International Journal of Control, Automation, and Systems (2014) 12(6):1283-1288 DOI 10.1007/s12555-013-0499-5
ISSN:1598-6446 eISSN:2005-4092http://www.springer.com/12555
Miniature Impact Actuator for Haptic Interaction with Mobile Devices
Sang-Youn Kim and Tae-Heon Yang*
Abstract: The crucial procedure in haptic interaction with mobile devices is to convey an appropriate
haptic signal to a user according to the device’s condition. This haptic effect is achieved by creating
vibrotactile signals with a large frequency bandwidth. However, it is quite challenging to generate vi-
brotactile signal with large frequency bandwidth in mobile devices because a vibration motor is fabri-
cated to be embedded into the mobile devices. This paper presents a tiny vibrotactile actuator that can
selectively stimulate human mechanoreceptors by creating a haptic signal with a large frequency
bandwidth. To maximize the haptic effect in limited size, we simulate the magneto-motive force
created by a solenoid by changing a wire diameter of the solenoid. In order to evaluate the haptic per-
formance of the proposed actuator, we construct an experimental setup for measuring the force and the
displacement of the proposed actuator. Using the experimental setup, the output force is measured by
varying input current. The experiment clearly shows that the proposed actuator creates enough output
force to stimulate human skin across a large frequency bandwidth and to convey a variety of vibrotac-
tile sensations to users. The proposed actuator not only offers more reliable input than plain onscreen
keypads in smaller spaces but also allows users to manipulate keypads more effectively.
Keywords: Haptics, mobile device, vibration, vibrotactile actuator.
1. INTRODUCTION
Recently, haptic feedback in mobile devices has been
regarded as one of the dominant factors to increase the
level of immersion because the visual display unit of a
mobile device is not large enough to provide realistic and
exciting sensations to users. A user can communicate
and/or interact efficiently with a mobile device by adding
haptic information to the auditory and visual information.
Haptic feeling consists of kinesthetic and tactile
sensations. Kinesthetic information refers to sensory data
obtained from receptors of joints, muscles, ligaments,
and etc. On the other hand, tactile sense refers to sensory
data acquired through receptors of skin. If both infor-
mations are conveyed to users at the same time, a user
can intuitively and immersively interacts with mobile
devices. However, most haptic actuators for creating
kinesthetic feedback are too bulky to be inserted into
mobile devices. Therefore, for haptic interaction with
mobile devices, researchers have focused on producing
tiny vibrotactile actuators that stimulate the skin.
There are four major mechanoreceptors (Meissner
corpuscle, Merkel’s disk, Ruffini ending, and Pacinian
corpuscle) in human glabrous skin [1]. Merkel’s disk
responds to quasi-static deformations of the skin, such as
force or displacement, in the frequency range of 0.3-3 Hz.
It plays an important role in detecting spatial structures
in static contact, such as an edge or a surface. The
Ruffini ending produces a buzzing sensation in the
frequency range of 40-500 Hz. The Meissner corpuscle,
which has a frequency range of 3-40 Hz, detects dynamic
deformations of the skin such as the sensation of flutter.
The Pacinian corpuscle, which has a frequency response
in the range of 40-500 Hz, is the most sensitive to
vibration amplitude and is particularly known to serve as
the detector of acceleration or vibration.
Judging from above facts, three mechanoreceptors,
except the Ruffini ending, can be selectively stimulated
if a haptic actuator has a large frequency bandwidth. For
example, 1 Hz vibrotactile signal can stimulate Merkel’s
disk and 20 Hz signal can stimulate the Meissner cor-
puscles. If a vibrotactile actuator creates 80 Hz vibration
signal, Pacinian corpuscle are responded. Therefore, it is
necessary to consider an actuator that can generate
vibrotactile information over a large frequency band-
width in order to provide a variety of sensations to users.
Many mobile devices have employed eccentric motors
with operating frequencies of 100-250 Hz. Since the
vibrational force created by an eccentric motor is
proportional to the square of the motor’s rotational speed,
© ICROS, KIEE and Springer 2014
__________
Manuscript received November 9, 2013; accepted May 9,2014. Recommended by Associate Editor Shinsuk Park under thedirection of Editor Hyouk Ryeol Choi. This research was supported by the National Research Founda-tion of Korea (NRF) funded by the Ministry of Education, Scienceand Technology (grant number: 2013M3C1A3059588). This workwas also supported by the “Development of pulse analysis systemfor personalized medicine by converging hemodynamics andpulse diagnostics” (K14310) funded by the Medical EngineeringR&D Group of Korea Institute of Oriental Medicine. Thisresearch has been done for research year (2013) in Koreatech. Sang-Youn Kim is with the Interaction Lab., Advanced Tech-nology Research Center, KoreaTech, Chungjello 1600, Byeongcheon-myeon, Cheonan, Chungnam 330-708, Korea (e-mail: [email protected]). Tae-Heon Yang is with the Center for Mass and RelatedQuantities, Korea Research Institute of Standards and Science,267 Gajeong-ro, Yuseong-gu, Daejeon 305-340, Korea (e-mail:[email protected]).
