Electronics and Communications in Japan, Part 1, Vol. 84, No.

1, 2001
Translated from Denshi Joho Tsushin Gakkai Ronbunshi, Vol. J82-B, No. 10, October 1999, pp. 1847–1856

A Study of the Effects of the Shoulder on the Effective Gain

Characteristics in the Multiple Radio Wave Environment of a
Dipole Antenna Close to a Human Head

Koichi Ogawa, Toshimitsu Matsuyoshi, and Kenji Monma

Matsushita Electric Industrial Co., Ltd., Device Engineering Development Center, Osaka, Japan 571-8501

SUMMARY phones are commonly near the horizontal orientation in the

usual talk position. © 2000 Scripta Technica, Electron
Comm Jpn Pt 1, 84(1): 21–30, 2001
When a portable telephone is used close to a human
head, the handset and the shoulder are in close proximity
because the shoulder is raised. Thus, the shoulder could
influence the antenna characteristics of the handset. How - Key words: Body effect; portable telephone; mo-
ever, there have been few reports dealing with such effects. ment method; radiation efficiency; mean effective gain.
This paper describes a basic study of the effects of the
shoulder on the impedance and radiation characteristics of
a half-wavelength dipole antenna located near the head at
900 MHz. Further study shows that there is a change in the 1. Introduction
mean effective gain (MEG) in a multiple radio wave envi -
ronment, and the mechanism by which this change is ef - Since portable telephones are used in multiple radio
fected in relation to antenna location and polarization is wave environments, the received signal level varies with
clarified. The wire-grid method is employed in modeling time due to multipath fadings. To estimate the antenna gain
the human body as a cylindrical head and a trapezoidal left in such an environment a statistical treatment of the incident
shoulder. Studies were also conducted using a dipole an - waves is necessary. To this end, the effective gain of a space
tenna close to a homogeneous human phantom and an adult diversity antenna comprising a pair of parallel half-wave-
man. The analysis of MEG is verified through indoor RFM length dipole antennas [1] and the effective gain of an
measurements. The analytical and empirical results indicate antenna comprising a whip antenna and a planar inverted F
that when the dipole antenna is in a horizontal position the antenna for portable telephones [2] have been analyzed to
radiation to the shoulder is suppressed and consequently a clarify the relationship between the effective gain, the an-
greater radiation intensity at the higher elevation angles of tenna structure, and the incident wave distribution. The
radio signal arrival is obtained due to the presence of the effective gain is the time-averaged gain when the antenna
shoulder. This effect results in MEG values 0.8 to 2 dB moves in a multiple radio wave environment [3]. In these
greater than those obtained from the human model without studies, an analysis of the effective gain characteristics in a
a shoulder. These facts suggest that the presence of the multipath environment has been made using the radiation
shoulder should be taken into account when an adequately power gain pattern in free space. However, in order to
accurate evaluation of the radiation characteristics of port- develop an antenna with a high performance when the radio
able telephone antennas is needed, since portable tele- is held by a human operator moving in a practical field, the

