|
|
Christophe Ramstein
Center for Information Technologies Innovation (CITI)
1575, Chomedey Blvd, Laval (Qc), Canada
Email: cramstein@citi.doc.ca
This article describes design issues for a bi-dimensional single cell braille display, called Pantobraille, combining a standard braille cell with a force feedback device developed as part of the CITI's PC-Access project. The Pantobraille, with a 10x16cm workspace, allows the user to place the pointer on a graphical interface, to perceive forms and textures using the sense of touch, and to read braille text on a bi-dimensional page. In order to determine the usability of such a device and to have a better understanding of the issues that may arise when manipulating it for actual interactive tasks, two visually handicapped persons were asked to use the device to follow reading patterns with one or two hands. Reading performance and comfort with the Pantobraille remain inferior to standard braille displays but significant improvments were observed while performing the complementary pointing and reading tasks using both hands.
Single cell braille display, haptic interface, force feedback device, braille display
As graphical user interfaces are becoming a standard in human-computer interaction, the problem of making computers accessible to the blind arises. Multimodal solutions consist of translating or enhancing graphical events with voice synthesis, non-speech sounds [15,19], force feedback [1,9,23] and tactile stimulations (e.g. braille). Incidentally, only a small percentage of blind people read braille; this is due to the availability of inexpensive text-to-speech synthesis systems on the market, but also to the diffculty many people have in learning it along with the prohibitive cost of braille displays. It is therefore important to develop new braille displays that are cheaper, easy to use and easy to integrate with graphical interfaces. This paper presents the conceptual and ergonomic elements of a single cell braille display which combines a braille module with a force feedback device: the Pantobraille.
A number of devices and systems have been designed to carry information through the skin. Among these, braille, vibrotactile stimulators (e.g. Optacon®) and computer mouses with tactile feedback [1,9] are attractive options because they demonstrate that the skin, and more specifically digital extremeties, can recognize complex shapes and quickly analyse the information.
Braille, a method of reading and writing invented by Louis Braille in 1840, defines the associations between a dot matrix, the sense of touch, and alphabetical characters. Each configuration of points produces a specific sensation on the finger tips.
Braille Slow Readers Fast Readers
Technology (<50wpm) (>50 wpm)
Printed m=41.4 m=99.6
braille (s=13.4) (s=22.9)
One cell m=12.0 m=46.1
(s=2.5) (s=19.2)
20 cells m=31.8 m=70.9
(s=9.8) (s=21.6)
40 cells m=35.2 m=83.3
(s=9.9) (s=25.5)
80 cells m=30.4 m=74.6
(s=8.2) (s=18.4)
Table 1. Braille mean (m) reading speed in words per minute (wpm) with standard deviation (s) according to the braille display used and the type of reader [21].
Many techniques can be used to produce braille text. Printed braille (on paper or plastic) consists of pages on which the braille dots - half spheres of about one millimetre in diameter - are embossed. Computerized braille information reading is made possible electronic braille displays (e.g. Navigator). These displays have one or more lines of 20, 40 or 80 braille cells; each cell is a mechanical device representing an 8- dot matrix (2 columns and 4 lines). The dot matrix cells of braille displays are programmable: the visually handicapped user scrolls through the text from one line to the next, in ascending or descending fashion. Although standardized and commonly used, these displays are both expensive and cumbersome.
Optacon® is a portable electronic device that converts visual images into tactile forms, which detectable using the finger-tips. In normal use by visually handicapped users, a camera moves across the printed page and the image is transmitted to the tactile array, which in turn reproduces the shape of letters (or graphics) in a vibrating pattern that can be felt by the finger-tip. Therefore, the braille language is not incorporated in Optacon technology [4].
In order to reduce the cost and size of standard electronic braille displays, the single cell approach replaces a 20-, 40-, or 80-cell braille line by one cell. Two alternatives have been considered: one is static and the other dynamic.
In the static case, the cell is steady and the words formed by the braille characters succeed one another under the finger. With the help of function keys, the characters' scrolling speed and direction are controlled manually. Although this static cell is less cumbersome and less costly to produce, it makes the reading process comparatively less efficient (see table 1) and more complex than the printed paper or larger displays, where the "same finger" controls speed, direction and reading.
