HAL Id: hal-01341981
https://hal.inria.fr/hal-01341981
Submitted on 8 Jul 2016
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Texture Rendering Strategies with a High Fidelity -
Capacitive Visual-Haptic Friction Control Device
Eric Vezzoli, Thomas Sednaoui, Michel Amberg, Frédéric Giraud, Betty
Lemaire-Semail
To cite this version:
Eric Vezzoli, Thomas Sednaoui, Michel Amberg, Frédéric Giraud, Betty Lemaire-Semail. Texture
Rendering Strategies with a High Fidelity - Capacitive Visual-Haptic Friction Control Device. Haptics:
Perception, Devices, Control, and Applications, Jul 2016, London, United Kingdom. �10.1007/978-3-
319-42321-0_23�. �hal-01341981�
Texture Rendering Strategies with a High Fidelity Capacitive
Visual-Haptic Friction Control Device
Eric Vezzoli
1
, Thomas Sednaoui
1,2
, Michel Amberg
1
, Frédéric Giraud
1
and Betty Lemaire-Semail
1
1
Univ. Lille, Centrale Lille, Arts et Metiers ParisTech, HEI, EA 2697 - L2EP –
Laboratoire d’Electrotechnique et d’Electronique de Puissance, F-59000 Lille, France
2
STMicroelectronics, Crolles F38920, France
*eric.vezzoli@ed.univ-lille1.fr; thomas.sednaoui@st.com;
Michel.Amberg@univ-lille1.fr;
{frederic.giraud,betty.semail}@polytech-lille.fr;
Abstract. Ultrasonic vibrations of a plate can modify the perception of the friction between a
surface and a sliding finger. This principle, coupled with modern position sensing tech-
niques, is able to reproduce textured materials. In this paper, an open loop control through
model inversion of the friction force between the finger and the plate is presented. The de-
vice incorporating the control system is described, and two different reproduction strategies
are formalized to address the reproduction of objects and textures. In the end, a psychophysi-
cal experiment evaluating the two control strategies is described.
Keywords: ultrasonic, tactile device, ultrasonic vibrations, friction control
1 Introduction
In recent years, a lot of interest raised around the implementation of haptic stimulation and
simulation. Many technologies are available to provide stimulation on a finger touching or sliding
on a surface. However, few of those shown a promising opportunity for a coupling with
touchscreen and the consequent aim for integration in the mobile world. Friction modulation
techniques, namely electrovibration [1], [2], electroadhesion [3] and ultrasonic vibrations [4] are
among this group. Their principle relies on the modulation of the friction between a finger sliding
on an active surface. Electrovibration and electroadhesion exploit the electrostatic attraction gen-
erated between the finger and the plate, the first through the application of a voltage and the se-
cond through the application of a current. Both of them enhance the friction between the finger
and the plate. On the other hand, ultrasonic vibrations of a plate reduce the friction between the
finger and the surface. The reduction of friction was firstly explained through the squeeze film
effect [4], later, a more mechanical explanation of the phenomenon was introduced [5], but a
reliable modelling is yet to be developed. With the availability of finger detection techniques, the
possibility to actively change the friction as a function of the finger position led to the ability to
simulate textures [6]. This study focused on the reproduction of a squared grating, but it is lacking
a broader approach on object reproduction. As introduced in [7], force plays a fundamental role
on the reproduction of shapes, and a similar approach on texture reproduction has not been intro-
duced yet.
In this paper, the development and the control of a device able to simulate textures thanks to
the friction coefficient modulation between the finger and the vibrating plate is described.
The paper is organized as follows: initially the characteristics of the friction reduction provided
by ultrasonic devices are recalled, following, an analysis of previously developed strategies for
friction coefficient control is performed. The structure and the open loop control scheme of the
device are then described followed by the formalization of the two haptic signal rendering
schemes. In the end, a psychophysical experiment exploring the advantages of the two rendering
techniques is proposed and the results discussed.
2 Ultrasonic Lubrication
The first explanation for the friction reduction in ultrasonic devices was found in the squeeze
film effect [8]. The effect relied on the generation of an ovepressured film of gas between the
ultrasonic vibrating plate in the range of micrometres and the interface of the finger. The underly-
ing physical phenomenon was firstly modelled by Biet et al [4]. The presence of the squeeze film
effect in these devices was lately questioned by Sednaoui et al [9]. The extended study on the
characteristics of the friction reduction, reported in [9], identified an empirical model to describe
the evolution of the friction modulation. The friction coefficient reduced by the ultrasonic vibra-
tion is unable to reach the zero value, and it is reduced with an exponential decay in function of
the vibration amplitude. The friction coefficient reaches an asymptotic value for high vibration
amplitude where 90% of the maximum reduction is reached for an amplitude around 2 microme-
ters, figure 1. It is possible to define the relative friction coefficient !