© 2000 Scripta Technica

21
analysis should involve the electromagnetic coupling be- The analytical results of the effective gain are verified by
tween the antenna and the human operator. an experiment conducted in an indoor radio wave propaga-
The operator’s influence on the impedance and radia- tion environment.
tion characteristics of dipole antennas and portable tele-
phone antennas has been investigated by the FDTD method
[4–7] and the moment method [8–11]. These studies indi- 2. Experimental Model and Analytical
cate that significant gain reduction occurs when the antenna Method
is used close to the human body, and approximately half of
the transmission power is absorbed in the human body,
2.1. Antenna structure and the human
resulting in a radiation efficiency of less than 0.5. In these
phantom
analyses a human head has been modeled in the shape of a
cube [4], a rectangular prism [6], a cylinder [11], a spheroid Figure 1 shows the experimental setup. The head is
[6, 9, 10], or a more realistic shape [5, 7, 8]. Recently, some approximated to a cylinder with a diameter of 18 cm and a
studies have modeled a hand holding a portable telephone height of 25 cm. The trapezoidal shoulder with the dimen-
[4–7]. The effect of the hand depends on the type of antenna sions indicated in Fig. 1 is located on the side of the head.
used; the effect is relatively small for an external whip These dimensions are determined from the average meas-
antenna whereas the effect becomes significantly large for ured values of 23- to 30-year-old men having a height of
a built-in antenna when the antenna approaches the hand [5, 170 to 173 cm and a weight of 53 to 75 kg. A half-wave-
6]. On the other hand, the shoulder is a part of the human length dipole antenna is placed at distance D from the head
body that can affect the antenna characteristics since the and at height H from the shoulder with an inclination angle
radio and the shoulder are in close proximity when the of D from the vertical. The human phantom used in the
shoulder is raised in the usual talk position. In such a experiment is made of polypropylene with a thickness of 3
situation, the shoulder could exert a different effect com- mm. A mixed solution comprising 0.6 N salt and normal
pared to the hand because the hand is regarded as a bulky ethylene glycol solutions with a weight ratio of 35:65 is
object located near the antenna whereas the shoulder is an introduced into the plastic phantom case. The dipole an-
extended plane at a distance of approximately a quarter- tenna is constructed by a sleeve balun [14]. The radiation
wavelength at 900 MHz from the antenna. For this reason, pattern was measured using a small oscillator device con-
the authors have presented an analysis of the diversity nected directly to the dipole antenna in order to eliminate
antenna for portable telephones considering the effect of the the influence of a coaxial cable.
shoulder [12]. However, there is little knowledge concern-
ing the basic mechanism of the shoulder effect, and further
research is needed. 2.2. Analytical model and the equation for
This paper presents the impedance and radiation radiation efficiency
characteristics of a half-wavelength dipole antenna located
The moment method is used in the analysis. As shown
near the human head in the 900-MHz band to clarify the
in Fig. 2, the head consists of 6 circles and 16 straight lines,
effect of the shoulder on the antenna characteristics. Further
and the shoulder is divided by lines into a mesh structure.
study demonstrates the variation found in the effective gain
The human body can be treated as a lossy dielectric material
in a multiple radio wave environment, which is anticipated
and the surface impedance is calculated from the biological
from the radiation pattern. The mechanism responsible for
parameters of human tissue as
the variation is also considered. From the results, an inves-
tigation examines how the effective gain changes due to the
geometrical relationship between the antenna and the (1)
shoulder. The dipole antenna is simple in structure, and thus
it is easy to understand the operating principle. In addition,
the radiation properties of a half-wavelength whip antenna
for portable telephones are similar to those of the half-
wavelength dipole antenna [13]; therefore, from the practi-
cal design point of view, it is important to know the
characteristics of the dipole antenna.
The analysis is made by the wire-grid method using
a human model comprising a head and a left shoulder. The
experiments employ a homogeneous human phantom of the
same configuration as the analytical model, and a real
human body with the dipole antenna at one side of the head. Fig. 1. Antenna and human phantom configuration.

22
 (4)

(5)

(6)

where Ph, Ps are the power absorbed in the head and

the shoulder, ZLh, ZLs are the load impedances, and
Ih, Is are the currents flowing through the load imped-
ances. From Eqs. (3) and (4), the radiation efficiency
can be calculated from:

Fig. 2. Wire-grid model. (7)

2.3. Theoretical equation for the mean

effective gain and incident wave modeling
where H and P are the electric permittivity and the magnetic
permeability, V is the conductivity, Z 2Sf, and f is the When an antenna moves in a multiple radio wave
frequency. When radio waves are applied to the human environment over a random route, the mean effective gain
body, induced currents flow on the surface of the skin up to (MEG) is obtained using the following equation [3]:
the skin depth with a surface impedance of ZS shown in Eq.
(1). In order to simulate the human body by the wire-grid
method, the diagonal elements in the impedance matrix are
changed to include lumped load impedances calculated
from the surface impedance. If the load impedances are (8)
arranged as in Fig. 3, the impedance ZL can be calculated
from the following equation [10]:
where : denotes the components (TM) in spherical coor-
dinates and d: sinTdTdM. GT: and GM: are the T and
(2)
M components of the antenna power gain pattern. PT:
and PM: are the T and M components of the angular
The input power Pin to the antenna is calculated as density functions of incoming plane waves. GT: and
GM: are calculated by the wire-grid model in Fig. 2.
PT: and PM: are from a statistical model in which the
(3) angular density functions are assumed to be Gaussian in
elevation and uniform in azimuth, and are given by
where Zin is the input impedance of the antenna and Iin is
the current at the input terminal. Power radiated from the
antenna is partially absorbed in the load impedances and
the radiation efficiency is lowered accordingly. The ab-
sorbed power is obtained from
(9)

(10)
where AT and AM are proportional constants, and mV and
mH are the mean elevation angles of the T- and M-component
wave distributions observed from the horizontal direction.
VV and VH are the standard deviations of the T- and M-com-
Fig. 3. Arrangement of load impedances. ponent wave distributions.

