Mobility of Blind and Elderly People Interacting with Computers

Dr. John Gill
April 1997



Contents

Glossary of terms
Summary
1 Objectives of the Project
2 Introduction
3 State of the Art
3.1 Electronic Travel Aids (ETAs)
3.2 Portable PC
3.3 GPS/DGPS
3.4 Other Positioning Systems
3.5 GSM
3.6 Software 4 Main Achievements and Results 4.1 Requirements and Initial Evaluations
4.2 System Design
4.3 Training
4.4 Evaluation
5 Publications


Glossary of Terms

DGPS Differential GPS

ETA Electronic Travel Aid

GDF Geographic Data File

GIC GPS /GLONASS Integrity Channel

GIS Geographic Information System

GLONASS Global Orbiting Navigation Satellite System

GNSS Global Navigational Satellite System

GPS Global Positioning System

GSM Global System for Mobile communications

MoODS MoBIC Outdoor System

MoPS MoBIC Pre-Journey System

MoTA MoBIC Travel Aid, consisting of MoPS and MoODS

OODBS Object-Oriented Data Base System

RDB Relational Data Base

RDS Radio Data System

SA Selective Availability

SMS Short Message Service


Summary

This is a report on the Mobility of Blind and Elderly People Interacting with Computers (MoBIC) project was carried out from 1994 to 1996 with the support of the European Union Technology for the Integration of Disabled and Elderly People (TIDE) programme of Directorate General XIII of the Commission of the European Union. The primary objective of the project was to increase the independent mobility of blind and elderly persons in unknown urban environments. Based on an intensive user needs studies, an interactive system, the MoBIC Travel Aid (MoTA) was developed with two parts:

  • the MoBIC Pre-Journey System (MoPS) which allows blind people to study maps and plan their journeys before setting out; an indoor system to enable a blind person to get information about an environment in preparation of a journey and to assist the user in route planning and
  • the MoBIC Outdoor System (MoODS) which assists blind people when they are out on their journeys using Global Positioning Satellite (GPS) technology.

The system has undergone extensive testing by blind and partially sighted people as well as several steps of technological refinement in the hardware and software based on the results. Particular attention was paid to aspects of training people in the use of the system.

This report serves to inform about the essential elements of the objectives of the project, the state of the art in the relevant technologies, and the main achievements.


1 Objectives of the MoBIC Project

The main objectives of the MoBIC project have been:

  • to increase the independent mobility of visually disabled and elderly people, particularly in unknown environments,
  • to equip individuals with their own aid which will enhance the use of other mobility aids and will avoid the need for expensive reactive equipment in the environment,
  • to use this aid to give visually disabled and elderly people access to information about the environment previously only available through visual channels,
  • to influence standardisation activities in order to stimulate a single market in rehabilitation technology in Europe and to integrate the results of this project with other CEC activities currently under way in navigation systems for non-disabled people.

The major thrust of MoBIC has been the development of a system to improve the independent mobility of blind persons.


2 Introduction

Visually disabled and elderly persons have problems in travelling independently. Firstly, they have difficulties in accessing the information they need to plan their journeys. Once on a journey they have problems in knowing where they are and keeping to the planned route.

The MoBIC project developed a map exploration and route planning system which was designed to allow a blind person access to electronic maps of the locality. The MoBIC Pre-journey Planning System (MoPS) helps blind persons to study and plan their routes in advance.

With the addition of devices to give the precise current position and orientation of the blind pedestrian, the system can then be used outdoors. The system is intended to be complementary to primary mobility aids such as the long cane and guide dog.

The MoBIC Outdoor positioning System (MoODS) is based on signals from satellites which give the longitude and latitude; the computer converts this data to a position on an electronic map of the locality. The output from the system is in the form of spoken messages.

To ensure the usability of this system, an extensive part of the design project focused on user requirements. This included the participation of visually disabled and elderly people in both the specification and implementation phase of the project. One outcome of the project is a detailed specification of user requirements which is independent of the specific technology. Two field trials were undertaken to evaluate the prototype systems against this specification.


3 State of the Art

Throughout the project the evolution of the state of the art in the fields of interest for MoBIC has been investigated. New technologies potentially usable for the hardware of the MoBIC Travel Aid were investigated. In addition, software products were examined to see if they would be suitable for integration within the overall software architecture.

3.1 Electronic Travel Aids (ETAs)

The main efforts to find technical solutions to blind people’s mobility and orientation problems have been undertaken during this century. They are of two principally different kinds, one modifying the environment through which the blind people travel (for instance tactile tiles) and the other providing the blind travellers with aids accompanying them. These two kinds of solutions are not in conflict with each other but may be complementary. Nor is there any conflict between the well-known low-tech aids, such as the long cane and the guide dog, and technically more advanced devices. In the way the two kinds of aids have developed they fulfil different information needs for the traveller. The two traditional aids inform mainly about the ground, while the more technical aids usually provide other types of information. Most of them inform about objects above waist level in the closest environment (1-2 metres).

There have been efforts when developing the electronic travel aids (ETAs) to utilise optical information via a TV-camera, such as the Electrophthalm and the Tactile Vision Substitution System. Typically, however, ETAs emit some kind of light or sound (laser, infrared light or ultrasonic sound) and use reflections of these beams for information about the environment. The most well-known of these devices is the Sonicguide which has been further refined into the Trisensor and, most recently, into the Kaspa. These ETAs are carried on the head, as is, among others, the Sonic Pathfinder. Other ETAs are built into the cane, for instance laser canes, or carried in the hand, such as the Mowat Sensor.

None of these kinds of ETAs, and there are dozens of them, are very widely used by blind people, possibly because of their adding too little information to the traditional aids and/or their form of information being difficult to pick up. Common to all of them is that they do not provide orientation and navigation information, such as information about where you are, from where you come and in what direction your destination is located. Sighted people receive get this kind of information directly from the environment, but also via maps and compasses.

Maps of many different kinds as well as compasses have been modified to allow blind people to read them, but they do not provide immediate information about the position of the traveller. MoBIC has developed an aid affording such information, as well as other kinds of information, to the blind user. This kind of aid starts a new era in the development of ETAs for the blind people.

3.2 Portable PC

3.2.1 Bodyworn PC

In June 1996 Rockwell introduced a new wearable, voice-controlled lightweight personal computer system called Trekker+ mobile information system. The Trekker system appears to be the smallest personal computer system currently available and is ideally suited for commercial development of the MoBIC outdoor positioning system.

The Trekker system is worn on a belt or shoulder strap and includes a head-up VGA display worn on a headband system. A mouse pad mounted on the computer is operated with the hand. The system allows users to take information into the field for real-time use without taking their eyes, hands or attention from the tasks they are performing. Users can get work done faster, and accomplish more each day. This commercially ruggedized system empowers users with new levels of information portability.

The Trekker system is a Windows-based computer system built around industry standards, allowing for expansion through use of off-the-shelf peripheral devices. Improvements over high-powered notebook computer capability include human interface, speed, power, multichip mode, weight, battery life, ease of use, and expandability.

The unit weighs less than three pounds and straps to the waist. The Pentium-based Trekker 2020 includes 16 MB of RAM, 1.2 GB hard drive, two PCMCIA ports, additional ports for VGA, optional keyboard, serial and parallel cables. The Trekker system is offered by Rockwell, which has drawn expertise for this market from its role as a leading worldwide supplier of high technology electronics and avionics.

