=Paper=
{{Paper
|id=Vol-1324/paper_1
|storemode=property
|title=Redesigning Input Controls of a Touch-Sensitive Pin-Matrix Device
|pdfUrl=https://ceur-ws.org/Vol-1324/paper_1.pdf
|volume=Vol-1324
}}
==Redesigning Input Controls of a Touch-Sensitive Pin-Matrix Device==
https://ceur-ws.org/Vol-1324/paper_1.pdf
Redesigning Input Controls of a Touch-Sensitive
Pin-Matrix Device
Denise Prescher
Human-Computer Interaction Research Group
Technische Universität Dresden
Nöthnitzer Str. 46, 01187 Dresden, Germany
denise.prescher@tu-dresden.de
ABSTRACT
To support blind people in dealing with graphical user inter-
faces some two-dimensional Braille devices were developed.
In order to ensure an efficient workflow not only the prepara-
tion of content, but also the handling of the device has to be
well usable. In this paper, the redesign of a touch-sensitive
pin-matrix device is discussed. At first, some mock-ups of
additional input controls were designed and evaluated with
six blind users. As the position of the keyboard has been
found to be important for ergonomic aspects, a second user
evaluation was conducted. Based on the results of these two
studies, a proposal for redesigning the device as well as some
basic design recommendations can be given. Furthermore, Figure 1. BrailleDis 9000 with touch-sensitive pin-
the new BrailleDis 7200 is described in detail. matrix area and Braille input keyboard on its front.
Categories and Subject Descriptors
H.5.2 [Information Interfaces and Presentation]: User keyboard should be reduced. For this, the device should of-
Interfaces—ergonomics, evaluation/methodology, haptic I/O, fer gestural input as well as some hardware keys allowing
input devices & strategies, prototyping, user-centered design the user to trigger most important functionalities directly
on the device.
General Terms The BrailleDis 9000, which is introduced by Völkel et al.
Design, Experimentation, Human Factors [7], consists of a touch-sensitive pin-matrix made out of 7200
pins and a Braille input keyboard on its front side (see Fig-
Keywords ure 1). In addition, a special screen reader, called Hyper-
Pin-matrix device, tactile display, Braille keyboard, blind Reader, was developed. With it, blind users are able to in-
users, mock-up test, ergonomic input teract with graphical user interfaces shown on the BrailleDis.
Based on further user studies with the pin-matrix device, e.g.
1. INTRODUCTION [5] and [4], some requirements for the design of the device
In order to enable an efficient usage of graphical user in- arose. For instance, while reading on the pin device users
terfaces for blind people, an ergonomic workstation has to often triggered functions of the HyperReader by activating
be provided. Therefore, two-dimensional Braille displays Braille input keys unintentionally. Furthermore, some users
were developed as they allow to explore graphical content would like to have some more keys which can be freely config-
and spatial relationships of interaction objects. Compared ured. For adapting the interface of the BrailleDis hardware
to single line Braille displays much more information can to these new requirements, a redesign of the device was nec-
be provided simultaneously. To reduce mental and physical essary. In the following, the design of some input control
workload, not only the preparation of content, but also the mock-ups as well as their evaluation with blind users is pre-
handling of the device has to be well usable. For example, sented. Based on the results, basic recommendations for an
unnecessary hand movements between Braille output and ergonomic tactile input device are given.
2. EXTENDING THE BRAILLEDIS WITH
Permission to make digital or hard copies of part or all of this work for
personal or classroom use is granted without fee provided that copies ADDITIONAL INPUT CONTROLS
are not made or distributed for profit or commercial advantage and that There are already numerous input controls that should allow
copies bear this notice and the full citation on the first page. Copyrights for graphical or haptic user interfaces to become more efficient
third-party components of this work must be honored. For all other uses, and intuitive. Taking up these ideas, it can be possible to
contact the Owner/Author. add some new input elements to the BrailleDis. By this
Copyright is held by the owner/author(s).
means, important interaction commands could be triggered
TacTT ’14 , Nov 16 2014, Dresden, Germany. more fast and intuitive.
