You are on page 1of 9

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 61, NO.

8, AUGUST 2013 4275

Adaptive Quad-Element Multi-Wideband Antenna


Array for User-Effective LTE MIMO Mobile
Terminals
Shuai Zhang, Kun Zhao, Zhinong Ying, Senior Member, IEEE, and Sailing He, Fellow, IEEE

Abstract—An LTE MIMO antenna array, which is adaptive to of each element has to be as high as possible [3], as also
the effects of the user’s body, is presented for mobile terminals. well-known for diversity in e.g. [4]. However, in practice we
The bands of 750–960 MHz and 1700–2700 MHz are covered need to solve the following two challenging problems in order
with a good efficiency in free space. Three kinds of user effects
are studied, namely, “SAM head and PDA hand,” “PDA hand,” to achieve these properties.
and “dual hands.” The array is formed by selecting the best two The first is the conflict between the limited space of mobile
elements out of four, with the two ports and grounding points of devices and the technical requirements for the handset MIMO
the other two elements open. The user effects on MIMO channel antennas (multiple wideband, low correlation and good effi-
capacity are reduced through the adaptive selection of two op-
timal antenna elements. The total efficiency, envelope correlation ciency). In recent years, several papers have reported specific
coefficient, and multiplexing efficiency are presented for the three designs that get round these problems. In [5] a compact MIMO
kinds of user effects. In the lower band, the decreased correlation antenna array with dual-band and dual-isolation properties can
due to the optimal selection improves the multiplexing efficiency. be found operating at WLAN and WiMAX bands. Single-band
The underlying physical mechanisms and some general rules are
discussed. The specific absorption rate (SAR) in the “SAM head
and dual-band MIMO arrays for the 700 MHz LTE band are in-
and PDA hand” case is studied. Experiments for the three kinds vestigated in [6] and [7], respectively. However, the designs in
of user effects are also carried out. [6] and [7] can only provide a low efficiency and quite narrow
Index Terms—Adaptive arrays, handset antennas, MIMO, spe- bandwidth, which cannot satisfy the needs for some mobile ter-
cific absorption rate, user effects. minal applications. In [8] the isolated antennas are introduced to
design a MIMO antenna with a very high total efficiency (e.g.,
better than 2 dB) at lower frequencies. However, this good iso-
I. INTRODUCTION lation is only realized for the MIMO array with one wideband

L ONG-TERM EVOLUTION (LTE) multiple-input and


multiple-output (MIMO) systems have drawn a great
amount of attention, due to their effectiveness in increasing
element and one very narrow-band element. In [9], an asym-
metrical dual-element antenna array with a big ground plane
and wideband property has been introduced. However, due to
channel capacity without requiring more spectrum efficiency the very strong mutual coupling in the lower band, the total effi-
and power. In an environment with strong fading, a multiple ciency becomes quite low. In addition, the asymmetrical struc-
antenna system can usually be used as MIMO mode or diversity ture in [9] might also result in different MIMO performances
mode according to the SNR level [1]. In a high SNR environ- depending whether the user utilizes the right or left hand.
ment, the MIMO mode will be selected and the highest data rate Another problem is the interaction between mobile terminals
can be achieved [2]. MIMO channel performance should be and users, which will reduce the total efficiency of handset an-
studied. If the SNR is low, a diversity mode will be utilized and tennas by shifting the resonant frequencies and absorbing some
diversity performance will be evaluated through mean effective of the radiated/received power. The user’s hand and head effects
gain (MEG), correlation coefficient, and diversity gains (DG). on mobile terminal antennas have been studied in [10]–[12].
In order to achieve a high MIMO channel performance, an LTE The user-hand effects on mobile terminal antennas operating in
MIMO system requires that envelope correlation coefficients lower UHF band can be found in [13]. Unfortunately, the re-
between antenna elements should be low and the efficiency searches in [10]–[13] only focus on the user effects on a single
antenna. A dual-element MIMO antenna array with hand effects
Manuscript received November 25, 2011; revised February 20, 2013; ac-
has been studied in [14] for the diversity performance at 2 GHz
cepted April 09, 2013. Date of publication April 30, 2013; date of current ver- in mobile terminals. Other research about actual diversity per-
sion July 31, 2013. The work was supported in part by a scholarship within the formance of a multiband diversity antenna with hand and head
EU Erasmus Mundus External Cooperation Window TANDEM.
S. Zhang and S. He are with the Centre for Optical and Electromagnetic
effects can be found in [15]. However, the diversity mobile an-
Research, Zhejiang University, Hangzhou 310058, China and also with the tenna elements utilized in [15] are simply located at the each end
School of Electrical Engineering, KTH-Royal Institute of Technology, SE-100 of the chassis. The cases of two closely-spaced antennas at the
44 Stockholm, Sweden (e-mail: sailing@ieee.org).
K. Zhao is with the School of Electrical Engineering, KTH-Royal Institute of
same end have not been studied. Furthermore, until now the user
Technology, SE-100 44 Stockholm, Sweden. effects on the MIMO channel performance of LTE MIMO an-
Z. Ying is with Research and Technology, Corporate Technology Office, Sony tennas have not been well investigated for the multi-wide bands
Mobile Communications AB, SE-221 88 Lund, Sweden.
of 750–960 MHz and 1700–2700 MHz.
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. In this paper a quad-element multi-wideband antenna array
Digital Object Identifier 10.1109/TAP.2013.2260714 is proposed for LTE MIMO mobile terminals. Utilizing an

