting in elevation do not affect the beamwidth in azimuth!

Fig.5.Azimuth pattern of 16×16 array simulated in MATLABwith Taylor weights in azimuth and quadratic phase weights in elevation

for different values

BWEL =75 x125 ) is used as antenna element for simulation purpose.

Fig.7.Azimuth pattern of 16×16 array simulated in CST for =00

Fig.8.Azimuth pattern of 16×16 array simulated in CST for =1800

Fig.7. and Fig.8 shows the narrow Azimuth pattern of 16×16

array simulated in CST for =00and =1800cases respectively.

Fig.9.Elevation pattern of 16×16 array simulated in CST for =00

Fig.9 is the naturally formed elevation pattern with all the

rows fed with equal phased (=00) RF inputs.

Fig.10.Elevation pattern of 16×16 array simulated in CST for =1800 Fig.10. demonstrates the effect of quadratic phase weighting on beam broadening of elevation beam as compared to

=00 case.

Fig.11.MATLAB simulated 3D pattern for =00

Fig.12.CST simulated 3D pattern for =00

For =00, Fig.11 is MATLAB simulated 3D pattern. Due

to Taylor weighting in azimuth, side lobe levels are very less

as compared to elevation side lobe levels and Fig.12 shows the CST simulated 3D pattern.

For =1800, Fig.13 is a MATLAB simulated 3D pattern.

Beam broadening in elevation is visible as expected .Fig.14 is

a CST simulated 3D pattern for the same case.

For both the cases, it is established that CST pattern follows the Matlab generated pattern.

Fig.14.CST simulated 3D pattern for =1800

Fig.13. MATLAB simulated 3D pattern for =1800

TABLE-II

Parameter

=00

MATLAB

CST

Gain at beam peak

28.14dBi

29.3dBi

Beamwidth Azimuth

4.50

4.50

Beamwidth Elevation

5.60

5.50

Sidelobe Level difference in Azimuth

30dB

29.5dB

Sidelobe Level differnce in Elevation

13.25dB

13.1dB

Ripple in pattern

No ripple

No ripple

Table-II.MATLAB and CST Simulation results comparison of 16×16 array for

=00

Comparison of MATLAB and CST results for =00 and

=1800 cases respectively is illustrated in Table-II and

Table-III respectively. It is observed the gain of the array

simulated in CST is more than MATLAB gain by ~1dB due to the use of folded dipole as radiation element in CST as compared to isotropic element considered in MATLAB. Elevation beamwidth is slightly lesser by 0.10 in CST as compared to MATLAB. This may be due to the change in pattern shape differences resulting in measurement error.

TABLE-III

VII CONCLUSION

A 16×16 rectangular array at L-band is considered for application and analysis of phase only beam broadening technique to widen the radar transmit beam in elevation. Quadratic phase weighting method is applied for N=16 elements in elevation and corresponding radiation patterns for

different cases of max(0,90,135,180&2250) are generated

using MATLAB. For validation of MATLAB generated

results, two cases(max = 00 and max = 1800) are

considered for EM simulation in CST software. All the

antenna array inputs are kept constant for CST simulation except that a folded dipole antenna is used as antenna element in CST.

It has been validated that by changing the phase excitations of individual elements in row by using phase only technique of "Quadratic phase weighting", elevation beamwidth of 16×16 array has been broadened from 5.60 to 19.20.A very good agreement of results is observed between MATLAB and CST simulations results. This gives good confidence in going towards implementation and fabrication of 16×16 array for L- band surveillance Radar applications where single antenna to be used to generate single wide beam in transmit and multiple narrow switchable beams on receive by only changing the phase excitation provided to each row of 16 elements.