Therefore, the following mock-ups for extending BrailleDis
9000 should use the mentioned ideas:
1. navigation bar in front of the pin-matrix area
2. TASO modules in front and on the right of the device
3. function keys behind the device
4. cursor keys in the middle of the Braille input keyboard
5. mouse wheels on the right and left of the pin-matrix
area
6. touch sliders5 in front and on the right of the pin-
matrix area
7. touch pad for gesture input on the right of the pin-
matrix area
8. finger mouse6
In Figure 2 all input control mock-ups are shown. Some of
these elements can be combined with others, too.
3. USER EVALUATION
Figure 2. Input control mock-ups evaluated in first To identify the most promising and helpful input controls
user study. and their positioning on the BrailleDis the designed mock-
ups were evaluated by six blind people. In a first step the
different interaction elements were rated separately. In a
Experiences of blind computer users include some special in- second step some new mock-ups, which combine the most
put controls that are common for using with standard Braille suitable input controls, were evaluated based on different
devices. For example, Braillex devices from Papenmeier1 kinds of working tasks.
inclose a joystick-like navigation bar (Easy Access Bar) al-
lowing a fast and intuitive moving of the display. The Vari- 3.1 Evaluation of Different Input Controls
oPro Braille device of Baum2 can be extended with Tactile A study with six blind subjects (three female and three male
Acoustic Side Orientation (TASO) modules. These consist with an average age of 34 years) was conducted to rate the
of vertical and horizontal sliders which can accelerate the applicability of the eight input control elements shown in
navigation on the two-dimensional screen through audible Figure 2. All participants are Braille readers dealing with a
signals. computer several hours a day. All of them have some experi-
ences with the pin-matrix device as they took part in former
Looking at a standard keyboard, function and cursor keys studies or were involved in the HyperReader development.
enrich the simple input of text. Besides, pointing devices,
such as a mouse, are necessary for efficiently dealing with
graphical user interfaces. In the world of sighted users there
3.1.1 Method
are also tools, such as mouse wheels and touch pads, which The test was divided into an introduction, a Wizard of Oz
allow to simplify important scrolling interactions. and a questionnaire phase. At first the HyperReader’s gen-
eral functions were introduced. Then the mock-ups and
Existing two-dimensional tactile displays, such as GWP3 their principles of operation were presented one after an-
and DotView graphics display4 , also use cursor keys or joy- other. Thereby, the user had to fulfill a certain task with
sticks for realizing panning or zooming functions as these the current input control element while thinking aloud7 .
are important in the exploration of graphical user interfaces.
The more often an input command is needed, the easier it The task for mock-up 1 to 6 was to explore a dialog window
should be triggered out. within symbol view of the HyperReader [5], in which the
spatial relationships of all GUI elements were obtained on
1 the pin-matrix device, but all texts were presented in Braille.
F.H. Papenmeier GmbH & Co. KG - Braille
devices, see http://www.papenmeier.de/en/reha- 5
For instance, SoftPot potentiometer sensors (see
division/products/overview-braille-devices.html http://www.spectrasymbol.com/potentiometer/softpot)
2
Baum Retec AG - VarioPro and TASO modules, see could be used to realize some touch sliders.
http://www.baum.de/cms/en/variopro/ 6
Finger mouse could be used to start gesture input instead
3
Maple GWP by Handy Tech GmbH, see of thumb keys of Braille input device, which is utilized at
http://handytech.de/produkte.php?produkt=58&lang=en present.
4 7
DotView Series tacile graphic display devices by KGS, see The think aloud method means that the user verbalizes his
http://www.kgs-america.com/gc.html reasoning and thoughts during a problem-solving task [3].