0018-926X/$31.00 © 2013 IEEE


4276 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 61, NO. 8, AUGUST 2013

Fig. 1. Geometry of the proposed quad-element LTE MIMO antenna array.

adaptive concept, the MIMO channel performance can be


enhanced for three kinds of user effects, namely, specific
anthropomorphic mannequin (SAM) head and personal digital Fig. 2. Comparison between the case with the ports and grounding points of
assistants (PDA) hand (talk mode), PDA hand (data mode), and Antenna 3 and Antenna 4 open, and the case without Antenna 3 and Antenna 4
dual hands (read mode). Some general rules and physical mech- when Antenna 1 and Antenna 2 are operating.
anism explanations for the user effects will also be provided,
which are not restricted only to our proposed antenna array.
operating elements. Consequently, the impedance bandwidths
II. THE PROPOSED ADAPTIVE QUAD-ELEMENT of the selected elements will be highly affected and cannot
MULTI-WIDEBAND ANTENNA ARRAY FOR cover the whole bands of 750–960 MHz and 1700–2700 MHz.
LTE MIMO MOBILE TERMINALS In addition, the grounding strips will also introduce some extra
losses and change the chassis mode. Therefore, the ports and
A. Configurations of the Adaptive LTE MIMO Antenna Array grounding points of the non-selected elements should stay
The configuration of the proposed LTE MIMO antenna array open simultaneously. In practice, this open condition for each
is shown in Fig. 1. The size of each antenna element is only element can be achieved by two single-pole double-through
. Each element consists of two parts: (SPDT) switches with a total typical insertion loss (for two
The first part is printed on the FR4 PCB substrate with the thick- switches) of 0.6 dB in the lower band, and 0.8 dB in the
ness of 0.8 mm and a permittivity of 4.4; the other part is a ver- higher band. If this kind of SPDT switches is based on
tical folded branch structure. Some detailed descriptions for the micro-electro-mechanical systems (MEMS) the total insertion
LTE MIMO antenna can be found in [16] and some other re- loss can be reduced to less than 0.3 dB in the lower band and
lated single LTE wideband antennas can be found in [17], [18]. less than 0.4 dB in the higher band. In order to describe the
A 1.5 mm- thick plastic mobile phone-cover with a total volume dual-element combinations, Antenna ij will be used to denote
of is used outside the antenna array the operating Antennas i and j with the grounding points and
(there are some extra gaps of 1 mm in length and width dimen- ports of the other two elements open.
sions). The choice for such a thick cover (1.5 mm) and extra In addition, with the open condition mentioned above the op-
gap (1 mm) is only to make the manual fabrication easier. If erating dual-element array will have very similar performances
we reduce the cover thickness from 1.5 mm to 0.5 mm with the (including the user effects) as the array without the non-oper-
1 mm extra gap unchanged, the increase of body loss is less than ating elements in the system. When Antenna 1 and Antenna 2
0.4 dB. The display screen will be placed on the other side of are operating, Fig. 2 shows the comparison between the case
the front chassis in Fig. 1. with the ports and grounding points of Antenna 3 and Antenna
In practical applications, according to different user effects 4 open, and the case without Antenna 3 and Antenna 4 but with
the best two elements out of four will be selected to form a a ground plane length of 108 mm (7 mm longer than that in
dual-element LTE MIMO antenna array with the ports and Fig. 1, due to the chassis mode effect). The S parameters are al-
grounding points (see Fig. 1) of the other two elements open. most the same. The comparison of Smith charts is presented in
The reasons for the open condition are as follows: When only Fig. 3, to show the impedance behavior of the two above-men-
the grounding points are open, the efficiencies of the operating tioned MIMO antenna arrays in free space.
elements will be reduced due to the coupling currents and
50 ohm terminations on the non-operating ports. If only the B. Performance of the Adaptive LTE MIMO Antenna Array
antenna ports are open, the long grounding strips will become The S parameters, total efficiency and envelope correlation
parasitic elements whose geometries are not optimized for the coefficients of the different dual-element combinations in the
ZHANG et al.: ADAPTIVE QUAD-ELEMENT MULTI-WIDEBAND ANTENNA ARRAY FOR USER-EFFECTIVE LTE MIMO MOBILE TERMINALS 4277

