Optimal Antenna Currents For Q, Superdirectivity, and Radiation Patterns Using Convex Optimization
Optimal Antenna Currents For Q, Superdirectivity, and Radiation Patterns Using Convex Optimization
Optimal Antenna Currents For Q, Superdirectivity, and Radiation Patterns Using Convex Optimization
Abstract—The high Q-factor (low bandwidth) and low efficiency Here, it is shown that convex optimization [15], [16] can be
make the design of small antennas challenging. Here, convex op- used as a tool to formulate and solve several fundamental ques-
timization is used to determine current distributions that provide tions for small antennas. The approach offers many opportuni-
upper bounds on the antenna performance. Optimization formula-
tions for maximal gain Q-factor quotient, minimal Q-factor for su- ties to derive new physical bounds on antennas. Here, we e.g.,
perdirectivity, and minimum Q for given far-fields are presented. present results for minimum of superdirective antennas and
The effects of antennas embedded in structures are also discussed. minimum for antennas with a prescribed far field. We also
The results are illustrated for planar geometries. illustrate how antennas embedded in metallic structures can be
Index Terms—Antenna Q, antenna theory, convex optimization, included in the bounds. This generalizes the results in [3]–[10]
physical limitations, small antennas, superdirectivity. to many new and important antenna problems. It also general-
izes the optimization formulation for in [10] to include the
maximum of the stored electric and magnetic energies.
I. INTRODUCTION
The presented results are for arbitrary shaped structures but
restricted to antennas composed of non-magnetic materials.
and , and is
the permeability of free space. The stored magnetic energy is
, where
(6)
are defined analogously. The matrices , and and the corresponding matrix representations for the integral
are real-valued and symmetric. It is observed that and operators in (13), see also Appendix A. This gives the convex
can be indefinite for electrically large structures [10], optimization problem
[14]. In the numerical examples in this paper, we restrict the
electrical size to be approximately less than half a wavelength.
The eigenvalues are also computed to verify that and
are positive semidefinite. Here, it is observed that there can be (14)
a few negative eigenvalues. These negative eigenvalues are
however due to the used finite numerical precision and their rel- This is a convex optimization problem in the form of a linearly
ative amplitude is compared to the positive eigenvalues. constrained quadratic program [15] that e.g., can be solved using
We transform the matrices to become positive semidefinite by [22]. It is also illustrative to use Lagrange multipliers [15],
setting these eigenvalues to zero. [26] to rewrite (14) as the linear system
(12)
(17)
subject to the constraint of zero total charge . Use the basis functions (8) to get the convex optimization
Use that the quotient is invariant for scaling to rewrite problem
the bound (12) as the optimization problem
(18)
(22)
(23)
In [10], it is also shown that the constraint on is relaxed
for combined electric and magnetic dipole antennas: giving , where
and a solution of (23).
(20) We consider currents confined to planar rectangles to illus-
trate the results. The bound on and its corresponding
and for lossless antennas are depicted in Fig. 3 for rectan-
We consider a planar rectangle to illustrate the physical bounds gles with side lengths and . The
on for small antennas, see Fig. 2. The rectangle has side partial gain is evaluated for the polarization and the di-
lengths and with the radius of the rections and . The two optimization formulations
smallest circumscribing sphere. The bound on the electric dipole (22) and (23) give similar results when solved using [22].
is identical to the results in [5], [6], [10] and many small dipole The bound on is normalized with the electrical size
antennas perform close to the bound [6]. The magnetic dipole to simplify comparison with the results in [6], [10], where
case is more constrained. In particular, the bound shows that it is denotes the radius of the smallest circumscribing sphere. The
difficult to utilize the magnetic dipole for elongated structures. low-frequency limit for is given by the polarizability as
The combined mode case (20) offers a substantially improved shown in [10]. It is also observed that is almost in-
performance [10]. dependent of the electrical size for or .