Bodyworn PCs are highly sought after. Several exhibitions and organisations working in this area confirm this impression.

3.2.2 Keyboard

A first example of a 3-Key-Keyboard with radio-link-interface has been examined. The keyboard makes use of the ISM-band at 433Mhz. One benefit is interfacing the keyboard without cables to the MoODS computer. For a product the necessary PC-Interface has to be developed. The keyboard meets the requirements fixed during the project.

3.2.3 Battery Technology

Li-Ion rechargeable batteries

Over the past ten years, there has been a massive increase in the use of portable electronic devices (e.g. notebook computers, camcorders and telephones) world-wide. These devices are more and more popular and thus a significant growth is expected. One of the problems facing the portable electronics industry is that the usability of many of the devices is limited by the battery. Current battery technology is heavy and bulky and tends to suffer from a number of performance and environmental problems.

Companies who seek to be world class suppliers of portable electronic equipment must be able to supply state-of-the-art battery technology in their products. Recently, several leading Portable Electronics companies have introduced a new, improved rechargeable battery technology called Li-ion (Lithium-ion). Li-ion technology represents the state-of-the-art in battery technology. It is lighter, smaller and environmentally more friendly than existing systems and many of the performance limitations (e.g. capacity, energy density, memory effect) have been eliminated. First engineering samples of Li-ion batteries are available for evaluation.

A world-wide collaboration for the production of Li-ion batteries was founded to develop these batteries. The know-how is spread all over the world (Asia / Europe / USA). AEA Technology Pplc., Micro Tech Inc. and the Engineering Division of Groupe Fribourg, which includes Metar and Falma, have entered a collaboration which combines the extensive expertise and know-how of each company to develop a pilot production line. AEA will develop the concept to manufacture rechargeable Li-ion batteries and will serve as a first platform for the pilot production line.

3.3 GPS/DGPS

3.3.1 Satellite technology

During the three year MoBIC project both American and Russian satellite positioning systems have become fully operational, with 24 satellites available in each system. In the same period receivers have become more prolific, especially the American GPS ones, and their accuracy and ‘ease of use’ have improved. In contrast, no Russian GLONASS receivers have been made available for the European civilian market. With the increase in volume sales of GPS receivers to a mass civilian market has come an the inevitable reduction in price. Handheld units which were more than 1000 ECUs at the start of the project are now less than 300 ECUs.

An improvement has also been seen in the overall performance of GPS receivers. At the start of the project most of the civilian receivers available were single channel with relatively poor ‘ time to first fix’ and accuracy characteristics. However, once the constellation of American satellites was complete, manufacturers invested more in developing either multichannel receivers or sophisticated algorithms for multi tracking single channel receivers, resulting in a significant improvement in performance by the end of the project. Also, in this period, new carrier phase receivers have been introduced giving cm accuracy to within centimetres, but although the prices of these units has fallen, they remain too expensive to be considered for MoODS.

MoBIC has also seen a reduction in the size of both GPS receivers and antennae. Bulky handheld units which started off with limited functionality and restricted battery operating times have been replaced by slimline units with vastly increased functionality and longer ‘mobile’ operating times. In addition, combined receiver/antenna units, the size of a PC mouse, have been produced which interface directly to the communications port of a PC. Manufacturers have even produced a PC card version with an integral antenna.

Tests on GLONASS receivers have shown them to have a nominal accuracy of 10 metres compared to 25 metres for GPS and, as the system does not have Selective Availability (SA), theoretically the raw receivers should have better accuracy. However, due to their scarcity these claims cannot be substantiated. There is a high probability that the American Department of Defence will remove SA in the near future as they have concluded that with the ending of the ‘Cold War’ and the dramatic increase in the civilian use of GPS, there appears to be little advantage in maintaining it. It has been estimated that with the removal of SA and the introduction of other GPS developments, the accuracy of a standard, single frequency GPS receiver could be improved to 6 metres, without DGPS.

Both satellite positioning systems will form the basis for the Global Navigation Satellite System (GNSS) and, in the future, combined GPS/GLONASS receivers will process data from both systems as well as other complimentary ones (e.g. Inmarsat GIC and the Wide Area Augmentation System).

Combined GPS/GLONASS receivers

In January 1996 the GLONASS constellation reached full deployment with 24 satellites transmitting valid navigation messages. This led to the development of combined GPS/GLONASS receivers. The ability to receive and use signals from not only the 24 GPS satellites, but also from the 24 GLONASS satellites, will ensure that it is integrated into a wide variety of high-precision land, sea, and air-based applications on a global scale.

Using a combined GPS/GLONASS receiver will yield several advantages over using GPS alone:

  • Enhanced satellite availability in obstructed, mountainous, and urban canyon environments
  • Improved accuracy - the absence of SA on GLONASS provides a three-fold improvement in range
  • Measurement versus GPS alone
  • Better coverage in high latitudes
  • Faster cold starts
  • Improved system integrity
  • Reduced effects of ionospheric delay

3.3.2 Differential GPS

Differential GPS (DGPS) was developed in conjunction with the American Coastguards who became frustrated with the unpredictable use of Selected Availability (SA) by the American Department of Defence. The DGPS correction signals can be produced at any DGPS Reference Site and then transmitted by the most convenient medium.

Most industrial DGPS users have made use of their own short range radio systems or, for more global operations, satellite communications to receive their corrections. However, these expensive systems have not been viable for civilian users and during MoBIC, several countries have introduced commercial or public service transmissions over long or medium wave radio or, more recently, Radio Data System encoding on national FM broadcast radio. Cellular Telephony, both analogue and digital, can also be used for the transmission of DGPS.

Early on in the MoBIC project it was decided to make use of a mobile phone to receive DGPS corrections, initially via analogue cellular and then, halfway through the project, over GSM, when data services became available. Throughout this period there have been dramatic changes in mobile phone technology, not least in the introduction of new GSM phones and services, which will aid the development of a product with greater efficiency.

The GSM method of receiving DGPS corrections proved to be highly reliable, outages being due mainly to call drop outs. Audio interference proved to be a source of irritation to many users but this was reduced by re-routing the audio cables.

During the last year a commercial DGPS service via RDS became available in the UK. The service, operated by Focus FM and broadcast on the national radio station CLASSIC FM, offers 1 metre and 10 metre accuracy and already covers the major part of the UK. Similar sources of DGPS via RDS are currently available in Sweden and Germany and more franchises are planned throughout Europe.

3.4 Other Positioning Systems

Research is currently being conducted into the use of GSM cellular radio signals to triangulate the position of the actual mobile phone. This technology could provide positioning accuracy equivalent to that of a satellite positioning system without the need for an additional radio receiver. In addition, it has the potential to produce enhanced performance in the built up urban areas where satellite receivers generally suffer degradation. As this technology is developed overall accuracy is likely to be improved but in the short term it could be used to supplement the satellite systems, especially in areas of poor coverage.

Changes of direction can be recognised using GyroChips . These chips are solid state rotation sensors which use a vibrating quartz tuning fork to send angular velocity. By using the Coriolis effect a rotational motion about the sensor's longitudinal axis produces a DC voltage proportional to the rate of rotation. These devices are normally used in military systems. They are very expensive. But mass production is expected in the next year. Thus, the price of these units will fall dramatically and they will become affordable for travel aids. It is conceivable to use GyroChips within the MoBIC device. In contrast to the compass they provide relative motion, are magnetically interference-free and can be used independently from their location.

