2914 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 63, NO. 7, JULY 2015

A Switched Beam Antenna With Shaped RadiationPattern and Interleaving Array Architecture

Han Wang, Zhijun Zhang, Fellow, IEEE, Yue Li, Member, IEEE, and Magdy F. Iskander, Fellow, IEEE

Abstract—A six-beam switched beam antenna (SBA), whosebeam patterns are shaped in the azimuth plane, is designed andfabricated. A synthesis method, which is based on a four-elementsarray and can generate a flat-top beam pattern with 46◦ flat gainregion (gain fluctuation within ±0.5 dB), 60◦ 3-dB beam width,and less than −20.1 dB side lobe level (SSL), is proposed. By com-bining six of those beams together, the proposed SBA can providenear consistent gain and good interference rejection in the azimuthplane. Meanwhile, an interleaving array architecture, which min-imizes the size of the proposed array, is also implemented. Toextend the beam coverage, a series-fed in-phase array with taperedamplitude distribution is introduced in the elevation plane. Thepeak gain of the beam reaches 16.6 dB, and the SLL is controlledunder −23.2 dB in this plane.

Index Terms—Antenna array, pattern synthesis, smart antenna,switched beam antenna (SBA).

I. INTRODUCTION

S MART antenna, as an array technique that can improvethe signal to interference/noise level and increase the fre-

quency reuse rate effectively, has been widely deployed inmodern wireless communication systems such as 3G and 4G[1]. The form of the implementation comprises two kinds ofarrays. One is the adaptive array that can trace the signal withits electrical adjustable beams, and the other is the switchedbeam antenna (SBA) that can choose the best beam for the userfrom its preset beams [2]. The former is more powerful with itsversatile beam generating ability, whereas the latter has its ownadvantage in that it can offer attractive performance boost in asimple and cost-effective ways [3]. The SBA is more suitablefor small-scale smart antenna applications [4], [5], especiallyfor those who has limit computing power and volume for theantenna system.

Manuscript received July 08, 2014; revised January 19, 2015; accepted April02, 2015. Date of publication April 14, 2015; date of current version July02, 2015. This work was supported in part by the National Basic ResearchProgram of China under Contract 2013CB329002, in part by the National HighTechnology Research and Development Program of China (863 Program) underContract 2011AA010202, in part by the National Natural Science Foundationof China under Contract 61271135, and in part by the National Science andTechnology Major Project of the Ministry of Science and Technology of Chinaunder Grant 2013ZX03003008-002.

H. Wang, Z. Zhang, and Y. Li are with the State Key Laboratoryof Microwave and Communications, Tsinghua National Laboratory forInformation Science and Technology, Tsinghua University, Beijing 100084,China (e-mail: zjzh@tsinghua.edu.cn).

M. F. Iskander is with the Hawaii Center for Advanced Communication(HCAC), University of Hawaii at Manoa, Honolulu, HI 96822 USA (e-mail:iskander@spectra.eng.hawaii.edu).

Color versions of one or more of the figures in this paper are available onlineat http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TAP.2015.2422838

Since the beam pattern of the SBA is fixed, the performanceof the SBA is largely determined by these preset beams and thebeam forming network (BFN) behind them. In classic designs,such as pencil beam-based SBA [6], [7], the designers mayemphasize more on designing high gain and low side lobe level(SLL) beam since it can extend the coverage and lower theinterference effectively. However, when combining these beamstogether to realize a full angular range coverage, the crossoverlevel, defining with the difference between the peak gain andthe gain at the intersection of the neighboring beams, is typi-cally high [8]. This causes problem in link budget estimationand may cause communication failure if the users are in thesezones.

To lower the crossover level, one solution is adding morebeams in a given angular range [9]. However, this will increasethe complexity of the BFN significantly, which will lose theSBA’s advantages in simplicity. Moreover, more beams meanmore overlapping area existing between adjacent beams. Theperformance will deteriorate [10] and the handoff rate willincrease, thus stressing the system [11]. If the designers usebeam with low gain and slow roll-off characteristics to solvethe crossover problem, the advantages of the SBA in interfer-ence filtering and range extension will be jeopardized, which isalso undesirable in real applications.