23
 3. Effect of the Shoulder on the Basic mately a wavelength. In contrast, the impedances decrease
with oscillation as the antenna approaches the head. In
Characteristics particular, the imaginary part decreases significantly in the
vicinity of the head. The results mentioned above are com-
3.1. Impedance characteristics pared to those of a previous study [9]. A relatively large
discrepancy was observed between the shoulder and non-
As for the impedance characteristics, the effect of the shoulder models for D < 15 cm. For example, when D = 10
shoulder was investigated by comparing the characteristics cm the real and imaginary part of the impedances for the
for the human model with the shoulder, shown in Fig. 1, shoulder model were 15 ohms greater than those of the
with those without the shoulder corresponding to the cylin- nonshoulder model. This discrepancy is prominent for dis-
drical head only in Fig. 1. These are referred to as “the tances from 5 to 15 cm, while it is very small for distances
shoulder model” and “the nonshoulder model,” respec- less than 5 cm. This phenomenon is accounted for by the
tively. fact that the influence of the head is dominant for small D
Figure 4 shows the real and imaginary part of the and decreased with increasing D, where the influence of the
input impedances for the horizontal dipole antenna with the shoulder begins to appear.
antenna-to-head distance as a parameter. The figure plots Figure 5 shows the calculated and measured results
the calculated and measured data at 900 MHz. The biologi- for the vertical dipole antenna. In comparison with the
cal parameters used in the calculation are Hr = 42 and V = horizontal dipole antenna in Fig. 4, a similar behavior of the
0.85 S/m. The measurement was conducted using the hu- impedance variation was observed as D was changed,
man phantom mentioned in Section 2.1. A good agreement whereas the influence of the shoulder was smaller for the
between the calculated and measured results was obtained vertical antenna. This is because of a weak electromagnetic
for the shoulder and nonshoulder models. The input imped- coupling between the antenna and the shoulder due to their
ances converge to the free space impedances when the normal orientation. It is clear from the results that the
distance D exceeds 30 cm, which corresponds to approxi- influence of the shoulder on the impedance characteristics

Fig. 4. Impedance versus antenna–head distance Fig. 5. Impedance versus antenna–head distance
(horizontal dipole). (vertical dipole).

24
is greater in the case of the horizontal dipole antenna
compared with the vertical antenna.

3.2. Radiation characteristics

In this section, the influence of the shoulder on the

radiation characteristics is examined. Figures 6 and 7 show
the radiation patterns in the vertical planes (z–x and z–y)
for the horizontal and vertical dipole antennas. The figures
plot the calculated data along with the measured ones. In
the figures, panels (a) and (b) show the characteristics for
the shoulder and nonshoulder models, respectively. The
human tissue parameters are Hr = 42 and V = 0.85 S/m. The
antenna is located at D = 5 cm and H = 10 cm.
The measured and calculated results show good Fig. 7. Radiation pattern changes due to shoulder
agreement in the upper hemispherical region (toward the (vertical dipole: D = 5 cm, H = 10 cm).
sky) with slight differences in the lower hemispherical
region (toward the shoulder). Incident radio waves coming
from a base station are known to be distributed in the
elevation angle range between 0° and 40° in a land mobile
communication environment in the 900-MHz band [15]. elevation angles occurs in the case of the horizontal dipole
Therefore, for accurate evaluation of portable telephone antenna. This effect can be deduced from the fact that the
antennas, it is essential to obtain accurate radiation patterns antenna is parallel with the shoulder, which functions as an
at those angles of signal arrival. efficient reflector, and the fact that the shoulder is spread
With the shoulder model for the horizontal dipole over a conformal plane at a distance of approximately a
antenna in Fig. 6(b), the radiation in the lower hemispheri- quarter-wavelength from the antenna.
cal region is suppressed and consequently a greater radia- Figure 8 shows the measured radiation pattern for a
tion intensity is obtained in the upper hemispherical region, 30-year-old man of 170 cm height and 60 kg weight. The
which includes the angles of signal arrival. This phenome- dipole antenna is located above the left shoulder horizon-
non is distinct in the z–y plane. On the other hand, there is tally at D = 5 cm and H = 10 cm. The measurement was
little change in radiation pattern for the vertical dipole made with the man laying on a rotation table [16]. Compar-
antenna as illustrated in Fig. 7. It is clear from Figs. 6 and ing Figs. 8 and 6(b), the two radiation patterns have similar
7 that gain enhancement due to the shoulder in the high characteristics such that the radiation in the shoulder direc-
tion is suppressed and that in the high elevation angles is
enhanced. However, there are a number of ripples in the
radiation pattern and significant differences are observed in
the cross-polarized radiation patterns in the data measured
using the real human body in Fig. 8. These differences are