To provide a travel aid system with the alteration of the distance accelerator sensors could be used. Three sensors with large scale calculation operation are necessary to determine the linear motion of the user. An alternative to the accelerator sensor is given by pedometers which use the number of strides and the users stride length for computing distance alteration.

A dead reckoning system can supplement the pure GPS to achieve a higher availability of positional information. Investigations were made into the various components required for a dead reckoning system. The units currently available were found to be expensive and specifically tailored for the automobile market. Consequently, they invariably relied upon wheel sensors for speed and direction indication. However, links to this potentially large market will eventually ensure that volume sales will bring the costs down.

The investigations found that a system like this is available on the market for a price of 30,000 to 50,000 ECUs with a stability of only a few minutes.

3.5 GSM

GSM started off as the European digital cellular standard but recently, particularly in the last 6 months, the standard has been adopted by many of the major industrial nations throughout the world. This has led to a significant increase in products, both voice and data, and a corresponding decrease in prices.

Nokia were the first company to produce an integral GSM phone and organiser (the Nokia 9000) but Motorola, Apple and other companies are reported to have similar products close to market. The Nokia 9000 has limited processing power but is capable of hot key calling of phone, Fax, SMS and internet services, as well as hands-free operation. It is envisaged that some of these products could be configured for use in a MoODS terminal.

Another new GSM product is a combined portable phone and data card which has a RS232 interface on the base of the phone.

3.6 Software

Geographic Information Systems (GIS) needed as a basis for MoBIC have gained further popularity within the recent years. Because of the availability of new technology new application fields have been explored for GIS's. The usability of satellite based position signals built up the basis for the development of navigation systems. Still, the main objective of GIS based navigation systems is the assistance of vehicle drivers, but with MoBIC we have expanded the application field of navigation and orientation systems, as the MoTA is designed as a navigation aid for visually-impaired (and elderly) disabled pedestrians.

Data Sources

The Geographic Data Files (GDF) standard has been developed to meet the needs of professionals and organizations involved in the creation, update, supply and application of referenced and structured road network data. It has been created to improve the efficiency of the capture, production and handling of road related geographic information by supplying a common reference model on which users and producers can base their requirements and product definition, respectively. The current version of the GDF standard is v2.2. GDF maps formed the basis of the maps used for the first field trial in Berlin but as no GDF maps were available for the Birmingham area of the UK, Ordnance Survey maps were used for the second field trial.

GDF maps now come under the jurisdiction of CEN/TC 278 (Road Transport and Traffic Telematics) and it is their remit to ensure that GDF maps are eventually available for the whole of Europe. These maps could form the basis for any commercialised MoTA but this GDF data would need to be supplemented with specific national information.

Data Bases

The storage of geographic data within GIS depends on database systems. In the past decade the field of database systems was dominated by relational databases (RDB). In contrast to the good features offered by relational database sytems there are severe limitations. For example, the modelling of complex objects to store spatial data as was needed within the MoBIC software is complicated and inefficient.

An alternative to relational database systems are object-oriented database systems (OODBS). They have become available on the market during the last years and address many of the limitations of relational database systems. After discussing the advantages and disadvantages of both object-oriented and relational systems it was decided to use an object-oriented database within MoBIC.

Competitive Systems

On systems which are equivalent to MoPS, the only product which is on the market, and then only in the USA, is Arkenstone’s Atlas Speaks. This was demonstrated at the C-SUN Conference in Los Angeles in March 1996, and at IMC8 in Norway in May 1996. Although the interface to the system has been improved, it has no route finding features at all, which may make its long term usefulness by blind people limited. Arkenstone are still publicizing the system they are developing which is equivalent to MoODs, to be called Strider, but this is not yet on the market.



4 Main Achievements and Results

4.1 Requirements and Initial Evaluations

The first task in designing the MoTA was to establish the current state of knowledge of the orientation and mobility problems of visually disabled and elderly people, and the efforts to solve them. A thorough review and theoretical analysis of the research so far was made. This included research on traditional mobility aids such as the long cane and the guide dog, and on the so-called electronic travel aids, as well as on efforts to find aids useful for the orientation of the blind traveller, that is problems of knowing where you are, from where you came and in what direction the goal is located. This review of current knowledge formed a firm basis for the continued work in the project.

The next step was an extensive study of what blind people might find useful from a macro-navigational system and how the information should be presented to the users of the systems. The philosophy was that the design should be driven as much as possible by potential users rather than the technology. A number of methods for eliciting user needs were employed. In-depth interviews with both potential users of the systems and mobility officers, rehabilitation officers and others involved in the training and rehabilitation of blind people were conducted. The latter group will subsequently be referred to collectively as mobility professionals. Group discussions with potential users also took place. Finally, simulations of how the MoODS system might function were made for potential users to try out to enable them to have a concrete experience of what the system might be like and enable them to comment on different possibilities.

4.1.1 Interviews with potential users of the MoTA

In-depth interviews were conducted with twenty-four blind and partially sighted people. Table 1 shows the categories of suggestions for information which could be provided by the MoTA system which were made by these interviewees, with specific examples within each category. An interesting suggestion was that roads and junctions could be graded for their level and speed of traffic. Of course, this information would only be a general indication of what a traveler might expect, as one can always encounter a speeding car even in a quiet street. This system of grading could also be used in relation to the amount of street furniture and general clutter (e.g. stands, displays from shops, parking meters, lampposts) on streets.

Potential users were asked to rate the acceptability of a number of suggested methods of making requests for information and receiving information from the MoTA. They were asked to make a rating of between 1 and 7, where 1 indicated they would definitely not want this option and 7 indicated that they would be perfectly happy to use this option. Figure 1 shows the mean ratings for 6 possible options. For requesting information or issuing commands to the system, the two possibilities considered were via a keyboard or via a voice input system (which would involve users wearing a small microphone, possibly on their clothing or on a shoulder strap). There was a significant preference for the keyboard which received an average rating of 5.62 in comparison to a rating of 4.45 for the voice input system (on a repeated measures F test: F 1,44 = 8.00, p < .0001). In discussing the use of a keyboard, potential users stressed the need for very simple key commands using a very small number of keys which would be feasible to undertake while walking. They felt that anything complex would be difficult to cope with in conjunction with a long cane or guide dog.

Table 1: Suggestions from in-depth interviews for information to be provided by the MoTA system

User's current location

63%

Directions to destination

71%

Layout of the environment

46%

Street information

67%

Roadworks

54%

Street furniture1

29%

Pedestrian crossings I

50%

Useful items in street

29%

Useful buildings and landmarks

58%

For receiving information from the MoTA, the three main possibilities considered were synthetic speech, non-speech sounds and vibrational information (for example, bands on the two wrists would vibrate to indicate whether the user should turn left or right). Synthetic speech received significantly higher ratings on the acceptability scale (mean rating for synthetic speech: 6.25; mean rating for non-speech sound: 3.45; mean rating for vibrational information: 3.39; on a repeated measures F test: F 2,44 = 627.43, p < .0001). Potential users were also asked to rate the acceptability of wearing headphones to receive the synthetic speech and non-speech sound information. This received the low mean rating of 2.42. A number of interviewees suggested that a device like a mobile telephone, which one could bring to the ear when one needed information, would be preferable to headphones. However, a limitation of a phone-like device is that it is much more difficult to quickly give users information (e.g. approaching potential hazards approaching).