Thus, having in mind the objectives of reducing cost and size but keeping the issues of performance and comfort in mind, Parreno et al. [20] designed a dynamic single cell braille display instead of a static one. This braille line consists of a standard braille cell incorporated into a carriage that, under the action of the reading finger, is able to move along a couple of straight guides. In order to improve the perception of laterally motions, Fricke et al. [8] designed a nonvibrating display comprising an array of tactile pins having a spacing of about 1 mm, at least in the x-direction. A scanned braille dot, for example, is presented by a smooth lateral "wave" of lifted and lowered pins.
Braille reading speed is typically three to four times slower than visual reading, with the average reading speed about 70 wpm (words per minute), but with considerable variation across readers [7]. Unlike visual reading, in which the two eyes always move together, braille may be scanned with either one or both hands. Most readers prefer to use two hands rather than one and bimanual reading is faster than unimanual reading [5]. When two hands are used, they may move together or independently: a good analysis of reading patterns can be found in [18].
Up until now, most research has concentrated on printed braille material, and not on electronic braille displays. The operation of the latter involves one hand controlling the terminal while the other explores the tactile field and adjusts the reading speed through special keys or, alternatively, using both hands to read the braille display (1, 20, 40 or 80 cells). Portalier [21] studied reading process on electronic braille displays and concluded that reading speed shows no appreciable increase when the reading range is either too large or too small: the limit seems to be the 40-cell display with 83.8 wpm for the fast readers and 35.2 for the slow readers (see table 1). However, reading on the printed page is still faster than on electronic displays (99.6 wpm for the fast readers and 41.4 wpm for the slow readers). The distinction of fast and slow readers is useful since, as stated above, there is considerable variation across readers.
Because multiple mechanical receptors are involved in the tactile perception mechanism, the displacement of the finger over a form or texture improves perception and facilitates acknowledgement [7,14]. As a result, in the single cell case, the finger will not feel the transition from one character to the next, but only the configurational changes in the cell's pins according to the following character. Since the finger is not moving on the braille characters, speed performance will of course be slower than on a 20-, 40- or 80-cell braille displays (46.1 wpm for fast readers and 12.0 wpm for slow readers).
Though a mesure of reading performance is not provided, one would assume that a dynamic single cell display (e.g. [20,8]) provides the reader with better performance than the static one-cell display: in effect, the reader regains the ability to move the finger over a line, thus the use of kinaesthetic perception of position, and direct motor control on reading direction (left or right) and speed. However, one assumes that performance and comfort would be significantly inferior to larger braille displays (20, 40, 80 or printed braille).
The term haptic refers to a combination of the tactile and kinaesthetic perception senses. The tactile sense brings awareness of a given stimulus over the body with the help of receptors found under the skin and linked to the nervous system whereas the kinaesthetic sense provides information regarding the position and displacement of the limbs [14].
An haptic device typically consists of a mechanical linkage in the form of a joystick or exoskeleton which couples the human operator to a source of mechanical power -- either electromagnetic or electrohydraulic actuators. The computation engines govern the behavior of the actuators and linkages of the haptic device as a function of kinematic and force or position measurments, according to algorithms and equations that describe a model to be simulated. One main disctinction between haptic devices and tactile devices should be mentioned: braille displays or vibrotactile arrays (e.g. Optacon) stimulate the tactile sense but don't have sensors - only small actuators - and stimulations occur on the skin and do not stimulate the perception of position or motion of the limbs (i.e. kinaesthetic sense).