The exploration should be done from left to right and from Only three of the participants could orient themselves on
top to bottom for getting a feeling of the input control’s us- the screen by using TASO modules as it seems to be unclear
age. In a second task the user should execute some zooming how much of the content was moved. Four of the subjects
functions within the pixel based layout view of the Hyper- think that vertical mouse wheels are not intuitive for scroll
Reader [5]. This task should be done with mock-up 3, 5, 7 operations, particularly for horizontal scrolling. For zoom-
and 8. Depending on the user’s input, the content shown on ing functionalities most of the subjects would prefer gesture
the pin-matrix display was modified by the test supervisor. input on the touch pad. The finger mouse was not liked
That means, he presented some predefined output by button very much as it not only needs getting used to, but also an-
command (Wizard of Oz experiment, see [2]). noys half of the users while reading the content. Regarding
some ergonomic aspects of the mock-ups, five of the six sub-
After each mock-up test some questions about the input con- jects prefer to work on an inclined pin-matrix display as it
trol element were asked, for example about its intuitiveness would be more comfortable. Half of the participants like the
or its positioning on the pin-matrix device. At the end, the keystroke of the Braille input keyboard and also the major
user should rate all mock-ups on a scale from 1 (very bad) to spacing to the outer Braille keys. The other would like to
5 (very good) and he also could give some comments. Fur- have a consistent spacing and either a more smooth or a
thermore, some general questions about ergonomic aspects, more rough keystroke.
for example, relating to the Braille input keyboard or the
slant of the pin-matrix display, were asked.
3.1.3 Discussion
Based on the results, the most suitable input controls can be
3.1.2 Results chosen to extend the BrailleDis 9000. Most of these special
The rating of input control mock-ups given by the six blind controls allow for a more intuitive usage than it is possible
participants was very heterogeneous for both tasks, panning with the Braille input keyboard. Like the function keys it is
and zooming (see Figure 3 and Figure 4). Only the naviga- important and desired by the user, but its key role should be
tion bar was accepted by every subject as it is very common the input of text. For efficiently working on the pin-matrix
for using as panning device on standard Braille displays. device, the goal of the redesign should be to avoid the fre-
Therefore, it is the most intuitive input control for panning quent handling with a standard keyboard. Therefore, the
commands. In contrast, most of the participants want to most important functionalities of graphical user interfaces,
have function keys, but they not seem to be very intuitive such as panning, zooming, text input and moving of focus or
for panning or zooming. Cursor keys could be used for pan- caret, should be available on the pin device. As mentioned
ning intuitively, but they are not sufficient as only possibility above, the navigation bar is the most intuitive tool for real-
for panning. izing panning operations. Also cursor keys were rated very
good, too. Therefore, they can complement the navigation
bar by realizing some further moving operations, e.g. of the
caret. Zooming can be implemented by gesture input. In-
stead of using a separate touch pad, this can be done much
better directly on the pin-matrix as it is touch-sensitive, too,
and it allows the user to keep a reference point. However, as
the gesture recognition has to be started by holding down
one of the Braille keys [6], a simple key pressing or scrolling
on a mouse wheel could be more efficient in some situations
where no reference point is necessary. Therefore, an addi-
tional mouse wheel can also be used on the pin device. Be-
sides these new input controls, the new pin device should be
reduced in its height and inclined in an angle between 10 and
Figure 3. Participants’ rating of input control mock-
15 degrees which allows for a more ergonomic working. Fur-
ups used for panning commands on a scale from 1
thermore, the Braille input keys should be lowered to avoid
(very bad) to 5 (very good).
accidental releases while reading on the pin-matrix. Com-
bining all these requirements and ideas, the new BrailleDis
could be designed as shown in Figure 5.