4.1.2 Interviews with orientation and mobility professionals

In-depth interviews were also conducted with eight officers involved in training and rehabilitation of blind and partially sighted people. Table 2 shows that a similar range of information was suggested by these officers as had been given by the potential users, though the mobility professionals tended to focus on the safety and wayfinding aspects of navigation whereas potential users included more things of general interest (e.g. shops, restaurants, libraries) The mobility professionals were also asked to rate the various methods for requesting/receiving information from the MoTA system. The overall pattern of the ratings was very similar to those given by potential users, with no significant difference between the two groups (on a two-way analysis of variance: F 1,29 = 1.21, n.s.).

The mobility professionals agreed that training in the use of the MoODS could be quite easily incorporated into current long cane and guide dog training practices. The use of the MoPS could also be taught in the same way that some institutions currently teach pre-vocational computer courses to small groups of blind users.

Table 2: Suggestions from in-depth interviews with mobility professionals for information to be provided by the MoTA system

User's Current Location including direction they are facing

Directions to destination for example number of turns (left or right), how many roads to cross

Layout of the environment for example steps (particularly going down)

Street information including numbers of houses and buildings, safe places to cross a road, bus stops

Roadworks including scaffolding, diversions

Street furniture especially on an unfamiliar route

Pedestrian crossings including type & location

Useful items in street including post boxes, telephone boxes

 Useful buildings and landmarks, for example filling stations (for dropped curbs), schools, banks

4.1.3 Discussion groups

Two discussion groups about the possibilities of the MoTA system were held to augment the information obtained from the in-depth interviews and Table 3 shows the suggestions which arose. These were again very similar to those which had been raised in the in-depth interviews. In the discussion groups, more general issues were aired including the circumstances in which someone would use the MoTA, and the safety of travelling with the system. For example, there was discussion of how MoPS could be used to plan a trip to a new city or locality and how MoODS could then be used for guidance on arrival at the locality.

Table 3: Suggestions from discussion groups for information to be provided by the MoTA system

Directions to destination including to a particular location, length of journey, option to follow different routes

Layout of the environment for example whether a street slopes, steps down

Street information including names of roads

Roadworks including accidents and other transient obstacles

Street furniture

Pedestrian crossings including type of crossing, location

Useful items in street for example pillar boxes

Useful buildings and landmarks for example bus stops, post offices, shops

When discussing the possibilities for the methods for requesting/receiving information from the MoTA, speech presentation was clearly thought to be the most preferable way of receiving information, although consideration was given to deafblind people who might benefit from vibrational information. Nearly all participants in both groups expressed serious reservations about headphones, with most being unable to accept the possibility that such equipment would not cut out vitally important environmental sounds. Touch tablets and touch screens were discussed as possibilities for the indoor map system, but generally thought to be too complex.

The results of the group discussions revealed many points consistent with the results obtained from the interviews with potential users. This is despite the fact that the mean age of the potential users was 40 years, whereas members of the first focus group were all over 64 years and current levels of independent mobility were much lower.

4.1.4 "Wizard of Oz" simulation of the MoODS system

Although the interviews and discussion groups yielded much useful information for the design of the MoTA, it was felt that many participants were having difficulty imagining what the systems might actually be like and how they could work. This is not surprising, as this is a completely new type of travel aid for blind people which no-one has had any experience with as yet. As an attempt to overcome this difficulty, a "Wizard of Oz" study of the MoODS system was conducted. Wizard of Oz studies are sometimes conducted in the design of complex computer systems (Landauer, 1987). In this type of study, the complex functionality of the system is simulated by a human being (hence the name of the technique which refers to the man behind the curtain in the book and film "The Wizard of Oz").

In the case of the MoODS system, when the studies were conducted, work on the integration of the GPS and GIS components was still being undertaken, so a human being acted as the GPS receiver for the blind traveller. Participants walked the route with their usual mobility aid, carrying a Walkman cassette player on which navigational instructions had been recorded, and accompanied by a sighted person. The pre-recorded instructions guided them from point to point along the route. At each point they paused and listened to the next instruction before proceeding. At the end of the route participants were asked a number of questions about their experience and asked to rate the system on a number of scales. It was stressed that what they were experiencing was only a simulation of part of the eventual MoODS system.

This procedure allowed us to achieve a number of aims: to give potential users a more concrete idea of what using the MoODS system while travelling might be like; to see whether the nature of the speech messages given was appropriate (e.g. level of detail, format etc); and to test a number of different headphones and earpieces to try to find a solution to the problem of needing to wear some kind of headphone or earpiece to quickly receive speech and sound signals while travelling.

4.1.5 Initial evaluations and recommendations

Based on the various studies conducted with users and others involved in the mobility of blind and partially sighted people, a set of designs for the MoPS and MoODS components of the MoTA was developed and initial evaluations conducted with various components of both systems. For example, eight potential users of the MoTA took part in the evaluation of a prototype keypad for use with the MoODS.

4.1.6 Conclusions

The design and implementation of the initial MoTA is now complete. Potential users of the system were consulted before any design decisions were made and throughout the life of the project their opinions were sought on a number of design options. The resulting comments and suggestions were then used to drive the design at all its stages. We believe that this extensive involvement of potential users and mobility professionals will mean that the MoBIC Travel Aid is designed to meet the real needs of blind travellers. It is also our hope that the information gathered will be of use to others working in the area of blind mobility.

4.2 System Design

The system design of the MoBIC Travel Aid (MoTA) has evolved throughout the lifetime of the project, the prime objective at all times being to match the user requirementas researched by other partners, with a realisable design using state-of-the-art technology. Throughout the project an iterative design process has been used. An initial design phase, when pre-prototype designs were considered based on current technology only, was followed by a redesign based on the recommendations which came out of the study of user requirements. The resultant prototype was evaluated during the first field trial and recommendations for improvements, including keypad functionality, together with the numerous advances in bodyworn computing, satellite positioning techniques and mobile communications, were taken into consideration when designing the final prototype. This design was used successfully for the six month period of the main field trial and culminated in further design and functional recommendations being made in preparation for the commercialisation of the two constituent parts of MoTA.

4.2.1 Functional description

The MoTA system consists of two components – the MoBIC Pre-journey System (MoPS) and the MoBIC Outdoor System (MoODS). The MoPS component facilitates the exploration of an previously unknown area and the selection and preparation of a route. It aids increasing the independent mobility of a visually-impaired person in a way that a user can get acquainted with an unknown area until he/she feels completely safe to actually walk out. For this purpose several presentation media are used, including verbal descriptions of routes and virtual exploration facilities. Natural language descriptions can be obtained for routes with varying detail enabling the user to customise the output to his preferences. In a virtual exploration mode the user can control an agent through the city. During his walk the environment is described. The user can stop the agent and inquire about his surroundings, e.g. receiving information about the location of shops around him and their opening hours or learning about schools or bus stops in the vicinity. As thorough evaluations revealed this exploration mode leads to a solid knowledge of an area and increases dramatically the confidence in being able to safely walk the route.