A number of haptic devices have been designed [12,13,16,17] (see [10] for a thorough discussion of haptic device requirements). Despite their respective qualities, none of them are adapted to direct manipulation on a graphical interface (e.g. small work space; not enough force feedback; too expensive; mechanical fragility). The Pantograph is a haptic pointing device intended for direct manipulation [11,24]. Its mechanical structure is based on a 5 bar linkage that guarantees stability, little friction and provides a comfortable 10x16cm workspace accessed by a button (the knob). By moving the knob, the four bars are dragged in a correlative and unidirectional way. At each moment, the ultimate position of the knob in the 0xy cartesian space is calculated from the angular positions recorded by two potentiometers, while two powerful motors, coupled to the inner bars, allow numerical control over the knob's behaviour, thus giving the user a haptic sensation of a physical nature (e.g. elasticity, rigidity). In order to generate force on the knob, the motors need to be supplied with adequate current, which in turn creates the torque that is transmitted to the rigid segments, synthesizing the mechanical force on the knob, with 10 Newton peak force.
The Pantograph has several potential applications. Its main application is a system called PC-Access, which is being developed at Industry Canada's CITI. Its aim is to provide visually handicapped people with a multimodal access to graphical interfaces [22,23]. In PC-Access, the Pantograph has two complementary functions: first, it serves as a device for pointing, moving, selecting and re-sizing icons, windows and pop-up menus; secondly, it translates objects and actions occurring in a graphical interface into objects which are perceptible through the sense of touch (complementing sound stimulations). We observed (see Dufresne & al [6]) that bimodality (non-linguistic sounds combined with haptic feedback) provided users with more comfort and performance than monomodal feedback (non-linguistic sounds only; or haptic feedback only).
The Pantograph is also being applied as an active help tool in computer-assisted learning of traditionnal graphical user interfaces. The goal of adaptive systems is to increase system suitability for specific tasks; facilitate handling the system for specific users, thus enhancing user productivity; and optimize workload and increase user satisfaction. Traditional adaptivity in the form of an adaptive system is based on the assumption that the system is able to adapt itself to the wishes and tasks of the user through an evaluation of user behaviour. Force feedback technology offers a new performance support system paradigm: since a force feedback device is used as the pointing device, it allows a system to physically guide the user's hand [25]. Other application projects are in progress: namely in microgravity environments and surgical environments.
A variety of applications have demonstrated the implementation of simple haptic objects. Essentially, these consist of walls modeled as massless plates backed by spring with stiffness and damper with viscosity [3]. It has been demonstrated that virtual objects can encode volume data in force sensation. The basic idea described in [13] is to map voxel data to force and/or torque: modeling methods are described based on scalar data and vector/tensor data. A more complex modeling method described in [2] is based on a physical modeling approach: a physical object is analysed and represented as a combination of masses connected through different categories of physical connexions (e.g. spring, damper, conditional link). Efficient algorithms were designed to control the simulation in real-time and to restitute these objects through visual, auditory and haptic feedback.
The Pantobraille integrates the Pantograph and tactile technologies. It offers three complementary functions in a human-computer interface: the Pantograph's two basic functions - pointing and force feedback - and the additional function of tactile stimulation for braille text or low-resolution graphics on bi-dimensional pages.
The initial objective was to design a device capable of reproducing a haptic finger-tip feedback similar to that produced by a printed braille page or more generally by a texture coating a two- or three-dimensional physical object. Let us consider following example: with eyes closed, use your hand to explore a physical object (e.g. clothes, a fruit, the keyboard of your computer). Normally, you will have a dual sensation: the micro-sensation of texture and the macro-sensation of the shape of the object. When you mentally separate the two feelings, it becomes obvious why tactile and haptic technologies are complementary and must be combined. What follows are the initial requirements underlying the Pantobraille design:
provide haptic stimulations to represent 2- or 3D object shapes; at least 10 Newton peak force; high resolution position (i.e. 1/50 of a millimetre) with low analogic noise; large workspace
provide tactile stimulations to simulate textures (i.e. braille) with discrete mechanical dot-arrays; at least 10Hz refreshing frequency ; 1 Newton peak force per dot; allow the perception of transition between two disjointed locations
enable movement of the finger over the workspace as one would move the finger over a surface with limited technical constraints. Achieves a reading performance which is similar to comparable technology.