Fig. 6. Radiation pattern changes due to shoulder Fig. 8. Radiation pattern of a horizontal dipole antenna
(horizontal dipole: D = 5 cm, H = 10 cm). located near a human head (D = 5 cm, H = 10 cm).

25
Table 1. Comparison between calculated and measured
MEGs (horizontal dipole: XPR = 6 dB, mV mH = 20°,
VV VH = 20°)

Fig. 9. MEG versus antenna–head distance.

a possible cause of error in the evaluation of the mean

effective gain which is dealt with in the next section. For Figures 9 and 10 plot the calculated MEG and radia-
this reason, the mean effective gain is evaluated from the tion efficiency as a function of the antennal–head distance
calculated and measured radiation patterns in Figs. 6 and 8, D. The incident radio wave parameters are determined to
and the validity of the human model and the analytical be typical values in urban areas; the cross-polarized power
method is confirmed. ratio (XPR) is 6 dB and the mean elevation angle and
Table 1 shows the MEGs calculated from the two standard deviation are 20°.
principal planes (z–x and z–y) [17]. Although the MEGs in From Fig. 9 the MEG decreases significantly when
Table 1 are different from the true MEG values, they are the antenna approaches the human body in the case of
considered to be effective in comparing MEGs. The MEG, both vertical and horizontal configurations. For example,
which is an averaged value from three successive measure- when D = 1 cm the MEG is as low as –9 dBd. Figure 9
ments, was repeatable within 0.5 dB. As can be seen from also shows that the MEG of the vertical antenna increases
Table 1, the differences between the calculated MEG and more rapidly than that of the horizontal antenna. For
the MEG measured using the human phantom, and between instance, the increment in the MEG from D = 1 cm to 5
the calculated MEG and the MEG measured using the real cm is 6 dB for the vertical antenna, whereas the incre-
human body are 0.5 and 1.8 dB, respectively. The greater ment is only 2 dB for the horizontal case. The reason for
difference for the human body is probably due to a simple this discrepancy is the greater gap existing between the
approximation in the analytical model. Table 1 shows that antenna and the head for the horizontal dipole since the
a good agreement between the changes in MEG for the head is modeled in the shape of cylinder, leading to weak
shoulder and nonshoulder models is obtained for the calcu-
electromagnetic couplings.
lated and the phantom results. This is supported by the fact
Comparing the radiation efficiency of the shoulder
that the measured radiation pattern in Fig. 6 accords with
and nonshoulder models in Fig. 10, a relatively large
the calculated one very well. Since one of the main points
of argument in this paper lies in the change in MEG due to
the presence of the shoulder, Table 1 indicates that the
calculated MEG values are adequately accurate for discuss-
ing the influence of the shoulder on the MEG charac-
teristics.

4. Effective Gain Characteristics and

Radiation Efficiency

In this section a detailed investigation is made into

how the change in the radiation characteristics affects the
MEG characteristics, and the mechanism by which the
MEG varies due to the geometrical relationship between the Fig. 10. Radiation efficiency versus antenna–head
antenna and the human body is considered. distance.