The system has been equipped with an adjustable human-computer interface to allow for the different disabilities and preferences of the users. Users can obtain synthetic speech output as well as textual information on the screen which can be mapped on a Braille line. To assist a user group with residual vision, the size and the colours of the visual presentation can be adjusted variedly. For the input, text mode or menu control can be used alternatively using a standard keyboard or a mouse depending on the degree of remaining vision. Several features ease the user dialogue, e.g. synonym handling ("open map" and "load map" result in the same action) and address databases support allowing the specification of commands like "search route from home to firm").

Once the route is set up, it may be downloaded in the portable component. The MoODS component gives assistance during the walk. It consists of a small wearable PC kernel the size of 16 x 11 x 7 cm. Speech information is given through a pair of special earphones which were selected to prevent the masking ambient sounds essential for echo location. A small and inconspicuous wrist-worn keypad can be used to inquire about information. The system provides on-route information about the current position of the traveller and his/her environment (e.g. "You are in Harborn Road near a bakery"). It helps the person navigate the chosen route. On pressing the designated button of the small keypad, the system responds with "Go ahead 50 metres to the next junction and turn right" for example. The positional information is received from the Global Position System (GPS). An electronic compass is used to give an indication of the direction of travel. As the user walks the selected route, clock type directional instructions can be obtained (e.g. "change direction to 3 o'clock and continue for 100 metres") based upon their computed position. For safety reasons, the system informs the traveller automatically when they are leaving the chosen route or if the accuracy of the system has degraded. For the same reason a mute button has been added to keypad to allow a sudden stop of the speech when reaction to the environmental sounds becomes necessary; the previous information can be regained using the repeat key.

4.2.2 Hardware

Portable PC

A portable travel aid like the MoBIC device requires specific characteristics of electronic components e.g. small size, low power, low price, plug and play technology, standardised format, high quality and reliability. So during the project several changes have been made to the selection of hardware components. These changes have been made to satisfy, on the one hand, the ongoing technology development and on the other hand the user requirements worked out during the project to flow into the different stages of prototyping.

Preprototypes I

The hardware design has been realised in different steps. In step 1, the MoODS and MoPS hardware was assembled using standard components available on the market. Exceptions were cables, adapters and other modifications to make the hardware work. This device was used as Preprototype I (first operational device) for initial studies.

The hardware components are shown in the table below, all of which were standard, ‘off the shelf’ components.

MoODS system

Notebook, Dock-Under-Station

RDS-decoder (DGPS-receiver)

Rockwell (GPS-receiver)

V24-Standard-Card (4xRS232)

Handy-Telephone (with Data Card)

MFII Standard Keyboard

Battery-Pack, Power-Management

Infovox (Speech output)

MoPS system

MoODS

LOGIN-Station

Prototype II (Berlin)

Personal computers in credit-card size became available on the market just before the Berlin field trail started. The manufacturer EPSON started production of a 486 PC in a case like a credit or PCMCIA card. The actual device used for the Berlin field trail was a 486/25 Mhz PC with 8MB RAM. So the selection of the PC104 standard in conjunction with the Epson Card-PC gave flexibility and the potential to simply update the hardware by changing PC-Cards. Carried in a kind of waist-coat this hardware was used for the Berlin field trail.

During this phase, having completed various tests on GPS receivers, it was decided to change to a Trimble Svee6 which gave a better overall performance and was more convenient to configure. This receiver was used for the rest of the project.

Prototype III (Birmingham)

Due to rapid changes in technology, the Birmingham hardware was designed with smaller and lighter components than before. It appears, for optimum use of space, that the PCMCIA technology was the right choice for the speech output and GSM handy communication. The use of both the PC-104 industrial standard for the PC kernel and the PCMCIA standard, reduced the size by about 20% of the Berlin hardware.

The number of available PCMCIA cards for different operations is increased, so that most interfaces have a PCMCIA solution (e.g. MODEM, GPS-Receiver) - PCMCIA boards are the solution for small and light weight components. Most of the notebooks sold today already contains slots for cards using this technology. The Personal Computer Memory Card International Association (PCMCIA) is responsible for the standardisation in this field.

Wearing the equipment

The MoBIC waist coat and the rucksack were specially created by an industrial designer. The drawing below shows the position location details for each of the components. The positions chosen for each of devices are strongly recommended as other locations could cause system malfunctions (the compass and the GPS antenna have to be mounted at the highest locations possible i.e. on the shoulders).

4.3 Training

During the field trials of MoBIC participants were trained in the use of both MoPS and MoODS. The training needs of potential users are important factors in the commercialisation of MoBIC, therefore a section is devoted to this topic. The following areas will be described:

  • The design of the training material
  • Training carried out in the Birmingham field trial
  • Summary of the findings
  • Recommendations

4.3.1 The design of the training material

The prepared structured training materials involve ten tutorials. Each tutorial takes no more than an hour unless travelling is required. Each tutorial has two sections: one for the user and the other for the trainer. The section for the user is provided in appropriate format (Braille, audio cassette, or large print). The ten tutorials are divide into three levels of difficulty - basic, intermediate and advanced. The difficulty reflects the complexity of the environment explored and the number of commands to accomplish the goals of the tutorial.

Each tutorial was designed to be delivered to a single participant by a trainer. An exception to this is an additional introductory tutorial which presents an overview of the MoBIC system which can be presented to a group.

It was anticipated that a typical training session would consist of review of progress, followed by instruction, followed by the agreement of practice objectives (which the participant would try to achieve before the following training session). Thus the training was planned as a cyclical process of review, teach, and practice. This would progress at an appropriate pace throughout the training until it was felt the user had reached an appropriate point.

4.3.2 Training carried out in the Birmingham field trial

In practice, we found that in the course of the training the content of the ten tutorials could in fact be captured in four general tutorials/stages:

Tutorial 1 Area exploration

Tutorial 2.1 Planning a route

Tutorial 2.2 More about planning a route

Tutorial 3 Using the outdoor system

It was felt this was adequate because additional complexity could be introduced by repeating the tutorials but using more complex and longer routes. Once the commands were mastered it was the route itself which determined the difficulty of the task. Additionally, the trainer encouraged the participant to specify routes and destinations, rather than the training program being determined wholly by the trainer.

Each of the tutorials was presented orally by the trainer in the first instance but copies of them were given to participants in Braille, large print or on audio cassette which they stored in a folder (thus constructing a reference guide).

Trainers visited each participant’s home weekly. These visits lasted at least one hour. The progress with which participants progressed varied a great deal. This progress was recorded by the trainer on a session record sheet. Participants were also encouraged to practise using MoBIC between training sessions. If this had taken place, the trainer and participant spent time discussing this practice.

Additional material was prepared during the trial as it was needed, e.g.

  • Tactile representations of T-junctions and cross-roads
  • Tactile representations of compass points
  • A reference sheet of local road names (in grade one Braille - to aid participants with poor spelling)
  • A reference sheet of commands (in grade one Braille)

Additionally, a number of strategies which were adopted by the trainer; for example, participants were encouraged to inspect junctions when they encountered them using MoPS (using the left and right keys to gain a spatial understanding of the decision point). Another example, is for participants to repeatedly listen to complete route descriptions in order to incrementally construct an overview of the route.