3.2 Evaluation of Input Control Placement
The first user study showed that the position of the key-
board is important for ergonomic aspects, too. Indeed, the
new prototype is designed for working independently from
a keyboard to some extent. In some situations a standard
keyboard will be necessary nevertheless. For example, some
users are not practiced in using a Braille keyboard or prefer
to use a standard keyboard, especially when writing a lot
of text. Therefore, a second study with five of the six blind
Figure 4. Participants’ rating of input control mock- participants was conducted. Aim of this evaluation was to
ups used for zooming commands on a scale from 1 find out how the keyboard should be placed in different sit-
(very bad) to 5 (very good). uations of working.
function keys
mouse pin-matrix
wheel
navigation
bar
Braille input
keys
cursor keys
navigation bar
pin-matrix
keys
top view side view
Figure 6. Four different mock-up scenarios consist-
Figure 5. Prototype of new BrailleDis design in top ing of pin device, Braille input and standard key-
and side view based on the first user study. board modules.
3.2.1 Method All of these three tasks were given to the user for each of the
In contrast to the first study, the BrailleDis hardware was four different mock-up scenarios. For each task a time slot
not included in the test. Instead, some paper-based mock- of 5 minutes was scheduled in which the subject should give
ups of the pin-matrix display complemented with prototypes his thoughts about the mock-up handling, e.g. ergonomics of
of navigation bar, mouse wheels and function keys were used. arms and hands while reading and interacting. The output
This pin device module was combined with a standard and on the pin device paper prototype was simulated by changing
a Braille input keyboard in four scenarios (see Figure 6): some Braille print-outs with different content (comparable
to Wizard of Oz experiment in the first study). After all,
some questions about the different positions of keyboard and
1. Braille input keys in front of the pin-matrix as pro-
Braille input keys were asked (see Table 1).
posed in the design shown in Figure 5, keyboard is
placed on a sliding carriage above the pin device
2. Braille input is integrated in the function keys, key- Table 1. The participants’ opinion about some pos-
board is placed in front of the pin device sible usage of Braille and standard keyboard
answer
3. Braille input keyboard is divided and placed next to question yes no
the pin device (left and right), standard keyboard is Could you imagine to integrate a standard
placed behind the pin device 2 3
keyboard into the BrailleDis?
4. Braille input keys are in front of the pin-matrix like in Do you need a Braille input keyboard? 1 4
scenario 1, standard keyboard is placed on the right of Would you accept to use a divided Braille
3 2
the pin device in the form of a corner seat scenario input keyboard?
Would you accept to assign the function
4 1
keys with Braille input commands?
The participants’ task was to simulate three different situa-
tions of working:
3.2.2 Results
In Figure 7 the results of user rating of some ergonomic as-
a) skimming over a text and scrolling by keyboard input pects for placing the keyboard in different positions relating
b) writing text by keyboard input and checking it in detail to the pin device are shown. The results for the Braille in-
area (lower line of BrailleDis output, see [5]) put keyboard are shown in Figure 8. Both keyboards got the
highest rating in all ergonomic aspects when they are placed
c) exploration of the pin-matrix content in different view in front of the pin-matrix device. In contrast, placing a key-
types by interacting with Braille input keys and gesture board behind was rated the worst as it is far away from the
input user. On the other side, all participants criticized a long
function keys with integrated Braille input
5 4,6 position of gesture key
4,4 4,6
keyboard
4 3,6 relating to
average rating
pin device:
2,8
3
2,4 on/above (S1) mouse
2,8 wheel
2 2,4 in front of (S2)
2,2 pin-matrix
2 2 2
behind (S3)
1 cursor keys
next to (S4)
reachability ergonomy (writing) ease of use
Figure 7. Average rating of some ergonomic as- navigation bar
pects for placing the keyboard on different positions tray for putting heel of hand
within the scenarios shown in Figure 6. The scale
reached from 1 (very bad) to 5 (very good), n=5.
Figure 9. Prototype of new BrailleDis design in top
view adapted to the results of the second user study.
