You are on page 1of 19

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/323174751

Power Density Measurements at 15 GHz for RF EMF Compliance Assessments


of 5G User Equipment

Article  in  IEEE Transactions on Antennas and Propagation · June 2017


DOI: 10.1109/TAP.2017.2712792

CITATIONS READS

31 1,046

7 authors, including:

Bo Xu Kun Zhao
Ericsson Sony Corporation
34 PUBLICATIONS   160 CITATIONS    61 PUBLICATIONS   782 CITATIONS   

SEE PROFILE SEE PROFILE

Björn Thors Davide Colombi


Ericsson Ericsson
28 PUBLICATIONS   497 CITATIONS    30 PUBLICATIONS   378 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

5G antenna system View project

RFID antenna View project

All content following this page was uploaded by Bo Xu on 14 February 2018.

The user has requested enhancement of the downloaded file.


Power Density Measurements at 15 GHz
for RF EMF Compliance Assessments
of 5G User Equipment
Bo Xu, Kun Zhao, Björn Thors, Davide Colombi, Oscar Lundberg,
Zhinong Ying, and Sailing He
April 2017

Abstract
In this paper, different measurement schemes are studied in order to investigate the possibilities
and limitations of scalar and vector-based measurement systems for radio frequency (RF) electromag-
netic fields (EMF) compliance assessments of 5G mobile communication user equipment (UE). Two
UE antenna array designs, transmitting at 15 GHz and employing patch and notch antenna elements,
are considered for different phase excitations. Using free space power density as the exposure metric,
the maximum permissible transmitted power (MPTP) of UE, compliant with the maximum permissi-
ble exposure (MPE) limits specified by the U.S. Federal Communications Commission (FCC) and the
basic restrictions of the International Commission on Non-Ionizing Radiation Protection (ICNIRP),
are determined. The accuracy of different measurement schemes is assessed using numerical simula-
tion. Verifying measurements are carried out in a semi-anechoic chamber. The results indicate that,
for UE employing array antennas and intended to be used in immediate vicinity of the human body,
scalar measurement systems used in combination with straightforward field combination techniques
will lead to overly-conservative results. A more accurate and less conservative approach for these
products is to conduct separate measurements for different excitations in order to span the space
of possible excitations. This will result in a more complicated measurement set-up and increase the
measurement time, which points to a need for very fast measurement systems.
Keywords: 5G mobile communication systems, antenna array, assessment method, RF EMF expo-
sure, maximum permissible transmitted power (MPTP), user equipment (UE), power density.

1 Introduction
Although the fourth generation (4G) cellular mobile systems are still being rolled out in many countries,
the telecommunication industry has already begun working on the fifth generation system (5G). While
many issues remain to be addressed, industry and academia have reached a consensus on a set of rough
requirements for a wide range of different applications. The most noteworthy requirements include data
rates of up to 10 Gb/s and latencies below 1 ms [1, 2]. Since the radio spectrum below 6 GHz is rather
crowded, one possibility to reach these data rates and low latency for 5G mobile communication systems
is to utilize frequency bands up to and including the millimeter-wave (mmWave) spectrum, which allows
for a much larger allocated signal bandwidth [3, 4]. One of the possible frequency band candidates for
This paper is published on IEEE Trans. Antennas Propag., vol. 65, no. 12, pp. 6584–6595, Dec. 2017. This is an author
produced preprint version. The final published paper can be found in http://ieeexplore.ieee.org/document/7942093/.
This work was supported in part by Swedish VR, AOARD, Fundamental Research Funds for Central Universities,
the Program of Zhejiang Leading Team of Science and Technology Innovation (No. 2010R50007), and China Scholarship
Council (CSC) under the Grant No. 201506320137.
B. Xu, K. Zhao, and S. He are with the Department of Electromagnetic Engineering, School of Electrical Engineering,
KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden (e-mail: sailing@kth.se). B. Xu and S. He are also with
Centre for Optical and Electromagnetic Research, Zhejiang Provincial Key Laboratory for Sensing Technologies, College of
Optical Science and Engineering, Zhejiang University, 310058 Hangzhou, Zhejiang, China.
B. Thors, D. Colombi, and O. Lundberg are with Ericsson Research, Ericsson AB, SE-164 80 Stockholm, Sweden. O.
Lundberg is also with Department of Engineering Sciences and Mathematics, Luleå University of Technology, SE-971 87
Luleå, Sweden.
Z. Ying is with Network Technology Lab, Research and Technology, Sony Mobile Communications AB, SE-221 88 Lund,
Sweden.