26
 Fig. 11. MEG versus antenna inclination angle D.

reduction in efficiency is observed for the shoulder model

in the case of a vertical dipole, whereas the reduction is very
small for a horizontal dipole. On the other hand, Fig. 9
shows that for the shoulder model, the reduction in the
MEG for the vertical dipole is almost the same as the
reduction in efficiency, while for the horizontal dipole
shown in Fig. 10, the efficiency remains unchanged, but the
MEG increases by 1.5 dB due to the presence of the
shoulder. It can be seen from Figs. 9 and 10 that the
reduction in the MEG of the vertical dipole for the shoulder
model is caused by the reduction in efficiency due to an Fig. 12. Absorbed power versus antenna–head distance.
increase in the power absorbed in the shoulder. This will be
dealt with in the next section. In contrast, the MEG en-
hancement for the horizontal dipole is caused by the radia-
tion pattern change in the direction of incident waves as
shown in Fig. 6. absorbed in the shoulder Ps from Eq. (6) for the vertical and
Figure 11 shows the MEG versus antenna inclination horizontal configurations. The absorbed power is normal-
angle D from the vertical with the mean elevation angle as a ized to the input power Pin. When the vertical dipole an-
parameter. In the figure, D = 0° and 90° correspond to the tenna is close to the head, Ph is dominant. However, Ps
vertical and horizontal configurations. It is seen that the reduc- increases with increasing antenna–head distance D. Note
tion in MEG is smaller for the shoulder model compared with that Ps / Pin is 15% at D = 5 cm. In the horizontal dipole,
the nonshoulder model with increasing D. It should be noted Ps / Pin is almost constant (5%) regardless of D. The dashed
that when the mean elevation angle is 40° the MEG increases lines in the figure show the power absorbed in the head for
with increasing D. As a result, in the case of the horizontal the nonshoulder model, which is almost the same as that for
dipole the increment in MEG due to the shoulder ranges from the shoulder model in the horizontal dipole. This corre-
0.8 to 2 dB for mean elevation angles from 0° to 40° which sponds to the radiation efficiency curve in Fig. 10. It can be
are values typical in an urban area [15]. seen from the results that the shoulder has less effect on the
The results imply that the shoulder should be taken into absorbed power.
account when evaluating the MEG of portable telephone
antennas with high accuracy, since they are commonly near
the horizontal orientation when used in the usual talk position. 6. Experimental Study in an Indoor
Multipath Environment

5. Absorbed Power In order to verify the validity of the analytical results

in Section 4, an experiment was conducted in an indoor
Figure 12 shows the total absorbed power Pa, the multiple radio wave environment to estimate the MEG
power absorbed in the head Ph from Eq. (5), and the power characteristics at 900 MHz. Figure 13 shows the experi-

27
 Table 2. Measurement results of incident wave
parameters

required for the MEG calculation is measured with the

dipole antenna skewed by 55° from the vertical. Since there
may be some deviation in the azimuthal angle of the inci-
dent waves, the MEG is calculated from an average of four
successive measured results, each measurement being
made for the human phantom at four different angular
positions at 90° intervals relative to the tangential direction
of the rotating arm. The detailed measurement procedure is
given in Ref. 18.
Table 2 shows the measured incident wave parame-
ters. The results indicate that the transmitted vertically
Fig. 13. Indoor experimental setup of multiple radio polarized component is dominant. This phenomenon can be
wave propagation. interpreted as being due to multiple radio waves being

mental setup. The experimental procedure and method for

incident wave parameter measurement for indoor environ-
ments are those reported by Taga [18]. The experiment was
conducted in a typical laboratory with concrete walls, and
plastic boards on a concrete base for the floor and ceiling.
The transmitting antenna is located vertically on the floor
and the receiving signal is sampled by an A/D converter
with the receiving dipole antenna close to the human phan-
tom (Fig. 1) which is moving on a rotating arm of 1.5 m. A
partition wall is placed between the transmitting and receiv-
ing antennas so that the out-of-sight condition is main-
tained. Both antennas are located at the same height of
1.5 m from the floor, which is the ceiling–floor center
position. In this configuration, the incident waves are as-
sumed to have a symmetrical distribution in elevation with
the maximum value in the horizontal direction. If we as-
sume the incident wave distribution to be Gaussian in
elevation, then the incident wave can be represented by
three parameters: the cross-polarized power ratio (XPR),
and the standard deviations of the vertically and horizon-
tally polarized components, VV and VH. XPR is measured
by the ratio of the average receiving power of a half-wave-
length dipole antenna to that of a cylindrical slot antenna
when the two antennas rotate once. The slot antenna is 28
mm in cylinder diameter, 246 mm in slot length and 1 mm
in slot width. The standard deviation can be found from the
MEG characteristics for the dipole and slot antennas lo- Fig. 14. Comparison between calculated and measured
cated vertically and horizontally. The total incident power MEGs (horizontal dipole).

