A system to aid blind people in the mobility:

A usability test and its results

Edoardo DAtri, Carlo Maria Medaglia, Alexandru Serbanati, Ugo Biader Ceipidor
CATTID - Universit di Roma La Sapienza Rome, Italy carlomaria.medaglia@uniroma1.it

Emanuele Panizzi
Dipartimento di Informatica Universit di Roma La Sapienza Rome, Italy panizzi@uniroma1.it

Alessandro DAtri
CeRSI Luiss Guido Carli University Rome, Italy datri@luiss.it

AbstractBlind people need to become as independent as possible in their daily life in order to guarantee a fully social inclusion. Mobility means the possibility of freely moving, without support of any accompanying person, at home, in public and private buildings, and in open spaces, as the streets of the town. Mobile and wireless technologies, and in particular the ones used to locate persons or objects, can be used to realize navigation systems in an intelligent environment. Such systems open new opportunities to improve the speed, easiness, and safety of the visually impaired persons mobility. Using these technologies together with Text To Speech systems and a mobilebased database the authors developed a cost effective, easy-to-use orientation and navigation system: RadioVirgilio/SesamoNet1. The cost effectiveness is due to the recovery of RFID identity tags from cattle slaughtering: these tags are then borrowed to create a grid used for navigation. In this paper the results of an usability analysis of this guide system are presented. A preliminary experiment involving a small group of experts and a blind person is described. In order to evaluate the usability, three cognitive walkthrough sessions have been done to discuss the systems basic functionality and to highlight the most critical aspects to be modified. Keywords- Blind mobility; Mobile communication and devices; RFID; Usability test;

B. Mobility and orientation by blind people Orientation is the subjective ability to know the own positioning in a space both in absolute sense and respect to the point of departure and the one of arrival; this is possible through a complex cognitive and perceptive process and by the acquisition of the sensory information coming from the environment and the body. A blind person can orient himself in already known places by recognising known features (absolute orientation) or by perceiving unknown but regular objects, like buildings walls in outdoor situations or the angles of a room walls in indoor spaces (relative orientation). This first kind of orientation is, for a blind person, the most difficult one due to the complexity of collecting information from the surrounding environment only through tact and hearing. In order to explore unknown places a blind person usually needs to be guided by another person or a dog. In order to move without a guide in an unknown space a blind needs to identify some points of reference that allow him to guess the optimal movements and the distance to cover to move from one place to an other one. A visually impaired decide her orientation from two types of information: the points of reference and the lines of reference [12]. Points of reference are every kind of environment information (tactile, acoustic or olfactory) a blind person easily perceive and/or remember in her memory: a visually impaired, has to knows his relative position with respects to these points and their meaning. Reference lines are based on a continuous environmental knowledge that allows a blind to orient himself and to maintain the right direction without the need of additional points of reference, example are the sidewalks, the walls, the changes of ground surface. RadioVirgilio/SesamoNet, through the voice synthesized instructions, supplies to the most important points of reference when the user cannot easily perceive these natural marks and generates sounds to identify additional and useful reference lines. C. State of Art Several solutions have been proposed in the recent years to increase the mobility and/or safety of disabled pedestrians. A first system is proposed by Reppucci [1]. This system is composed by: an information broadcast service provided by a transmitting station covering the area of interest and a

I.

INTRODUCTION

A. Perceiving spaces by blind people Blind people compensate the lack of sight augmenting the capabilities of the other senses. In this way they are able to create mental maps from sensory images [5]. In fact all the remaining senses, mainly tact and hearing, but also in some cases the olfaction, are used to replace, even if an incomplete way, the ability to perceive the spaces. In order to know an environment a visually impaired tries to build mentally the spaces by identifying obstacles around him and progressively searching environmental information. RadioVirgilio/SesamoNet provides a suitable audio output indicating path borders, possible danger or other important information; this audio helps the visually impaired in creating his/her mental maps.
Radiovirgilio/Sesamonet (SEcure and SAfe MObility NETwork) is a joint project between SILAB-IPSC of the EC Joint Research Centre (JRC) and the RFID LAB of the University of Rome Sapienza
1