1
5G lies in the 15 GHz region [5]. The free space path loss is much higher in the mmWave band compared
with existing cellular communication bands (700 MHz–6 GHz). For example, the free space path loss at
15 GHz is 17.5 dB higher than at 2 GHz, according to the Friis transmission equation. To alleviate these
problems and to improve the link budget, beam-steering phased arrays, enabled by the lower required
physical footprint, can be used in user equipment (UE) [2, 6].
Products emitting radio frequency (RF) electromagnetic fields (EMF) need to comply with relevant
regulatory requirements on RF EMF exposure of humans. In most countries, the RF EMF exposure
guidelines specified either by the International Commission on Non-Ionizing Radiation Protection (IC-
NIRP) [7] or by the U.S. Federal Communication Commission (FCC) [8] are applicable. Another set
of RF EMF exposure limits is specified by the IEEE [9, 10], but this has not yet been adopted in any
national regulations. The restrictions on RF EMF exposure determine the maximum permissible trans-
mitted power (MPTP) of UE and constitute an important boundary condition for the design of the mobile
communication systems [11]. For most current wireless communication systems operating below certain
transition frequencies (3 GHz for IEEE, 10 GHz for ICNIRP, and 6 GHz for FCC), the basic restrictions
on RF EMF exposure are specified in terms of Specific Absorption Rate (SAR), which is defined as the
power absorbed per mass of human tissue (W/kg) [7–10]. A large number of studies on SAR for UE
transmitting in existing cellular bands are available (see e.g. [12–20]). At frequencies above the transition
frequencies mentioned above, the restriction metric changes from SAR to free space power density. This
is because the penetration depth of the field into tissues is small at higher frequencies, which makes
energy absorption difficult to measure. Instead, incident power density is a more appropriate dosimetric
quantity [7].
Above 10 GHz, ICNIRP specifies 10 W/m2 as the basic restriction for general public exposure, taken
as an average over any 20 cm2 of exposed area. Furthermore, the spatial maximum power density averaged
over any 1 cm2 shall not exceed 200 W/m2 . Above 6 GHz, the FCC stipulates a maximum permissible
exposure (MPE) limit of 10 W/m2 for uncontrolled environments (general population) taken as a spatial
peak value. However, the spatial peak power density is not a well-defined quantity, and the obtained
results will depend on the exposure assessment method [21]. Recently, in a Notice of Proposed Rulemaking
(FCC 15-138) [22], the FCC proposed the MPE limits for an averaging area of 1 cm2 above 6 GHz in
order to be consistent with one gram averaging of SAR. This interpretation of the spatial peak power
density has not yet formally made its way into the FCC regulations, but as it is likely to be adopted, it
is considered in this paper. Reference to this interpretation of the spatial peak power density is made
using the term “proposed FCC limit” in the present paper.
Studies on RF EMF exposure for phased arrays in UE operating above 10 GHz are available in the
literature. Colombi et al. [11] studied the output power levels to comply with the MPE limits for 5G
devices above 6 GHz. It is shown that the current regulatory requirements on RF EMF exposure lead
to a non-physical discontinuity of several dB in MPTP because of the transition from SAR to power-
density-based basic restrictions. Thus, the MPTP for UE operating above 6 GHz may need to be several
dB below the power levels used for current cellular technologies. Zhao et al. [23, 24] studied RF EMF
exposure of phased arrays in handset devices at 15 GHz and 28 GHz with respect to the current FCC
limits. Thors et al. [21] presented a systematic study of the MPTP and maximum effective isotropic
radiated power (EIRP) to comply with the MPE limits specified by the ICNIRP, FCC, and IEEE for
canonical dipole arrays, considering the effects of frequency, array size, distance to the human body, scan
range, and array topology.
To define accurate power density measurements for the purpose of assessing RF EMF exposure of
future 5G UE intended to be used in the immediate vicinity of the human body is a challenge, and not
much information can be found in the literature. In [25], near-field test practices for the measurement of
antenna properties are provided. Conventional exposure measurement systems often assess the electric
field magnitude point-by-point using a movable single probe, from which a plane-wave equivalent power
density is determined in the assessment domain. The accuracy of this approach may decrease with
decreasing separation distance, possibly requiring additional knowledge of the magnetic field to obtain an
accurate assessment of the power density. Furthermore, RF EMF compliance assessments of phased arrays
require that compliance with the MPE limits is demonstrated for all possible excitations. Straightforward
measurements for all possible excitations using single probe systems are most likely not feasible. This
problem may be addressed either by using fast probe-array systems or by utilizing conservative field
combining methods based on embedded power density measurements. In this paper, the accuracy of
different measurement approaches and conservative field combining methods is investigated for two UE
phased array antenna mock-ups transmitting at 15 GHz.
This paper is organized as follows. Section II briefly introduces the two UE mock-ups using an 8 × 1

2
patch array and an 8×1 notch array. Section III presents the exposure metrics and defines the assessment
schemes considered. Section IV presents the simulation environment and the compliance measurement
scheme. The results from the numerical simulations and verifying measurements are presented in Section
V. Finally, a discussion and conclusions are provided in Sections VI and VII, respectively.

2 Antenna Configuration
In this paper, a uniform linear patch array (see Figure 1) and a uniform linear notch array [6, 26] (see
Figure 2) with 8 × 1 elements printed on Rogers RO4003 substrate (r = 3.55) are designed for the RF
EMF exposure study. The radiation pattern of the notch element is quite omnidirectional, while the
patch element is more directive. They represent two typical antenna solutions and thus provide a good
basis for comparison in the study.
Both antenna arrays make use of a three-layer structure mounted on a 140 mm × 78 mm metal plate
(phone chassis). The thickness of the substrates is 0.508 mm (between Layer 1 and 2) and 0.305 mm
(between Layer 2 and 3) (see Figure 1(a) and Figure 2(a)). The patch elements are fed by aperture
coupling (see Figure 1(b)–(d)), whereas the notch elements are coupled-fed through bended microstrip
lines (see Figure 2(b)–(d)). To reduce mutual coupling, parasitic notches are cut between adjacent
radiating notches (see Figure 2(c)). Both antenna arrays operate at 15 GHz with an element separation
distance of slightly less than half a wavelength. A plastic box (163 mm × 82 mm × 5 mm, r = 3) is
used to simulate the phone casing. The distance between the antenna array and the phone casing is less
than 1 mm for both arrays. The key parameters of the two array configurations are listed in Table 1.
For simplicity, both arrays are excited uniformly in amplitude and fed with a progressive phase shift ϕi
between adjacent antenna elements (see Figure 3).

3 Exposure Metric and Assessment Scheme Definitions


RF EMF exposure is considered in terms of spatial peak, 1 cm2 spatially-averaged, and 20 cm2 spatially-
averaged incident power densities, corresponding to the current FCC limits, the proposed FCC limits,
and the ICNIRP limits, respectively. The power densities should normally be time-averaged over 30
minutes and 68/f 1.05 minutes1 for the FCC and ICNIRP guidelines, respectively. The output power,
and thereby the RF EMF exposure, of a UE in a real operation is usually much less than the maximum
possible, see e.g. [27–30]. This is a consequence of a time-varying transmission due to factors such as
the traffic variation, power control mechanisms, discontinuous transmission (DTX) and soft handover.
In this paper, a time-invariant transmission is assumed because RF EMF compliance testing is normally
required for maximum power configurations. For simplicity, all spatial averaging was conducted using
square-shaped averaging areas.

3.1 Exposure Metrics


The time-averaged power density is defined as [31]
1
S (x, y, z) = Re [E (x, y, z) × H ∗ (x, y, z)] , (1)
2
in which S, E, and H ∗ denote the time-averaged power density (Poynting vector), the electric field, and
the complex-conjugated magnetic field, respectively. The employed coordinate system is illustrated in
Figure 1 and Figure 2 for the two antenna arrays considered.
To study compliance with a spatially-averaged power density limit in the plane z = d, the maximum
power density averaged over any square-shaped area, Aav , in this plane, can be expressed as
 ZZ 
1
S av (z = d) = max S (x, y, d) · ẑ dx dy , (2)
x,y Aav Aav

where the scalar multiplication is made to obtain the total flux of power that passes through the averaging
area. The maximum spatial-peak power density, S p (z = d), is defined analogously by letting Aav → 0.
Note, in order to state compliance with the exposure limits at a distance d, compliance must also be
1 f is the operating frequency in GHz according to the ICNIRP guidelines. For example, “68/f 1.05 minutes” is about 4

minutes when the frequency is 15 GHz (i.e. f = 15).