A piezoelectric braille module (model SC3, manufactured by KGS Corporation) is placed at the outer edge of the Pantograph's extremity, in free articulation on the knob's axis. The SC3 model is presently one of the smallest: its vertical architecture has two juxtaposed cells of 8 pins each (2 columns and 4 lines). The spacing of the cell's pins is 2.4mm, generating a minimal force of 10gf. The diameter of a pin is 1.3mm, while the maximum elevation of each pin is 0.8mm. The size of the entire module is 12.8x42x 26.5mm; its weight is 10.5g.
To integrate the braille module into the Pantograph, the SC3 module was installed in a plastic frame that is attached to the back of the Pantograph's knob. The position of the braille module is assimilated to the knob's position and calculated by the Pantograph. The device allows forces to be synthesized within the module itself and makes it possible to program rapidly (>100Hz) braille characters on one or the other of the two cells included in the module.
Each of the Pantobraille's display features can be programmed: the distance between two braille characters; the number of characters per line; the number of lines par page; haptic sensations that translate text structure and facilitate reading.
A 40-cell on-line braille display has a width of 256mm, which corresponds to 40 juxtaposed cells each of which is 6.4mm wide. The distance between two cells is constant. The Pantobraille enables one to program the distance between two characters and two lines, as the distances are entirely virtual and determined at run-time. Within the 10x16cm workspace, a text of 25 6.4mm characters can be programmed on each line; one can also increase the density and program 40 characters with a 4mm interval between each; or vary the distance between characters using text alignment of the right or left margin (this function, while commonly used in word processing by sighted people, is not available to visually handicapped people using standard braille displays).
With a 2D monocell braille display, following a line steadily can be a problem as the braille module is free and only the pins' clicking really gives any information on changes between characters and lines. Unlike the printed page, where multiple fingers and both hands can be used, the braille display user can only identify or set locations with one finger; following a single line from beginning to end can thus be difficult. It is, therefore, important to supply a haptically detectable mark around the current line so that the braille module can follow it without slipping into the next. We have made the use of just such a function optional in the Pantobraille. If this option is chosen, each line is associated with a "gutter", or block of text currently being treated: the cursor's current position is thus limited to this block and moves within it. This keeps the cursor from moving off the current line, while moving from one line to the next (up or down) requires the application of an extra bit of strength. Also, programming small transients to define each character is beneficial because the tactile information produced by the braille cell is reinforced. This simulation principle is based on the physical metaphor and modelling approach described in [22].
This novel approach, which combines both tactile and haptic technologies using the Pantobraille concurrently as a pointing and braille reading device was previously unexplored. Therefore, we permitted (and continue to consider) several hypotheses and avenues of investigation. Some of the questions raised were: Is it easy for a braille reader to use the device? Are there multiple reading patterns and, if so, which one is the most efficient and comfortable? Then, based on the answers to these questions, how can we improve the initial design?
Two visually handicapped people were sollicited to perform experiments on the Pantobraille. One reads braille with the index finger of his right hand. The second reads braille fluently with both hands. Neither had used the Pantobraille before and therefore had no previous training on it. In order to gather some quantitative information concerning reading speed and ease of use, subjects were asked to read a passage of English technical prose (210 words in length) using a traditional 40-cell braille display and then read the same passage using the Pantobraille. For the latter, three possible reading patterns were analysed which focused on the use of one or two hands (see figures)
Scenario A: monomanual use. The cell is mounted on the Pantograph, and the reader is instructed to "Use only one hand to read the braille text."
Scenario B: concentrated bimanual use. The cell is mounted on the Pantograph, and the reader is instructed to "Use both hands to read the braille text. "
Scenario C: dedicated bimanual use. The braille cell is dismounted from the Pantograph and set aside. The reader is instructed to "Use two hands to read the braille text: one hand to move the cursor within the text using the Pantograph and the other hand to read the braille information on the cell"
Assuming that monomanual reading would achieve slower speeds than bimanual reading, and in an effort to reduce the training effect, different orders for the three scenarios were proposed to the two readers: ABC was proposed to one reader while ACB was proposed to the other. The density of characters per line was 45 (3.5mm between each character), half the amount of the standard displays (6.4mm) but matching the user's intuitive preferences, who had the choice of 25, 45 or 60 words per line. The Pantobraille information feedback is reinforced by gutters defining the lines and vertical transients between characters. Before the user read a paragraph, the experimenter adjusted the physical parameters delimiting the characters and lines to the users' liking.