(1)

where and are the total efficiency of the first and second
MIMO antenna elements, respectively. is the envelope corre-
lation coefficient between the two elements. In the LTE MIMO
system the body loss (BL) of 95% throughput and BL of multi-
plexing efficiency at certain frequency are comparable (almost
the same) based on the measured results [19].
The multiplexing efficiency of the different dual-element
combinations in free space can be found in Fig. 5(c). In the
lower band the MIMO performance of Antennas 12 is better
than those of the other dual-element combinations. This is
because the chassis mode is very strong if the MIMO antenna
elements are placed at different ends of the mobile terminal,
which will increase the mutual coupling (reduce the total
efficiency) and the envelope correlation coefficient. However,
when the user effects are introduced, in some situations the
antenna elements at each end of the chassis can also provide
a better multiplexing efficiency by dramatically reducing the
correlation between the elements.
Fig. 3. Comparison of Smith charts between the case of Antenna 3 and Antenna
4 open, and the case without Antenna 3 and Antenna 4 when Antenna 1 and III. USER EFFECTS STUDY AND REDUCTION THROUGH THE
Antenna 2 are operating.
PROPOSED ADAPTIVE QUAD-ELEMENT MULTI-WIDEBAND
ANTENNA ARRAY
In this section three kinds of user effects (“SAM head and
PDA hand,” PDA hand, and dual hands) will be studied and re-
duced through the proposed adaptive quad-element multi-wide-
band antenna array. Some general rules and physical mecha-
nism explanations for the user effects will be presented in the
lower-band and higher-band, and these are not restricted only
to our proposed quad-element LTE MIMO antenna array.
The positions of the whole LTE MIMO antenna array and
the user’s body (“SAM head and PDA hand,” PDA hand) are in
accordance with the CTIA revision 3.1 [20]. For the case of dual
hands, since there is no standard yet, the whole antenna array
and dual hands are arranged in a common handsets-holding way.
The dielectric properties of the human tissue used in this paper
can be found in [20].
Since the structure of our designed array is symmetrical,
using the left or right hand will not cause any difference. In
Fig. 4. S parameters of the different dual-element combinations in the adaptive the simulation, we assume that the user uses the right hand (as
quad-element LTE MIMO antenna array. most people do). All the simulations in this paper are carried
out by the commercial software CST [21]. Since CST does
not support the function of rotating fingers, the figures within
adaptive quad-element LTE MIMO antenna array are presented the cover area (see Fig. 6) have been totally subtracted. This
in Fig. 4, Fig. 5(a) and (b), respectively. It can be observed method is quite commonly used in industrial simulations and
that each combination can cover the bands of 750–960 and will affect very little the final results.
1700–2700 MHz with a good efficiency in free space. In addi-
tion, all the envelope correlation coefficients in this paper are A. User Effects-Talk Mode (SAM Head and PDA Hand)
calculated through the 3D E-field radiation patterns, as show in The tissue model and antenna elements locations in talk mode
Formula (2). (SAM head and PDA hand) case are shown in Fig. 6(a). The
In [3], a parameter known as multiplexing efficiency has been display screen is facing the SAM head.
introduced in order to estimate the MIMO channel performance The total efficiency of each port in different dual-element ar-
simply through the efficiency and envelope correlation coeffi- rays is shown in Fig. 7(a). The ranking of the total efficiencies
cient of the MIMO antenna elements. This formula is based on in lower band is: port 3, port 4, port 2 and port 1; in higher band
the assumption of high SNR and isotropic environment, and can the ranking is port 3, port 1, port 2 and port 4. Port 3 always has
be found as follows: the best total efficiency. Thus if the environment is not optimal
4278 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 61, NO. 8, AUGUST 2013

Fig. 5. (a) Total efficiency, (b) Envelope correlation coefficients and (c) Multiplexing efficiency, of the different dual-element combinations in the adaptive quad-
element LTE MIMO antenna array.

Fig. 6. Model and antenna elements locations of: (a) SAM head and PDA hand (talk mode), (b) PDA hand (data mode), and (c) dual hand (read mode).

Fig. 7. (a) Total efficiency, (b) Envelope correlation coefficient and (c) Multiplexing efficiency, of the SAM head and PDA hand (talk mode).