3
Figure 1: UE mock-up using patch array. (a) Side view. (b) Layer 1 with chassis and patch array. (c)
Layer 2 with slot apertures. (d) Layer 3 with microstrip lines.

Figure 2: UE mock-up using notch array. (a) Side view. (b) Layer 1 with chassis. (c) Layer 2 with notch
array. (d) Layer 3 with microstrip lines.

Table 1: Parameters of Antenna Array Configurations (Unit: mm)


lp wp dp gn dn ln
5.7 4.2 9.75 1 2 5.25

ensured for all distances larger than or equal to d. Especially for array antennas, where the maximum
exposure may occur some distance away from the array face due to the focusing of energy [21], this is
important to consider. In this work, when MPTP is determined for a specific distance d, it is understood
that this implies compliance for all distances greater than or equal to d.

3.2 Power Density Assessment Methods


Traditional laboratory measurements of power density using robot-controlled single probes are compli-
cated and time-consuming. To determine the power density in (1) and (2), knowledge of both the electric

4
Figure 3: Phased array power distribution network illustrating the progressive phase shift utilized in the
beam-steering application.

field and magnetic field is required. The electric field is usually measured through dipole probes, and
the magnetic field is usually measured through small loop probes [32]. Measuring both fields, however,
would add complexity to the measurement set-up and increase the measurement time. Fast and accurate
assessment methods are required to simplify the measurement set-up and procedures. As mentioned
before, some systems are scalar in nature and only provide the magnitude of the electric field, from which
a plane-wave equivalent power density can be determined. Others may provide both the amplitude and
phase of the electric field, from which the corresponding magnetic field can be deduced by the principle
of modal expansion when the scanning plane is much larger than the radiating aperture [32, 33]. This
would potentially allow for more accurate assessments of power density.
If a mono-axial probe is used, three separate measurements, where the sensor axis is set up along three
mutually orthogonal directions, are required to completely characterize the field strength distribution.
In many measurement systems, however, isotropic (tri-axial) probes are used to simultaneously measure
the fields of three perpendicular polarizations. Some systems are based on the characterization of the
tangential field components over some surface. For these systems, a probe with two orthogonal sensors
is suitable.
In this part, approximate methods to assess the power density for a specific excitation ϕi are de-
scribed, corresponding to different measurement system capabilities. This is in contrast with Section
III.C, in which methods for compliance assessments are discussed, i.e. where the maximum exposure for
all possible excitations is of interest.

3.2.1 Power density evaluations based on amplitude-only measurements

For a measurement system which provides the magnitude of the electric field only, the time-averaged
plane-wave equivalent power density may be determined as
1 2
SPW (x, y, z) = |E (x, y, z)| , (3)
2η0
in which η0 denotes the impedance of free space. Equation (3) gives an accurate estimation of the power
density magnitude in the far-field region. For small distances, the near field detected is dominated by
the closest antenna element since the other elements are located much further away. Even in the reactive
near field, the Poynting vector still represents the directional energy flux density and its direction gives
the local direction of wave propagation. Given the definition of power density in (2), where the real
part of the Poynting vector is projected on the normal to the assessment plane, the maximum point-wise
exposure will be obtained in front of the antenna element where the energy flow is perpendicular to
this plane. For array antennas, a periodic power density pattern will be obtained over the assessment
plane where, at least in the regions of maximum exposure, the wave propagation will be essentially
perpendicular to the array face. For this case, it is reasonable to approximate the power density based on
the tangential components of the electric field, Et , since this corresponds to a power flow perpendicular
to the assessment plane.
Thus, for a measurement system which provides the magnitude of the electric field, the spatially-
averaged power density for comparison against the corresponding limit may be approximated using either
of the following expressions
av
SPW,mag (z = d) =
(4)
 ZZ 
1 2
max |E (x, y, d)| dx dy ,
x,y 2η0 Aav Aav

5
or
av
SPW,tan (z = d) =
(5)
 ZZ 
1 2
max |Et (x, y, d)| dx dy ,
x,y 2η0 Aav Aav

where the accuracy of the approximations will depend on the antenna type, frequency, and distance from
the antenna array.
Although no plane-wave equivalent assessment method can precisely determine the power density
distribution in the near field, our target is to assess the maximum spatial-peak and spatially-averaged
power densities for comparison with the corresponding limit values. As such, the achievable accuracy of
the different approximations is of interest to investigate.

3.2.2 Power density evaluations based on amplitude and phase measurements

For a measurement system able to measure both the magnitude and phase of the electric field of
each antenna element, an embedded measurement approach may be utilized where each port is measured
separately to allow for accurate field combining in a post-processing step [34]. While still making use
of the electric field only, the spatially-averaged power density for comparison against the corresponding
limit may be written as
av
SPW,compl (z = d) =
 N 2 
(6)
ZZ
1 X
max  E i (x, y, d) e−jϕi dx dy  ,

2η0 Aav

x,y Aav i=1

where E i denotes the electric field produced by the ith antenna element, and ϕi denotes the phase
excitation of the ith port. The combined power density SPW,compl av
is a true vector summation and
requires that a proper phase calibration technique is employed for the UE. Below, the approximation in
(6) is referred to as the Complex Method for assessing plane-wave equivalent power density.
As mentioned before, another possibility with a measurement system providing both the magnitude
and the phase of the electric field is to retrieve the magnetic field [33] and calculate the true power density
according to (2).

3.3 Compliance Assessment Methods


An RF EMF compliance assessment shall provide evidence that the tested device is compliant for all
possible transmission modes and excitations. For the beam-steering devices considered in this paper, this
implies that compliance with exposure limits needs to be demonstrated for all possible phase excitations.
With knowledge of the transmitted field from each antenna, conservative field combining methods
have been developed where the exposure is maximized by assuming that the fields are totally correlated
at all assessment points [35]. These methods, denoted the Components field combining method (CFCM)
and the Magnitude field combining method (MFCM), respectively, are suitable for use in combination
with scalar measurement systems and have been used to assess exposure from MIMO-enabled networks
and terminals (see e.g. [34, 36]). Descriptions of the CFCM and MFCM are provided in Section III.C.1
and III.C.2, below.

3.3.1 Components Field Combining Method

With the assumption that the fields are temporally “in phase”, a conservative estimate of the spatially-
averaged power density is obtained by summing the magnitude of the field components according to:
av
SPW,CFCM (z = d) =

N
!2 
(7)
ZZ
1 X X i
max  Eτ (x, y, d) dx dy  ,
x,y 2η0 Aav Aav τ =x,y,z i=1

6
in which Eτi is the x, y, or z component of the electric field produced by the ith antenna element. Note
that CFCM is not invariant under coordinate transformations [36].