* Corresponding author.
Sang-Youn Kim and Tae-Heon Yang
1284
we can hardly control the magnitude of the vibrational
force without changing the vibrational frequency, and we
can hardly produce a satisfactory low-frequency
vibrotactile signal under 100 Hz. Another critical
problem is that its response time is too slow to create
haptic feedback in real-time. Hence, the vibrotactile
effect obtained from an eccentric motor is limited to
creating an alert signal.
To improve response time of the haptic actuator,
Samsung Electro-Mechanics has developed a linear
resonance actuator (LRA) consisting of an elastic spring,
a permanent magnet with a flux path, and a solenoid coil
[2]. When an alternating electric current is applied to the
solenoid coil, the permanent magnet with the flux path
attached beneath the spring is linearly actuated and this
linear actuation causes vibration at a particular frequency.
Due to this structure, LRA can have fast vibration
response characteristic. However, the strategy of
vibration near the resonant frequency brought a new
issue for creating a variety of haptic sensations. The
limited frequency bandwidth of the LRA precludes it
from the creation of a variety of haptic sensations.
Therefore, it is necessary to consider a vibration actuator
which generates appropriate vibrations over a large
frequency bandwidth.
Piezo ceramic actuators have been developed for
producing vibrations with a wide frequency range from a
small device [3-6]. I. Poupyrev et al. developed a piezo-
actuator and embedded it to mobile devices for haptic
feedback [3]. M. Wagner et al. developed a helically
wound piezo actuator, helimorph, to increase the stroke
of the piezo actuator [4]. Since these actuators are
actuated over a broad bandwidth, they can convey
various tactile sensations (such as the clicking of a button,
and the surfing of menus) to users. Cruz and D. Grant
developed a piezoelectric actuator and a design tool for
delivering a variety of haptic effect to users [5]. J.
Lylykangas et al. designed a tactile stimulation system,
which is a platform for creating vibrotactile sensation by
piezo actuators, in order to investigate preferred
perception of tactile feedback from non-physical buttons
[6]. Research in Motion released a one button haptic
feedback system with four separate piezoelectric
actuators in order to improve the manipulability of a
mobile phone [7].
Even though these actuators based on piezo-ceramic
materials can selectively stimulate mechanoreceptors, it
is vulnerable to shocks. Another drawback to piezo
actuators is that their vibrational force is not strong
enough to stimulate mechanoreceptors except at their
resonant frequencies. Therefore, two or more piezo
actuators are operated simultaneously to create a strong
vibrotactile sensation.
For overcoming the weakness to shocks, impact type
motors have been proposed for haptic interaction with
mobile devices. The rapid responsiveness of the
vibration generated from the impact type motors makes
the motor create numerous vibration patterns according
to surfing and scrolling menus [8,9]. Furthermore,
impact type vibrotactile actuator can produce realistic
button sensation on the touch panel of mobile devices
[8,9]. ALPS Electric co. developed an impact actuator
(we call it Force Reactor™) [10]. Although, s-type
(small type) of the Force Reactor™ is small enough to be
incorporated into mobile devices, its vibration strength is
not strong enough to oscillate the whole body of mobile
devices. Engineering Acoustics Inc. developed a tiny
vibrator, a C-2 Tactor, to create a strong and localized
sensation on the body [11]. Tactile Labs Inc. developed a
vibrotactile actuator with a bandwidth of 50-500 Hz
capable of producing up to 3G of acceleration [12]. Even
though those vibrotactile actuators produce strong
vibration over wide range of frequency, it is not easy to
create vibrotactile sensation at frequency range below
50 Hz. Another minor problem is that the size of the
actuators is large to be embedded into mobile devices.
Therefore, this paper presents a miniature vibrotactile
actuator that creates a strong haptic effect sufficient to
stimulate human skin over a wide working frequency
range including low frequency.
