Today, there are approximately 75 million wheelchair users worldwide. From
this collective, there is a percentage of people, the severely disabled, whose motor
capabilities below the neck are damaged, and despite the fact that they are
also wheelchair users, they can’t really move it their own. Even though research
has been attempting to bring a reliable hands-free control for them for nearly two
decades, still no commercial models exist, as no system has proven to be reliable
enough for a ...
Today, there are approximately 75 million wheelchair users worldwide. From
this collective, there is a percentage of people, the severely disabled, whose motor
capabilities below the neck are damaged, and despite the fact that they are
also wheelchair users, they can’t really move it their own. Even though research
has been attempting to bring a reliable hands-free control for them for nearly two
decades, still no commercial models exist, as no system has proven to be reliable
enough for a real environment situation. In this thesis, I study the most successful
remote control modalities so far involving head and eye tracking, and develop
from scratch a collection of 3 enhanced navigation systems inspired from the best
of each. These three systems allow to successfully control a powered wheelchair
using exclusively the head via IMU sensors, the eyes via a head-mounted eyetracker,
and a combination of both. Finally, I test the first two systems on myself
and also on different healthy people to contrast their performance with respect to
the standard joystick navigation. From the results collected, I propose future improvements
that could step up level of these hands-free controllers to the market
level.
+