3.3.2 Magnitude Field Combining Method

With the assumption that the fields are both spatially and temporally “in phase”, a conservative
estimate of the spatially-averaged power density is obtained by summing the magnitude of the fields
according to:
av
SPW,MFCM (z = d) =

N
!2 
(8)
ZZ
1 X i
max  E (x, y, d) dx dy  .
x,y 2η0 Aav Aav i=1
av av
The MFCM is more conservative than the CFCM, i.e. SPW,MFCM ≥ SPW,CFCM , since the electric fields
excited by the individual antenna elements are not only temporally in phase, but also are assumed to
have the same polarization.

3.4 Maximum Permissible Transmitted Power


With the simulation results normalized to a total stimulated power of Ps = 100 mW (20 dBm) and an
RF EMF exposure limit of Slim = 10 W/m2 , the MPTP in dBm is determined according to
 
P s Slim
M P T P (d) = 10 log10  , (9)
max S av/p (d)
ϕi

av/p
in which max S (d) denotes the maximum obtained spatially-averaged or spatial-peak power density
ϕi
in a plane located a distance d above the UE for any possible excitation.

4 Exposure Assessment Methods


Numerical simulations were carried out using the time-domain module of the commercial electromagnetic
solver CST Microwave Studio (CST AG, Germany) based on the Finite Integration Technique (FIT) [37].
CAD models of the antenna arrays presented in Section II were developed and subsequently meshed using
a minimum resolution of at least 10 mesh cells per wavelength. In the vicinity of conducting edges the
mesh was further refined with a factor of 6. A description of the sampling scheme used to obtain the
spatially averaged power density is given in the Appendix.
Two 5G UE mock-ups with notch and patch array antennas were fabricated, as shown in Figure 4.
Measurements of electric field strength magnitude were conducted in a semi-anechoic chamber using a
DASY5 near-field measurement system together with an isotropic electric field probe EF3DV3 (SPEAG,
Switzerland), see Figure 5. The probe calibration was conducted using the free-space standard-field
method [38].
For the notch array, the scanned volume was 180 mm × 110 mm × 100 mm with a grid separation
of 5 mm in the horizontal plane, corresponding to λ/4 at 15 GHz, where λ denotes the wavelength.
The grid separation in the vertical direction was 10 mm. The scanned volume over the patch array was
180 mm×120 mm×100 mm with the same sampling distance as for the notch antenna in the horizontal and
vertical planes. The ports were excited and measured individually, resulting in one set of measurements for
each port. Non-excited ports were terminated in matched loads as indicated in Figure 4. The stimulated
power was scaled to a total of 20 dBm based on power measurements conducted with a Keysight N8488A
power meter and a Pulsar Microwave CS20-55-435/4 directional coupler.

5 Results
5.1 UE characteristics
A uniform amplitude excitation was used for the antenna elements. The actual transmitted power of
the ports, however, was found to be slightly different due to mutual coupling effects. All power density

7
results in this paper are normalized to a total stimulated power of 20 dBm, on par with the maximum
transmitted power level of Bluetooth class 1 devices but lower than the maximum transmit power level
for existing 4G UE (23–24 dBm).
Initially, all ports are excited with the same amplitude and phase, resulting in main lobes directed in
the +z- and +x-directions for the patch and notch arrays, respectively, as shown in Figure 6. In the near
field, power density magnitude distributions in planes located d = 5 mm above the UEs in the +z- and
+x-directions are shown in Figure 7. From Figure 7(a) and Figure 7(c), it is evident that the exposure
of the patch array in the +x-direction is quite small. For the notch array, however, the power density
magnitude is similar between the two planes, as shown in Figure 7(b) and Figure 7(d). In order to not

Figure 4: Fabricated 5G UE mock-ups with front covers removed. (Left) Notch array. (Right) Patch
array. The non-excited ports were terminated in matched loads during the measurements.

(a) (b)

Figure 5: Near-field measurement set-up. (a) Notch array, measurement in yz-plane. (b) Patch array,
measurement in xy-plane.

(a) (b)

Figure 6: 3D far-field patterns when ϕi = 0◦ . (a) Patch array. (b) Notch array.

8
Figure 7: Power density magnitude distributions in the near field in different planes located d = 5 mm
above UE in the +z- and +x-directions when ϕi = 0◦ . The white lines indicate the position of UE. (a)
Patch array, xy-plane. (b) Notch array, xy-plane. (c) Patch array, yz-plane. (d) Notch array, yz-plane.

15
= 0°
i

10 = 30°
i
Realized Gain (dBi)

= 60°
i
5
= 90°
i

0 = 120°
i

= 150°
i
-5
= 180°
i

-10

-15
-90 -60 -30 0 30 60 90
Scan Angle (degree)

(a)

15
= 0°
i

10 = 30°
i
Realized Gain (dBi)

= 60°
i
5
= 90°
i

0 = 120°
i

= 150°
i
-5
= 180°
i

-10

-15
-90 -60 -30 0 30 60 90
Scan Angle (degree)

(b)

Figure 8: Radiation patterns for different progressive phase shifts ϕi = 0◦ , 30◦ , 60◦ , 90◦ , 120◦ , 150◦ , and
180◦ with uniform amplitude. (a) Patch array, yz-plane. (b) Notch array, xy-plane.

9
4 4
1 0 p
1 0 p
S P W .C F C M S im . M e a s. S P W .M F C M S im . M e a s.

1 c m 2 1 c m 2
S P W .C F C M S im . M e a s. S P W .M F C M S im . M e a s.