28
created by reflections from the walls, floor, and ceiling in a 5. Gain enhancement due to the shoulder in the high
regular manner since there is no furniture such as desks or elevation angles occurs more markedly in the case of a
bookshelves in the room. horizontal dipole antenna than a vertical dipole antenna.
Figure 14 plots the measured MEGs together with the 6. The MEG decreases to –9 dBd when the antenna
calculated ones for the shoulder and nonshoulder models. approaches the human head for both vertical and horizontal
The receiving dipole antenna is located horizontally. The configurations in a typical radio wave environment in an
figure shows the average receiving powers measured with urban area.
the human phantom directed to 0°, 90°, 180°, and 270° 7. The MEG enhancement for the horizontal dipole
relative to the tangential direction of the rotating arm. The is due to the radiation pattern change in the direction of
average value for the four measurements is also shown. incident waves.
There is some discrepancy between the receiving powers in
the four directions, implying that incident waves were Field experimentation in cellular propagation envi-
nonuniform in azimuth. However, even in such a situation ronments is left for future studies.
it has been reported that the averaging process for the
receiving power for azimuthally nonuniform incident
waves agrees very well with that for a uniform model [19].
Acknowledgments. The authors wish to express
Thus, it is concluded that the measured result in Fig. 14 that
their appreciation to Mr. Toru Ishida, director of the Device
should be compared with the analytical result is the average
Engineering Development Center, for his support and en-
value of the four directions. The average values show good
couragement, and to Dr. Tomoki Uwano, manager of the
agreement with the calculated ones. In addition, it can be
Central Research Laboratory, for useful suggestions.
seen from Fig. 14 that the MEG increment due to the
presence of the shoulder is confirmed by experiment. These
results indicate that the MEG analysis detailed in this paper
is of sufficiently high accuracy.
REFERENCES

1. Taga T. Characteristics of space diversity branch us-

7. Conclusion ing parallel dipole antennas in mobile radio commu-
nications. Trans IEICE 1992;J75-B-II:370–378. (in
The shoulder’s effect on the impedance and radiation Japanese)
characteristics of a half-wavelength dipole antenna located 2. Ogawa K, Uwano T. Mean effective gain analysis of
near the head at 900 MHz has been examined. Further study a diversity antenna for portable telephones in mobile
demonstrates a change in the mean effective gain (MEG) in communication environments. Trans IEICE
a multiple radio wave environment, and clarifies the mecha- 1998;J81-B-II:897–905. (in Japanese) (Translation
nism of the change in relation to the antenna location and in Electron Commun Japan Part 1 2000;83:88–97.)
polarization. The main results are as follows. 3. Taga T. Analysis for mean effective gain of mobile
antennas in land mobile radio environments. IEEE
1. The input impedances converge to the free space Trans 1990;VT-39:117–131.
impedances when the antenna–head distance D exceeds 4. Toftgard J, Hornsleth SN, Andersen JB. Effects on
30 cm, which corresponds to approximately a wavelength. portable antennas of the presence of a person. IEEE
In contrast, the impedances decrease with oscillation when Trans Antennas Propagat 1993;AP-41:739–746.
the antenna approaches the head. 5. Jensen MA, Rahmat-Sami Y. EM interaction of hand-
2. The real and imaginary part of the impedances for set antennas and a human in personal communica-
the shoulder model are 10 to 15 ohms greater than those of tions. Proc IEEE 1995;83.
the nonshoulder model. This discrepancy is prominent for 6. Sato K, Nishikawa K, Suzuki N, Ogawa A. Analysis
distances from 5 to 15 cm, while it is very small for of antennas mounted on portable equipment near
distances less than 5 cm. human body. Trans IEICE 1996;J79-B-II:892–900.
3. The shoulder has a greater influence on the imped- (in Japanese)
ance characteristics in the case of a horizontal dipole an- 7. Watanabe S, Taki M, Nojima T. Effects of the inter-
tenna than a vertical antenna. action between a human head and a hand-held port-
4. In the horizontal dipole antenna, the radiation in able radio on the antenna input impedance. Trans
the lower hemispherical region is suppressed, and conse- IEICE 1996;J79-B-II:557–565. (in Japanese)
quently the radiation intensity is greater in the upper hemi- 8. Chuang HR. Human operator coupling effects on
spherical region. This is due to the presence of the shoulder. radiation characteristics of a portable communication