Second International Conference on Systems (ICONS'07) 0-7695-2807-4/07 $20.00 2007

portable receiving station allowing the user to poll the transmitting station that deliver intelligible information about this area. In this way, an unsighted person receives relevant local information, such as the status of a traffic light signal. Another widespread approach for outdoor navigation relies on the Global Positioning System (GPS), which is based on geostationary satellite signals. In the recent years several systems have been developed on this technology and oriented to visually impaired persons. These tools often integrated other technological services such as Geographical Information Systems (GIS). Remarkable blind navigation systems based on GPS, GIS and wireless technologies are presented in [7], [9] and [18]. Unfortunately, the resolution of a GPS localization for civil purposes is limited (few meters) and there are relevant delays in the answer when the receivers are used. Due to such a lack of accuracy and the low real time responsiveness, the needed level of safety for a blind pedestrian cannot be currently provided by such a kind of systems. Another example of GPS based system has been developed by Mori and Totani [13]. The system is composed by an automatic wheelchair, a GPS receiver and a portable GIS. Authors affirm that the obtained quality of mobility is better than the one provided by a guide dog, but on the other side this solution seems not easy to be extended to walking mobility. The solution proposed in [1] is based on a portable device having a radio transceiver and fixed radio transceivers located as reference points. When a command is issued by the user, the portable device receives from the base transmitter a location message indicating the information to be synthesized. To improve the blind user orientation the portable transceiver should also include an electronic compass. Another approach described by Hancock [6] proposes a portable navigation unit and a plurality of location beacons based on infrared emitting diodes. A direct line of sight is required in this case between the navigation unit and the beacons. A similar solution, proposed by Shojima et al. [17], uses a portable guidance apparatus and a plurality of markers which output a beam carrying location information and direction information of this beam. Some authors [11,14,20] propose the usage of RFID passive tags for navigation system for blind people. RFID tags provide location markers and a mobile receiver provides user navigation information based on the code stored in each tag. These systems does not require a power supply for location markers and overcomes the GPS navigation limits in indoor environments. On the other side, the tags placement cost could be high and usually tags are not easy to adapt for an outdoor, harsh environment. II. GENERAL DESCRIPTION AND SYSTEM OVERVIEW

3.

PDA

Figure 1. The physical architecture composed by: a tag grid, the RFID reader cane, and the PDA

Figure 2. Tags Rfid used in RadioVirgilio/SesamoNet

Passive RFID Tags -- The final tag grid, by which absolute orientation is achieved, will be made of tags burrowed into the ground up to 4cm. These tags were used for animal identification and, after slaughtering, they are recovered for this purpose, reducing the cost of both the transponder and its encapsulation, due to the fact they are already in a plastic or ceramics envelop. Figure 2 shows two typical transponders. In the current study the134.2 KHz channel was used because of the minimal noise from the environment (i.e. concrete, water, etc) by which the signal is affected. This allows a reading range of up to 15cm. Transponders have associated data and ID memory blocks. The uniqueness of the number stored in the read-only ID memory block is guaranteed by the European law for animals identification. More specifically, in this embodiment the capsule has the form of a bolus, as described in EP 0 849 989 [2] for use with ruminants such as cattle. Since RFID tags are quite cheap, the packaging is a relevant cost factor. Recycling of such animal identification tags significantly reduces the cost of the encapsulated transponders. Rfid Cane Reader -- The RFID antenna is placed inside and near the tip of a white cane, designed for this purpose (see Figure 3), having a shape similar to the classic blinds cane. The current version of the cane is made of plastic and needs an external RFID controller, Bluetooth interface, and batteries2.
2

The system is composed by the following parts: 1. 2. RFID tag grid RFID Cane Reader

For protoyping purposes we used a standalone battery and an RS232-toBluetottoh commercial converter which are bulky compared to the final prototype. A new model will be ready in December 2006.

Second International Conference on Systems (ICONS'07) 0-7695-2807-4/07 $20.00 2007

for tag ID string transmission. The user is not supposed to perform any task when navigation session starts or the BT manager is resumed after any loss of connection with the cane. Navigation Data Interface: it retrieves navigational data from a local database, providing the Navigation Logic with extended data related to a tag ID. Navigation Logic: this is the core software which handles navigation and tag data in order to provide the user with mandatory (safety-related) or on demand (environmental) navigation informations. It also checks if the direction is right and not reverted. This module tells the user he is probing the central tag or one on the right or the left hand side of the path3. Eventually it can send a text string containing more complex navigation or environmental information to the TTS component [19].