)
3 3
1 0 1 0
2

2
P o w e r D e n s ity (W /m

P o w e r D e n s ity (W /m
2 0 c m 2 2 0 c m 2
S P W .C F C M S im . M e a s. S P W .M F C M S im . M e a s.
2 2
1 0 1 0

1 1
1 0 1 0

0 P a tc h A rra y 0 P a tc h A rra y
1 0 1 0
0 2 0 4 0 6 0 8 0 1 0 0 0 2 0 4 0 6 0 8 0 1 0 0
d (m m ) d (m m )
(a ) (b )
4 4
1 0 p
1 0 p
S P W .C F C M S im . M e a s. S P W .M F C M S im . M e a s.
1 c m 2 1 c m 2
S P W .C F C M S im . M e a s. S P W .M F C M S im . M e a s.
)

)
3 3
1 0 1 0
2

2
P o w e r D e n s ity (W /m

P o w e r D e n s ity (W /m
2 0 c m 2 2 0 c m 2
S P W .C F C M S im . M e a s. S P W .M F C M S im . M e a s.
2 2
1 0 1 0

1 1
1 0 1 0

0 N o tc h A rra y 0 N o tc h A rra y
1 0 1 0
0 2 0 4 0 6 0 8 0 1 0 0 0 2 0 4 0 6 0 8 0 1 0 0
d (m m ) d (m m )
(c ) (d )

Figure 9: Simulated and measured results as function of d above the arrays. (a) CFCM, patch array. (b)
MFCM, patch array. (c) CFCM, notch array. (d) MFCM, notch array.

clutter the paper with too many results, a choice was made to limit the RF EMF exposure assessments
presented here to the +z- and +x-directions for the patch and notch arrays, respectively.
Figure 8 shows the far-field patterns of both arrays when the progressive phase-shift angle ϕi (see
Figure 3) changes from 0◦ to 180◦ , resulting in the main beam scanning from 0◦ (in the +z-direction for
the patch array, in the +x-direction for the notch array) to −80◦ (in the −y-direction for both arrays).
As expected, the realized gain drops quickly when ϕi ≥ 150◦ , as the main beam is steered towards very
large scan angles. Generally, the realized gain of the patch array is several dB higher than that of the
notch array because the patch element is more directive than the notch element.

5.2 Verifying measurement results


A comparison between simulated and measured results is provided in Figure 9 for the two array antennas
using both the CFCM and MFCM. The overall agreement between simulations and measurements is very
good. A slight difference between the CFCM and MFCM is observed.

5.3 Plane-wave equivalent power density results


Figure 10 compares the true power density in (2) with the estimated plane-wave equivalent power density
(4)–(6) in terms of the spatial peak and spatially-averaged exposure metrics of FCC and ICNIRP as
function of the distance when ϕi = 0◦ . Furthermore, the corresponding power density values at d = 5 mm
are listed in Table 2. The true power density is found to agree quite well with all estimations and for all
exposure metrics when d > 20 mm. When d < 20 mm, the accuracy of the approximate methods when
used to determine the spatially averaged power density over 20 cm2 (ICNIRP) is reduced. For the spatial
peak power density and spatially averaged power density over 1 cm2 , a quite good accuracy is obtained
also below 20 mm. Although plane-wave equivalent power density cannot be used to estimate the true
power density when d is small and their spatial peak values may not appear at the same position, their
spatial peak values can still be very close to each other (see Figure 10(a) and (d)). As the averaging
area increases, the discrepancy between the plane-wave equivalent power density and the true power
density cannot be neglected, and these errors are essentially accumulated point by point. Thus, as d

10
103 103
S
2

S
p 1cm
Current FCC Proposed FCC

S
2

S
p 1cm
Patch Array Patch Array
Power Density (W/m )

Power Density (W/m )


2

2
PW.mag PW.mag

102 S
p
102 2

PW.tan
S
1cm

PW.tan

S
p 2

PW.compl S
1cm

PW.compl

101 101

100 100
0 20 40 60 80 100 0 20 40 60 80 100
d (mm) d (mm)
(a) (b)

3 3
10 10
S
2

S
20cm p
ICNIRP Current FCC

S
2

S
20cm p
Patch Array Notch Array
Power Density (W/m )

Power Density (W/m )


2

PW.mag PW.mag

102 2
102 S
S
p
20cm

PW.tan
PW.tan

S
p
2

S
20cm

PW.compl
PW.compl

101 101

100 100
0 20 40 60 80 100 0 20 40 60 80 100
d (mm) d (mm)
(c) (d)

3 3
10 2 10 2

Proposed FCC S
1cm
ICNIRP S
20cm

2 2

Notch Array
S
1cm
Notch Array
S
20cm
Power Density (W/m )

Power Density (W/m )


2

PW.mag PW.mag

102 2
102 2

S
1cm

PW.tan
S
20cm

PW.tan

2 2

S
1cm

PW.compl
S
20cm

PW.compl

101 101

100 100
0 20 40 60 80 100 0 20 40 60 80 100
d (mm) d (mm)
(e) (f)

Figure 10: True power density compared with various plane-wave equivalent approximations when ϕi =
0◦ . (a) Patch array, current FCC. (b) Patch array, proposed FCC. (c) Patch array, ICNIRP. (d) Notch
array, current FCC. (e) Notch array, proposed FCC. (f) Notch array, ICNIRP

11
25 25
True Power Density True Power Density

20 CFCM MFCM
20 CFCM MFCM
MPTP (dBm)

MPTP (dBm)
15 15

10 10

5 5
Current FCC, Patch Array Proposed FCC, Patch Array
0 0
0 30 60 90 120 150 180 0 30 60 90 120 150 180
(degree) (degree)
i i

(a) (b)

25 25
True Power Density True Power Density

20 CFCM MFCM
20 CFCM MFCM
MPTP (dBm)

MPTP (dBm)

15 15

10 10

5 5
ICNIRP, Patch Array Current FCC, Notch Array
0 0
0 30 60 90 120 150 180 0 30 60 90 120 150 180
(degree) (degree)
i i

(c) (d)

25 25
True Power Density True Power Density

20 CFCM MFCM
20 CFCM MFCM
MPTP (dBm)

MPTP (dBm)

15 15

10 10

5 5
Proposed FCC, Notch Array ICNIRP, Notch Array
0 0
0 30 60 90 120 150 180 0 30 60 90 120 150 180
(degree) (degree)
i i

(e) (f)

Figure 11: MPTP to comply with the FCC and ICNIRP exposure metrics at a distance of 5 mm from
UE as function of ϕi . (a) Patch array, current FCC. (b) Patch array, proposed FCC. (c) Patch array,
ICNIRP. (d) Notch array, current FCC. (e) Notch array, proposed FCC. (f) Notch array, ICNIRP.