5 position of
Braille input
4,2
4 3,8 keys relating
average rating
4
3,4 to pin device: a keyboard, e.g. for writing long text, he can put it in front
3
3
integrated in of the pin device or in any other position. Otherwise, in
2,6 function keys (S2)
exploration tasks the use of hardware keys placed on the pin
in front of (S1 + S4)
2
2 2 2
device, namely the navigation bar, mouse wheels as well as
1,2 1,2 behind (S2)
1
1
next to (S3)
function, cursor and gesture keys, is sufficient. This allows to
reachability ergonomy (writing) ease of use individualize the work station much more as the old design
could do.
Figure 8. Average rating of some ergonomic aspects
for placing the Braille input keys on different posi-
4. THE NEW BRAILLEDIS 7200
tions within the scenarios shown in Figure 6. The The design proposed in this paper was included into the
scale reached from 1 (very bad) to 5 (very good), process of constructing a new version of BrailleDis (called
n=5. BrailleDis 7200, see Figure 10). Thereby, all basic ideas were
realized.
distance to the pin-matrix area as it appeared in scenario 2 4.1 Description
(see S2 in Figure 6), too. Instead of the suggested mouse wheels two toggle switches
were embedded. Besides the new input controls, the height
Even before the study it was clear that compromises must of the device could be reduced from about 95 to 60 mm.
be made. Therefore, some further questions about possible Apart from that, the new BrailleDis has a dimension of 410
usages of the Braille and the standard keyboard were asked. x 255 mm and a weight of about 5.5 kg. To minimize the
In Table 1 the opinion of the five participants relating to risk of injury, all edges of the body are rounded. On the
these questions is shown. left and right bottom side there is a notch facilitating the
carrying of the device. Additionally, there are two keyboard
3.2.3 Discussion feet allowing to increase the tilt of the display by about six
Expectedly, positioning the input device in front of the pin- degrees.
matrix device is liked the most by the participants, regard-
less of whether the standard or the Braille input keyboard is
used. For writing text this would allow to have the keyboard
and the detail area (for checking the written words) next to
each other. This shortens the distance which is necessary
for movements of the hand and, therefore, can increase the
efficiency. Nevertheless, if both keyboards would be placed
in front of the BrailleDis, the pin-matrix output would be
far away from the user. Especially for exploration tasks this
is not very comfortable. On the other side, most of the par-
ticipants could do without an extra Braille input keyboard,
but would accept to assign the function keys with text input
commands.
Based on these results, a new prototype design was created
(see Figure 9). For shortening the distance between the user Figure 10. BrailleDis 7200 with touch-sensitive pin-
and the pin-matrix area with all input keys behind it, the matrix area and additional input controls (navi-
Braille input keyboard on its front side was omitted. Only gation bar in the front; cursor key pads, toggle
an ergonomic tray for putting the heel of hand on it was switches and gesture keys on left and right side;
added enabling a more comfortable usage. If the user needs Braille input and function keys behind).
Like the previous version of BrailleDis it also has a 150 x In addition, some further recommendations for an ergonomic
300 mm sized tactile area built up of 720 Braille modules, tactile input/output device can be given:
each consisting of 2 x 5 pins (on the whole 120 x 60 pins).
However, the amount of touch sensors was doubled from
one per module to a total of 1440 sensors [1]. The modules 1. Keys that can be easily triggered should not be placed
are covered with some replaceable gray caps. This allows for directly in front of the reading area to avoid an unin-
more color contrast against the white pins. In the same way, tentional activation.
the black input keys on a silver-gray background can support 2. Unneeded hand movements over a long distance should
some remaining vision abilities of the users. BrailleDis 7200 be avoided.
allows for a refreshment of touch and pins each 50 ms, the
keys can be readout each 10 ms. It is connected via USB 3. Additional input controls placed on the device can in-
and needs a separate power adapter. crease the user’s efficiency. Therefore, they should be
mapped to a specific function as intuitively as possible.