4.3.3 Summary of the findings

The content and process of training have two main determinants: (1) the objectives or skills to be taught, and (2) the skills the participants already possess. With this in mind, a useful distinction which can be made when discussing the training is between the following inter-related areas:

  • Learning to use the ‘technology’
  • Learning to carry out the underlying task

It is inevitably difficult to make such a distinction in this way because the technology makes the task (planning and making a journey) possible; however it is still a useful distinction for summary purposes.

In the case of the technology, participants must learn (or already know) general skills related to technology - e.g. understanding speech synthesisers, understanding the process of loading a file, being able to touch type, etc. Even with all these skills already in place, learning a new system with its new commands, syntax, etc will always have some training overhead. Indeed some participants found learning to use the technology very difficult, and much time was spent teaching them these skills.

In the case of learning to carry out the underlying task (i.e. plan and make journeys), it was also found that there was a large variance among individuals. This ranged from participants who appeared unable to develop these skills at all with training of one hour per week, to those who were able to develop the skills with relatively little training.

It is worth also discussing more explicitly the overlap between technology and underlying task. There are a number clear instances where the limitations (or configurations) of the current prototype have important implications for training. For example, the analyses of the MoODS positioning system, demonstrated that in particular built environments the positioning system was inaccurate. As well as implications for technical developments, it also has implications for the teaching of strategies to overcome these difficulties. Those who commercialise MoBIC must decide which difficulties encountered by participants can adequately be overcome by training, and which need additional technical development.

4.3.4 Recommendations

To account for the large differences among individuals described above, it is clear that different types of training will be required. As a general recommendation, this might fall into two categories which will be discussed in turn.

(1) Rapid intensive training

This type of training would involve quite intensive courses in the use of MoBIC, e.g. several single day courses, a large amount of practice alone, a help line. This would cater for potential users who are both comfortable with technology and have well developed mobility skills.

(2) Longitudinal training

This type of training would involve a large amount of training over a prolonged period of time. For example, it might be tied in with general mobility training within a school, college, or rehabilitation training centres. This would cater for the potential user with poor mobility skills, or with mobility skills still developing.

In both cases (particularly the latter) there is a need for much training input. Since this is expensive it is likely that this must be supported by various national organisations with relevant interest, or who in any case have specific responsibility for providing instruction in orientation and mobility skills. In the case of the UK this would include the national charities (e.g. Royal National Institute of the Blind, Guide Dogs for the Blind Association), and such statutory bodies as the Local Education Authorities and Social Service Departments.

4.4 Evaluation

User involvement is vital for the successful development of complex high technology products and services. The needs of the user must be introduced into the design process at the earliest possible stage. This is so that technology decisions do not lock the system development into solutions that are unsuitable for the customer. Retrospective changes are costly and complex if a usability barrier is not discovered until development is completed and a product is in the marketplace.

Within the MoBIC project the introduction of the needs of the user has been approached in three ways:

1. User requirements elicitation studies capture

2. Usability studies of system components and prototypes

3. Field trials

Formative evaluation has played a key role in the usability studies and field trials within the MoBIC project. The usability studies have been reported in this document as part of the requirements capture activities.

There were two field trials were carried out within the MoBIC project. The first field trial was held in Berlin (for 3 months) to evaluate the first technical integration of the hardware and software components into a functioning system. The second, in Birmingham (for 6 months), to evaluate the use of fully functioning prototype aids by a group of blind and visually disabled volunteers at home and in their local neighbourhood.

At all times during the MoBIC project there has been a very close relationship between the evaluation and engineering teams in the project. This has enabled formative feedback from the trial sites to be quickly included into software updates and hardware selection and configuration. The role of evaluation in the project has at all time been to provide a formative path from user experience into the technical design of the system.

4.4.1 Berlin Trial

A field trial at the Free University of Berlin was used to test the technical system integration of the first design, and gain initial feedback on the core functionality of the MoBIC prototype system.

An area for the Berlin trial was surveyed and test routes selected. Contact with potential clients was established through the Berlin Blind Association. A total of six blind and visually impaired people were accepted for participation in the field trial. Training materials were translated into German and prepared for use in the Berlin trial site. Digital map databases were checked on the trial test routes and some additional information was added manually.

In total, 25 trial sessions were held with the six participants. Written notes were taken and video recording of tutorial sessions were made for analysis of human-computer interaction, for recommendations and for error reporting on MoTA components. Since the software was modified during the trial to deal with the problems that were uncovered, the training materials were modified to keep in step with the upgrades.

Members of the MoBIC project team from the Universities of Uppsala, Birmingham and Hertfordshire collected data on six of the Berlin field trial sessions.

The Berlin field trial was a major milestone in the development of the MoBIC travel aid. It was the first time that the MoBIC system components were integrated into a functioning unit that blind people could use. It was the first time that a European self positioning navigation aid was available for visually disabled people to try out and directly contribute to the development of its design.

The Berlin trial resulted in improvements to the prototype MoBIC travel aid and to the training materials and evaluation methodology used in the Birmingham trial.

4.4.2 Birmingham trial

The Birmingham field trial was the second and final field trial in the MoBIC project. It made use of fully functioning prototype systems that participants were loaned to use in their own home. The field trial lasted six months (March to September 1996) and involved 13 visually disabled participants.

The trial participants were allocated into one of two cohorts (cohort one covering the first three months of the trial, or cohort two covering the second three months of the trial). A single participant was involved for the entire six month trial period. Seven of the participants took part in intensive evaluation sessions in which they explored and walked two specified routes while number of measures were taken.

The trial demonstrated that MoPS was successful in enabling blind people to increase their knowledge of a given environment, in terms of them learning about:

1) places and services,

2) names and layouts of roads and junctions,

3) the spatial relationship between different locations,

4) new routes.

MoODS successfully enabled the participants to exploit this new knowledge by walking to new places along new routes. MoODS helped people in a number of ways, some of these were:

  1. to reassure users of their current position (e.g. the system would tell participants that they were on a given road and it was a given distance to the next junction).
  2. to help participants make decisions at relevant places along a route (e.g. participants would ask the system which way they should go when reaching a side road).
  3. to alert participants when they have gone off route (e.g. participants would be told they were leaving the route if they walked past a relevant side road).
  4. to reduce the burden of remembering a route.

Within the trial the relative success of MoPS and MoODS varied greatly from individual to individual. Indeed, a key issue that came through from case studies into the use of the prototypes, was the importance of individual differences between participants. The skills and interests that some participants brought to the trial helped them overcome some of the technical difficulties of they encountered. Whereas, other participants were much less successful.

4.4.3 Planning a journey using MoPS

Trial participants were able to break down their journey sub-goals with relative ease, e.g. go to the bus stop, go from the bus stop to the supermarket, etc. However, it was not easy to specify destinations to MoPS; the user was only able to specify destinations by typing an address. People typically have a need or interest, e.g. they need groceries, or they are interested in records rather than a specific destination address. A simple model of the way in which MoPS was used by trial participant is as follows:

1) The user must first want to make a journey.

2) They must then generate a number of problems they need to solve in order to do this.

3) They must then present this information to MoPS.

4) MoPS must then process this and present solutions to the user.

5) The user must understand the mode of presentation, and

6) interpret its content.

7) Having understood the solution, the user must decide whether it is adequate before,

8) making the journey.