Ideally, the most suited pattern type for the SBA should bea rectangular-shaped pattern. Its flat top can provide identi-cal gain for all direction, where no crossover problem existsbetween neighbor beams. Moreover, its instantaneous transi-tion can provide ideal interference rejection outside the beam,and least number of beams are required for full angular cover-age since no overlapping exists between the adjacent beams.In practice, this kind of pattern can be approached with thepattern synthesized method proposed in [12]–[14]. It can beviewed as an extension of digital filter design in that the rela-tionship between the synthesis pattern and the excitation of auniform linear array (ULA) is Fourier transformation, whichis similar to the relationship between the frequency responseand impulse response of the finite impulse response (FIR) filter.Even though this kind of method is very effective to generateflat-top pattern with low SLL, the space freedom between thearray elements are not utilized, which means the number of ele-ments is not optimized. Thus, most of these works request tensor hundreds elements, which is not feasible in small-scale SBA.To improve the performance, Sabharwal et al. [15] proposed anew optimized method targeted to lower the cross-beam inter-ference, which utilizes this space freedom between elements.Nonuniformly spaced array is further introduced, which use

0018-926X © 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

WANG et al.: SBA WITH SHAPED RADIATION PATTERN AND INTERLEAVING ARRAY ARCHITECTURE 2915

analytical method [16], Bayesian compressive sampling tech-nique [17], and extended matrix pencil method [18], [19] todecrease the number of elements and minimize its effects onarray performance. However, these works are still based onlarge array with around tens or more elements, which is notoptimal for small-scale SBA applications.

In this paper, a six-beam SBA is designed and fabricated,whose beams are shaped in the azimuth plane. A synthesismethod based on a four-element array is proposed, which cangenerate flat-top pattern with low SLL. In this synthesizedpattern, the flat gain region (gain fluctuation within ±0.5 dB)reaches 46◦, and the 3-dB beam width covers 60◦. A fastroll-off characteristic with less than −20.1 dB SLL is alsoachieved, which can be viewed as a good approximation to therectangular-shaped pattern. By combining six of these beamstogether, the proposed SBA can provide near consistent gain inthe azimuth plane with least number of beams, and can real-ize good interference rejection outside the selected beam. Toreduce the size of the SBA and improve the aperture efficiency,an interleaving array architecture is proposed, with which onearray panel, the physical printed circuit board that supports thearrays, is shared by three adjacent arrays. In the elevation plane,an eight-element series-fed nonuniform amplitude distributedin-phase array is introduced, which extends the beam coverageeffectively. The gain of the beam reaches 16.6 dB, and the SSLis also controlled under −23.2 dB in this plane.

The paper is organized as follows. Section II gives the basicidea of the synthesis method and the interleaving array archi-tecture applied in the azimuth plane. The BFN and the double-layer feeding structure are also introduced, which is designedto satisfy the amplitude/phase requirement of the proposed syn-thesis method. Section III provides a detailed description ofthe series-fed nonuniform amplitude distributed in-phase arraydesign in the elevation plane. In Section IV, the built prototypeis described, and the measurement results are provided to verifythe performance of the proposed array.

II. SYNTHESIS METHOD IN THE AZIMUTH PLANE

AND THE BFN DESIGN

A. Synthesis Method

In previous work [20], we proposed a dual-beam SBA withshaped radiation pattern. The synthesis target is to generate two45◦ flat-top beams aiming at ±22.5◦, respectively. Since thosetwo beams share one aperture, a hybrid network is applied inthat design. Even though high aperture efficiency is achieved,the SLL of that design is relatively high (around −8 dB).