for the MIMO performance and only one port will be utilized, in Fig. 7(c). As expected, due to the high efficiency and low cor-
port 3 (or port 4 for left hand) may be a good choice. relation in lower band, Antenna 34 has a better multiplexing effi-
The envelope correlation coefficient is shown in Fig. 7(b), ciency than the other combinations of antennas. In higher band,
where some interesting results are found: the correlation of An- since the correlations for different combinations are quite sim-
tenna 34 in lower band is much lower than those of the other ilar, Antenna 13, with the highest efficiency in the higher band,
combinations. Efficiency reduction is not the main reason, be- gives the best multiplexing efficiency. However, since the used
cause Antenna 12 has nearly the largest total efficiency loss in antenna structure is an off-ground type, if the specific absorp-
lower frequencies (see Fig. 7(a)) but the correlation decreases tion rate (SAR) is considered, Antenna 34 will have the lowest
only a little. The Antenna 34 case has two characteristics: large SAR value compared with the other combinations. Therefore,
human body coverage of the whole dual-element antenna array from the aspects of the MIMO channel performance and SAR
and the user’s hand is placed approximately symmetrically be- values, Antenna 34 can be used in the whole band for the talk
tween two ports. mode.
Utilizing formula (1), the multiplexing efficiencies of dif- The SAR value of Antenna 34 has been investigated ac-
ferent dual-element combinations are calculated and presented cording to the CTIA standards [20] (SAM head and cheek
ZHANG et al.: ADAPTIVE QUAD-ELEMENT MULTI-WIDEBAND ANTENNA ARRAY FOR USER-EFFECTIVE LTE MIMO MOBILE TERMINALS 4279

TABLE I
SAR VALUE OF ANTENNA 34

Fig. 8. SAR value investigation: (a) CITA head, (b) Original antenna array
location in the phone cover, (c) Antenna array location in the phone cover with
reduced SAR.

touch) as shown in Fig. 8(a). The American Federal Communi-


cation Commission (FCC) standard requires 1.6 W/kg average
1 g tissues, while the European standard requires 2 W/kg
average 10 g tissues. The maximum emission power for LTE
system is 24 dBm (0.25 W) regardless of the used frequency.
Therefore, the input power is 0.25 W for both the lower band
and higher band. All the SAR calculations include the effects of
the S parameters (mutual coupling and mismatching). The SAR
values of Antenna 34 are calculated and shown in Table I. With
the original location, Antenna 3 and Antenna 4 always have low
SAR values in lower band for the FCC and European standards.
However, in higher bands (1700–2700 MHz), the SAR of
Antenna 4 is quite high, especially for the FCC standard. The
high value is mainly due to the high maximum emission power
for LTE system and the small distance between the feeding
Antenna 12 are severely mismatched at 900 MHz, due to the
point and SAM head. If we change the whole LTE MIMO
hand effects causing a shift of the lowest resonance out of the
antenna array from the original position in Fig. 8(b) into that in
lower band.
Fig. 8(c) (and also place the display on the front chassis side),
The envelope correlation coefficient is shown in Fig. 9(b).
the SAR values of both Antenna 3 and Antenna 4 are reduced
The correlation of Antenna 34 has the lowest correlation and
very significantly, as shown in Table I. This change will have
its characteristics are the same as the SAM head and PDA hand
little effect on the total efficiencies and envelope correlation
case. The multiplexing efficiency is also given in Fig. 9(c). It can
coefficients in the three cases (SAM head and PDA hand, PDA
hand, and dual hands). Therefore, in the following study we be observed that in the lower band, due to the lower correlation,
will still use the original antenna array position (placing the Antenna 34 (which has the most coverage) is a little better than
display on the back side of chassis) for analysis. In addition, Antenna 12, but not significantly because of the high loss at port
some other LTE MIMO SAR studies can be found in [22], [23]. 4. Therefore, for the data mode, in the lower band Antenna 12
or Antenna 34 can be used, and in the higher band Antenna 12
will be utilized.
B. User Effects- PDA Hand (Data Mode)
In Fig. 6(b), the tissue model and antenna elements locations C. User Effects- Dual Hand (Read Mode)
are presented for the PDA hand (data mode) case. According to The dual hand model is shown in Fig. 6(c), and the total ef-
the CTIA standard in [20], the hand model and holding rules are ficiency is shown in Fig. 10(a). In lower frequencies, the dif-
exactly the same as in the SAM head and PDA hand case. ferences in the total efficiency between different dual-element
The total efficiencies for the different combinations are combinations are quite small due to the chassis mode, while in
shown in Fig. 9(a). Except for the efficiency of port 4, the the higher band port 2 and port 3 with less user hand coverage
efficiency difference between all the ports is not as significant gives higher efficiency. The envelope correlation in Fig. 10(b)
as that in the “SAM head and PDA hand” case. In other words, shows that Antenna 14, with the largest hand coverage and sym-
the SAM head causes larger effects (including the frequency metrically placed hand, has the lowest value. In the dual hand
shifting and body loss) to port 1 and port 2 positions than to the case, due to the similar efficiency in lower band the correlation
other two port positions. In the data mode, port 4 is the worst is the dominant factor for the high multiplexing efficiency, and
position in terms of the efficiency. Additionally, in Fig. 9(a) thus Antenna 14 gives the best multiplexing efficiency, as illus-
one can observe that at 900 MHz the total efficiency of Antenna trated in Fig. 10(c). In the higher band, because of the similar
12 is somehow smaller than the efficiency of Antenna 34. This correlation, the efficiency becomes very important and conse-
is because compared with Antenna 34 both two elements of quently Antenna 23 has a higher multiplexing efficiency than
4280 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 61, NO. 8, AUGUST 2013

Fig. 9.(a) Total efficiency, (b) Envelope correlation coefficient and (c) Multiplexing efficiency, of the PDA hand (data mode).

