john Gill technology header image

Inclusive Design of Public Access Terminals

John Gill

 

Increasingly individuals are required to use self-service public terminals if they wish to access services.  This includes automatic teller machines (ATMs) as well as ticket machines for public transport and public telephones.  At the same time many of these machines are increasing the range and types of service they offer.  However many designers have neglected to consider the needs of people with special needs (see Table 1) until the systems are in use, and have then found it prohibitively expensive to make them accessible by retrofitting new features.

Table 1:           People with special needs in the UK

Children (<16 years)                                          20%
Older people (>65 years)                                    15%
Disabled (with respect to using ICT)                 10%
Primary language not English                              5%
Left-handed                                                       10%

 

The estimates of the number of people with disabilities will depend of the definitions being used.  There are a number of medical definitions which have the advantage that they involve measurement of some aspect (eg visual acuity) but they tend to be a poor indicator of the number of people who are likely to experience difficulties doing a particular task.  Therefore functional definitions are often used, but the numbers obtained will be dependent on the nature of functional task.  There are also estimates of the numbers used by the fund-raising departments of some organisations, but the scientific basis for the statistics is often not quoted and may be questionable.

Using a functional definition of those who have problems using information and communication technology systems, the estimates of the percentage of the population in Europe are shown in Table 2

Table 2:           User with problems using ICT

Wheelchair users                                               0.4%
Cannot walk without an aid                                 5%

           Reduced strength                                               2.8%
Reduced co-ordination                                      1.4%

           Speech impaired                                               0.25%
Language impaired                                            0.6%

           Dyslexic                                                                1%
Intellectually impaired                                          3%

           Deaf                                                                   0.1%
Hard of hearing                                                     6%

           Blind                                                                  0.4%
Low vision                                                         1.5%

 

Within these groups there are considerable variations from one individual to the next.  For instance people with low vision includes those with:

 

(a)        Macular degeneration

Macular degeneration accounts for about half of all registerable blindness in the UK; it is particularly common among the older population.  Typically it results in the loss of central vision.  This group often benefits from larger than normal print.

 

 



(b)        Cataracts

A cataract is an opaqueness of the lens at the front of the eye.  It has similarities to driving a car with a dirty windscreen; if the sun is in your eyes, it is difficult to see anything, but with the sun behind you there are relatively few problems.  Fortunately cataracts can usually be operated on successfully, with the lens being replaced by a plastic lens.

 

 



(c)        Diabetic retinopathy

Haemorrhages occur on the retina at the back of the eye.  One effect of diabetes tends to be a very poor sense of touch; therefore few people with diabetic retinopathy can read braille.

 

 

(d)       Tunnel vision

Tunnel vision can be associated with some forms of retinitis pigmentosa or a late stage of glaucoma.  Retinitis pigmentosa (RP) is the name for a group of conditions which are genetic and result in night blindness.  Glaucoma is from increased pressure in the eye; if detected early enough, then it can be treated by taking pills.

 

 

In addition total colour blindness affects about 0.0025% of the population.  However about 8% of males and 0.5% of females have problems distinguishing red and green.

The ageing process affects vision.  In a sixty year old only about one third of the light reaches their retina compared to when they were twenty.  There is also a decline in visual accommodation (the ability of the eye to change focal distance) coupled with a deterioration in the speed of adapting to changes in illumination.  In addition many older people have a combination of impairments.  Also multi-tasking becomes less easy.  The effect of all these factors is that many older people may have problems in using a biometric terminal at the same speed as their younger counterparts.

More than half of people with a disability have a significant additional impairment.  This number is increasing since more people are living to an older age.  However it is difficult to predict the effect of multiple impairments since two individuals who appear to have the same combination of impairments may function very differently.  Therefore there is a need to design systems whose user interface can be tailored to the needs of different individuals.

If a selection has to be made from a large number of functions it is possible to use a large number of buttons with each button being for a specific function.  This is not inherently a bad arrangement as long as the buttons are arranged in a logical order and can be easily differentiated by the user.  Unfortunately what is ‘logical’ to the designer may not be ‘logical’ to the user.  Another possibility is to have fewer buttons but each button having more than one function associated with it; this can be very confusing for the occasional user.  Another possibility is to select from menus; if the user can clearly read the screen and understands what is required of them, this approach can work well.

The choice of font can affect the legibility of text both printed and on a screen. The character shape, text format and layout can enhance or detract from the intended meaning. In general legibility is higher with simple 'open' typefaces.
Arial is a poor typeface for numerals such as it is preferable to use more open numerals such as This example of some open characters shows how the ends of the strokes can appear to close upon the letter and make them less easy to distinguish. Legibility may be reduced further in a bold format as the characters become more 'closed' and the white space within them decreases.
In many sans serif typefaces the lower case is visually similar to the capital For instance the meaning of is more difficult to read in Arial than in Tiresias .
Increasingly password and email addresses use both letters and numbers. For such applications it is essential to use a typeface which clearly differentiates the numeral and lower case For example:

In some applications where the context does not make the meaning obvious, it is essential to be able to differentiate the zero and the capital ; in this case it may be necessary to use a cancelled zero

The user could select their preferred user interface (eg larger characters on the screen) by pressing a button or selecting from a menu.  Alternatively this requirement could be stored on a central database, but frequently this is not possible since the user is anonymous (eg ticket selling machine at a railway station).  However it is possible to store details of the user’s preferred interface on their card; there is a European standard (EN 1332-4) for how to code this information.