Figure 3. RFID reading cane

The reader is set to either operate on demand or in continuous read mode, while holding a button on the cane handle. In continuous read mode the battery lifetime is about three hours. In the new version the cane will be closer to the one used by blinds: smaller, lighter and easier to use. It will be white, like the classic blind cane. Inside the tip of the cane, as in the prototype, an RFID antenna will be integrated. the RFID controller will be included in the handle, as the rechargeable batteries and the Bluetooth terminal that now are external devices. PDA System development and running usability tests were done using a HP iPAQ hw6500 PDA having a built-in Bluetooth 1.2. This device has 64MB of main memory and a Secure Digital (SD) expansion slot, used as data storage to contain the database file, in order to have the maximum main memory available to run the RadioVirgilio/SesamoNet application. Standard batteries provide at least three hours of continuous use while Bluetooth communication is on. This time will be further increased when the Bluetooth 2.0 or ZigBee interfaces will become available. RadioVirgilio/SesamoNet can also be supported by any Windows CE based portable device having a Bluetooth antenna to communicate with the Rfid cane. A. Logical Architecture The software, still under development, is based on the Microsoft .Net architecture (but we are also ready to start a JAVA version of the project) and can be divided in the following blocks: Bluetooth Cane Connection Manager: a BT connection channel is open between the RFID reader and the PDA

Figure 4. The overall architecture

B. System Functionality A grid of tags has been designed to provide the needed information about the path the blind should follow in order to be safe . Such a grid, and the related acoustic signals, guides him along a route alerting him when he exits the safe grid area. This system is a great improvement compared to the classical
3

A sweep of the cane usually probes an arc in front of the user and it takes about one second. This means that, if three tags are probed while drawing a single arc, three output should be generated in this time. Thus we decided that spoken messages are not suitable (too long) so tones are used for the main three (and more frequent) navigation signals.

Second International Conference on Systems (ICONS'07) 0-7695-2807-4/07 $20.00 2007

assisting technology for blind peoples mobility, like the white cane, the guide dog, and the tactile ground surface indicators [10,16]. RadioVirgilio/SesamoNet offers various aids for the autonomous deambulation of visually impaired: Keep the user on a safe path Information about turns and obstacles in the path Checking of the right direction (useful if user falls or is disoriented) General and specific environment information on demand Check and trace his position in the tag grid On-line help and assistance (via GSM) III. A. Usability analysis A deep analysis of the interaction of visually impaired persons with the system is fundamental in order to have a usable system that enhances the standard cane and gives an added value to the blind person. For this reason, we decided to study the usability of the system, analyzing the different functionalities, both with an expert-based study and with real users. First, we performed three cognitive walkthrough sessions (Fig 5), in order to discuss the systems basic functionalities and to highlight the most critical aspects to be modified or to be tested with users. The first two walkthrough sessions were participated by the authors plus some students, while, in the third one, a blind person joined the group. In all the walkthrough sessions one or more participants executed some basic tasks within a typical scenario, while the other participants observed; during this execution, as well as after it, a discussion arose among participants with the goal of: assess tasks completion analyze encountered problems, and single and recurring errors make an expert usability evaluation RadioVirgilio/SesamoNet basic features propose and discuss solutions discuss about further testing of the same or similar tasks (preparation for usability tests) of THE USABILITY TEST

the cane at an appropriate distance from the floor and trying to find a central tag and placing over it.

Figure 5. Task execution during the cognitive walkthrough

Figure 6. The grass like carpets as they were disposed on the room floor. Only the two gray carpets were equipped with RFID tags.

In the scenario in which the tasks were performed, a student (the blind person in the third session) tried to orientate and find his way using RadioVirgilio/SesamoNet along the path represented in Fig. 6. Each student was taught about the system tones and the tags disposition in the path; they were then instructed to approach the RadioVirgilio/SesamoNet pathway by scanning with the cane and moving on, keeping

We positioned 5 grass-like carpets in the room. Two of them, the ones represented in grey in fig 6, were equipped with RFID tags that were configured as a path with starting point on the left, with a straight part followed by a 90 right turn. We juxtaposed the other inactive carpets to them in order to avoid tactile feedback that the student could exploit to identify the path. The visually impaired user confirmed that it was not possible to identify the path border by touching, as he could not find the carpet junctions due to the grass-like consistency of the carpets. The disposition of tags in the path is reported in Fig. 7, the direction of arrows representing the configuration of sound and messages for each tag (left border, right border, central path and direction, right turn).

Second International Conference on Systems (ICONS'07) 0-7695-2807-4/07 $20.00 2007

Three tasks were executed and discussed in each cognitive walkthrough session:

Fig 7. Layout and configuration of RFID tags in the path.