Some of the trial participants were able to use MoPS successfully, i.e. they were able to successfully carry out all the above actions. For example, one trial participant decided he wanted to go the a local record shop. He planned the journey and understood the route he had explored using the system. There are many examples of participants using MoPS in this way.

There were trial participants who had problems issuing commands, and understanding and interpreting system responses. It is likely that improvements to key aspects of the existing user interface will enable more people to benefit from access to this device. Even so, it is predicted that some participants will still find the use of a verbal description to represent routes too difficult. Additional functions, for example to improve the method of specifying destination points or increase the safety of planned routes, will be required to broaden the accessibility of MoPS technology amongst blind people.

4.4.4 Making a journey using MoODS

Generally, participants wanted to use MoODS. In some cases participants were anxious about their appearance, the weight of the device, and the attention it might draw, but these concerns were rarely raised. When using the system, participants generally wanted to first establish where they were on the route, and then what the next part of their journey should be. A simple model of the way MoODS was used is as follows:

1) The user must first want to walk a journey using MoODS.

2) They must then generate a number of sub-goals they need to solve in order to do this.

3) They must then issue commands to MoODS (in fact MoODS can be used effectively without the user issuing many commands, because the system will give relevant information automatically as the user moves through the environment).

4) MoODS must then process this and present solutions to the user. The user must

5) understand the mode of presentation, and

6) interpret its content.

7) Having understood the solution, the user must decide whether it is adequate before,

8) acting upon it.

Some of the field trial participants were able to use MoODS successfully, i.e. they were able to successfully carry out the actions given above. For example, one participant used MoODS to help him walk to a record shop, a route he had planned using MoPS. There were other examples of participants successfully using MoODS in this way.

Field trial participants generally did not have any difficulty understanding how to use MoODS or how to issue commands. However, some participants found it difficult to understand the speech messages, and to interpret the compass directions, or the information range. The main problem the participants experienced was the unreliability in the availability of GPS self- positioning. It is likely that, if these problems are addressed in the future, then blind people will successfully use MoODS type devices on a regular basis.

Analysis of positioning data

When MoODS is used, position data for that journey is stored in a log file. These log files can be analysed to establish the availability and accuracy of the positioning signal at a given time. As part of the Birmingham field trial the following two analyses of the MoODS positioning data were carried out.

Analysis of log data collected from a test route

In this analysis, a 3 km test route was walked with a MoODS device on six separate occasions. Analysis of the resulting log files showed that both the availability and the accuracy of the GPS signal were affected by the nature of the surrounding built environment - with poorer signal in shopping streets with tall buildings. And, that the GPS signal was not affected by smaller environmental features such as overhanging branches of trees. There were also indications that satellite availability changes with time (different numbers of satellites being available in the same position on different days).

These results throw some light on the problems which trial participants encountered. They suggest both that there are limitations on the use of GPS and DGPS for positioning blind pedestrians and that there are strategies which MoODS users can be taught to overcome them.

Analysis of log data collected throughout the trial by participants

In the second analysis, over 7000 fix attempts were studied. Although overall availability of GPS and DGPS signals was high (90.7% of the time GPS was available), this may be an artefact of the analysis, since a number of short journeys were used for simplicity of analysis.

Interestingly, although overall availability was high, there were many short outages. On one typical longer journey, there was an outage every 85 seconds. Not all of these outages would be long enough to be notified to the MoODS users, but may affect the reliability of the system.

It was also found that the number of satellites in view was lower than expected (typically 3); this may be due to the urban environment used for the trial. The accuracy of the signal was also somewhat disappointing, being "very good" or "adequate" on only 61.1% of occasions. Further analyses are necessary to clarify these issues.

4.4.5 Discussion

Participants responded positively to MoPS. There are many examples of participants learning about shops and services in the local area, learning about the names and shapes of streets and junctions, and learning routes they did not previously know. It is likely that blind people will be able to gain benefit from the provision of access to electronic maps on their own.

There are also examples of participants using MoODS very effectively when walking a route; using it to make decisions about directions to go at junctions, being alerted when leaving routes, and being reassured when on routes. However, the units self positioning ability was not always available when it was needed by the trial participants.

A key technical issue for MoODS is satellite availability and position accuracy in urban environments (particularly when the user is in close proximity to tall buildings). Obvious areas where further development is needed in MoODS are the technical performance of the self positioning system and user strategies for dealing with position loss.

Typical of the comments by the Birmingham trial participants about the prototypes they used are the following; each quotation comes from a different participant in the Birmingham trial:

". . .is a fantastic aid"

"Without it you are disadvantaged"

"I think it would be more important to myself than any other equipment"

"Put it this way, I want one"

The trial participants were generally very positive about the facilities offered by a MoBIC travel aid. For many this type of system, if reliable and robust in operation, was a desirable product to own or have provided for personal use.

Within the interviews in the Birmingham trial participants were asked about the price they would be willing to pay for products resulting from the MoBIC project. There was a great deal of variation between individuals, however if a single price value is needed to express the typical view of the participants, it is:

Purchase a full system:

1250 ECU

Rent a full system (per calendar month):

30 ECU

Purchase map access software:

125 ECU

Purchasing decisions for assistive technology in the UK are subject to assistance from statutory and voluntary bodies. As such, the cost may need to reflect their cost limits rather than those of the end user. However, this form of assistance may only be available in specialist situations such as it being a necessity for work or for education.

If production systems can be provided at the cost level of conventional domestic technology, more people in the UK may be able to benefit from increased mobility due to MoBIC than from current UK voluntary or statutory provision alone.

4.4.6 Conclusions

The field trials have been very successful in feeding through user experience into the design of MoPS and MoODS. Project members from different parts of Europe have worked at the field trial sites to contribute to the evaluation of MoBIC prototypes. This has been a multi-disciplinary effort with computer scientists, engineers and social scientists taking part in the field trial activities.

As a result of this close co-operation within the MoBIC project it was possible to develop fully functioning prototypes and to train blind people to use them in their own homes. Thus, for the first time in Europe it was possible for a group of blind people to have access to complex state of the art self positioning technology for personal navigation and orientation. And, for some of these people, despite the technical problems that were encountered, to be able to use these prototypes to plan and travel journeys in their local neighbourhood.

Blind people can benefit from access to electronic maps from a MoPS type system. This has the potential for rapid exploitation as a product from the MoBIC project. Availability of GPS in urban areas is a problem that requires further technical work to be carried out if these systems are to be used near objects that may obscure sight of the satellites from which the units position is determined.

4.4.7 Recommendations

The Birmingham field trial has demonstrated that there are blind people who, with appropriate skills, motivation and training can successfully use MoPS and MoODS to plan and navigate journeys. Participants were generally very positive and excited about the MoBIC project and the potential that an orientation and navigation system had for them. It is strongly recommended that MoPS and MoODS technology be developed further to find the technical solutions necessary for them to become available as products for use by blind people in Europe.

Initially an electronic map access product should be commercialised from the work of this project. If the product is intended for used in urban environments, after further development to deal with the GPS availability problem, this should be followed with an outdoor unit.


5 Publications

Bozic, N., Douglas, G. & Murdoch, H. (1996) Using the concept of the Functional System to help structure the implementation of educational technology in intervensions with children and adults who have disabilities. The 1996 Annual Conference of the British Psychological Society: Piaget-Vygotsky Centenary Conference, Brighton, 11-12th April 1996.