In total, there are 14 standard push-buttons, two toggle
4. A comfortable arm positioning should be ensured. For
switches, two cursor key pads consisting of five buttons and
instance, in front of the reading area some kind of tray
one navigation bar. The cursor key pads are made of one
for putting the heel of hand should be available.
piece (comparable to a swash plate) allowing to press them
in the middle or on one of the four directions. The naviga-
tion bar enables eight input levels, two in every direction. These recommendations arose out of two user evaluations in
To avoid an unintentional activation it is embedded into the the process of extending the BrailleDis 9000 with additional
device and inclined a little bit. Under the wrist rest, which input controls. Depending on the tasks and context of using
narrows to the front, there is some space for placing a key- a special two-dimensional pin-matrix device some other or
board about 2 cm closer to the pin device. supplemental requirements can be necessary.
4.2 First Evaluation 6. ACKNOWLEDGMENTS
Based on the new input controls, a more intuitive concept I thank all blind subjects for participating in the evalua-
for using the HyperReader was developed. For example, tion and giving a lot of valuable feedback. The Hyper-
scrolling operations can be triggered with the navigation bar, Braille project was sponsored by the German Ministry of
zooming functions with the toggle switches and the cursor Economy and Technology (BMWi) under the grant number
keys on the pin device can emulate that of the standard key- 01MT07004.
board. To allow for a first evaluation of the new BrailleDis
7200, some questions about the device were given to twelve
visually impaired users after they had taken part in a user
7. REFERENCES
[1] Data sheet of hyperbraille display 7200, Metec AG,
study dealing with the HyperReader on this device. Seven
Stuttgart, Germany. URL =
of the participants already had some experiences with the
http://web.metec-ag.de/display%207200%20en.pdf
previous version, namely BrailleDis 9000.
(last access on 21-august-2014).
Some important results of the interviews are presented in [2] N. O. Bernsen, H. Dybkjær, and L. Dybkjær. Wizard of
the following. For seven users the height of the device is oz prototyping: How and when. Proc. CCI Working
adequate, while five of the participants think it is still too Papers Cognitive Science and HCI, Roskilde, Denmark,
high. The form of the device is liked by ten users, only 1994.
two people think it is too chunky. Eleven participants think [3] M. E. Fonteyn, B. Kuipers, and S. J. Grobe. A
that none of the input controls is needless. The input by description of think aloud method and protocol
navigation bar, cursor key pad and toogle switches is rated analysis. Qualitative Health Research, 3(4):430–441,
as easy. Only the handling of the function keys appears a 1993.
bit more complicated as these are farthest from the user. [4] D. Prescher, O. Nadig, and G. Weber. Reading braille
In summary, the BrailleDis 7200 was liked by the users and and tactile ink-print on a planar tactile display. In
could allow for a more intuitive handling of the HyperReader Proceedings of the 12th international conference on
functionalities. Computers Helping People with Special Needs, pages
482–489. Springer, 2010.
5. CONCLUSION [5] D. Prescher, G. Weber, and M. Spindler. A tactile
Depending on the current working task there are very dif- windowing system for blind users. In Proceedings of the
ferent requirements a two-dimensional Braille device has to 12th international ACM SIGACCESS conference on
fulfill. For example, in reading tasks the pin-matrix area Computers and accessibility, pages 91–98. ACM, 2010.
showing the content has to be close to the user. Thereby, im- [6] M. Spindler, M. Kraus, and G. Weber. A graphical
portant functions for zooming and panning should be avail- tactile screen-explorer. In Proceedings of the 12th
able directly on the device for avoiding hand movements international conference on Computers Helping People
over a long distance. In contrast, writing tasks require a with Special Needs, pages 474–481. Springer, 2010.
comfortable access to the keyboard, either a Braille input [7] T. Völkel, G. Weber, and U. Baumann. Tactile graphics
or a standard keyboard. Besides, the distance between key- revised: The novel brailledis 9000 pin-matrix device
board and Braille content area for proofreading should be with multitouch input. In Proceedings of the 11th
minimized, too. Keeping this in mind, allowing the adapt- international conference on Computers Helping People
ability of a workstation is very important. with Special Needs, pages 835–842. Springer, 2008.