In this paper, specifically to lower the SLL and realize a sym-metric flat-top pattern, each array in the SBA is designed tosupport only one beam. The structure of the array is shown inthe top right corner of Fig. 1, which is composed of four patchelements. The central two elements are named as the “centralgroup” and the outer two elements are named as the “shap-ing group.” In each group, the elements are fed with the equalamplitude and phase to generate the pattern at broadside direc-tion. Their patterns, namely the original pattern and shapingpattern, respectively, are plotted in Fig. 1. Since the distance

Fig. 1. Basic concept of the synthesis method and the structure of the shapedarray.

between the elements in the shaping group is large, lobes withflipping phase appear in its pattern. By choosing proper phaseand amplitude relationship between the central and shapinggroup, these lobes in the shaping pattern can have differentimpacts on the original pattern. Thus, the shaping can be per-formed and the desired pattern can be achieved as shown inFig. 1. In this figure, the plus sign noted on the lobes of theshaping pattern means that it will have positive effect on theoriginal pattern while the minus sign means the contrary. Asa result, ripples can be observed in the main beam of the syn-thesized pattern, in which the gain in the central region of theoriginal pattern decreases and the falling edge of the main beamis sharpen. This pattern can be viewed as a good approximationto the rectangular-shaped pattern, and can be further optimizedwith the target shown as⎧⎪⎨⎪⎩

Pgoal1(ϕ): {max(P (ϕ))−min(P (ϕ))}<1 dB |−23◦<ϕ<23◦

Pgoal2(ϕ): {max(P (ϕ))−min(P (ϕ))}<3 dB |−30◦<ϕ<30◦

Pgoal3(ϕ): SLL < −20dB.

(1)

In this target, a 46◦ flat gain region and a 60◦ 3-dB coverageare requested to provide a near consistent gain in its main beamregion and −20 dB SLL constraint is proposed to realize goodinterference rejection in its side lobe region.

To better describe the optimization process, the synthesismethod described above can be expressed numerically as

P (ϕ) =

⎡⎢⎢⎢⎣

Pe2(ϕ)

Pe1(ϕ)

Pe1(ϕ)

Pe2(ϕ)

⎤⎥⎥⎥⎦

T

×

⎡⎢⎢⎢⎢⎣

Aejφ · ej·2π(−d2/2λ) sin(ϕ)

ej·2π(−d1/2λ) sin(ϕ)

ej·2π(d1/2λ) sin(ϕ)

Aejφ · ej·2π(d2/2λ) sin(ϕ)

⎤⎥⎥⎥⎥⎦

(2)

where Pe1(ϕ) and Pe2(ϕ) are the element patterns of the centraland shaping group in angular position ϕ; A and φ representrelative amplitude and phase excitation of the shaping group tothe central group; and d1, d2 are the elements distance in thecentral and shaping group, respectively.

In the optimization process, the d1 is tuning first to findproper original pattern profile. After that, the d2 is adjusted to

2916 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 63, NO. 7, JULY 2015

Fig. 2. Beam pattern and structure comparison between the conventional arrayand the shaped array.

align the lobes of the shaping pattern to the desired impact areaon the original pattern. Finally, the A and φ are optimized, andthe shaping is performed with a proper extend. Fig. 2 showsthe optimized synthesized pattern. It can be observed that theripples on the main beam are minimized, and the 3-dB beamwidth along with the SLL reach 60◦ and −27.9 dB, respectively,which satisfy all three targets as listed in (1).

To demonstrate the effectiveness of the proposed synthesismethod, the pattern of a two-element conventional array, whose3-dB coverage is optimized to 60◦, is also provided in Fig. 2.It can be observed that the fluctuation of the gain in the mainbeam region, the falling speed of the main beam outside the3-dB region, and the SLL have been significantly improved byimplementing the proposed synthesis method.

However, as indicated by the dimension comparison shownin the bottom of Fig. 2, these advantages are achieved at theexpense of a larger array size. If directly arrange the elementsof this shaped array on one array panel and combine six ofthis array panel together to realize 360◦ azimuth coverage, thediameter of the SBA will be quite large as shown in Fig. 3(a).This is undesirable for small-scale SBA applications. Thus,interleaving array architecture is proposed, which can realizesimilar performance with a much compact array panel size.