Both users are right-handed. In Scenario A, they held the Pantobraille cell between the thumb and middle finger of the right hand and moved their index finger over the braille text. The other hand was not used. While reading the text, the users mentioned that physical gutters and transients between characters helped them to follow lines steadily and to enhance the perception of successive characters. Comfort was reduced when the non-dominant (i.e. left) hand was used to do the reading task. At the end of Scenario A, users reported muscular fatigue.
In Scenario B, users worked with both hands on the cell mounted on the Pantograph: the right hand (dominant hand) being used to move the cell and the left hand index finger to read the braille text. The use of both hands seemed more natural to the subjects. However, comfort was reduced when the roles of the two hands were reversed, i.e. if dominant hand is used for reading while the other is used for moving the cell. No fatigue or discomfort was reported.
In Scenario C, the cell was removed from the Pantograph and put aside. Users moved the Pantograph (without the cell) with the right hand; lines and characters were still reinforced by force feedback. The left hand index finger read from the braille cell, which was no longer attached to the Pantograph but set aside. Subjects could not tell me if they read faster than in Scenario B but they felt more at ease with this scenario. Comfort is reduced if the roles of the two hands are reversed, i.e. if the dominant hand is used for reading while the other is used for moving the cell.
Scenario Monomanual Bimanual
reader reader
(ACB) (ABC)
Braille 30.6 64.3
Display
A 7.0 9.3
B 11.8 17.6
C 13.1 18.3
Table 2. Mean speeds in words per minute (wpm) according to different reading scenarios with the best reading pattern for each scenario, i.e. dominant hand is used for moving cell.
Table 2 shows means reading speed in words per minute for the two readers: recall that the first subject uses one finger to read on a daily basis and the other reads fluently with both hands. Both began with the standard 40-cell display followed by the three scenarios: A, B and C. For the monomanual reader, the order of scenarios was ACB whereas ABC was proposed to the second reader. The values shown in Table 2 correspond to the optimal reading patterns: the use of the dominant hand in Scenario A, and the dominant hand being used to move the cell in Scenarios B and C.
Compared to the large braille displays (20, 40 and 80 cells), the Pantobraille registered lower reading speeds but had a performance comparable to that of single cell braille displays (see Table 1). Thus we can assume that with training, users could significantly improve their reading speed with the Pantobraille.
When using only one hand (Scenario A), the Pantobraille registers low reading speed and comfort. This is due to the sensori-motor complexity of the task. The user must hold the braille cell between the thumb and middle finger. It is necessary to grasp the braille cell firmly; this is due to the cell's mobility, placed as it is on a rotation axle which keeps it from maintaining a constant direction. Holding the cell, the user moves it along horizontal lines in a slow motor operation that involves the perception of a position; this perception in turn depends on the forces created by the Pantograph and the braille characters. Finally, the index finger of the same hand reads the braille character, going back and forth in this way on the cell.
This three-part sensori-motor activity is shared between the two hands in Scenarios B and C: the grip and displacement of the cell is performed by one hand while the index finger of the other reads from the cell. This accounts for enhanced performance and reading comfort. If the cell is separated from the Pantograph (Scenario C), performance and comfort show substantial further improvement. By eliminating the left hand's carrier function, the left hand no longer follows the right.
Finally, users preferred to use their dominant hand for displacing the cursor and the other for reading. They reported that this configuration ensures precise control while the other is more sensitive to tactile stimulations. In effect, we tested the scenarios with reversed roles (i.e. dominant hand for reading and the other for displacement) and readers did not like it. Of course, since only two visually handicapped subjects were involved in this pilot study, none of the results described herein is statistically significant; in order to obtain such results we would need to repeat the experiment with a larger sample of subjects.