The bounds in Fig. 2 are computed using piecewise linear The case with radiation in the direction offers an
basis functions on rectangular elements. We use a non-equidis- increased . In particular, the case increases
tant mesh for the electric dipole case, where the mesh is con- from 0.29 to 0.63. Here, we note that the bound
structed to have approximately equal charge on each element in [5], [6], [27] is sharper for this case. It is also important to re-
for improved convergence, see also [27]. The magnetic dipole alize that the bound in [5], [6], [27] are for the bandwidth of the
case is computed on an equidistant mesh. We also use the con- antenna and it is not guaranteed that the optimal current distri-
straint to reduce the size of the linear system. butions considered here can be generated from single-port an-
tennas with a half-power fractional bandwidth , see
B. Maximal also [28]. In Fig. 3(a), it is further seen that the improvement
Maximization of for finite sized antennas is formulated for the direction diminishes as decreases. The resulting
as a convex optimization problem. Combine (3) and (4) to ex- current distribution on a coarse mesh is depicted for
press the gain Q-factor quotient as , and . The current is composed of one -di-
rected component and one loop type component, see also Fig. 4.
It is similar for the thinner structures but the loop current gets
weaker. The corresponding currents for the cases are sim-
(21)
ilar to the -directed component. It is also known that the cur-
rent distribution is not unique [10]. There are antenna designs
In [10], the quotient is maximized under the assumption that perform close to the bound for the cases [6]. The
using a Lagrangian formulation. To solve the gen- cases are more involved and there are, to our knowledge,
eral case (21), we rewrite the quotient as a convex op- presently no design that reach these bounds. It is also known
timization problem. We follow [10] and note that is in- that the loop current is associated with low radiation resistance.
variant for multiplicative scaling with arbitrary com- The corresponding partial directivities, , and Q-fac-
plex valued . It is hence sufficient to consider real-valued tors for lossless structures are depicted in Figs. 3(b), (c), respec-
quantities , see (9). Moreover, maximization of tively. Here, it is observed that the directivity differs between
GUSTAFSSON AND NORDEBO: OPTIMAL ANTENNA CURRENTS FOR Q, SUPERDIRECTIVITY, AND RADIATION PATTERNS 1113
(24)
(25)
(29)
(30)
to but not exact dipoles. The radiation patterns are depicted for
where quantifies the deviation of the desired radiation pattern . It is hard to distinguish between the patterns for
and the least-square norm is used for simplicity. It is conve- the cases, but the partial directivity for the projection case
nient to expand the radiated field in spherical vector waves and and is lower than for for the case
rewrite the deviation as in the region around . The radiation patterns differ
more for the case, where again the projection formulation
(26) offers the lowest .
(31)
D. Maximal for
where The Chu bound [3] shows that the radiation is dominated by
dipole modes for small antennas . Consequently, the
directivity is low, i.e., for single mode antennas and
in general bounded as for mixed electric and magnetic
dipole modes. Higher directivity requires higher order modes
that imply a higher , e.g., the of quadrupole modes is propor-
tional to for . It is known to be difficult to design
(32) and utilize high (or super-) directivity for small antennas [2],
[30]–[33]. It is hence interesting to investigate the bound
for antennas with directivities for some .
This gives the optimization problem These bounds give an estimate of the increased Q-factor for su-
perdirective antennas.
The partial directivity (3) is included in the optimization
problem (23) with the constraint giving
(33)
(34)
It is noted that arbitrary weight functions and norms can be used where the factor is due to the normalization of and
in (33). .
A planar rectangle with is considered to illustrate The bounds are illustrated in Fig. 6 for planar rectangles with
the results for given far-fields (26) and (33). The factors are and for the polarization . The constraints
depicted in Fig. 5 for projections (26) and norm bounds (33) for and are considered for
the cases of the -directed electric dipole and -directed and , respectively. These values for are chosen
magnetic dipole patterns, see Fig. 4. as they require excitation of higher order modes. The addition
We observe that the is lower for the electric dipole mode of the constraints reduces for small structures when the
than the magnetic dipole mode. The Q is also lowest for the constraint is active. The resulting partial directivities are seen in
projection cases (26). Moreover, tend to increase as de- Fig. 6(b) together with the mode distribution for .