2. DESIGN AND SIMUATION OF A NEW IMPACT
ACTAUTOR
Fig. 1 shows the components of the newly proposed
impact haptic actuator. The proposed impact actuator
consists of a steel housing, two solenoid coils fixed in a
steel housing, and two permanent magnets passing in and
out of the solenoid coils. The proposed actuator also
includes a link bar that connects the two permanent
magnets. A steel flux path is installed for each permanent
magnet in order to concentrate the magnetic field
strength in the gap between the steel flux path and the
steel housing. There are steel ball bearings in the
proposed actuator to decrease friction between the steel
housing and the steel flux path. Furthermore, we attached
silicon to both sides of the solenoid coils to minimize the
noise from collisions between the steel flux path and the
solenoid coil.
Fig. 2(a) shows the design of solenoid coil. The size of
the solenoid coil was determined by considering the size
of the housing and the permanent magnets. The inner
width and length of the solenoid coil was determined to
be 2 mm × 2 mm, and the outer width and length was
chosen as 2.5 mm × 2.5 mm. The height of the coil was
chosen as 3.4 mm. In order to maximize the force of an
actuator with limited size, the magneto-motive force
Fig. 1. Componenet of a new actautor and its assembled
figure.
Miniature Impact Actuator for Haptic Interaction with Mobile Devices
1285
(A·turns) generated from the coil was simulated by
changing the wire diameter of the coil as shown in Fig.
2(b). The magneto-motive force (NI) and the power
consumption (VI) were calculated using (1) and (2). The
power consumption of a solenoid coil to be incorporated
into mobile devices was determined to be 0.3 W. As
shown in Fig. 2(b), as the wire diameter of the solenoid
increases, the magneto-motive force increases. On the
contrary, the magneto-motive force decreases as the wire
diameter increases while keeping the power consumption
fixed at 0.3 W, 0.6 W and 0.9 W. Thus, the maximum
magnetomotive force by the solenoid coil should be
intersection points of the graph. Since the power
consumption is fixed at 0.3 W, the wire diameter of the
solenoid coil and the number of turns were determined as
0.05 mm and 40 A·turns, respectively (Fig. 2). The
chosen magneto-motive force (40 A·turns) was applied
in FEM simulation to obtain the output force of the
actuator. Fig. 2(c) shows the fabricated solenoid coil.
tot
N V PNI
ρ
× ×
=
×�[A turns],⋅ (1)
2 2
tot
V V PVI
R ρ
×
= =
× �[W], (2)
2( .),
4
Wire DiaP
π ×
= (3)
2 ( ),tot m in out
r N N d dπ π= = +� (4)
8 234.5 . .1.724 10 ,
234.5 . .
Max Temp
Min Tempρ
−
⎡ ⎤+= × × ⎢ ⎥
+⎣ ⎦ (5)
where
NI: Magneto-motive force VI: Power consumption
N: Number of turns V: Input voltage
P: Pure conductor area ρ: Resistivity
:tot� Total length of coil
Fig. 3. Working principle with FEM simulation of a new
impact actuator.
Fig. 4. Result for FEM Simulation.
Fig. 3 shows the working principle of the proposed
actuator and its simulation result when the magneto-
motive force of 40 A·turns was provided to the two
solenoid coils. The magnetic field from the upper
permanent magnet goes by the upper steel flux path and
then passes through the steel housing. After that, the
magnetic field returns to its original position. Both
permanent magnets can be moved up and down
according to the direction of the applied current. In order
to create a strong impact at the downside of the proposed
actuator, the upper and lower permanent magnets are
both pulled down by the upper and lower solenoid coils,
respectively. For generating a strong impact at the upside,
the two permanent magnets are pushed up by the
respective solenoid coils. In this manner, repulsive and
attractive forces are created by the Lorentz force between
the permanent magnets and the solenoid coils. The two
permanent magnets and the solenoid coils produce linear
Lorentz force according to the direction of the stroke. Fig.
4 shows the haptic result for FEM simulation. The output
force was around 40 mN when the input current of 0.1 A
is provided to the actuator.
3. DEVELOPMENT OF THE IMPACT
ACTUATOR
Fig. 5 shows the constructed actuator prototype and its
detailed components. As mentioned above, we attached
the steel flux paths to the ends of the permanent magnets.
The steel flux paths and the steel housing were made of
1008 carbon steel (0.008% or less carbon and 99.8-
99.9% iron), which is a good ferromagnetic material, to
conduct magnetic flux from the permanent magnet with
very low resistance. The moving part consists of two
(a) (b)
(c)
Fig. 2. Parametric design of a solenoid coil and its
picture.
Sang-Youn Kim and Tae-Heon Yang
1286
permanent magnets, two steel flux paths, and a link bar.