A common request from older users is for more time. On an ATM this may mean more time before being timed out and the money sucked back into the machine.  On a public telephone it could change the operation to ‘compose and send’ where the user can dial the number at their own speed and then press a ‘send’ button (as is done on many mobile phones).  With an automatic gate on public transport, more time would permit both guide dog and owner to get through before the gate slams shut.

Contactless smart cards can be proximity (operate at less than 10 cm) or vicinity (10 cm up to 2 metres).  Proximity cards offer the advantage of no card insertion which helps blind people and people with poor manual dexterity (eg hand tremors).  Vicinity cards offer the potential of triggering an audible message to help a blind user locate the bus stop, or trigger a message at the door of the bus giving the route and destination of this bus.

It is also possible to connect assistive devices to public terminals.  Some ATMs are in use which have a socket for headphones for speech output.  However it is the wireless systems, such as Wi-Fi, Bluetooth and ZigBee, which appear to have significant potential in this area.  For instance the user could use their mobile phone handset connected by Bluetooth to get speech output from an ATM.

It is becoming increasingly important to prove one’s identity to obtain access to certain services.  There are three elements of a person’s identity:
1.         Things which you ‘are’ i.e. your biometric identity.  These are attributes that are unique to an individual (e.g. fingerprints).
2.         Things are given to you i.e. your attributed identity.  These include full name, date and place of birth.
3.         Things which happen to you during your life, i.e. your biographical identity.  This includes educational qualifications, electoral register entries, and history of interaction with organisations such as banks.

Biometrics permits the automatic identification of an individual based on his or her distinguishing physiological and/or behavioural characteristics. Biometric identification involves comparing with a database of templates to find out who you are, but biometric verification is where the template is compared to the one supplied with your claimed identity. Some biometric systems cannot do identification but can only verify the claimed identity of a person.

Table 3:           Common biometrics

            Hand geometry
Dynamic signature
Voice
Vein geometry
Keystroke
Facial recognition
Fingerprints
Iris recognition

 

For the user, it should be easy and comfortable to use the system. Many users would prefer methods which do not require physical contact between the individual and the device. Consumers need confidence that the system will reliably correctly identify them while not permitting other users access; no current biometric system achieves 100% success in both these aspects.

Two different systems need to be designed so that they can be used comfortably by as many people as is reasonably possible.  Firstly the system which captures the biometric.  This will normally be manned by experienced personnel who should have received some training in working with people with disabilities.

Secondly there are the biometric readers which may be unmanned or have limited supervision.  It is essential that these terminals have a consistent user interface.  They should incorporate standard icons, symbols and pictograms to help the user operate the terminal.  This will be particularly important for applications such as passports where there is a significant probability that the user is not familiar with the local language. 

It would also be desirable for the user to be able to personalise the user interface.  This can be done by pressing a button or selecting from a menu, but this is time consuming for anything other a simple change.  Alternatively the user’s preferences could be stored on a central database or on the user’s card or smart media.  There is a European standard (EN 1332-4) for how to code this information.

Facial recognition can have an unacceptable level of either false positives or false negatives. It is technically best used to say “is this the same person” rather than “who is this person”. Thus it is an appropriate technology when used with a secure token such as a smart card. From the users perspective it's non-intrusive nature is a major advantage and users are likely to accept such a system if it can provide a decision quickly, and is seen to be protecting their interests.  In passport applications, a false rejection will only result in a referral to an immigration officer who can handle problems such as changes in facial appearance.

Fingerprint systems are good for the low number of false acceptances, but can be problematic for those with damaged fingers or with prosthetic hands. Some users will associate fingerprints with criminal investigations, so may be reluctant to use the system.

Iris recognition is a secure system, but the user may have to position their eye in relation to a camera. This can give problems for users who are very tall, very short, or in a wheelchair. There are obvious problems for users who are blind or have a visual prosthesis. In addition some ethnic and religious groups may consider such a system unacceptable.

The biometric information can be stored in a central database or on a smart card. Users are likely to prefer the information to be stored on their card rather than on a remote database. However, it is easier to regularly update the database with revised biometric data as the user’s characteristics change.

Users should have the facility to choose an alternative verification system even if it is a PIN. However this choice may be subject to regulatory or legal requirements imposed on the service provider. The user should be advised if the alternative is less secure, but the decision to use an alternative system should be left to the user.

The obvious advantage of biometric systems is that the user no longer has to remember PINs (personal identification numbers) and keep this number secret.  So many people with a cognitive impairment will find most biometric systems much easier to use and provide a greater level of security.

However there are a number of unknown effects such as nystagmus on iris recognition systems.  Also concerns have been expressed about the effect of an eye operation between registering for an iris scan system and using a terminal.  Another unquantified aspect is the effect of a severe hand tremor on fingerprint recognition systems.  Therefore we need  research to determine how many people will be excluded and how to inclusively design the terminals.

In any new project for use by the general public it is worthwhile considering the needs of disabled users from the outset, since retrofitting tends to be expensive.  Also good design for people with disabilities is frequently good design for everyone.

 

 

 



John Gill Technology Limited Footer
John Gill Technology Limited Footer