creases, i.e., the lowest is for radiation patterns that are close The superdirectivity for the cases are due to
GUSTAFSSON AND NORDEBO: OPTIMAL ANTENNA CURRENTS FOR Q, SUPERDIRECTIVITY, AND RADIATION PATTERNS 1115
(35)
where the similarities with (5), (6), (7), and (11) are noted. The
integration in (35) is over the PEC surface of the structure. The
driving sources of the EFIE are confined to the region and
they are unknown. Moreover the EFIE is not necessarily valid
Fig. 6. Upper bounds on the partial gain Q-factor quotient for antennas with in . Decompose the current density as , where
for a planar rectangle with side lengths is the current density in . The EFIE gives two
and . (a) . (b), (c) resulting and for lossless antennas. The rows corresponding to test functions in and . Here, the first
radiation patterns with their mode expansions (48) are depicted for ,
where it is seen that higher order modes appear in addition to for
row is unknown but the second row gives the constraint
the superdirective cases.
(36)
the excitation of electric quadrupole terms. This also explains that can be added as a constraint to the convex optimization
the increased Q-factors. problems in this paper, e.g., the bound in (22).
Fig. 7 illustrates the corresponding results for using It is convenient to use (36) to express the induced current den-
and radiation in the directions with polariza- sities in , i.e., and eliminate
tions for a lossless structure. Here, the cost of in the optimization problem. Decompose the matrices and
superdirectivity is clearly seen. The Q-factor is highest for the according to
case where the symmetry causes the current to radiate
in both the and directions. The mode expansion indicates (37)
that many higher order modes are excited. The cases
have pencil beams and much lower factors. The end-fire case
that gives
has the lowest with for . The re-
sulting current distributions are oscillatory as for superdirective
arrays [2].
APPENDIX A APPENDIX B
POLARIZABILITY MODE EXPANSION
The physical bounds on in [5], [6] are expressed in The radiated electromagnetic field is expanded in
the polarizability of the antenna structure. In [10], it is also spherical vector waves [29] (or modes) outside a circum-
shown that the bound on small antennas (12) can be expressed scribing sphere. The corresponding far field is expanded in
in the polarizability. Here, we further show that the solution of spherical vector harmonics as
the convex optimization problem (14) using (15) is identical to
computing the high-contrast polarizability [10] using Galerkin’s (48)
method [26].
The high-contrast polarizability for the polarization can be
determined from the first moment of the induced normalized giving for the expansion coefficients in (32),
charge density as where is the spatial coordinate, , and the
wavenumber. The multi index for
(41) , and is introduced to simplify
the notation. The index, , is also ordered such that
, see [35].
Here, we keep the notation with complex conjugates on to There are a few alternative definitions of the spherical vector
simplify the comparison with (15), although it is sufficient to waves in the literature [29], [36], [37]. Here, we follow [36] and
consider real valued unit vectors to determine the electrostatic use and as basis functions in the azimuthal co-
polarizability. The charge density is the solution of the integral ordinate. This choice is motivated by the interpretation of the
equation fields related to the first 6 modes as the fields from different
Hertzian dipoles. For , we use spherical vector har-
(42) monics
(49)
where the constant is determined from the constraint of zero
total charge . It turns out that it is convenient
and where denotes the spher-
to set for comparison with (15). Expand the charge
ical harmonics [37]. The modes labeled by (odd ) iden-
density in basis functions , where
tify TE modes (or magnetic -poles) and the terms labeled by
and are matrices, to rewrite (41) as
(even ) correspond to TM modes (or electric -poles).
Moreover, the dipoles corresponding to are directed in
(43)
the -direction, in the -direction, and in the
-direction having the explicit representation
Solving the integral equation (42) with the Galerkin’s method
[26] gives the linear system of equations
(50)
(44)
[6] M. Gustafsson, C. Sohl, and G. Kristensson, “Illustrations of new phys- [29] Spherical Near-Field Antenna Measurements, ser. IEE electro-
ical bounds on linearly polarized antennas,” IEEE Trans. Antennas magnetic waves series, J. E. Hansen, Ed. Stevenage, U.K.: Peter
Propag., vol. 57, no. 5, pp. 1319–1327, May 2009. Peregrinus, 1988, no. 26, ISBN: 0-86341-110-X.