12
Table 2: Power Density as Obtained Using Different Approximate Methods for a Total Transmitted Power of 20 dBm at a Distance of 5 mm Away From UE
Current FCC Proposed FCC ICNIRP
Assessment Methods Power Density Power Density Power Density
Relative Error Relative Error Relative Error
( W/m2 ) ( W/m2 ) ( W/m2 )
True S (ref.) 81.6 - 59.2 - 19.5 -
SPW,mag 82.0 0.5% 68.6 15.9% 29.4 50.8%

13
Patch Array
SPW,tan 81.8 0.3% 58.6 1.0% 16.2 16.9%
SPW,compl 82.0 0.5% 68.6 15.9% 29.4 50.8%
True S (ref.) 56.8 - 48.2 - 20.9 -
SPW,mag 55.3 2.6% 49.4 2.4% 27.8 33%
Notch Array
SPW,tan 54.1 4.8% 48.4 0.4% 27.5 32%
SPW,compl 55.3 2.6% 49.4 2.4% 27.8 33%
decreases, the discrepancy between the true S av and different kinds of SPW
av
becomes large, especially for
2
Aav = 20 cm (see Figure 10(c) and (f)).

5.4 RF EMF Compliance Assessment Schemes


To compare the compliance assessment schemes (7) and (8) with the true power density in (2), the
corresponding MPTP at d = 5 mm is shown in Figure 11 and Table 3 for the FCC and ICNIRP exposure
metrics. The level of conservativeness for the CFCM and MFCM is found to range between 5–10 dB,
where the largest deviation is found for the ICNIRP exposure metric due to the larger averaging area.

Table 3: MPTP and Its Conservative Estimation Complying With Different Limits at Distance of 5 mm
(Unit: dBm)
Assessment Current Proposed
ICNIRP
Methods FCC FCC
True S 8.8 11.6 16.5
Patch Array SCFCM 3.6 4.3 7.0
SMFCM 3.6 4.3 7.0
True S 10.8 12.7 17.0
Notch Array SCFCM 4.4 5.2 7.7
SMFCM 4.0 4.5 7.0

6 Discussions
As shown in Figure 11, RF EMF exposure will vary with the array excitation. As a consequence,
compliance with RF EMF exposure limits needs to be verified for all possible excitations. In this paper,
the possibilities and restrictions of different power density measurement systems, for the purpose of
demonstrating compliance, have been investigated.
The simplest measurement approach is to make use of a scalar measurement system with an isotropic
electric field probe. By measuring the electric field excited by each individual port, a conservative
estimate of the maximum exposure for all possible phase excitations may be obtained by utilizing either
the CFCM or MFCM. Despite its attractive simplicity, this approach suffers from severe drawbacks. The
accuracy of CFCM and MFCM is reduced for small test separation distances. Furthermore, the level of
conservativeness observed in Figure 11 for the investigated antennas was found to range between 5–10 dB.
This will give a lower maximum transmitted power, which will severely limit the applicability of these
approaches for RF EMF compliance assessment of devices with array antennas used in the immediate
vicinity of the body.
The conservativeness problem of CFCM and MFCM may be alleviated by first conducting field
strength measurements for all possible excitations separately, based on (4) or (5), and then select the
maximum RF EMF exposure among all these measurements. Although more time-consuming, this could
be an option for UE when the excitations are chosen from a code book with a limited number of possible
excitations. For an exposure scenario where a large number of phase excitations need to be assessed,
and if the measurement system is able to measure both the magnitude and phase of the electric field, an
embedded measurement approach may be adopted where only one measurement is required per antenna
port, with the rest of the analysis left to post-processing according to (6). However, this would require
that additional phase calibration methods be employed for the UE.
For UE employing reciprocity-based beamforming [39,40], the number of possible excitations is infinite.
An RF EMF assessment scheme based on selection and evaluation of a large set of random phases may
then be employed [21]. This approach is well-suited for integration with the Complex Method in (6). A
brief summary of the basic requirements for different RF EMF compliance measurement systems is listed
in Table 4.
One interesting finding of this paper is that knowledge of the magnetic field is not always required
for accurate RF EMF exposure measurement, as shown in Figure 10 and Table 2. For the investigated
antenna arrays transmitting at 15 GHz, a very good accuracy in peak power density was obtained for the
plane-wave equivalent power density approximations also at d = 5 mm, which at 15 GHz is at the boundary

14
Table 4: Basic Requirements for RF EMF Compliance Measurement Methods
S SPW,mag/tan SPW,compl SPW,CFCM/MFCM
E-Field Amplitude X X X X
UE Phase Calibration X
Feeding Network X X
H-Field X

of the reactive near-field region2 of the antenna arrays. A possible reason is that the spatial peak power
density is assessed, rather than the power density distribution. This is supported by the observation that,
for small distances, the accuracy in general is reduced as the averaging area is increased. The obtained
accuracy for small separation distances is most likely also antenna-dependent.
Given that the probes considered are constructed to sample the electric field, the validity of using
the power density estimates in (3)–(8) for antenna elements with a dominant magnetic field component
may be discussed. The considered notch antenna element may be regarded as half a slot where the upper
part has been removed. Therefore, a stronger magnetic field may be expected in the reactive near-field
region. As noted above, at 15 GHz a separation distance of 5 mm will be at the boundary of the reactive
near-field region. As shown in Table 2, quite accurate results are obtained for the maximum peak power
density.
Note that, at higher frequencies, d = 5 mm may belong to the radiating near-field region or even
the far-field region for small array antennas3 . An implication of this is that if frequency bands above
15 GHz are utilized for 5G mobile communication systems, a better accuracy of the discussed RF EMF
compliance assessment methods may be expected.
One of the purposes of this study is to provide an indication of RF EMF exposure and MPTP for
realistic array antennas to be used in future 5G UE in order to comply with FCC and ICNIRP exposure
limits. Compared with existing wireless UE with a maximum transmitted power of 23–24 dBm, the true
MPTP in Table 3 is quite low. This is in agreement with earlier findings for canonical single element
and array antennas [11, 21]. To achieve a reasonable link budget in future 5G systems, assuming that
current RF EMF exposure limits above 6 GHz will be unchanged, one approach to raise the MPTP levels
is to distribute the antenna elements over a larger surface by either increasing the number of elements or
slightly adjusting the inter-element separation distance [21, 24]. Other approaches include utilization of
technical solutions whereby energy is always directed away from the human body.
As presented above, the patch array and the notch array, representing two typical antenna solutions,
have different radiation directions and far-field patterns. In terms of RF EMF near-field exposure,
however, there is no significant difference between the investigated array designs.
As mentioned in Section IV, the measurement sampling distance was λ/4. This sampling rate is twice
the Nyquist sampling rate which should lead to accurate results, especially for comparison against the
proposed FCC limits and the ICNIRP limits. In the immediate vicinity of the array, however, where
power density maxima may appear with a period of about λ/2 as shown in Figure 7(a), there is some risk
for underestimation of the spatial peak power density when comparing against the current FCC limits.
This may be alleviated by interpolating between the measurement values. As noted in the introduction,
spatial peak power density is not a well-defined quantity and the obtained results depend on the chosen
assessment method, including probe type, sampling rate, interpolation techniques, etc.