Fig. 10. (a) Total efficiency, (b) Envelope correlation coefficient, and (c) Multiplexing efficiency, of the dual hand (read mode).

the others. Therefore, for the read mode, Antenna 14 can be uti- MIMO antenna elements and consequently achieve a low
lized in the lower band, and in the higher band Antenna 23 will correlation.
be used. 4) In the lower band, the multiplexing efficiency of the
dual-element combination with the largest coverage of the
D. Some Rules and Physical Explanations of User Effects user’s body can sometimes perform better than the other
From the total efficiency, envelope correlation coefficient and combinations. This is some what counterintuitive. Low
multiplexing efficiency for the three kinds of user effects, some body coverage does not always give a good result. In the
general rules and physical explanations can be given as follows higher band, the best performance should belong to the
(not restricted only to our proposed MIMO antenna array): combination with the least body coverage. The mechanism
1) In the talk mode, compared to the PDA hand, the SAM behind this phenomenon is as follows: the multiplexing ef-
head will bring more efficiency losses for the position of ficiency is determined by two factors: total efficiency and
Antenna 1 and Antenna 2 than the position of Antenna 3 the envelope correlation coefficient. In the lower band, the
and Antenna 4. correlation between two antennas is relatively high. This
2) The total efficiency variations between different combina- is the main reason for the decrease of the MIMO channel
tions in the lower band (750–960 MHz) are relatively small capacity. As mentioned in items 2 and 3, the large body
compared with those in the higher band (1.7–2.7 GHz). In coverage can reduce the correlation very effectively and
other words, the lower band efficiency is less sensitive to the efficiency will not decrease much. Thus the large body
the body coverage than the higher band. This is because coverage can sometimes give a better multiplexing effi-
in the higher band the array is the only main radiating part, ciency than the others. In the higher band, the correlation
while in the lower band, besides the dual-element array, the has fallen to a quite low level and the total efficiency is the
ground plane is also a main radiating part [24]. factor which mainly affects the multiplexing efficiency.
3) The envelope correlation coefficient in the lower band can Reducing the body coverage (the loss of efficiency) can
be reduced effectively if the dual-element combination directly increase the MIMO performance. Additionally, if
has the following two characteristics: large human body the MIMO antenna elements have a quite low envelope
coverage of the whole dual-element antenna array, and the coefficient in the lower band, it can be expected that the
user hand is placed approximately symmetrically between low body coverage will also give a higher multiplexing
the two ports. The reason is as follows: when the dual-ele- efficiency than a high body coverage. However, due to the
ment array satisfies these two characteristics the user hand limited space in mobile terminals and the strong chassis
can be viewed as a source of scattering. This scattering mode in the lower band, it is very difficult to make sure all
will efficiently separate the radiation patterns of the two the dual-element combinations have a low correlation.
ZHANG et al.: ADAPTIVE QUAD-ELEMENT MULTI-WIDEBAND ANTENNA ARRAY FOR USER-EFFECTIVE LTE MIMO MOBILE TERMINALS 4281