Douglas, G. (1996) MoBIC: a navigation aid for the blind. Optician, 212, 16-17, October 11.

Douglas, G. (1996) The MoBIC Project. Centre Software Newsletter, No. 40, 8-13.

Fritz, S., Michel, R., Raab, A., Petrie, H. & Strothotte, T. (1996) MoBIC: designing an travel aid for blind people. Journal of Navigation, 49(1), 45-49.

Fritz, S., Michel, R., Raab, A. & Strothotte, T. (1995) User interface design for a travel aid for blind people. Workshop on Non-visual graphical user interfaces for blind people. Software Ergonomie. Darmstadt, March 1995.

Gill, J. M. (1996) An orientation and navigation system for blind pedestrians. London: Royal National Institute of the Blind.

Holmes, E., Michel, R. & Raab, A. (1995) Computer-supported exploration of digital maps by blind people. Kolloquium Taktile Medien, Technische Universität Dresden, 24-26 November 1995, 81-87.

Holmes, E., Jansson, G. & Olsson, E. (1996) Tactile enhancement of the reading of a digital map presented via synthetic speech. Proceedings of maps and diagrams for blind and visually impaired people: needs, solutions, developments, Slovenia, 21-25 October 1996. London: International Cartographic Association, Commission on Maps and Graphics for Blind and Visually Impaired People.

Holmes, E., Jansson, G. & Jansson, A. (1996) Exploring auditorily enhanced maps for travel in new environments. In: Burger, D. ed. New technologies in the education of the visually handicapped. France: John Libbey Eurotext. pp 191-196.

Holmes, E. & Jansson, G. (1997) A touch tablet enhanced with synthetic speech as a display for visually impaired people's reading of virtual maps. California State University Conference on technology and persons with disabilities, USA, 18-22 March 1997.

Jansson, G. ed. (1995) Requirements for effective orientation and mobility by blind people. London: Royal National Institute of the Blind.

Jansson, G. (1995) MoBIC - utveckling av en ny typ av orienteringshjälpmedel för gravt synskadade. Rapport från forskningskonferensen Människa Handikapp Livsvillkor, Örebro, 5-6 April 1995, 17-20. Örebro: Örebro L ä ns Landsting, Habiliteringsförvaltningen.

Jansson, G. (1995) Independent reading of tactile pictures by severely visually impaired computer users. Proceedings of the ECART3 Conference, Portugal, 10-13 October 1995, 56-58. Lisbon: National Secretariat of Rehabilitation.

Jansson, G. (1995) The MoBIC project: a navigation system for blind and elderly people. International Hi-Tech Forum Osaka '95: Advanced Navigation Technology - from car navigation to escort system, Japan, 27-28 November 1995, 57-63.

Jansson, G. (1996) Old and new efforts to find technical solutions to the orientation and mobility problems of the blind. In: Tellevik, J. M. & Haugum, G. E. eds. Proceedings of the International Mobility Conference No 8, Norway, 10-18 May 1996, 148-158. Melhus: Tambartun National Resource Center.

Jansson, G. (1996) Some properties of touch and haptics important for the usefulness of tactile displays. Rank Prize Funds Symposium on technology to assist the blind and visually impaired, England, 25-28 March 1996.

Jansson, G. (1996) Possibilities and problems with tactile maps from a perceptual point of view. Abstracts, maps and diagrams for blind and visually impaired people: needs, solutions, developments, Slovenia, 21-25 October 1996. London: International Cartographic Association, Commission on Maps and Graphics for Blind and Visually Impaired People.

Jansson, G., Holmes, E. & Jansson, A. (1995) Spatial information during locomotion without sight. Poster at the Eighth International Conference on Perception and Action (ICPA8), Marseille, 10-14 July 1995.

Jansson, G. (1997) Gravt synskadades orienteringsproblem - Ett försök att lösa dem i europeiskt samarbete mellan tekniker, datavetare, psykologer och pedagoger. (Orientation problems of severely visually impaired people - An effort to solve them in a European co-operation between specialists on technology, computer software, psychology and pedagogics). Forskningskonferens i Örebro, 9-10 April 1997. Människa Handikapp Livsvillkor 97.

Jansson, G. (1997) Spatial orientation and mobility of the visually handicapped. In: Silverstone, B., Lang, M.A., Rosenthal, B. & Faye, E. E. eds, The Lighthouse handbook on visual impairment and rehabilitation. New York: The Lighthouse and Oxford University Press.

Johnson, V. & Petrie, H. (1996) Evaluation methodologies for navigational aids for blind people. In: Klaus, J., Auff, E., Kremser, W. & Zagler, W. eds. Interdisciplinary aspects on computers helping people with special needs. Vienna: R. Oldenbourg.

Petrie, H. (1996) Psychology and technology for people with disabilities: report of a symposium held at the BPS London Conference, December 1995. The Psychologist , 9(2), 80.

Petrie, H. & Johnson, V. (1996) MoBIC: a new type of navigational device for blind travellers. Access IT.

Petrie, H. & Johnson, V. (1995) MoBIC: an aid to increase the mobility of blind and elderly travellers. Proceedings of Third European Conference for the Advancement of Rehabilitation Technology (ECART3), Portugal, 247-249. Lisbon: National Secretariat of Rehabilitation.

Petrie, H. & Johnson, V. (1996) Methodologies for the evaluation of navigational aids for blind persons. Proceedings of International Mobility Conference 8, Norway.

Petrie, H. (1997) User-centred design and evaluation of adaptive and assistive technology for disabled and elderly users. Informationstechnik und Technische Informatik 39(2), 7.

Petrie, H., Johnson, V., Holmes, E., Jansson, G., Strothotte, T., Michel, R. & Raab, A. (1996) Access to digital maps for blind people: results from the MoBIC Project. Proceedings of Conference on maps and diagrams for blind and visually-impaired people: Needs, solutions, developments. Slovenia, 21-25 October 1996. London: International Cartographic Association, Commission on Maps and Graphics for Blind and Visually Impaired People.

Petrie, H., Johnson, V., Strothotte, T., Michel, R., Raab, A., Reichert, L. & Schalt, A. (1997) User-centred design in the development of a navigational aid for blind travellers. Proceedings of INTERACT 97- IFIP TC13 Sixth International Conference on Human-Computer Interaction, Australia, 14-18 July 1997.

Petrie, H., Johnson, V., McNally, P., Morley, S., O'Neill, A-M. & Majoe, D. (1997) Inexpensive tactile interaction for blind computer users: two application domains . IEE Colloquium on developments in tactile displays, London, January 1997.

Strothotte, T., Petrie, H., Johnson, V. & Reichert , L. (1995) MoBIC: user needs and preliminary design for a mobility aid for blind and elderly travellers. In: Placencia Porrero, I. & Puig de la Bellacasa, R. eds. The European Context for Assistive Technology Proceedings of 2nd Tide Congress, Paris, 26-28 April 1995. Amsterdam: IOS Press. pp348-351.

Strothotte, T., Fritz, S., Michel, R., Raab, A., Petrie, H., Johnson, V., Reichert, L. & Schalt, A. (1996) Development of dialogue systems for a mobility aid for blind people: initial design and usability testing. Proceedings of ASSETS ‘96: The Second Annual Conference on Assistive Technology, Canada. 11-12 April 1996. New York: ACM Press.