29
 dipole antenna. IEEE Trans Antennas Propagat 15. Lee WCY, Brandt RH. The elevation angle of mobile
1994;AP-42:556–560. radio signal arrival. IEEE Trans Commun
9. Yamada T, Mashiko T, Koeba S, Sawaya K, Adachi 1973;COM-21:1194–1197.
A. Interactions between a dipole antenna and a man 16. Ogawa K, Matsuyoshi T, Monma K. Antenna-pattern
model—Analysis by surface impedance method. measurement at angles of signal arrival of portable
Trans IEICE 1988;J71-B:246–255. (in Japanese) radios near a human operator. APMC98 Int Symp Dig
10. Tsunekawa K, Ando A. Advanced wire grid method p 461–464, 1998.
for solving the scattered field of a lossy dielectric 17. Saito Y, Ogawa K, Matsuyoshi T. An estimation
object. IEEE AP-S Int Symp Dig 1992;2:797–800. method for the mean effective gain of a handset
11. Koyanagi Y, Ogawa K. Analysis of a dipole antenna antenna by principal plane radiation patterns. Conv
located near the human body in the VHF/UHF band. Rec IEICE Japan p 647–648, 1999. (in Japanese)
Conv Rec IEICE Japan, 1996; p 80. (in Japanese) 18. Taga T. Indoor measurement method for evaluating
12. Ogawa K, Matsuyoshi T, Monma K. An analysis of statistical distribution of incident waves under out-of-
the performance of a handset diversity antenna influ- sight condition and experimental studies of charac-
enced by head, hand and shoulder effects at 900 MHz. teristics of mobile station polarization diversity.
Tech Rep IEICE 1998;AP98-88, RCS98-133: 89–96. Trans IEICE 1991;J74-B-II:608–615. (in Japanese)
13. Ogawa K, Uwano T. A diversity antenna for very 19. Taga T. A study on mean effective gain of mobile
small 800-MHz band portable telephones. IEEE antennas in effective line-of-sight propagation envi-
Trans Antennas Propagat 1994;AP-42:1342–1345. ronments. Conv Rec IEICE Japan, p 2–23, 1990. (in
14. Kraus JD. Antennas. McGraw–Hill; 1988. p 742. Japanese)

AUTHORS (from left to right)

Koichi Ogawa (member) received his B.S. and M.S. degrees in electrical engineering from Shizuoka University in 1979
and 1981. He joined Matsushita Electric Industrial Co., Ltd. in 1981, and engaged in research and development on a 50-GHz
millimeter-wave integrated circuit and a 12/14-GHz very small aperture terminal (VSAT) satellite communication system. He
is currently a research group leader of Mobile Communication RF-Devices. His research interests include diversity antennas
for portable handsets, compact antennas for mobile communication systems, and other related areas of radio propagation. He
received the OHM Technology Award from the promotion foundation for electrical science and engineering in 1990. He was
elected a member of the IEICE Committee on Electromagnetic Human Phantom. He is a member of IEEE. He is listed in Who’s
Who in the World.

Toshimitsu Matsuyoshi (member) received his B.S. and M.S. degrees in communication engineering from Osaka
University in 1992 and 1994. He joined Matsushita Electric Industrial Co., Ltd. in 1994, and has been engaged in research and
development on microwave filters and amplifiers. He is presently engaged in research and development on antennas for mobile
communication systems. He is a member of IEICE.

Kenji Monma received his B.S. degree in communication engineering from Tohoku University in 1996. He then joined
Matsushita Electric Industrial Co., Ltd., and has been engaged in research and development on microwave circuits and antennas
for mobile communication systems.

30