B. Interleaving Array Architecture

By observing the structure of the four elements shaped arrayshown in Fig. 2, it may be noticed that there is sufficient spaceto fit another element between the central and the shaping ele-ment at both sides of the array panel. Thus, interleaving arrayarchitecture is proposed, which utilizes this space to reduce thesize of the array panel as shown in Fig. 3(b).

In this figure, the elements for specific array, or beam here-after, are noted with the same color. For any beam in the SBA,its central group elements are fixed at the original positionbut its shaping group elements are swapped with the one ofits neighboring beams. Thus, its largely distanced shaping ele-ments are relocated to its adjacent beams’ array panels, and thespace between its central and shaping elements is now utilized

Fig. 3. Interleaving array architecture. (a) Size comparison between the arraypanel and the SBA with/without interleaving architecture. (b) Proposed inter-leaving array architecture.

TABLE IOPTIMIZED ARRAY PARAMETERS OF THE SHAPED ARRAY

Fig. 4. Pattern and structure comparison between the array with/withoutinterleaving.

by the shaping elements of its neighboring beams. As a result,one array panel is now being shared by three beams, and its sizeis reduced from 3.34λ to 2.36λ as shown in Fig. 3(a).

Since the only difference between the shaped arraywith/without interleaving is the orientation of the shapingelements, the synthesis method proposed above is still applica-ble. Table I provides the optimized array parameters of thesetwo shaped arrays, and Fig. 4 shows their synthesized pat-terns. It can be observed that the SLL is increased but its

WANG et al.: SBA WITH SHAPED RADIATION PATTERN AND INTERLEAVING ARRAY ARCHITECTURE 2917

Fig. 5. Schematic diagram of the BFN for one beam.

Fig. 6. Circuit-level implementation of the BFN for one beam.

falling edge is shaper with interleaving, and all the targets areachieved as listed in (1).

C. BFN and Feeding Structure Design

To implement these optimized phase/amplitude distributionprovided in Table I, the BFN is designed and shown in Fig. 5.It is composed of two directional couplers to control the rela-tive amplitude (A) of the shaping group to the central group,and microstrip feeding lines with different electrical lengthsto realize the phase difference (φ) between these two groups.Fig. 6 shows its circuit-level implementation, in which a λ/4impedance transformation line is added to provide 50-Ω inter-face to the array element.

Since interleaving array architecture is applied, a double-layer feeding structure built with two hexagonal-shaped feedingpanels is introduced as shown in Fig. 7. By allocating the BFNof the even number beams on the upper feeding panel and theodd number beams on the lower feeding panel, this double-layer feeding structure can combine the BFN of six beamstogether without using crossover.

III. PATTERN DESIGN IN THE ELEVATION PLANE

In Section II, a flat-top beam pattern is generated with theproposed synthesis method in the azimuth plane. In this section,a high-gain low SLL pattern is achieved in the elevation plane.This pattern is generated by an eight-element series-fed nonuni-form amplitude distributed in-phase array [21], and the SLLis controlled by optimizing the amplitude distribution of thisin-phase array.

Fig. 7. Double-layer feeding structure designed for interleaving arrayarchitecture.

Fig. 8. Structure of the array and its basic element in elevation plane.

Fig. 8 shows the structure of the array, which is also apatch elements-based array. It is fabricated on an 1.5-mm-thick Teflon-based substrate (εr ≈ 2.65 and tan δ ≈ 0.002), inwhich the effective wavelength is λg . An enlarged view of thebasic element is shown in the left top of Fig. 8. It is composedof a λg/2 radiating patch and two λg/4 microstrip transmissionlines. The total equivalent electrical length of this basic elementis close to λg. Thus the in-phase characteristic is guaranteed inthis array near the central resonant frequency.