where
, and where denotes the elec-
tric far field of the antenna and is the ratio of the average
vertical power to the time-averaged horizontal power in the
fading environment in a linear form [26], [27]. The second
method is to use a reverberation chamber and directly measure
the correlation between two ports. However, the accuracy
highly depends on the number of samples in the measurement.
Usually, in the lower band at least 4000 samples are needed for
each frequency point. It would take a very long time to carry
out these samples in all the bands. We also use the reference
MIMO antenna to compare the correlations in both ways. It
is found that both methods can give correlation coefficients
very similar to that of the reference antenna, if the number of
samples in the second method is large enough. All the measured
correlation coefficients in this paper are obtained through the
first method, and with a uniform, uncorrelated, and full-sphere
Fig. 11. Four different scenarios for MIMO performance evaluation, (a) Free scenario. The simulated and measured efficiency, envelope
space (no user effect), (b) SAM head and PDA hand (talk mode), (c) PDA hand correlation coefficient and multiplexing efficiency of Antenna
(data mode), (d) Dual hand (read mode). 12/34 are given in Fig. 5(a), (b) and (c), respectively. One
can see that the measured correlation in the band of 850–960
MHz is a little higher than the simulated one, and the measured
IV. EXPERIMENT RESULTS efficiency (and multiplexing efficiency) is somewhat lower
In order to verify the simulation results, some experiments than the simulation results. This will introduce some differ-
have been carried out. In Section II-A, it has been analyzed that ences between the simulated and measured total efficiency and
when Antenna 12 is operating the performances are quite similar multiplexing efficiency in the three user-effect cases. In order
between the case with the ports and grounding points of Antenna to remove these effects in the user-effect measurement, we
34 open, and the case without Antenna 34. From our further use normalized multiplexing efficiency (NME) and normalized
studies, we found that the performances of user effects are also total efficiency (NTE) defined as follows:
quite similar if the operating dual elements in the above-men-
tioned two cases have the same antenna element positions for
the three different kinds of user effects. In order to simplify the
manual fabrication, a dual-element LTE MIMO array in the An- (3)
tenna 12 position (see Fig. 1), with a ground plane length of 108
mm, has been made. During the experiments the fabricated array (4)
is placed exactly in the same positions of Antenna 12 or Antenna
34. Four different scenarios for MIMO performance evaluation
are illustrated in Fig. 11(a)–(d). B. User Effects
The measured and simulated envelope correlation coeffi-
A. Free Space cients of Antenna 12 and 34 in the talk mode, data mode and
read mode are shown in Fig. 7(b), 9(b) and 10(b), respectively.
Fig. 2 shows the comparisons between the simulated S param- The results in the higher band are almost the same. The small
eters of Antenna 12/34 in the quad-element array and the mea- differences in the lower band are mainly due to the difference
sured S parameters of Antenna 12/34 in the fabricated dual-el- of correlations in free space. As we previously mentioned, the
ement array. The results agree well with each other. measured correlations of Antenna 34 in the talk mode and data
In practice, the measured envelope correlation coefficient mode are quite low due to the scattering of the user’s hand.
can be obtained through two methods. The first one is to measure The measured normalized total efficiencies of the talk mode
the 3D E-field radiation patterns of two antennas and thereafter and read mode are illustrated in Fig. 12(a) and (c), respectively.
put the 3D patterns into the correlation definition formula in [25] Fig. 13(a) and (c) show the measured normalized multiplexing
to calculate efficiencies of these two modes, respectively. We observe that
the results match the simulations quite well. For the data mode
in Fig. 12(b) and Fig. 13(b), in the lower band the measured
results for Antenna 12 are similar to the simulation results. In
practice Antenna 34 becomes worse due to the larger losses of