[7] A. D. Yaghjian and H. R. Stuart, “Lower bounds on the Q of electrically [30] D. Margetis, G. Fikioris, J. Myers, and T. Wu, “Highly directive current
small dipole antennas,” IEEE Trans. Antennas Propag., vol. 58, no. 10, distributions: General theory,” Phys. Rev. E, vol. 58, no. 2, p. 2531,
pp. 3114–3121, 2010. 1998.
[8] G. A. E. Vandenbosch, “Simple procedure to derive lower bounds for [31] S. R. Best, E. E. Altshuler, A. D. Yaghjian, J. M. McGinthy, and T. H.
radiation Q of electrically small devices of arbitrary topology,” IEEE O’Donnell, “An impedance-matched 2-element superdirective array,”
Trans. Antennas Propag., vol. 59, no. 6, pp. 2217–2225, 2011. IEEE Antennas Wireless Propag. Lett., vol. 7, pp. 302–305, 2008.
[9] J. Chalas, K. Sertel, and J. L. Volakis, “Computation of the Q limits [32] D. Arceo and C. A. Balanis, “A compact Yagi-Uda antenna with en-
for arbitrary-shaped antennas using characteristic modes,” in Proc. hanced bandwidth,” IEEE Antennas Wireless Propag. Lett., vol. 10, pp.
IEEE Int. Symp. on Antennas and Propagation (APSURSI), 2011, pp. 442–445, 2011.
772–774. [33] O. Kim, S. Pivnenko, and O. Breinbjerg, “Superdirective magnetic
[10] M. Gustafsson, M. Cismasu, and B. L. G. Jonsson, “Physical bounds dipole array as a first-order probe for spherical near-field antenna
and optimal currents on antennas,” IEEE Trans. Antennas Propag., vol. measurements,” IEEE Trans. Antennas Propag., vol. 60, no. 10, pp.
60, no. 6, pp. 2672–2681, 2012. 4670–4676, 2012.
[11] G. A. E. Vandenbosch, “Reactive energies, impedance, and Q factor of [34] M. Cabedo-Fabres, E. Antonino-Daviu, A. Valero-Nogueira, and M.
radiating structures,” IEEE Trans. Antennas Propag., vol. 58, no. 4, pp. Bataller, “The theory of characteristic modes revisited: A contribution
1112–1127, 2010. to the design of antennas for modern applications,” IEEE Antennas
[12] W. Geyi, “A method for the evaluation of small antenna Q,” IEEE Propag. Mag., vol. 49, no. 5, pp. 52–68, 2007.
Trans. Antennas Propag., vol. 51, no. 8, pp. 2124–2129, 2003. [35] M. Gustafsson and S. Nordebo, “Characterization of MIMO antennas
[13] P. Hazdra, M. Capek, and J. Eichler, “Radiation Q-factors of thin-wire using spherical vector waves,” IEEE Trans. Antennas Propag., vol. 54,
dipole arrangements,” IEEE Antennas Wireless Propag. Lett., vol. 10, no. 9, pp. 2679–2682, 2006.
pp. 556–560, 2011. [36] R. F. Harrington, Time Harmonic Electromagnetic Fields. New York,
[14] M. Gustafsson and B. L. G. Jonsson, “Stored electromagnetic energy NY, USA: McGraw-Hill, 1961.
and antenna Q,” 2012 [Online]. Available: http://arxiv.org/abs/1211. [37] A. Boström, G. Kristensson, and S. Ström, “Transformation proper-
5521 ties of plane, spherical and cylindrical scalar and vector wave func-
[15] S. P. Boyd and L. Vandenberghe, Convex Optimization. Cambridge, tions,” in Field Representations and Introduction to Scattering, ser.