The Pantobraille can be used as it stands. Our subjects preferred it to the Optacon, although there is no appreciable improvement in performance over a large braille display. On the other hand, visually handicapped people now have options available to them which the braille display did not provide, such as direct bi-dimensional page reading, the editing of various text display features (e.g., alignment, centre) as well as combining reading with a graphical interface pointer, as well as low-resolution graphics (nothing had previously been done with respect to graphics). To ensure progress, the following three factors must be taken into account:
Adaptation to the tool. Conditioned reflexes used in standard braille displays cannot be taken into account in our evaluation of the Pantobraille. Subjects had no training with the Pantobraille and therefore had to develop new reflexes and new reading habits.
Transition simulations. It is well known that finger-tip motion improves tactile perception [7]. Moving the finger on a standard large-scale braille display provides the reader not only with fine-grained details about braille characters but also about transitions between characters. With one braille cell this possibility is not available to the reader and s/he has to explore the cell (e.g. from front to back). This forces displacement and exploration capabilites to be independent (see [8] for an example of a movable dynamic tactile display where braille is presented by a smooth lateral "wave" of lifted and lowered pins)
Hardware and/or gestural decomposition. The combination of three types of sensori-motor activities is costly. To reduce cost, two solutions must be considered. The first is to eliminate the grasping action applied to the cell so that the it's displacement would be on the same axis and at the same rate as the forearm or hand. The second solution consists of separating the cell from the Pantograph, which would require a bimanual operation.
Further effort should be invested in order to improve the abovementioned technical and ergonomic aspects, enabling a user to naturally and quickly feel both the texture and shape of physical objects with one finger. However, the use of both hands should be explored further in an effort to optimize the use of each during the human-computer interaction process.
This project is partly funded by Industry Canada's CITI, the Public Service Commission of Canada's Special Measures Innovation Fund, and Visuaide. I am grateful to Matthew Mather who carefully integrated the braille cell with the Pantograph and devised the electronics to control it. I thank the visually handicapped persons who helped me to realize the pilot study. This work was also influenced by stimulating discussions with Vincent Hayward (McGill University) and members of the PC-Access and Intelligent Multimodal Interface groups.
1. Akamatsu, M., Sato, S., MacKenzie, S. (1994) "Multimodal Mouse: a Mouse-Type Device with Tactile and Force Display", in Presence, Vol. 3, No.1, Winter 1994, MIT Press.
2. Cadoz, C., Lizowski, L., Florens, J. L. (1990). “Modular Feedback Keyboard.” Computer Music Journal, 14, N2, p.47-5. MIT Press, Cambridge Mass (1990).
3. Colgate, J.E., Grafing, P.E., Stanley, M.C., Schenkel, G., (1993) "Implementation of stiff virtual walls in force-reflecting interfaces", in proc. of IEEE Virtual Reality Annual International Symposium, sept. 18-22, 1993, Seattle, Washington
4. Craig, J.C, Sherrick, C.E (1982) "Dynamic tactile displays", in W.Shiff & E.Foulke (Eds.) Tactual Perception - a source book. Cambridge: Cambridge University Press, 1982, p.209-233
5. Davidson, P., Appelle, S., Haber, R. (1992) "Haptic Scanning of Braille Cells by Low- and High-Proficiency Blind Readers", in Research in Developmental Disabilities, Vol13, pp.99-111, Pergamon Press Ltd.
6. Dufresne, A., Martial, O., Ramstein C. (1994) “Multimodal User Interface System for Blind and "Visually Occupied" Users: ergonomic evaluation of the haptic and auditive dimensions”, in proc. of IFIP Interact'95, Lillehammer, Norway.
7. Foulke, E. (1991) "Braille". In The psychology of Touch edited by Morton A. Heller and W.Schiff, 1991
8. Fricke, J., Baehring, H. (1994)"Displaying Laterally Moving tactile Information", in proc. of 4th international conference ICCHP'94 on computers for handicapped persons. pp.461-468.
9. Göbel M., Springer J., Hedicke V., Rötting M., Luczak H. (1994) "Tactile Feedback applied to Computer-mice". Manuscript for the International Journal of Human-Computer-Interaction.