port3 and port 4. Thus, in practical applications, Antenna 12
can be selected for the data mode. Furthermore, the losses in
all measurements are somewhat higher than those in the sim-
(2) ulations (as expected), but the relative loss levels between dif-
4282 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 61, NO. 8, AUGUST 2013

Fig. 12. Simulated and measured normalized total efficiency of Antenna 12 and Antenna 34 for (a) talk mode, (b) data mode, and (c) read mode.

Fig. 13. Simulated and measured normalized multiplexing efficiency of Antenna 12 and Antenna 34 for (a) talk mode, (b) data mode, and (c) read mode.

ferent ports and array combinations are still similar to those in ACKNOWLEDGMENT
the simulations. The authors would like to thank Sony Mobile Communica-
Generally speaking, the measurements and the simulations tions AB for providing the measurement equipment used for this
agree quite well, though sometimes the actual losses are higher paper. The partial support of a Swedish VR grant (2011-4620)
than expected, but still acceptable. is also gratefully acknowledged.
V. CONCLUSIONS REFERENCES
An adaptive quad-element multi-wideband antenna array has [1] L. Zheng and D. Tse, “Diversity and multiplexing: A fundamental
been introduced for user-effective LTE MIMO mobile termi- tradeoff in multiple-antenna channels,” IEEE Trans. Inf. Theory, vol.
49, no. 5, pp. 1073–1096, May. 2003.
nals. With the designed antenna array, three kinds of user ef- [2] B. Holter, “On the capacity of the MIMO channel: A tutorial introduc-
fects on the MIMO channel performance have been investi- tion,” in Proc. IEEE Norwegian Symp. Signal Processing, 2001, pp.
gated, based on the total efficiency, envelope correlation coeffi- 167–172.
cient and multiplexing efficiency. The user effects have been re- [3] R. Tian, B. K. Lau, and Z. Ying, “Multiplexing efficiency of MIMO
antennas,” IEEE Antennas Wireless Propag. Lett., vol. 10, pp. 183–186,
duced efficiently through the adaptive selection of two optimal 2011.
antenna elements. Some general rules and physical mechanisms [4] W. C. Jake, Microwave Mobile Communications. New York, NY,
have been provided and discussed. All the measured results have USA: Wiley-IEEE Press, 1994.
been analyzed. To summarize the simulation and experimental [5] S. Zhang, B. K. Lau, Y. Tan, Z. Ying, and S. He, “Mutual coupling
reduction of two PIFAs with a T-shape slot impedance transformer for
results, in practical applications, Antenna pair (3, 4) and An- MIMO mobile terminals,” IEEE Trans. Antennas Propag., vol. 60, no.
tenna pair (1, 2) can be utilized for the talk mode (with the con- 3, pp. 1521–1531, Mar. 2012.
sideration of the SAR) and data mode, respectively. For the read [6] J. Lee et al., “Miniature Long-Term Evolution (LTE) MIMO ferrite
antenna,” IEEE Antennas Wireless Propag. Lett., vol. 10, pp. 603–606,
mode case, Antenna pair (1, 4) and Antenna pair (2, 3) can be
2011.
used in the lower and higher bands, respectively. Due to the page [7] M. Han and J. Choi, “Small-size printed strip MIMO antenna for next
limitation, the diversity performance is not studied here. How- generation mobile handset applications,” Microw. Opt. Technol. Lett.,
ever, in an isotropic environment, mean effective gain (MEG) vol. 53, no. 2, pp. 348–352, Feb. 2011.
[8] J. Ilvonen, O. Kivekäs, A. A. H. Azremi, R. Valkonen, J. Holopainen,
can be simplified and is equal to half of the total efficiency in
and P. Vainikainen, “Isolation improvement method for mobile ter-
value. The diversity gain (DG) can be estimated through the minal antennas at lower UHF band,” in Proc. 5th Eur. Conf. Antennas
provided correlation and MEG. Furthermore, in this paper we and Propagation (EUCAP), Rome, Italy, 2011, pp. 1307–1311.
have introduced the distributed MIMO concept in the uniform [9] L. Huitema, M. Koubeissi, M. Mouhamadou, E. Arnaud, C. Decroze,
and T. Monediere, “Compact and multiband dielectric resonator
environment. This concept can be extended to other environ- antenna with pattern diversity for multistandard mobile handheld de-
ments for the improvement of MIMO channel capacity with user vices,” IEEE Trans. Antennas Propag., vol. 59, no. 11, pp. 4201–4208,
effects. Nov. 2011.
ZHANG et al.: ADAPTIVE QUAD-ELEMENT MULTI-WIDEBAND ANTENNA ARRAY FOR USER-EFFECTIVE LTE MIMO MOBILE TERMINALS 4283