Task 1 was related to finding the path in the room. The student was first deliberately disoriented, then he was asked to find the path by just hearing the tones of the RadioVirgilio/SesamoNet system. His starting position (which he was not aware of) was either on the left or on the right of it, and he did not know where the path was located (either in front of him, or rear, or aside). Task 2 was about orientating in the path, i.e. finding path direction and identifying left and right borders. Finally in task 3 the student was positioned at the starting point of the RadioVirgilio/SesamoNet path and he was asked to move along towards the exit point.

B. Results Several information came both from the task execution and from the discussion that followed it. Results can be organized into different areas. 1. System tones. Participants noticed that the system tones and messages were not appropriate for several reasons: i) the message go straight on as central tag tone is too long and it is repeated too often due to the cane usage (see point 3 below); ii) right border tone is too low; iii) left border is not advertised in the curve as it is superseeded by the turn right message, and it is likely to exit the path in that case; iv) the message you are going backwards, that is pronounced when the user reads tags which he already passed-by, should be delayed to after some occurrences of backward tags, in order to avoid false alarms. It then should be repeated only two or three times; v) the blind user suggested to use more tones, which could be learned by the user, as a sort of road sign (as he said). 2. Cane length. The cane length (that should be proportionate to user height) highlighted some problems related to tags finding and to the right turning:

in fact, if the cane is too long (and that was the case for all the three users), it is not possible to scan very close to the user body; one of the students in task 2, standing with her body on the right border tags, could not hear their tones and she only found the other two lines of tags; the blind user, in task 3, turned right too early because the cane was sensing tags too far away and thus the system pronounced the turn-right message; 3. Cane usage. Cane constituted a problem for students, as they tend to move it too fast, loosing some tags; in some cases the system played the tone related to a new tag before the tone of the previous was finished, leading to unrecognition of border positions. 4. Borders vs. central path. It was interesting to note that while the border signals were used to locate the path direction (task 2) they were not so used to follow the path in task 3. The blind user, especially, followed the central line of tags and was able to follow the path walking quickly; for this reason we repeated this task three times with him just using one tag line and also placing an obstacle on the path; lateral tags allow for redundancy and guarantee for fault tolerance, they can help to find again a lost path and they can help in closed building (that have no lodges) and in case an obstacle (e.g. a bag) is placed on the central path line; in other cases, however, it seems that central line tags could be embedded in the lodges with no need for lateral tags; 5. Look ahead. Synchronization is quite important when a turning approaches. We discussed two possible solutions to the problem of turning too early or too late based on the possibility of explicitly saying something like two steps from now, turn right or configuring the turn right messages on tags placed in the final part of the straight path only, in order to let the user turn as soon as he receives the message, using it as a virtual wall. 6. Augmented reality. The whole discussion confirmed that RadioVirgilio/SesamoNet is to be used as an augmentation of the information that the normal cane already gives (physical objects identification) and that its usage must be integrated in the cane usage. User tests will stress the analysis of this aspect. Based on this input, we are now calibrating the system (modifying some system tones and the moments when they must be played) and we are planning a usability test with blind users that will be conducted in the next weeks and that is aimed to assess the validity of our assumptions and of the whole system. This test, that will be based on standard methodology (scenario, tasks, observer, questionnaires, etc.) and with the management software Techsmith Morae (www.techsmith.com/morae.asp), will allow to evaluate the usability of the system, the error rates, and a learning curve of RadioVirgilio/SesamoNet, and to highlight and propose solutions for potential other problems that may eventually arise.

Second International Conference on Systems (ICONS'07) 0-7695-2807-4/07 $20.00 2007

IV. FUTURE DEVELOPMENTS The RFID technology is still under development and test in several application fields. There are many applications already in the market (e.g., for tracing the products). Probably RFID technology, when production costs will be lower than now, will replace magnetic bands and barcodes in identifying and tracing the products. In the near future there could be additional applications of the RadioVirgilio/SesamoNet. For example blind could also be able to buy products in shops without the need of external aid: the cane, will help them to move in the shopping mall, while the system (according to what is suggested in [4]) will supply the necessary information about the products on the shelves: e.g. the name, price, description, and other types of information. In a domestic environment, RFID tags, combined with the cane, could also be useful for visually impaired to find objects. Applying tags for example on keys or on the wallet they could be able to find them in a easy way [8]. A. Mapping spaces Incrementing the number of the tagged paths will allow the e-cane to be a useful resource for a blind in exploring the living spaces. This system could provide a complete mapping for every kind of space and the visually impaired could be able to have mental maps of the spaces around him. B. Personalized Assistant A software for the management of the system is under development and will allow its customization according to user preferences and choices. It will be able, for instance to define the preferred path, as it is now in GPS navigators . Moreover it will be possible to store in the systems memory the user preferences and past selections in order to automatically suggest the destinations and the most frequently used paths. REFERENCES
[1] L. W. Alonzi, D. C. Smith, G. J. Burlak, M. Mirowski, (1992). Radio frequency message apparatus for aiding ambulatory travel of visually impaired persons, U.S. Patent 5,144,294 issued Sept. 1, 1992. L. G. Caja; J. F. Vilaseca Vintro; C. Korn, Ruminal bolus for electronic identification of a ruminant, 2002 http://www.freepatentsonline.com/6474263.html S.Cardin, D.Thalmann and F. Vexo, Wearable System for Mobility Improvement of Visually Impaired People, VR Workshop on Haptic and Tactile Perception of Deformable Objects, Hannover, Germany, December 2005.

