\newcommand{\degr}[1]{{$#1^\circ$}\xspace}
+% P values
+\newcommand{\p}[1]{\mbox{$p<#1$}}
+\newcommand{\pEq}[1]{\mbox{$p=#1$}}
+\newcommand{\pSup}[1]{\mbox{$p>#1$}}
+
+\newcommand{\f}[4]{\mbox{$F_{#1,#2}=#3$}}
+\newcommand{\fone}[2]{mbox{$F_{#1}=#2$}}
+
+\newcommand{\chisq}[2]{\mbox{$\chi^2\left(#1\right)=#2$}}
+\newcommand{\etasq}[2]{\mbox{$\eta^2_G=#1$}}
+
% ANOVAS
\newcommand{\anovat}[0]{{\small {\sc Anova}}\xspace}
-\newcommand{\owanova}[3]{{\small $F_{#1}\!=\!#2$, $p\!=\!#3$}} % One-Way Anova : DF, DFDen, F et p
-\newcommand{\anova}[4]{{\small $F_{#1,#2}\!=\!#3$, $p\!=\!#4$}} % Anova classique : DF, DFDen, F et p
-\newcommand{\anovae}[5]{{\small $F_{#1,#2}\!=\!#3$, $p\!=\!#4$, $\eta^2_G\!=\!#5$}} % Anova classique : DF, DFDen, F et p
-
-\newcommand{\anovas}[3]{{\small $F_{#1,#2}\!=\!#3$, $p\!<\!0.0001$}} % Anova "s-ignificatif" : DF, DFDen, F, en considerant que p < 0,0001
-\newcommand{\anovase}[4]{{\small $F_{#1,#2}\!=\!#3$, $p\!<\!0.0001$, $\eta^2_G\!=\!#4$}} % Anova "s-ignificatif" : DF, DFDen, F, en considerant que p < 0,0001
-\newcommand{\anovapi}[4]{{\small $F_{#1,#2}\!=\!#3$, $p\!<\!#4$}} % Anova "s-ignificatif" : DF, DFDen, F, en considerant que p < #4
-\newcommand{\anovapie}[5]{{\small $F_{#1,#2}\!=\!#3$, $p\!<\!#4$, $\eta^2_G\!=\!#5$}} % Anova "s-ignificatif" : DF, DFDen, F, en considerant que p < #4
-\newcommand{\anovap}[3]{{\small $F_{#1,#2}\!=\!#3$, $p\!>\!0.05$}} % Anova "non s-ignificatif" : DF, DFDen, F, en considerant que F > 1 et p > 0,05
-\newcommand{\anovans}[3]{{\small $F_{#1,#2}\!=\!#3$, ns}} % Anova "non s-ignificatif" : DF, DFDen, F, en considerant que F < 1
-\newcommand{\chis}[1]{{\small $\chi^2\!=\!#1$}} % chi square seul
-\newcommand{\chisquare}[3]{{\small $\chi^2\!\left(#1\right)\!=\!#2$, $p\!=\!#3$}} % chi square classique : DF, chi^2, et p
-\newcommand{\chisquares}[2]{{\small $\chi^2\!\left(#1\right)\!=\!#2$, $p\!<\!0.0001$}} % chi square "s-ignificatif" : DF, chi^2, avec p < 0,0001
-
-\newcommand{\p}[1]{{\small $p\!<\!#1$}}
-\newcommand{\pEq}[1]{{\small $p\!=\!#1$}}
+\newcommand{\owanova}[3]{{\small \fone{#1,#2}, \pEq{#3}}} % One-Way Anova : DF, DFDen, F et p
+\newcommand{\anova}[4]{{\small \f{#1,#2,#3}, \pEq{#4}}} % Anova classique : DF, DFDen, F et p
+\newcommand{\anovae}[5]{{\small \f{#1,#2,#3}, \pEq{#4}, \etasq{#5}}} % Anova classique : DF, DFDen, F et p
+
+\newcommand{\anovas}[3]{{\small \f{#1,#2,#3}, \p{0.0001}}} % Anova "s-ignificatif" : DF, DFDen, F, en considerant que p < 0,0001
+\newcommand{\anovase}[4]{{\small \f{#1,#2,#3}, \p{0.0001}, \etasq{#4}}} % Anova "s-ignificatif" : DF, DFDen, F, en considerant que p < 0,0001
+\newcommand{\anovapi}[4]{{\small \f{#1,#2,#3}, \p{#4}}} % Anova "s-ignificatif" : DF, DFDen, F, en considerant que p < #4
+\newcommand{\anovapie}[5]{{\small \f{#1,#2,#3}, \p{#4}, \etasq{#5}}} % Anova "s-ignificatif" : DF, DFDen, F, en considerant que p < #4
+\newcommand{\anovap}[3]{{\small \f{#1,#2,#3}, \pSupp{0.05}}} % Anova "non s-ignificatif" : DF, DFDen, F, en considerant que F > 1 et p > 0,05
+\newcommand{\anovans}[3]{{\small \f{#1,#2,#3}, ns}} % Anova "non s-ignificatif" : DF, DFDen, F, en considerant que F < 1
+\newcommand{\chis}[1]{{\small \mbox{$\chi^2=#1$}}} % chi square seul
+\newcommand{\chisquare}[3]{{\small \chisq{#1}{#2}, \pEq{#3}}} % chi square classique : DF, chi^2, et p