10. Hayward, V., Astley, O.R (1995), "Performance Measures for Haptic Interfaces", preprints of the 7th Int. Symp. on Robotics Research, Herrsching, Germany, Oct. 21--24, 1995. (To be published by Springer Verlag).
11. Hayward, V., Choksi, J., Lanvin, G., Ramstein, C. (1994) "Design and Multi-objective Optimization of a Linkage for a Haptic Device." in proc. of the 4th workshop on Advances in Robot Kinematics, ARK'94, Ljubljana, Slovenia, Kluver academic.
12. Hunter, I.W, Doukoglou, T.D., Lafontaine, S.R, Charette, P.G., Jones, L.A., Sagar, M.A., Mallinson G.D., Hunter P.J. (1994) "A Tele-operated Microsurgical Robot and Associated Virtual Environment for Eye Surgery" in Presence, Vol.2, No.4, Fall 1993, pp.265-289, MIT Press.
13. Iwata, H., Noma, H. (1993), "Volume haptization", in proc. of IEEE Symposium on Research Frontiers in Virtual Reality, October 23-26 1993, San Jose.
14. Lederman, S.J (1990) "The perception of Texture by Touch" in W.Schiff, E.Foulke (eds.): Tactual Perception. Cambridge University Press 1982, pp.142-155.
15. Martial O., Dufresne, A. (1993) "Audicon: Easy Access to Graphical User Interface for Blind Persons - Designing for and with People", in Human Computer Interaction, pp.808-813, Orlando: Elsevier
16. Massis (1995) "Phantom", in O Electronic Magazine April 1995.
17. Minsky, M., Ouh-Young, M., Steele, O., Brooks F.P.Jr. (1990) "Feeling and Seeing: Issue in Force Display", in Computer Graphics 24(2), April 1990, pp.235-243.
18. Mousty, P. (1986) "La lecture de l'écriture braille - Patrons d'exploration et fonctions des mains", mémoire de thèse de doctorat, Faculté des Sciences Psychologiques et Pédagogiques de l'Université Libre de Bruxelle, Octobre 1986
19. Mynatt, E.D., Weber, G. (1994) "Nonvisual presentation of graphical user interfaces: contrasting two approaches, in proc. of the ACM CHI'94 conference on human factors in computing systems, 1994, Boston, p.166-172
20. Parreno A., Magallon P. (1991) "A Scanning Braille Line", proceedings of the fifth annual conference Technology and persons with disabilities, Los Angeles, Mars 1991, pp.725-736.
21. Portalier, S. (1990) "Prise d'information par les aveugles sur les terminaux d'ordinateur", in annual conf. on "La lecture: processus, apprentissage, évaluation, troubles", Lille, march 1990.
22. Ramstein, C. (1995). "MUIS: Multimodal User Interface System with Force Feedback and Physical Models.", in proc. of IFIP International Conference on Human-Computer Interaction Interact'95, Lillehammer, Norway 1995.
23. Ramstein, C., Martial, O., Dufresne, A., Carignan, M., Chasse, P., Mabilleau, P. (1996) "Touching and Hearing GUIs. Design Issues in PC-Access System", in proc. of ACM ASSETS'96 conference, Vancouver, April 14-17 1996.
24. Ramstein, C., Hayward, V. (1994). "The Pantograph: a Large Workspace Haptic Device for Multimodal Interaction." in proc. of ACM CHI'94, Interactive Experience.
25. Ramstein, C., Arcand, J.F, Deveault, M. (1996) "Adaptive User Interfaces with Force Feedback", in proc. of ACM CHI'96 on Human Factors in Computing Systems, Vancouver, BC, 14-18 April 1996
Scenario A: monomanual use. The cell is mounted on the
Pantograph, and the reader is instructed to "Use only one
hand to read the braille text."
Scenario B: concentrated bimanual use. The cell is mounted
on the Pantograph, and the reader is instructed to "Use both
hands to read the braille text. "
Scenario C: dedicated bimanual use. The braille cell is
dismounted from the Pantograph and set aside. The reader is instructed
to "Use two hands to read the braille text: one hand to move
the cursor within the text using the Pantograph and the other
hand to read the braille information on the cell"