7 Conclusion
In this paper, possibilities and limitations of scalar and vector-based measurement systems for RF EMF
compliance assessment of 5G UE were investigated. The study was based on numerical simulations of 8×1
patch and notch arrays employing beam-steering and transmitting at 15 GHz. The numerical simulations
were verified with measurements, and the obtained results were found to agree very well.
As RF EMF exposure will vary with the array excitation, compliance with the relevant limits needs
to be demonstrated for all possible excitations. Two conservative methods, i.e. the Components field
2 In the literature different criteria have been used to define the extent of the reactive near-field region depending on

application. There is no well-defined sharp boundary, but rather a transition region where the effects of the reactive fields
gradually increase as the distance is reduced. In this paper λ/4 was used to indicate an approximate distance below which
noticeable effects of the reactive fields may be expected.
3 In this context, the number of antenna elements and the electrical inter-element distance is assumed to be fixed, resulting

in an array size which scales with the wavelength and produces a constant antenna gain.

15
combining method and Magnitude field combining method, were investigated. Although requiring a
relatively simple measurement system and procedure, the level of conservativeness for devices intended
to be used in the immediate vicinity of the body is most likely too high.
A more accurate approach would be to conduct separate measurements in order to span the space
of possible excitations. This will result in a more complicated measurement set-up and increase the
measurement time, which indicates a need for very fast measurement systems.
RF EMF exposure performance of phased arrays in UE for 5G applications is a nontrivial research
topic with consequences not only for the UE design, but also for the entire wireless communication
system. Different regulatory requirements in different parts of the world result in different restrictions on
the MPTP of UE. If the exposure limits are not harmonized, this may have large consequences for the
development of future 5G mobile communication systems.
Figure 12 shows the sampling scheme used in the numerical assessments. For the patch array mock-up
plane, the complex electric and magnetic fields are extracted in the plane z = d using a sampling interval
of 1 mm in the x- and y-directions. The dimension of the sampling plane is 363 mm × 282 mm. The
maximum power density assessed according to (2)–(8) was obtained by considering all possible averaging
area positions in the sampling plane. For the notch array, an analogous sampling scheme was used in the
+x-direction.

Figure 12: Sampling scheme used in the numerical assessments. The arrows indicate that the square-
shaped averaging area Aav moves around in the assessment plane.

References
[1] A. Osseiran, F. Boccardi, V. Braun, K. Kusume, P. Marsch, M. Maternia, O. Queseth, M. Schell-
mann, H. Schotten, H. Taoka, H. Tullberg, M. A. Uusitalo, B. Timus, and M. Fallgren, “Scenarios
for 5G mobile and wireless communications: the vision of the METIS project,” IEEE Commun.
Mag., vol. 52, no. 5, pp. 26–35, May 2014.
[2] W. Hong, K. H. Baek, Y. Lee, Y. Kim, and S. T. Ko, “Study and prototyping of practically large-
scale mmWave antenna systems for 5G cellular devices,” IEEE Commun. Mag., vol. 52, no. 9, pp.
63–69, Sep. 2014.
[3] T. S. Rappaport, S. Sun, R. Mayzus, H. Zhao, Y. Azar, K. Wang, G. N. Wong, J. K. Schulz,
M. Samimi, and F. Gutierrez, “Millimeter wave mobile communications for 5G cellular: It will
work!” IEEE Access, vol. 1, pp. 335–349, 2013.

[4] W. Roh, J. Y. Seol, J. Park, B. Lee, J. Lee, Y. Kim, J. Cho, K. Cheun, and F. Aryanfar, “Millimeter-
wave beamforming as an enabling technology for 5G cellular communications: theoretical feasibility
and prototype results,” IEEE Commun. Mag., vol. 52, no. 2, pp. 106–113, Feb. 2014.
[5] “DOCOMO 5G white paper,” NTT DOCOMO Inc., Jul. 2014. [Online]. Available: https:
//www.nttdocomo.co.jp/english/corporate/technology/whitepaper 5g/

16
[6] J. Helander, K. Zhao, Z. Ying, and D. Sjöberg, “Performance analysis of millimeter-wave phased
array antennas in cellular handsets,” IEEE Antennas Wireless Propag. Lett., vol. 15, pp. 504–507,
2016.

[7] International Commission on Non-Ionizing Radiation Protection, “Guidelines for limiting exposure
to time-varying electric, magnetic, and electromagnetic fields (up to 300 GHz),” Health Phys., vol. 74,
no. 4, pp. 494–522, 1998.
[8] Code of Federal Regulations, Title 47, Part 1.1310, FCC Std., 2016.

[9] IEEE Standard for Safety Levels with Respect to Human Exposure to Radio Frequency Electromag-
netic Fields, 3 kHz to 300 GHz, IEEE Std. C95.1-2005 (Revision of IEEE Std. C95.1-1991), 2006.
[10] IEEE Standard for Safety Levels with Respect to Human Exposure to Radio Frequency Electromag-
netic Fields, 3 kHz to 300 GHz Amendment 1: Specifies Ceiling Limits for Induced and Contact
Current, Clarifies Distinctions between Localized Exposure and Spatial Peak Power Density, IEEE
Std. C95.1a-2010 (Amendment to IEEE Std. C95.1-2005), 2010.
[11] D. Colombi, B. Thors, and C. Törnevik, “Implications of EMF exposure limits on output power
levels for 5G devices above 6 GHz,” IEEE Antennas Wireless Propag. Lett., vol. 14, pp. 1247–1249,
2015.
[12] N. Kuster and Q. Balzano, “Energy absorption mechanism by biological bodies in the near field of
dipole antennas above 300 MHz,” IEEE Trans. Veh. Technol., vol. 41, no. 1, pp. 17–23, Feb. 1992.
[13] T. Schmid, O. Egger, and N. Kuster, “Automated E-field scanning system for dosimetric assess-
ments,” IEEE Trans. Microw. Theory Techn., vol. 44, no. 1, pp. 105–113, Jan. 1996.
[14] P. Vainikainen, J. Ollikainen, O. Kivekas, and K. Kelander, “Resonator-based analysis of the combi-
nation of mobile handset antenna and chassis,” IEEE Trans. Antennas Propag., vol. 50, no. 10, pp.
1433–1444, Oct. 2002.
[15] O. Kivekas, J. Ollikainen, T. Lehtiniemi, and P. Vainikainen, “Bandwidth, SAR, and efficiency of
internal mobile phone antennas,” IEEE Trans. Electromagn. Compat., vol. 46, no. 1, pp. 71–86, Feb.
2004.