While resonating, this array can be equivalent to a series cir-cuit shown in the bottom right of Fig. 8. The amplitude of eachbasic element is decided by its resonant impedance and can betuned by changing the width (Wf ) of its two λ/4 microstriptransmission lines. By concatenating the basic elements withdifferent Wf together, a tapered amplitude distribution can beachieved in the array.

Since only the Wf is tuned, the dimensions of the radiationpatch (Lp = 16.5mm and Wp = 20.0mm) are identical for allbasic elements. Thus, minimized effect is applied on the radi-ation pattern. Fig. 9 shows the amplitude pattern versus Wf atthe central resonant frequency (5.22 GHz). It can be observedthat the normalized amplitude tuning range is between 0.67 and1. By optimizing the array’s amplitude distribution within thisrange, −23.2 dB SLL is achieved with a peak gain of 16.6 dB.The simulated pattern is provided in Section IV, and the opti-mized amplitude and the Wf of these eight basic elements arelisted in Table II, in which the order of elements is in accordwith the numbers shown in Fig. 8.

2918 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 63, NO. 7, JULY 2015

Fig. 9. Relationship between the amplitude pattern and the width of themicrostrip transmission line.

TABLE IIOPTIMIZED AMPLITUDE DISTRIBUTION AND THE ARRAY PARAMETERS

Fig. 10. Structure and dimensions of the array panel.

Because the amplitude distribution in this array is symmet-ric, the feeding point can also be mirrored to integrate with thedouble-layer feeding structure as proposed in Section II. Fig. 10shows the array panel by combining the array designs in theazimuth and the elevation plane together. It can be observedthat the feeding points of the central group and shaping groupare symmetrically located on the panel.

With the array panel and the double-layer feeding structureproposed in Section II, the six beams SBA can now be assem-bled as shown in Figs. 11 and 12. The array panel and thefeeding panel are connected via MMCX connectors, and thedouble-layer feeding structure provides a good supporting tothe arrays via these connectors.

IV. ARRAY PROTOTYPE AND THE MEASUREMENT RESULT

To verify the performance of the proposed array, a prototypeis built and measured in this paper. Fig. 13 shows the photos ofthe prototype, and its S-parameters, radiation patterns, and gain

Fig. 11. Connection between the feeding panel and the array panel.

Fig. 12. Explosive view of the proposed SBA.

Fig. 13. Photos of the fabricated prototype. (a) Assembled prototype of theproposed SBA. (b) Array panel. (c) Double layer feeding structure. (d) Feedingpanel.

WANG et al.: SBA WITH SHAPED RADIATION PATTERN AND INTERLEAVING ARRAY ARCHITECTURE 2919

Fig. 14. Simulated and measured S-parameters of the proposed SBA.

are measured and compared with the simulation results in thissection.

Consider that the array is rotational symmetric and similarresults can be observed for all ports, the results of port one areprovided below.

A. S-Parameters

The central radiation frequency of the fabricated prototypeis designed as 5.22 GHz, and the S-parameters are measuredwith Agilent Vector Network Analyzer E5071B. The measuredand simulation return loss and the crossbeam coupling level areprovided in Fig. 14. It can be observed that the measured returnloss matches the simulation result well, and the mutual couplinglevel between adjacent beams is low.

B. Radiation Pattern

The radiation patterns are measured in the ETS anechoicchamber AMS8500. Fig. 15 provides the normalized measuredand simulated patterns in both the azimuth plane (H-plane)and the elevation plane (E-plane). It can be observed that themeasured patterns fit the simulation results well.

In the H-plane, the measured 3-dB beam width is 56◦, andthe stable gain region covers from −20◦ to 22◦. A sharp roll-off can be observed at the edge of the main beam, and the SLLreaches −17.8 dB. In the E-plane, a pencil beam is achievedas expected, and the measured SLL is −20.8 dB. The dif-ference between the simulation and measurement results maydue to fabrication and assembly error or inaccurate phase andamplitude generation by the BFN.