The link bar connects the two permanent magnets, and it
linearly guides the moving part. This moving part is
vertically actuated to create strong impacts at both ends
of the housing. Each solenoid coil is made out of SS41
steel and it is fabricated by a precision machining (wire
cutting) process. The solenoid coil was wound using a
coil winding machine. The solenoid diameter is 2.5 mm
and its height is 3.4 mm. The two permanent magnets
were installed at the ends of the link bar with their north
poles facing each other. The two steel flux paths were
mounted at the outside ends of the corresponding
permanent magnets. The moving part travels linearly
inside the two solenoid coils due to the Lorentz force.
The two solenoid coils and the moving part were located
inside the steel housing that has two covers (the upper
and lower covers).
The wires from the two solenoid coils were connected
with a flexible PCB. A copper wire with a diameter of
0.05 mm is wound around the core of the solenoid. The
solenoid coil was designed to have 40 A·turns as we
mentioned in Section 2. When current is applied to the
solenoids, the permanent magnet moves from the initial
position to the other end and collides with a silicon
bumper attached to the end of the solenoid coils. This
collision generates strong and sharp impact vibration.
The size of the developed impact vibration actuator is
3 mm × 3 mm × 15.7 mm. Since the volume of the
proposed actuator (141.3 mm3) is smaller than that of
commercial linear resonance actuators (360 mm3), the
proposed actuator can be easily embedded in mobile
devices.
4. PERFORMANCE OF THE IMPACT
ACTUATOR
Fig. 7 shows an experimental setup to evaluate the
performance of the proposed actuator. The experimental
setup includes a microstage, equipped with a single load
cell (CAS BCL-1L). A digital indicator (CAS CI-510A)
was used to display the output. A function generator
(Agilent 33220A) was connected to the impact actuator
to provide current input. The impact actuator was placed
on the micro stage that moves in the vertical direction
(Z-axis). In order to delicately measure the impact force
using the load cell, we attached a contactor to the end of
the moving part as shown in Fig. 7. The forces generated
by the impact actuator were measured by varying the
input frequency and the position of the micro stage. The
indicator displayed the force generated by the impact
actuator. We measured forces in six trials and took their
average.
The average output force was around 70 mN. The
output force is large enough to generate strong impact
vibration for mobile devices [13]. The most important
factor in the vibrotactile actuator is to create a full stroke
in a wide frequency range, because maximum impact
vibration is generated with a full stroke. Therefore, we
needed to investigate whether the actuator’s stroke
decreases as the frequency varies. Fig. 8 shows a
schematic diagram of the measurement system for
sensing the displacement of the proposed actuator as a
function of the operating frequency.
Fig. 7. Experimental setup for measuring the force of the
impact actuator.
Fig. 5. Fabricated components and assembling process.
Fig. 6. Developed impact actautor.
Miniature Impact Actuator for Haptic Interaction with Mobile Devices
1287
Fig. 8. Experimental setup for measuring the displace-
ment of the impact actuator.
Fig. 9. Frequency response of the actuator.
Fig. 9 shows the frequency response when voltage
input is provided to the proposed actuator. We produced
square waves with a function generator and supplied the
square waves to the proposed actuator. The contactor’s
displacement was measured by a laser Doppler
vibrometer. The measured signal was displayed on an
oscilloscope. As the input current increases from 60mA
to 120mA, the operating frequency range expands. The
proposed vibrotactile actuator produced its full stroke in
the frequency range of 0 to about 100 Hz. Therefore, we
can conclude that the proposed actuator selectively
stimulates mechanoreceptors with large enough feedback
vibration to convey haptic sensation to a user. Further-
more, we did not find any resonance effect because there
is no spring in the proposed vibrotactile actuator.
5. CONCLUSION
Although there are a lot of commercial haptic
actuators which are inserted into mobile devices for
creating tactile sensation, it is not easy to selectively
stimulate human’s mechanoreceptors. This paper
proposed a miniature impact actuator that provides
enough working frequency and output force to stimulate
all mechanoreceptors. The proposed impact actuator
could be embedded in mobile devices, and it could
selectively stimulate mechanoreceptors in the human
somatosensory system by adjusting frequency. The
proposed actuator was composed of the moving part
(which consists of two permanent magnets, two steel flux
paths, and a link bar), two solenoids, a steel housing, and
two covers. In the moving part, the two permanent
magnets was installed at the ends of the link bar with
their north poles facing each other and the magnets
passed in and out of the solenoid coils. According to the
current input, the moving part ran from the initial
position to the other end and collided with a silicon
bumper attached to the end of the solenoid coils in order
to generate vibration. A parametric design for a solenoid
was conducted for maximizing the impact force of the
proposed actuators. In order to demonstrate the
feasibility of the proposed actuator, we constructed an
experimental setup and verified that the proposed
actuator created sufficient output force to generate strong
impact vibration, and furthermore, it generated a wide
range of frequencies to express a variety of haptic
sensations. Our work underscores the importance of the
proposed haptic actuator to enable users to experience
immersion while interacting with mobile devices.
