Back
Michael Gafert
UNIKATE

Work from 7 years ago (10/3/2019) at Michael Gafert

Assistive computer

Every year UNIQA is holding an idea competition named UNIKATE in cooperation with the Behindertenrat, and the technical university of vienna (TU Wien). The goal of the competition is to come up and build an assistive solution for people with disabilities.

Every idea should be linked to a specific person. In our case Natascha Toman required a new assistive computer which is able to help her communicate better. Our goal was to build a new version of the assistive computer she currently had. This included creating sentences with as little keystrokes as possible and presenting them to other people as text and voice. Together with Fabian Steiner and Natascha Toman we began to draft first concepts.

Behindertenrat UNIKATE 2019

Concept

The concept includes two screens. One primary facing the user and one secondary facing the person with whom the user communicates. This secondary screen should show the completed sentence the user has entered. A space below the primary screen is allocated for a TOBI eye tracking bar and a special keyboard.

Hardware Development

The main chassis of the device is 3D printed, excluding the side panels which are laser cut pieces of acrylic. The computer is a Latte Panda V1.0 4GB running Windows 10. It is connected to one primary 7” IPS touch display facing the user and a secondary 7” LCD facing the away from the user. The secondary display is mirroring the primary display and is used as an assistive aid to help the user be more easily understood by showing the creating of sentences.

Other features include stereo speakers, a Tobi eye tracker for gaze control, a battery with indicator and a special keyboard. The keyboard and computer are connected by a steel plate at the bottom which can be mounted to a wheelchair.

Software

To allow better sentence creation a custom user interface was developed. The program can be controlled by keyboard shortcuts and learns and predicts sentences based on the user’s input. If the prediction is incorrect new sentences can be added. When writing new words an internal dictionary tries to predict the user’s input. Once a sentence is completed it can reproduced via voice output.

Software Source Code

Final assistive computer

The final computer without the mounting steel plate for the wheelchair.

Presentation at UNIQA

The results were presented at the UNIQA tower in vienna.

Last two images © Lukas Ilgner / UNIQA