As for the cross polarization component, the simulated valuein E-Plane is not displayed since it is beyond the lower bound,and the difference between the measured and simulated resultsis due to the dynamic range limitation in the measuring system.

C. Gain

The gain is also measured in the ETS anechoic chamber.Fig. 16 shows the comparison between the simulation and

Fig. 15. Simulated and measured radiation patterns of the beam in the proposedSBA. (a) Simulated and measured radiation pattern in E-plane. (b) Simulatedand measured radiation pattern in H-plane.

Fig. 16. Simulated and measured gain of the proposed SBA.

measured results. It can be observed that the peak gain reaches16.4 dB in the measurement, which matches well with thesimulation result near the resonant frequency.

2920 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 63, NO. 7, JULY 2015

V. CONCLUSION

In this paper, a six-beam SBA is proposed and fabricated,whose beam pattern is shaped that can provide near consis-tent gain in the azimuth plane. A synthesis method based ona four-element array is described, and a flat-top beam patternis generated with this method. This flat-top pattern is charac-terized with 60◦ 3-dB beam width, 46◦ stable gain region, andfast falling edge with −20.1 dB SSL, which can be viewed as agood approximation to the rectangular-shaped pattern. By com-bining six of these shaped beams together, this proposed SBAnot only provides full azimuth coverage with least number ofbeams, but also achieves good interference rejection out of theselected beam.

To minimize the size of the SBA and improve the apertureefficiency, interleaving array architecture is proposed, whichshares one array panel with three adjacent beams. As a result,the size of the proposed array panel in the azimuth plane isreduced to 2.36λ, which is comparable to traditional equallyspaced four-element array.

To extend the beam coverage, a nonuniform amplitude dis-tributed series-fed in-phase array is designed and optimized inthe elevation plane. The peak gain reaches 16.6 dB, and the SLLis controlled under −23.2 dB.

REFERENCES

[1] A. Osseiran and A. Logothetis, “Smart antennas in a WCDMA radio net-work system: Modeling and evaluations,” IEEE Trans. Antennas Propag.,vol. 54, no. 11, pp. 3302–3316, Nov. 2006.

[2] A. El-Zooghby, Smart Antenna Engineering. Norwood, MA, USA:Artech House, 2005, pp. 8–11.

[3] S. W. Choi, D. H. Shim, and T. K. Sarkar, “A comparison of tracking-beam arrays and switching-beam arrays operating in a CDMA mobilecommunication channel,” IEEE Antennas Propag. Mag., vol. 41, no. 6,pp. 10–56, Dec. 1999.

[4] M. Maqsood et al., “Low-cost dual-band circularly polarizedswitched-beam array for global navigation satellite system,” IEEETrans. Antennas Propag., vol. 62, no. 4, pp. 1975–1982, Apr.2014.

[5] H. Liu, S. Gao, and T. H. Loh, “Small director array for low-profile smartantennas achieving higher gain,” IEEE Trans. Antennas Propag., vol. 61,no. 1, pp. 162–168, Jan. 2013.

[6] J. Butler and R. Lowe, “Beam-forming matrix simplifies design ofelectronically scanned antennas,” Electron. Des., vol. 9, pp. 170–173,1961.

[7] J. Shelton and K. S. Kelleher, “Multiple beams from linear arrays,”IRE Trans. Antennas Propag., vol. 9, no. 2, pp. 154–161, Mar.1961.

[8] H. Novak, “Switched-beam adaptive antenna system,” Ph.D. disserta-tion, Inst. Telecommunications and High-Frequency Engineering, ViennaUniv. Technology, Vienna, Austria, 1999, pp. 30–33.

[9] S. Mosca, F. Bilotti, A. Toscano, and L. Vegni “A novel design method forBlass matrix beam-forming networks,” IEEE Trans. Antennas Propag.,vol. 50, no. 2, pp. 225–232, Feb. 2002.