+\newcommand{\chisquares}[2]{{\small \chisq{#1}{#2}, \p{0.0001}}} % chi square "s-ignificatif" : DF, chi^2, avec p < 0,0001
\newcommand{\cor}{correlation ($\rho$, p)}
He has a background in cognitive sciences, and what he refers to as input and output is the opposite of the convention I use.
This chapter will cover haptics as a way to stimulate the sense of touch.
-The sense of touch is the primary sense of newborns.
-Their vision takes months just to perceive shapes and colors.
-Therefore they start exploring the world by touching with their hands and mouth.
-Several years are necessary to reach their adult visual accuracy.
-However, children quickly use vision as their primary source of information.
-\fixme{trouver ref pour tout ça}
+%The sense of touch is the primary sense of newborns.
+%Their vision takes months just to perceive shapes and colors.
+%Therefore they start exploring the world by touching with their hands and mouth.
+%Several years are necessary to reach their adult visual accuracy.
+%However, children quickly use vision as their primary source of information.
+%\fixme{trouver ref pour tout ça}
There is a large diversity of haptic sensations and perceptions.
-When I use vocabulary related to haptic, I use the definitions Oakley \etal wrote in this article~\cite{oakley00}.
+When I use vocabulary related to haptic, I use Oakley \etal's' definitions~\cite{oakley00}.
As far as I know, this was the first time these terms were clearly defined in the Human-Computer Interaction community.
Following these definitions, \defword{haptics} is a general term that refers to anything related to the sense of touch.
\defword{Kinesthetic} refers to the feeling of motion, resulting from sensations produced in muscles, tendons, and joints.
This is the system side, out of the physical world.
Here we consider, relatively speaking, that the software is the computer's mind.
These commands activate various kinds of actuators to produce a mechanical effect.
-Since there is a dichotomy between kinesthetic and tactile sensation, there is also a dichotomy between force-feedback and tactile systems.
+Since there is a dichotomy between kinesthetic and tactile sensation, there is also a dichotomy between force-feedback and tactile systems.
Force-feedback systems typically use one of two major types of controls.
The usual way to control a force-feedback system is to measure motion and compute a force, this is called \defword{impedance control}~\cite{ruspini97-2}.
This is the most common technique, mostly because there are many easy and cheap ways to measure motion.
A function that computes a force depending on movement is called a force model.
Other systems sense forces and compute an output motion, this is called \defword{admittance control}~\cite{vanderlinde02}.
-This is less used because forces are harder to measures, especially if precision is required.
+This is less used because forces are harder to measure, especially if precision is required.
The advantage of such systems is their much higher stiffness than inductance systems.
It is also worth mentioning Electrical Muscular Stimulation (EMS) which directly stimulates muscles with electrical signals~\cite{tamaki10}.
