Assessment of the Human-Machine Association on a Smart Wheelchair

In this communication, we present a method to assess a mobility assistance developed on a smart wheelchair. This assistance provides to the users a functionality of wall following, as well as a functionality of automatic passing through narrow passages, usable via a deictic interface. The aim here is to compare a semi autonomous driving and the typical manual driving with a joystick. We would like that this assessment focuses on human. To do this, we have implemented a test protocol using a dual task method which allows to measure the attentional load of subjects for each type of driving. We also carry out a finer analysis of attentional load of subjects to identify the types of actions in which it is the highest during the driving with and without assistance.