"
as:
!
"
# $
! #
! %
(1)
Where ! # is the friction coefficient for a given vibration amplitude &#. In the reported study,
!' was identified as the invariant for similar conditions between different subjects, figure 1b. The
heig
ht of
the
as-
ymp
tote
of
the
fric-
tion
re-
duc-
tion
is
line-
Fig. 1. Experimental results measured for the friction reduction in ultrasonic device. On the left, the
friction coefficient measured through the tribological assessment. On the right, the relative friction
coefficient of the same measures.
arly dependent on the normal force applied. The reduction of relative friction coefficient, () * in
function of the vibration amplitude can be expressed as:
() # $ + , ()
-
.
/01
2 ()
-
(2)
With ()
-
$ +3+45
6
Where 7 is an empirical constant with 1.67 µm
-1-
as value, 5
6
is the normal force applied by the
finger.
3 Friction Coefficient Control
In the present section, an analysis of previous implementation of friction coefficient control in
ultrasonic devices is presented.
The friction felt by the finger while exploring different surfaces is greatly variable for different
subjects: many different parameters influence this force, like the nature of the surface, lipid and
water content or the mechanics properties of the finger. Moreover, the friction coefficient may
dramatically change within seconds after the contact due to generated occlusion and non-coulomb
friction [10]. A closed loop approach of the friction coefficient control was previously developed
with the SMARTTAC [11]. This device incorporated a broad bandwidth tribometer around an
ultrasonic tactile plate. The lateral and normal force sensors implemented allowed the real time
recording and processing of the normal and friction force permitting the implementation of a
closed-loop control of the friction coefficient. The system works smoothly for a step of friction
coefficient and it is able to maintain the control in the steady state. In this device, the participants’
friction coefficient is recorded along the first stroke over the plate and then controlled based on
the previous measure.
One issue with this architecture is that a relevant change of the friction due to the generated oc-
clusion or deposited moisture could lead to the saturation of the control. Moreover, the bandwidth
of the device is intrinsically limited by the mechanical properties of the fingertip. To effectively
control the friction coefficient, the lateral force sensor needs to be able to measure the friction
force. Based on the measurements performed in the cited work on different participants, an aver-
age rise time of the lateral force of 3.84ms was recorded. This value is reflected in a bandwidth of
around 90Hz for the signal reproduction. The reported value matches the mechanical properties of
the fingertip highlighted in [12] and it is in accord with similar measurements performed in [13]
and validated for both electrovibration and ultrasonic vibrations in the cited work. The elastic
response of the fingertip imposes a bandwidth limit to the closed loop friction coefficient control
far below the perceptual bandwidth of the finger. Due to the intrinsic limitation of a direct meas-
urement and control of the lateral force, an open-loop approach will be described in the following
sections.
4 High Fidelity Texture Rendering
In this section, the developed device is described and the strategies of texture reproduction are
introduced.
4.1 Device
T
he
de-
vice
is
built
arou
nd
the
ba-
nana
pi
(She
nzhe
n
LeMaker Technology Co. Ltd, China) single board computer featuring a 1 GHz ARM Cortex-A7
dual-core CPU with 1 GB of ram working in parallel with a microcontroller (stm32f4, STMicroe-
lectronics, France). The single board computer is connected to a 5 inches flat capacitive touch
screen (Banana-LCD-5"-TS, Marel, China) providing the finger position input and display output,
where the sampling frequency of the finger position is 50 Hz. The communication between the
microcontroller and the single board pc is provided by an SPI bus working at 10 kframes/second.
4 flat resistive force sensors (CP 150, IEE, Luxemburg) are placed under the corners of the dis-
play and provide the normal force value to the microcontroller. The microcontroller synthesizes a
PWM signals to pilot a voltage inverter the motor piezoceramics. In figure 2, the structure of the
device is represented.
4.2 U
ltrason
ic
Vibrati
ng
Plate
The
ultra-
sonic
vibrat-
ing plate implemented in the device is a glass plate 154x81x1.6mm
3
resonating at 60750Hz,
where the half wavelength of the vibration mode is 8 mm. 22 piezoceramics, 14x6x0.5mm
3
, are
mounted on the sides of the plate along a full wavelength, 20 of which are used as motors and 2 as
vibration sensors. Their unglued electrode was split along the nodal line and both halves connect-
Fig. 2. Device structure.
Fig. 3. a, cartography of the ultrasonic vibrating plate. b, Bode diagram of the vibration re-
sponse of the plate. The dashed line at -3dB indicates the bandwidth of the plate up to 400 Hz.