Each pin is controlled individually, either up or down.
\defword{Variable friction} technologies change the perceived friction of a surface.
Two methods can produce this effect: \defword{electro-vibration} and \defword{squeeze-film effect}.
-The electro-vibration mechanism uses a high voltage (hundreds to thousand volts) to stick the user onto the interactive surface~\cite{strong70}.
-The signal is a sinusoid (even though other shapes are possible), with controllable amplitude and frequency~\cite{bau10}.
+The electro-vibration mechanism uses a high voltage (hundreds to thousand Volts) to stick the user's finger onto the interactive surface~\cite{strong70}.
+The signal is a sinusoid with controllable amplitude and frequency~\cite{bau10}.
The squeeze film effect uses a high-frequency signal (tens of thousand Hertz) that we cannot perceive directly.{}
This vibration creates an air cushion between the finger and the surface so that this surface feels smoother~\cite{biet07}.
-
Finally, there are non-contact tactile technologies that transmit tactile feedback through the air.
One technique uses an array of ultrasound actuators.
The interferences of ultrasound waves create a stress field that triggers the sense of touch~\cite{hoshi10,carter13}
The objective when producing a physical effect with a haptic system is to stimulate the users' sense of touch.
On \reffig{fig:hapticpath}, I separate the pure sensing part from the interpretation part.
The body has sensors in the skin, muscles, tendons, and articulations~\cite{bolanowski88}.
-These sensors are different type sof nerves and similarly to electronic sensors, they transform physical effects into electric signals.
+These sensors are different types of nerves and similarly to electronic sensors, they transform physical effects into electric signals.
This document will not cover this in detail.
The interested reader will find explanations in many Ph.D. manuscripts about haptics~\cite{brown07,crossan03,hoggan10,oakley03}, including mine~\cite{pietrzak08}.
-We will just mention cornerstone studies about the perception of touch that established the limits of tactile perception~\cite{goff67}, which depend on the body part, sex, and laterality~\cite{weinstein68}.
+We will just mention cornerstone studies about the perception of touch that established the limits of tactile perception~\cite{goff67}.
+Interestingly, the perception of touch depends on the body part, sex, and laterality~\cite{weinstein68}.
In particular, the authors established the range of frequencies humans can perceive (roughly up to 1000Hz), with a peak perception around 250Hz.
This study is the motivation why most vibrotactile systems use a 250 Hz signal, sometimes modulated with another frequency~\cite{brown06}.
My opinion on this choice is that other frequencies provide different and interesting sensations.
Gibson showed that people are more efficient at tactile exploration with active touch~\cite{gibson62}.
This means the brain combines sensations with other information, including exploratory movements that resulted in these sensations.
Given the diversity of touch sensations: weight, shape, temperature, etc., Lederman and Klatzky describe specific exploratory procedures that enable people to perceive these types of sensations~\cite{lederman96}.
-This chapter will not get into details regarding the relation between action and perception since we will cover this specific topic in chapter~\ref{chap:loop}.
+This chapter will not get into details regarding the relation between action and perception, and the \reffig{fig:hapticpath} ignores this phenomenon.
+However we will cover this specific topic in chapter~\ref{chap:loop}.