U.K.: Cambridge Univ. Press, 2004. Acoustic, Electromagnetic and Elastic Wave Scattering, V. V. Varadan,
[16] M. Gustafsson, M. Cismasu, and S. Nordebo, “Physical bounds on A. Lakhtakia, and V. K. Varadan, Eds. Amsterdam: Elsevier Science,
small antennas as convex optimization problems,” in Proc. IEEE An- 1991, ch. 4, pp. 165–210.
tennas and Propagation Society Int. Symp., 2012, pp. 1–2.
[17] Y. Rahmat-Samii and E. Michielssen, Electromagnetic Optimization
by Genetic Algorithms. New York, NY, USA: Wiley-Interscience,
1999.
[18] O. M. Bucci, G. D’Elia, G. Mazzarella, and G. Panariello, “Antenna
pattern synthesis: A new general approach,” Proc. IEEE, vol. 82, no.
3, pp. 358–371, 1994. Mats Gustafsson (M’05) received the M.Sc. degree
[19] J. C.-E. Sten and E. A. Marengo, “Inverse source problem in the sphe- in engineering physics and the Ph.D. degree in elec-
roidal geometry: Vector formulation,” IEEE Trans. Antennas Propag., tromagnetic theory from Lund University, Sweden,
vol. 56, no. 4, pp. 961–969, 2008. in 1994 and 2000, respectively.
[20] P. Rocca, M. Benedetti, M. Donelli, D. Franceschini, and A. Massa, In 2005, he was appointed was appointed Docent
“Evolutionary optimization as applied to inverse scattering problems,” and, in 2011, Professor of electromagnetic theory
Inverse Prob., vol. 25, p. 123003, 2009. at Lund University, Sweden. He co-founded the
[21] Y. Nesterov and A. Nemirovsky, Interior Point Polynomial Methods company Phase Holographic Imaging AB in 2004.
in Convex Programming. Philadelphia, PA, USA: Society for Indus- His research interests are in scattering and antenna
trial and Applied Mathematics, 1994, vol. 13, Studies in Applied Math- theory and inverse scattering and imaging with
ematics. applications in microwave tomography and digital
[22] M. Grant and S. Boyd, “CVX: Matlab software for disciplined convex holography. He has written over 60 peer reviewed journal papers and over 75
programming, version 1.21,” [Online]. Available: cvxr.com/cvx Apr. conference papers.
2011 Prof. Gustafsson received the best antenna poster prize at EuCAP 2007 and
[23] S. J. Orfanidis, “Electromagnetic waves and antennas,” 2002 [Online]. the IEEE Schelkunoff Transactions Prize Paper Award 2010. He serves as an
Available: www.ece.rutgers.edu/~orfanidi/ewa, revision date Jun. 21, AP-S Distinguished Lecturer for 2013–2015.
2004
[24] IEEE Standard Definitions of Terms for Antennas, IEEE Std 145-1993,
1993, Antenna Standards Committee of the IEEE Antennas and Prop-
agation Society. Sven Nordebo (SM’05) received the M.S. degree
[25] A. F. Peterson, S. L. Ray, and R. Mittra, Computational Methods for in electrical engineering from the Royal Institute
Electromagnetics. New York, NY, USA: IEEE Press, 1998. of Technology, Stockholm, Sweden, in 1989 and
[26] G. Strang, Introduction to Applied Mathematics. Wellesley, MA, the Ph.D. degree in signal processing from Luleå
USA: Wellesley-Cambridge Press, 1986. University of Technology, Luleå, Sweden, in 1995.
[27] M. Gustafsson, “Physical bounds on antennas of arbitrary shape,” in Since 2002, he is a Professor of signal processing
Proc. IEEE Antennas and Propagation Conf. (LAPC), Loughborough, at the School of Computer Science, Physics and
2011, pp. 1–5. Mathematics, Linnæus University. His research
[28] M. Gustafsson and S. Nordebo, “Bandwidth, Q factor, and resonance interests are in statistical signal processing, electro-
models of antennas,” Progr. Electromagn. Res., vol. 62, pp. 1–20, magnetic wave propagation, inverse problems and
2006. imaging.