[10] M. G. Jansen and R. Prasad, “Capacity, throughput, and delay analysis ofa cellular DS CDMA system with imperfect power control and imperfectsectorization,” IEEE Trans. Veh. Technol., vol. 44, no. 1, pp. 67–75, Feb.1995.

[11] J. H. Yea, “Smart antennas for multiple sectorization in CDMA cellsites,”RF Des. Mag., pp. 28–38, Apr. 2001.

[12] J. E. Evans, “Synthesis of equiripple sector antenna patterns,” IEEETrans. Antennas Propag., vol. 24, no. 3, pp. 347–353, May 1976.

[13] A. Ksienski, “Maximally flat and quasi-smooth sector beams,” IRE Trans.Antennas Propag., vol. 8, no. 5, pp. 476–484, Sep. 1960.

[14] A. Lopez, “Sharp cutoff radiation patterns,” IEEE Trans. AntennasPropag., vol. 27, no. 6, pp. 820–824, Nov. 1979.

[15] A. Sabharwal, D. Avidor, and L. Potter, “Sector beam synthesis for cel-lular systems using phased antenna arrays,” IEEE Trans. Veh. Technol.,vol. 49, no. 5, pp. 1784–1792, Sep. 2000.

[16] B. P. Kumar and G. R. Branner, “Design of unequally spaced arraysfor performance improvement,” IEEE Trans. Antennas Propag., vol. 47,no. 3, pp. 511–523, Mar. 1999.

[17] G. Oliveri and A. Massa, “Bayesian compressive sampling for patternsynthesis with maximally sparse non-uniform linear arrays,” IEEE Trans.Antennas Propag., vol. 59, no. 2, pp. 467–481, Feb. 2011.

[18] S. Yang, Y. Liu, and Q. H. Liu, “Combined strategies based on matrixpencil method and tabu search algorithm to minimize elements of non-uniform antenna array,” Prog. Electromagn. Res. B, vol. 18, pp. 259–277,2009.

[19] Y. H. Liu, Q. H. Liu, and Z. P. Nie, “Reducing the number of elementsin multiple-pattern linear arrays by the extended matrix pencil methods,”IEEE Trans. Antennas Propag., vol. 62, no. 2, pp. 652–660, Feb. 2014.

[20] H. Wang, Z. J. Zhang, and Z. H. Feng, “A beam-switching antenna arraywith shaped radiation patterns,” IEEE Antennas Wireless Propag. Lett.,vol. 11, pp. 818–821, Jun. 2012.

[21] Z. Chen and S. Otto, “A taper optimization for pattern synthesis ofmicrostrip series-fed patch array antennas,” in Proc. IEEE Eur. WirelessTechnol. Conf. (EuWIT), 2009, pp. 160–163.

Han Wang received the B.S. degree in appliedphysics from Beijing University of Posts andTelecommunications, Beijing, China, in 2010. Heis currently pursuing the Ph.D. degree in electricalengineering at Tsinghua University, Beijing, China.

His research interests include antenna design andtheory, particularly in smart antenna, MIMO antenna,and antenna measurement.

Zhijun Zhang (M’00–SM’04–F’15) received theB.S. and M.S. degrees in electronics engineer-ing from the University of Electronic Science andTechnology of China, Chengdu, China, in 1992 and1995, respectively, and the Ph.D. degree in electron-ics engineering from Tsinghua University, Beijing,China, in 1999.

In 1999, he was a Postdoctoral Fellow with theDepartment of Electrical Engineering, University ofUtah, Salt Lake City, UT, USA, where he wasappointed as Research Assistant Professor in 2001.

In May 2002, he was an Assistant Researcher with the University of Hawaii atManoa, Honolulu, HI, USA. In November 2002, he joined Amphenol T&MAntennas, Vernon Hills, IL, USA, as a Senior Staff Antenna DevelopmentEngineer and was then promoted as Antenna Engineer Manager. In 2004, hejoined Nokia Inc., San Diego, CA, USA, as a Senior Antenna Design Engineer.In 2006, he joined Apple Inc., Cupertino, CA, USA, as a Senior Antenna DesignEngineer and was then promoted as Principal Antenna Engineer. Since August2007, he has been a Professor with the Department of Electronic Engineering,Tsinghua University. He is the author of Antenna Design for Mobile Devices(Wiley, 2011).

