Page 204 - Handbook of Electronic Assistive Technology

192 HANDBOOK OF ELECTRONIC ASSISTIVE TECHNOLOGY

to make use of such resources to save time and provide a starting point for the creation of
a vocabulary, it is important to remember that all vocabularies will require customisation to
make them specific to the needs of the individual. This is especially the case for vocabularies
based on categorisation, where careful thought should be given to the contents of each cat-
egory to ensure the user has quick access to words that are relevant and motivating to them.
When considering the organisation of a selection set, particularly if this is to be pre-
sented in a grid format, the consistency of layout should be a primary consideration for
clinicians and those designing systems. Particularly when working with dynamic display
devices that move between pages, having language concepts ordered in a consistent way
with consistent visual indicators, such as colouring the components of language accord-
ing to a consistent ‘key’, will facilitate the user’s learning to operate the system more effi-
ciently (Light and Drager, 2007). As with the positioning of an input method, consistency
in the placement and identification of elements in a selection set will reduce the cogni-
tive load and promote automaticity. Positioning the nonlanguage functions such as speak,
clear and delete consistently throughout the system will further enhance the automaticity
of a user’s interaction with the selection set.
A selection set may also include one or more methods of rate enhancement. Since users
of AAC systems typically communicate at an output rate significantly below spoken lan-
guage at between 2 and 15 words per minute (wpm) compared (Beukelman and Mirenda,
1998) to rates of around 120–150 wpm for conversational speech, several techniques to
increase the rate at which language is generated may be considered. These may include
word-prediction or word-completion, based on frequency models that may adapt over
time to the user’s way of speaking. Additionally, prestored phrases can significantly reduce
the number of selections required to generate full sentences. Such phrases are particularly
useful for social language such as greetings or frequently used questions or comments.
Phrases can be prestored by the user themselves or by their support team. Similarly, abbre-
viation expansion may allow a user to type a small group of letters that are then automati-
cally expanded by the software (e.g., inputting NP could be expanded into the phrase ‘no
problem’). Sentence or phrase banking, where the user can retrieve a full prestored utter-
ance by searching for keywords, can also enhance conversation rate.
In recent years, advances in technology have meant that additional methods of rate
enhancement are becoming available to AAC users. For example, the addition of location-
aware software to many AAC users means that language specific to a context or location
can be provided dynamically for faster retrieval by a user (Black et al., 2016). Similarly, new
technologies such as facial recognition may offer additional opportunities by providing
language tailored for a specific communication partner.

Output Methods

Any AAC system will also need to include a method of output for the user to transmit mes-
sages or commands. A device may use synthesised or digital (recorded) speech to transmit
a message to a conversation partner and may also produce text, either to support the voice
output or for recording the language generated.