Communication is recognised as a human right by the United Nations. Currently there are millions of people who for a variety of reasons cannot communicate comfortably. For example, in the UK alone there are 250,000 people who as a result of a stroke are now unable to communicate and are affected by a condition known as Aphasia. These people are said to have Complex Communication Needs. With the proliferation of smart devices like tablets and Smartphone, people with Complex Communication Needs are discovering the assistive potential of these devices to aid them either in the recuperation of their communicative abilities or to assist them in their daily lives. These systems are called Alternative and Augmentative Communication Systems.However current AAC systems suffer from the fact that they are cumbersome to use and users require a long time to form sentences, with the result that they cannot confidently communicate and therefore are left isolated and frustrated. Even though much work has been done in the area these systems are remain slow and communication is not effective.This thesis investigates whether the inclusion of Natural Language Processing and Natural Language Generation techniques into Augmentative and Alternative Communication systems on mobile devices can improve the ease and speed of use for users with Complex Communication Needs by implementing “Dictum” an AAC app which makes use of NLG/NLP techniques.The work followed the approach of Action Research in which the target users help the investigator by identifying the problem, sanctioning the research and evaluating the results. Therefore, users were actively involved in the design of the application from the very start and gave feedback after each iteration leading to the final application.This work has found that the inclusion of NLP and NLG in AAC does indeed improve ease and speed of use when compared to popular apps available today.Dictum improves speed by doing two things: reducing the set space of words by providing words that are relevant to the last word inserted by using a Semantic Network of nouns and allowing the user to build sentences by requiring selection of key words only and delegating the responsibility of sentence formation to the application itself. In addition, during the course of this work, an effective mechanism of capturing requirements for users with Complex Communication Needs discovered by looking at how users adapt the functionality of their devices.The app was evaluated both quantitatively, by computing keystrokes savings and evaluating the interface using well-established HCI laws, and qualitatively by asking for the feedback of potential users and Speech and Language Therapists, following the practice of Action Research to involve those touched by the problem.
|Date of Award||1 Apr 2016|
- University Of Strathclyde
|Sponsors||EPSRC (Engineering and Physical Sciences Research Council)|
|Supervisor||David McMenemy (Supervisor) & Mark Dunlop (Supervisor)|