-\begin{figure}[!b]
+\begin{figure}[htb]
\definecolor{cellred}{rgb} {0.98,0.17,0.15}
\definecolor{virt}{rgb} {0.98,0.17,0.15}
\definecolor{cellblue}{rgb} {0.17,0.60,0.99}
\begin{figure}[htb]
\tikzexternalenable
\def\sx{1mm}
- \def\sy{0.4mm}
+ \def\sy{0.6mm}
\begin{tikzpicture}
%\draw[loosely dotted] (0,-1) grid (17,1);
\draw[x=\sx,y=\sy, ultra thick]
\begin{figure}[htb]
\tikzexternalenable
\def\sx{1mm}
-\def\sy{0.4mm}
+\def\sy{0.6mm}
\begin{tikzpicture}
%\draw[loosely dotted] (0,-1) grid (17,1);
\draw[x=0.5mm,y=\sy, ultra thick]
\node[minimum width=3.0cm, minimum height=1cm,fill=cellred, text=white,text width=3.5cm, align=center, rounded corners=2ex, outer sep=0](#1) {#2};
}
\newcommand{\mediumbluecell}[2]{
- \node[minimum width=2.0cm, minimum height=5mm,fill=cellblue, text=white,text width=2.0cm, align=center, rounded corners=2mm, outer sep=0](#1) {\footnotesize #2};
+ \node[minimum width=2.0cm, minimum height=6mm,fill=cellblue, text=white,text width=2.0cm, align=center, rounded corners=2mm, outer sep=0](#1) {\footnotesize #2};
}
\newcommand{\mediumredcell}[2]{
- \node[minimum width=2.0cm, minimum height=5mm,fill=cellred, text=white,text width=2.0cm, align=center, rounded corners=2mm, outer sep=0](#1) {\footnotesize #2};
+ \node[minimum width=2.0cm, minimum height=6mm,fill=cellred, text=white,text width=2.0cm, align=center, rounded corners=2mm, outer sep=0](#1) {\footnotesize #2};
}
\newcommand{\smallbluecell}[2]{
- \node[minimum width=8mm, minimum height=4mm,fill=cellblue, text=white,text width=8mm, align=center, rounded corners=1mm, outer sep=0](#1) {\scriptsize #2};
+ \node[minimum width=8mm, minimum height=5mm,fill=cellblue, text=white,text width=8mm, align=center, rounded corners=1mm, outer sep=0](#1) {\scriptsize #2};
}
\newcommand{\smallredcell}[2]{
- \node[minimum width=8mm, minimum height=4mm,fill=cellred, text=white,text width=8mm, align=center, rounded corners=1mm, outer sep=0](#1) {\scriptsize #2};
+ \node[minimum width=8mm, minimum height=5mm,fill=cellred, text=white,text width=8mm, align=center, rounded corners=1mm, outer sep=0](#1) {\scriptsize #2};
}
\tikzexternalenable
\begin{tikzpicture}
In section~\ref{sec:activibe}, I will address another issue with the semantic level of the haptic vocabulary.
Most, if not all, haptic devices we use daily such as mobile phones or smartwatches have low-quality vibrotactile actuators, and typically only have one.
-In this study I will discuss there we addressed two challenges.
+In this study we addressed two challenges.
First, we designed Tactons with an off-the-shelf smartwatch, which has only one actuator of poor quality.
Second, we studied the interpretation of Tactons while users performed their daily activities.
Not only this haptic feedback improves subjective assessment of comfort, but it also improves performance in some cases.
The clicking sensation on the Apple Magic Trackpad is impressive.
It feels like a physical button.
-But when it is powered off, this haptic effect disappears.
+But we notice it is generated with a vibrotactile actuator because this haptic effect disappears when the trackpad is powered off.
Further, studies show that vibrotactile feedback increases typing performance~\cite{hoggan08} on touch keyboards.
I will discuss in \ref{sec:printgets} the design of vibrotactile widgets for a touch dashboard.
I rather discuss the contribution of these studies to the research questions above.
% Haptic variables, vocabulary \texttt{=>} Tactons. Diversity of sensations \cite{lederman87} \texttt{=>} diversity of devices~\cite{seifi19}.
-
% The features are: linguistic/nonlinguistic, analogue/non-analogue, arbitrary/non-arbitrary, static/dynamic\cite{bernsen93a}
-
% 8. Touch language Touch letters, numerals, words, other touch language related signs, text, list, and table orderings.
% Example: Braille
% 18. Real-world touch Single touch representations, touch sequences.
% 24. Arbitrary touch Touch signals of different sorts.
% 28. Touch structures Form fields, frames, grids, line separations, trees.
-
\subsection{Tactons for activity monitoring}
\label{sec:activibe}
\end{table}
Table~\ref{tab:activibedesignspace} summarizes the design space of the vibration pattern sets we developed for our laboratory studies.
-We were interested in knowing which sets of icons were most suitable for representing progression values, and what is the best precision we can obtain using these representations.