[10] W. Yu, S. Yang, C. L. Tang, and D. Tu, “Accurate simulation of the sity, Lund, Sweden, and Sony Mobile Communication AB, Sweden. His re-
radiation performance of a mobile slide phone in a hand-head position,” search interests include ultrawideband (UWB) antennas, MIMO antenna sys-
IEEE Antennas Propag. Mag, vol. 52, no. 2, pp. 168–177, Apr. 2010. tems, body centric wireless communications, mm-wave antennas, RFID an-
[11] M. Pelosi, O. Franek, M. B. Knudsen, G. F. Pedersen, and J. B. An- tennas and multiple antennas-user interactions.
dersen, “Antenna proximity effects for talk and data modes in mobile
phones,” IEEE Antennas Propag. Mag, vol. 52, no. 3, pp. 15–26, Jun.
2010.
[12] J. IIvonen, O. Kivekas, J. Holopainen, R. Valkonen, K. Rasilainen, and Kun Zhao was born in Zhejiang, China, in 1987.
P. Vainikainen, “Mobile terminal antenna performance with the user’s He received the B.S. degree in communication
hand: Effect of antenna dimensioning and location,” IEEE Antennas engineering from Beijing University of Posts and
Wireless Propag. Lett., vol. 10, pp. 772–775, 2011. Telecommunications, Beijing, China, in 2010, and
[13] J. Holopainen, O. Kivekas, J. IIvonen, R. Valkonen, C. Icheln, and P. the M.S. degree in wireless systems from the Royal
Vainikainen, “Effect of the user’s hands on the operation of lower UHF- Institute of Technology (KTH), Stockholm, Sweden,
band mobile terminal antennas: Focus on digital television receiver,” in 2012, where he is currently working toward the
IEEE Trans. Electromagn. Compat., vol. 53, no. 3, pp. 831–841, Aug. Ph.D. degree.
2011. He has been a Visiting Researcher at Sony Mobile
[14] A. A. Azremi, J. Ilvonen, R. Valkonen, J. Holopainen, O. Kivekäs, Communication AB, Sweden. His current research
C. Icheln, and P. Vainikainen, “Coupling element-based dual-antenna interests include MIMO antenna design, multiple an-
tennas-user interactions and body centric wireless communications.
structures for mobile terminal with hand effects,” Int. J. Wireless Inf.
Netw., vol. 18, no. 3, pp. 146–157, 2011.
[15] V. Plicanic, B. K. Lau, A. Derneryd, and Z. Ying, “Actual diversity
performance of a multiband diversity Antenna with hand and head Ef-
fects,” IEEE Trans. Antennas Propag., vol. 57, no. 5, pp. 1547–1556, Zhinong Ying (SM’05) is a Principal Engineer
May 2009. (senior expert) of antenna technology in Network
[16] S. Zhang, S. He, K. Zhao, and Z. Ying, “Multi-band wireless terminal Research Lab. Research and Technology, Sony Mo-
with multiple antennas along an end portion of a backplate,” US patent bile Communication AB within Sony Group, Lund,
Sweden. He joined Ericsson AB in 1995. He became
EP 2575208 A1, Apr. 3, 2013.
Senior Specialist in 1997 and Expert in 2003 in his
[17] Z. Ying, Simple Compact LC Loading Wideband Antenna for LTE and
engineer career at Ericsson, and Principal Engineer
Other Bands Sony Mobile patent pending and filed by US, Oct. 2010.
at Sony Group. He has been a Guest Professor at
[18] Z. Ying, Ground Free Type Compact LC Loading Dual Loop More Zhejiang University, China, since 2002. His main
Than 10 Bands Antenna for Cellular and Non-Cellular Wireless Ter- research interests are small antennas, broad and
minals Sony Mobile patent pending and filed by US, Mar. 2010. multi-band antenna, multi-channel antenna (MIMO)
[19] Z. Zhang and Z. Ying, Downlink Throughput Measurements Report of system, near-field and human body effects and measurement techniques. He
Original Mint vs. Gemini Monopole Mint Sony Mobile internal report, has authored and co-authored over 90 papers in various journal, conference and
May, 2012. industry publications. He holds more than 80 patents and patents pending in the
[20] Test plan for Mobile Station Over the Air Performance CTIA revision antenna and mobile terminal areas. He contributed a chapter to the well-known
3.1, Jan. 2011. book Mobile Antenna Systems Handbook, 3rd edition (Artech House, 2008).
[21] CST Microwave Studio 2011 by Computer Simulation Technology He invented and designed various types of multi-band antennas and compact
[Online]. Available: http://www.cst.com MIMO antennas for the mobile industry. One of his contributions in the
[22] K. Zhao, S. Zhang, Z. Ying, T. Bolin, and S. He, “SAR study of 1990s was the development of a non-uniform helical antenna. The innovative
different MIMO antenna designs for LTE application in smart mo- designs are widely used in the mobile terminal industry. His patented designs
bile handsets,” IEEE Trans. Antennas Propag., vol. 61, no. 6, pp. have reached a commercial penetration of more than several hundred million
3270–3279, June 2013. products worldwide.
[23] D. Shi, Y. Gao, and X. Du, “The SAR value analysis of LTE terminals,” He received the Best Invention Award at Ericsson Mobile in 1996 and the
in Proc. Int. Symp. Electromagnetic Compatibility, Rome, Italy, 2012, Key Performer Award at Sony Ericsson in 2002. He was nominated for the
pp. 1–4. President Award at Sony Ericsson in 2004 for his innovative contributions. He
[24] P. Vainikainen, J. Ollikainen, O. Kivekäs, and I. Kelander, “Res- served as TPC Co-Chairmen in International Symposium on Antenna Tech-
onator- based analysis of the combination of mobile handset antenna nology (iWAT), 2007, and served as session organizer of several international
and chassis,” IEEE Trans. Antennas Propag., vol. 50, no. 10, pp. conferences including IEEE APS, and was a reviewer for several academic jour-
nals. He was a member of the scientific board of the ACE program (Antenna
1433–1444, Oct. 2002.
Centre of Excellent in European 6th frame) from 2004 to 2007.
[25] R. G. Vaughan and J. B. Andersen, “Antenna diversity in mobile com-
munications,” IEEE Trans. Veh. Technol., vol. 36, no. 4, pp. 149–172,
Nov. 1987.
[26] R. Clarke, “A statistical theory of mobile radio reception,” Bell Syst.
Tech. J., no. 2, pp. 957–1000, 1996. Sailing He (M’92–SM’98–F’13) received the
[27] Channels, Propagation and Antennas for Mobile Communications, R. Licentiate of Technology and the Ph.D. degree in
Vaughan and J. B. Andersen, Eds. London, U.K.: Inst. Elect. Eng., electromagnetic theory from the Royal Institute of
Technology (KTH), Stockholm, Sweden, in 1991
2003.
and 1992, respectively.
Since then, he has worked at the same division of
Shuai Zhang was born in Liaoning, China, in 1983. the Royal Institute of Technology as an Assistant Pro-
He received the B.E. degree from the University fessor, an Associate Professor, and a full Professor.
of Electronic Science and Technology of China He also serves as Director for a joint research center
(UESTC), Chengdu, in 2007 and the Ph.D. degree between KTH and Zhejiang University (China). His
in electromagnetic engineering from the Royal current research interests include applied electromag-
Institute of Technology (KTH), Stockholm, Sweden, netics, electromagnetic metamaterials, optoelectronics, microwave photonics
in February 2013. and biomedical applications. He has first-authored one monograph (Oxford
He is currently a Research Associate with Depart- University Press) and authored/co-authored about 500 papers in refereed in-
ment of Electromagnetic Engineering, KTH. He has ternational journals. He has given many invited/plenary talks in international
been a Visiting Researcher with the Department of conferences, and has served in the leadership for many international confer-
Electrical and Information Technology, Lund Univer- ences.

You might also like