Dr. Zhang is serving as an Associate Editor of the IEEE TRANSACTIONS

ON ANTENNAS AND PROPAGATION and the IEEE Antennas and WirelessPropagation Letters.

Yue Li (S’11–M’12) received the B.S. degree in com-munication engineering from Zhejiang University,Zhejiang, China, in 2007, and the Ph.D. degree inelectronic science and technology from TsinghuaUniversity, Beijing, China, in 2012.

In 2012, he was a Postdoctoral Fellow withthe Department of Electronic Engineering, TsinghuaUniversity. Since January 2014, he has beena Postdoctoral Fellow with the Department ofElectrical and Systems Engineering, University ofPennsylvania, Philadelphia, PA, USA. He was also

a Visiting Scholar at the Institute for Infocomm Research (I2R), A*STAR,Singapore, and Hawaii Center of Advanced Communication (HCAC),University of Hawaii, Hilo, HI, USA. He has authored and coauthored over50 journal papers, and holds over 10 granted and filed Chinese patents. Hisresearch interests include metamaterials, electromagnetics, and antennas.

Dr. Li is serving as a Reviewer of the IEEE TRANSACTIONS ON ANTENNAS

AND PROPAGATION and the IEEE Antennas and Wireless Propagation Letters.

WANG et al.: SBA WITH SHAPED RADIATION PATTERN AND INTERLEAVING ARRAY ARCHITECTURE 2921

Magdy F. Iskander (F’93–LF’12) is a Professor ofelectrical engineering and the Director of the HawaiiCenter for Advanced Communications (HCAC),College of Engineering, University of Hawaii atManoa, Honolulu, HI, USA. He is Co-Director ofthe NSF Industry/University Cooperative ResearchCenter with four other universities. He joined theUniversity of Hawaii in 2002 and prior to that he wasa Professor of electrical and computer engineeringand the Engineering Clinic Endowed Chair Professorat the University of Utah. He has authored over 250

papers in technical journals, holds nine patents, and has made numerous presen-tations at national/international conferences. He authored/edited several booksincluding the textbook Electromagnetic Fields and Waves (Prentice Hall, 1992,and Waveland Press, 2001; second edition 2012), and four books published bythe Materials Research Society (MRS) on Microwave Processing of Materials.He is the Founding Editor of the Computer Applications in EngineeringEducation (CAE) journal (Wiley, 1992–present). His research in computationaland biomedical electromagnetics and wireless communications is funded bythe National Science Foundation, National Institute of Health, Army ResearchOffice, U.S. Army CERDEC, Office of Naval Research, and several corporatesponsors.

Dr. Iskander was the 2002 President of the IEEE Antennas and PropagationSociety, Distinguished Lecturer, and a Program Director in the Electrical,Communications, and Cyber Systems Division at the National ScienceFoundation. He was the recipient of many awards for excellence in researchand teaching including the University of Hawaii Board of Regents’ Medalfor Excellence in Research (2013), the Board of Regents Medal for TeachingExcellence (2010), and the Hi Chang Chai Outstanding Teaching Award (2011,2014) which is based on votes by graduating seniors. He was also the recipientof the IEEE MTT-S Distinguished Educator Award (2013), IEEE AP-S Chen-To Tai Distinguished Educator Award (2012), and Richard R. Stoddard Awardfrom the IEEE EMC Society in 1992. He received the Northrop GrummanExcellence in Teaching Award in 2010, the American Society for EngineeringEducation (ASEE) Curtis W. McGraw National Research Award in 1985, andin 1991 the ASEE George Westinghouse National Award for Excellence inEducation.