+We were interested in knowing which sets of icons were the most suitable for representing progression values, and what is the best precision we can obtain using these representations.
We can clearly see that we did not explore all the possible combinations.
This is mostly due to the limited number of sets we could reasonably evaluate in a laboratory study.
Other dimensions are also relevant, like adding a warning vibration before the actual code, to help users paying attention to the signal while doing their daily activities.
This next contribution is also about the extension of the haptic output vocabulary.
However, contrary to the previous contribution, in this project we studied the output vocabulary for a new kind of tactile technology: programmable friction.
Two different technologies exist for changing the perceived friction of a surface.
-The first one is electrovibration,and uses a high voltage signal on an
-electrode to stick the user's finger on the surface~\cite{strong70}.
+The first one is electrovibration.
+It uses a high voltage signal on an electrode to stick the user's finger on the surface~\cite{strong70}.
Therefore this technology increases the perceived friction.
The second technology leverages the squeeze film effect to reduce the perceived friction~\cite{biet07}.
Technically, the whole surface vibrates at a high frequency, typically tens of thousand Hertz~\cite{amberg11}.
This project has two orthogonal challenges.
First, the technology was new and developed by colleagues with expertise in electrical engineering.
-Therefore we had to work with research prototypes, called STIMTAC, that were evolving in parallel to our own studies.
+Therefore we had to work with research prototypes, called Stimtac, that were evolving in parallel to our own studies.
Second, we had to characterize the tactile sensations produced with this technology in order to define the output vocabulary.
Our results guided the design of the devices and the improvements of the devices enabled better sensations.
The co-evolution of both research activities pushed forward this technology, and it is today developed by Hap2U\footnote{\href{https://www.hap2u.net/}{https://www.hap2u.net/}}.
The connection between the command and what users perceive through their fingers is not trivial though.
This is a typical example of the phenomenon discussed at the beginning of this chapter and illustrated on \reffig{fig:hapticpath}.
-Both the relation between the command and the vibration amplitude and the relation between the amplitude and the sensation of slickness is not linear~\cite{biet07}.
+Both the relation between the command and the vibration amplitude and the relation between the amplitude and the sensation of slickness are not linear~\cite{biet07}.
They are affected by many environmental factors such as air moisture, temperature; surface properties such as material, cleanliness; and fingers aspect: tribology, cleanliness, moisture.
There are differences in performance between prototypes.
Some of these differences are due to their hand-made nature, especially the ceramics gluing.
The first study in which I was involved consisted in evaluating the Just Noticeable Difference (JND) of friction between two adjacent zones.
The objective is to measure the useful resolution of the device.
-With this information, our colleagues can optimize the power consumption of the device, and we can design more complex patterns such as textures.
+With this information, our colleagues could optimize the power consumption of the device, and we could design more complex patterns such as textures.
This work was not published, therefore I will get in more details below with a new statistical analysis.
\paragraph{Methodology}
%We were interested
We used a one-up/two-down adaptative method, which enables a faster convergence to the JND value~\cite{leek01}.
-It consists in reducing the intensity of the signal after two good answers and augmenting it after one wrong answer (\reffig{fig:adaptativeprocedure}).
+This method consists in reducing the intensity of the signal after two good answers and augmenting it after one wrong answer (\reffig{fig:adaptativeprocedure}).
The evaluation of each trial used a 3-alternatives forced-choice (3-AFC).
This means we presented participants with 3 configurations.
One of them had a signal: a difference of friction between the left and right sides of the device.
%A randomly chosen side presented the tested reference level, and the other presented another value that depend on the tested intensity.
The friction was uniform on the whole surface for the two other configurations.
Participants had to tell which configuration presented a signal.
-This procedure reduces the chance level and avoids anticipation biases.
+This procedure reduces the chance level to $33\%$ and avoids anticipation biases.
The position of the signal among the 3 configurations was random, and the side presenting the reference level was random as well.
\begin{figure}[htb]
We estimated the threshold by averaging the values corresponding to the last 10 reversals.
%Participants typically performed between 30 to 50 trials per block.
-We experimented with six reference levels, one for each block.