REFERENCES
[1] R. S. Johansson and A. B. Vallbo, “Tactile sensibil-
ity in the human hand: relative and absolute densi-
ties of four types of mechanoreceptive units in
glabrous skin,” Journal of Physiology, vol. 286, pp.
283-300, 1979.
[2] S. D. Kweon, I. O. Park, Y. H. Son, J. Choi, and H.
Y. Oh, “Linear vibration motor using resonant fre-
quency,” US PATENT, 7, 358633 B2, 2008.
[3] I. Poupyrev, S. Maruyama, S., and J. Rekimoto,
“Ambient touch: designing tactile interfaces for
handheld devices,” Proc. of ACM Symposium on
User Interface Software and Technology, pp. 51-60,
2002.
[4] M. Wagner, A. Roosen, H. Oostra, R. Hoppener,
and M. D. Moya, “Novel low voltage piezoactua-
tors for high displacements,” Journal of Electroce-
ramics, vol. 14, no. 3, pp. 231-238, 2005.
[5] M. Cruz and D. Grant, “High definition haptics for
consumer electronics,” Proc. of the IEEE Int. Conf.
on Consumer Electronics, Las Vegas, USA, pp. 57-
58, 2011.
[6] J. Lylykangas, V. Surakka, K. Salminen, J. Raisa-
mo, P. Laitinen, K. Rönning, and R. Raisamo, “De-
signing tactile feedback for piezo buttons,” Proc. of
ACM CHI Conference on Human Factors in Com-
puting Systems, pp. 3281-3284, 2011.
[7] Blackberry storm2 touchscreen phone. http://us.
blackberry.com/smartphones/blackberry-storm-2.
html
[8] K. U. Kyung, J. Y. Lee, and M. A. Srinivasan,
“Precise manipulation of GUI on a touch screen
with haptic cues,” Proc. of IEEE World Haptic
Conference, pp, 202-207, 2009.
[9] K. U. Kyung and J. Y. Lee, “Ubi-Pen: a haptic in-
terface with texture and vibrotactile display,” IEEE
Computer Graphics and Applications, vol. 29, no. 1,
Sang-Youn Kim and Tae-Heon Yang
1288
pp. 24-32, 2009. [10] ALPS, Alps Electric Press Release Force Reactor
http://www.alps.com/e/news release/2005/060801. html, 2005.
[11] Engineering Acoustics Inc. http://www.atactech. com/PR_tactors.html
[12] H. Y. Yao and V. Hayward, “Design and analysis of a recoil-type vibrotactile transducer,” Journal of Acoustic Society of America, vol. 128, no. 2, pp. 619-627, 2010.
[13] T.-H. Yang, D. Pyo, S.-Y. Kim, Y.-J. Cho, Y. D. Bae, Y. M. Lee, J. S. Lee, E. H. Lee, and D.-S. Kwon, “A new subminiature impact actuator for mobile devices,” Proc. of IEEE World Haptic Con-ference, Istanbul, Turkey, pp. 95-100, 21-24 June, 2011.
Sang-Youn Kim received his B.S. degree from Korea University, Seoul, Korea and an M.S.E (1996) and a Ph.D. (2004) in Mechanical Engineering at Korea Advanced Institute of Science and Technology (KAIST). He was a researcher in Virtual Reality Research Center in 2002. From 2004 to 2005, he was a researcher at Human Welfare
Robot System Research Center. In 2005, he was a research staff member at Samsung Advanced Institute of Technology (SAIT). He is currently an assistant professor of internet-media engineering at Korea University of Technology and Education. His current research interests include Human-Computer Interaction, 3D object modeling, Virtual Reality, and Haptics (haptic rendering, tactile display).
Tae-Heon Yang received his B.S. degree in Mechanical Engineering from Yonsei University, Seoul, Korea, and his M.S. and Ph.D. degrees in Mechanical Engi-neering from Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2006, 2008, and 2012, respectively. He is currently a senior research scientist at Korea Research In-
stitute of Standards and Science (KRISS), Daejeon, Korea. His research interests include Haptics (tactile actuator, tactile sen-sor and haptic interface), polymer devices (electroactive poly-mer and pressure resistive material), and national measurement standards of pressure and vacuum.