VIII. ACKNOWLEDGEMENTS

The authors would like to thank G.M. NS2, Shri.Umesh K.S,

      1. Smt. Dharani.B and DGM Shri. Muralidhara.N f or their continuous support and guidance in publishing this research work

        Table- III.MATLAB and CST Simulation results comparison of

        Parameter

        =1800

        MATLAB

        CST

        Gain at beam peak

        22.83dBi

        23.88dBi

        Beamwidth Azimuth

        4.50

        4.50

        Beamwidth Elevation

        19.20

        19.10

        Sidelobe Level difference in Azimuth

        30dB

        29.5dB

        Sidelobe Level difference in Elevation

        11.19dB

        11.44dB

        Ripple in pattern

        ±0.36dB

        ±0.43dB

        16×16 array for =1800

        While Azimuth Side lobe level difference between beam peak

        and side lobes is observed almost same, Elevation side lobe levels are degraded by 0.2-0.3dB in CST simulation. Also, there is around ~2dB degradation of sidelobe levels in

        =1800 case as compared to max=00 case. But this 2dB

        degradation at beamwidth of ~19.20would not affect the

        detection or processing performance as much as the 2dB degradation in SLL if the main beamwidth is narrow. Interestingly, 5.42dB loss in beam peak gain is observed in

        CST simulation as shown in Table-3 for =1800 with respect to =00 case. Though it looks like a many fold

        reduction in gain, the actual effect on ranges will not be very

        large and will be compensated by thoughtful waveform planning and timing. Finally, a ripple of ±0.4dB is observed

        for max=1800 case, which can be comfortably factored

        under beam shape loss in radar link budget.

        REFERENCES

        [1] Section 7.6, Taylor Line source and 7.7,Taylor design-, Antenna Theory Analysis and Design, III edition, by Constantine. A. Balanis

        [2] Section 6.3, N-element Linear Array, Uniform Amplitude and spacing, Antenna Theory Analysis and Design, III edition, by Constantine. A. Balanis

        [3] Paper titled "Synthesis of wide beam array patterns using Quadratic- Phase Excitations"by Khalil H. Sayidmarie, QusaiH.Sultan, International Journal of Electromagnetics and Applications 2013,

        [4] Section 7.9,Quadratic phase distributions, Line source distributions, Antenna Theory Analysis and Design, III edition, by Constantine.

        1. Balanis

[5] MATLAB, generated by MATLAB 9.9 and Phased Array System Toolbox 4.4, ver R2020b, Mathworks,Inc.,

[6] CST Microwave studio ver 2021, 3D EM simulation and full array simulation, Dassault systems

AUTHORSINDEX

Shashi Shekhar received his B.E. degree in Electronics and communication from National Institute of Technology, Patna, India. He joined BEL Bangalore in Oct 2004.He has contributed in development of various Surveillance and Air-borne Radars.

His field of interest includes Active phased array radars, RADAR system design, and system integration. He is presently involved in design and development of Active phased surveillance radars in C/D band and E/F band. He is also a qualified PMP from PMI, USA and certified DFSS greenbelt professional.

Sarath Chandra received his B.Tech degree in Electronics and Communications Engineering from MVGR College of Engineering, Vizianagaram, AP, India in 2012. He joined Tata Consultancy Services (TCS),

Hyderabad in 2013 and worked on telecom software integration and verification. He joined Bharat Electronics, Hyderabad in 2014 as Deputy Engineer and worked in Integration and development of ESM and ELINT systems. Presently, he is working as Senior Engineer in Development and Engineering of naval 3D surveillance Radars. He is a practicing ASQ Certified Reliability Professional (CRE) since 2018.His areas of expertise include Antenna design, System integration and Reliability planning.

Harshvardhan Paithane received his B.Tech degree in Electronics and communication engineering from National Institute of Technology(NIT), Goa, India in 2020. He joined Jio Platforms Lim as engineer in ECPC WiFi team and later moved to Bharat

Electronics Limited in February 2021. He is presently working as Deputy Engineer in Development and Engineering of Naval Radar systems and involved in design and development of Active phased surveillance radars in E/F band. His interest lies in RADARs, communication networks, system design and prototyping. He also loves to explore futuristic technologies.

Leave a Reply

Your email address will not be published.