[16] B. B. Beard, W. Kainz, T. Onishi, T. Iyama, S. Watanabe, O. Fujiwara, J. Wang, G. Bit-Babik,


A. Faraone, J. Wiart, A. Christ, N. Kuster, A.-K. Lee, H. Kroeze, M. Siegbahn, J. Keshvari, H. Abr-
ishamkar, W. Simon, D. Manteuffel, and N. Nikoloski, “Comparisons of computed mobile phone
induced SAR in the SAM phantom to that in anatomically correct models of the human head,”
IEEE Trans. Electromagn. Compat., vol. 48, no. 2, pp. 397–407, May 2006.

[17] Z. Ying, “Antennas in cellular phones for mobile communications,” Proc. IEEE, vol. 100, no. 7, pp.
2286–2296, Jul. 2012.
[18] K. Zhao, S. Zhang, Z. Ying, T. Bolin, and S. He, “SAR study of different MIMO antenna designs
for LTE application in smart mobile handsets,” IEEE Trans. Antennas Propag., vol. 61, no. 6, pp.
3270–3279, Jun. 2013.

[19] K. Zhao, S. Zhang, C. Y. Chiu, Z. Ying, and S. He, “SAR study for smart watch applications,” in
2014 IEEE Antennas and Propagation Society International Symposium (APSURSI), Jul. 2014, pp.
1198–1199.
[20] H. Li, A. Tsiaras, and B. K. Lau, “Analysis and estimation of MIMO-SAR for multi-antenna mobile
handsets,” IEEE Trans. Antennas Propag., vol. 65, no. 3, pp. 1522–1527, Mar. 2017.

[21] B. Thors, D. Colombi, Z. Ying, T. Bolin, and C. Törnevik, “Exposure to RF EMF from array
antennas in 5G mobile communication equipment,” IEEE Access, vol. 4, pp. 7469–7478, 2016.
[22] “Notice of proposed rulemaking,” FCC, Commission Documents 15-138, Oct. 2015.
[23] K. Zhao, Z. Ying, and S. He, “Human exposure to mmWave phased array antennas in mobile terminal
for 5G mobile system,” in 2015 IEEE 81st Vehicular Technology Conference (VTC Spring), May
2015, pp. 1–2.

17
[24] ——, “EMF exposure study concerning mmWave phased array in mobile devices for 5G communi-
cation,” IEEE Antennas Wireless Propag. Lett., vol. 15, pp. 1132–1135, 2016.
[25] IEEE Recommended Practice for Near-Field Antenna Measurements, IEEE Std. 1720-2012, Dec.
2012.
[26] J. Helander and Z. Ying, “Stripline coupled antenna with periodic slots for wireless electronic de-
vices,” U.S. Patent WO2 016 072 035, May 12, 2016.
[27] J. Wiart, C. Dale, A. V. Bosisio, and A. L. Cornec, “Analysis of the influence of the power control and
discontinuous transmission on RF exposure with GSM mobile phones,” IEEE Trans. Electromagn.
Compat., vol. 42, no. 4, pp. 376–385, Nov. 2000.
[28] T. Persson, C. Törnevik, L.-E. Larsson, and J. Lovén, “Output power distributions of terminals in
a 3G mobile communication network,” Bioelectromagnetics, vol. 33, no. 4, pp. 320–325, 2012.
[29] P. Joshi, M. Agrawal, B. Thors, D. Colombi, A. Kumar, and C. Törnevik, “Power level distributions
of radio base station equipment and user devices in a 3G mobile communication network in India
and the impact on assessments of realistic RF EMF exposure,” IEEE Access, vol. 3, pp. 1051–1059,
2015.
[30] P. Joshi, D. Colombi, B. Thors, L. E. Larsson, and C. Törnevik, “Output power levels of 4G user
equipment and implications on realistic RF EMF exposure assessments,” IEEE Access, vol. 5, pp.
4545–4550, 2017.
[31] C. A. Balanis, Advanced engineering electromagnetics, 2nd ed. New York: Wiley, 2012, ch. 1, pp.
1–38.
[32] F. Ferrara, C. Gennarelli, and R. Guerriero, “Near-field antenna measurement techniques,” in Hand-
book of Antenna Technologies. Singapore: Springer, 2015.
[33] C. Scott, The Spectral Domain Method in Electromagnetics. Artech House, 1989.
[34] B. Thors, A.Thielens, J. Friden, D. Colombi, C. Törnevik, G. Vermeeren, L. Martens, and W. Joseph,
“Radio frequency electromagnetic field compliance assessment of multi-band and MIMO equipped
radio base stations,” Bioelectromagnetics, vol. 35, no. 4, pp. 296–308, May 2014.

[35] A. Faraone, H. Heinrich, T. Harrington, J. Keshvari, T. Onishi, J.-K. Pack, J. Pledl, J. Prats,
M. Wood, and P. Zollman, “Guidance for evaluating exposure from multiple electromagnetic sources-
guidance for evaluating exposure from multiple electromagnetic sources,” International Electrotech-
nical Commission, Tech. Rep. IEC TR 62630:2010, Oct. 2010.
[36] N. Perentos, S. Iskra, A. Faraone, R. J. McKenzie, G. Bit-Babik, and V. Anderson, “Exposure com-
pliance methodologies for multiple input multiple output (MIMO) enabled networks and terminals,”
IEEE Trans. Antennas Propag., vol. 60, no. 2, pp. 644–653, Feb. 2012.
[37] T. Wieland, “A discretization method for the solution of Maxwell’s equations for six-components
fields,” Electonics and Communications AEU, vol. 31, no. 3, pp. 116–120, 1977.

[38] IEEE Recommended Practice for Measurements and Computations of Radio Frequency Electromag-
netic Fields With Respect to Human Exposure to Such Fields, 100 kHz-300 GHz, IEEE Std. C95.3-
2002, 2002.
[39] L. D. DiDomenico and G. M. Rebeiz, “Digital communications using self-phased arrays,” IEEE
Trans. Microw. Theory Techn., vol. 49, no. 4, pp. 677–684, Apr. 2001.

[40] R. Y. Miyamoto and T. Itoh, “Retrodirective arrays for wireless communications,” IEEE Microw.
Mag., vol. 3, no. 1, pp. 71–79, Mar. 2002.

18

View publication stats

You might also like