[4]

[5]

[6] [7]

[8]

[9]

[10]

[11]

[12]

[13]

[14] [15] [16] [17]

[18]

[19] [20]

[2]

[3]

H. Donker, P. Klante, P. Gorny, The design of auditory user interfaces for blind users. Proceedings of the second Nordic conference on Human-computer interaction, October 19-23, 2002, Aarhus, Denmark M.-A. Espinosa, S. Ungar, E. Ochata, M. Blades, C. Spencer, Comparing methods for introducing blind and visually impaired people to unfamiliar urban environments. Journal of Environmental Psychology, 18, 277-287, 1998 M.B. Hancock, Electronic autorouting navigation system for visually impaired persons. U.S. Patent 5,806,017 issued September 8, 1998. A. Helal, S. E. Moore, B. Ramachandran, Drishti: An Integrated Navigation System for Visually Impaired and Disabled , Proceedings of the Fifth International Symposium on Wearable Computers, 2001, 149 156. A. Hub, J. Diepstraten, T.Ertl, Design and development of an indoor navigation and object identification system for the blind, proceedings of the 6th international ACM SIGACCESS Conference on Accessibility and Computers , 2004 Z. Hunaiti, V, Garaj, W. Balachandran, F. Cecelja, An Assessment of 3G link in a Navigation Systems for Visually Impaired Pedestrians, Proceedings of the 15th International Conference on Electronics, Communications and Computers, 2005, 180 -186. Y. Kobayashi, T. Takashima, T. Hayashi, H. Fujimoto, Gait analysis of people walking on tactile ground surface indicators, in IEEE Transaction on Rehabilitation Engineering, 2005, 13:1, pp 53-59 V. Kulyukin, C. Gharpure, J. Nicholson, S. Pavithran, RFID in RobotAssisted Indoor Navigation for the Visually Impaired, Proceedings of 2004 lEEE/RSJ International Conference on Intelligent Robots and Systems, September 28 -October 2,2004, Sendai, Japan O. Lahav, D. Mioduser, Multisensory virtual environment for supporting blind persons acquisition of spatial cognitive mapping, orientation, and mobility skills, Proc. 4th Intl Conf. Disability, Virtual Reality & Assoc. Tech., Veszprm, Hungary, 2002 H. Mori, S. Totani, Robotic Travel Aid for the Blind: HARUNOBU-6. In Second European Conference On Disability, Virtual Reality, and Assistive Technology, Svde, Sweden, 1998. T. E. Piotrowski, RFID navigation system. EP patent 1 313 079, 2003 A. L. Reppucci, Computer controlled information broadcasting system for aiding blind persons in town walking, EP patent 0 338 997, 1989. M. J. Scherer, MPT-Matching persons and technology, http://hometown.aol.com/IMPT97/MPT.html, 1994 H. Shojima, T. Yashiki (2000). Information processing apparatus and pedestrian navigation system using the same, U.S. Patent 6,259,990 issued January 28, 2000 T.Strothotte, S. Fritz, R. Michel, A. Raab, H. Petrie, V. Johnson, L. Reichert, A. Shalt, Development of dialogue systems for a mobility aid for blind people: initial design and usability testing, ACM SIGACCESS Conference on Assistive Technologies, Proceedings of the second annual ACM conference on Assistive technologies, 1996. M. Tatham, K.Morton, Developments in Speech Synthesis, Digital, October 2005. S.Willis, S. Helal, A Passive RFID Information Grid for Location and Proximity Sensing for the Blind User, University of Florida Technical Report number TR04-009. http://www.cise.ufl.edu/tech_reports/tr04/tr04-009.pdf

Second International Conference on Systems (ICONS'07) 0-7695-2807-4/07 $20.00 2007