+We used six reference levels, one for each block.
The values were spread out linearly between the minimum and maximum command (\degr{0} to \degr{180} phase shift).
For the first three values (\degr{0}, \degr{36}, and \degr{72}) we searched the minimum greater value for which participant could feel a step.
For the last three values (\degr{108}, \degr{144}, and \degr{180}) we searched the minimum lower value for which participant could feel a step.
The first one is a sequence of tactile patterns of different types: field, constant, gradient, and random.
Users can most likely identify different parts in the textures, without necessarily locate their boundaries.
The second one repeats a field-type pattern, which is itself a repetition of a shape.
-We can also see it as frequency modulation of two signals.
+We can also see this texture as the frequency modulation of two signals.
We can obviously use the same structure and parameters than with vibrotactile Tactons (see \reffig{fig:lexical} and \ref{fig:syntactic}).
However, the range of friction we can produce with this device is not as wide as the range of vibration amplitudes we can produce with state-of-the-art vibrotactile actuators.
-Therefore we cannot simply translate vibrotactile Tactons into tactile textures so far.
-It would be interesting to perform JND studies with another device that produces a higher quality variable friction feedback, such as devices made by Hap2U\footnote{\href{https://www.hap2u.net/}{https://www.hap2u.net/}}.
+Therefore we cannot simply translate vibrotactile Tactons into tactile textures with current programmable friction devices.
+It would be interesting to perform JND studies with newer devices that produces a higher quality variable friction feedback, such as devices made by Hap2U\footnote{\href{https://www.hap2u.net/}{https://www.hap2u.net/}}.
\begin{figure}[htb]
% \node[x=1mm,y=1mm, anchor=center] () at (160,-36){Set F};
\end{tikzpicture}
\tikzexternaldisable
- \caption[Examples of tactile texture.]{Examples of tactile texture. The first one is a repetition of a field-type tactile pattern. The second one is the combination of a series of various pattern types. It starts with a field, then a constant, a decreasing gradient, a random pattern, another field and finishes with an increasing gradient.}
+ \caption[Examples of tactile textures.]{Examples of tactile textures. The first one is the combination of a series of various pattern types. It starts with a field, then a constant, a decreasing gradient, a random pattern, another field and finishes with an increasing gradient. The second one is a repetition of a field-type tactile pattern. }
\label{fig:tactiletexture}
\end{figure}
Then we compute the number of required dimensions and map each item to a position in space so that the distance between two items is proportional to their dissimilarity.
Finally, we identify clusters of similar patterns.
This method was already successfully used to evaluate Tactons~\cite{enriquez06}.
-\reffig{fig:stimtacpatterns} shows the patterns we evaluated~\cite{potier16}.
+\reffig{fig:stimtacpatterns} shows the patterns we evaluated in our study~\cite{potier16}.
They are made of 5 \textsc{shapes} and 7 \textsc{densities}, for a total of 34 patterns (the lowest density vertical line and squares are identical).
\begin{figure}[htb]
We first worked on the lexical level and evaluated the JND of steps for six reference values of friction.
We observed differences of perception between reference values.
We attributed these differences to both the non-linearity of the mechanical effect resulting from the command and the perception of the mechanical effect.
-As a result, the commands were adapted to the device.
+As a result, the signal commands on the device were adapted.
Then we worked on the syntactic level and proposed definitions of tactile patterns and textures.
We compared the users' estimation of similarity of tactile patterns made of different shapes and densities implemented with a Stimtac device and with coated paper cards.
-We observed differences in grouping strategy across conditions, suggesting differences of perception between the haptic device and the paper cards.
-It means that results about perception made with research prototypes are hard to generalize and that using physical props is an alternative to evaluate best-case scenarios.
+We observed differences in grouping strategy across conditions, suggesting differences of perception between the haptic device and a similar physical representation of textures.
+It means that results about perception made with research prototypes are hard to generalize.
+Using physical props is an alternative to evaluate best-case scenarios.
+However, devices cannot reproduce all haptic properties of the physical world.
%\cite{bertin83}
projects / hdr.git / commitdiff