Information Technology Reference
In-Depth Information
human activity, and not on the contrary replacement for it (i.e. where the activity of
the driver would be defined “by default”, leaving the human with residual tasks
outside the automaton's reach).
After a concise introduction of the context and the stakes of ergonomic research
in terms of human-machine cooperation (section 5.2), we present a few examples of
cooperative assistance devices on which the different authors in this chapter are
working (section 5.3).
The first of these devices (section 5.3.1) aims to prevent collision risks. The
specific challenge of this research project is to design an
adaptive
technology that is
able to jointly take into account external driving conditions (how critical a situation
is, as it can be measured by time to collision with an obstacle, for example) and the
driving behaviors implemented by the human driver, in order to determine in real
time whether the latter is properly managing the risk or needs assistance.
The second device (section 5.3.2) is concerned with the prevention of risks of
exiting lanes and aims to guarantee a
secure driving trajectory
from the point of
view of lateral control. This research enabled different philosophies of human-
machine cooperation to be investigated from a
perceptive
mode centered on the
diffusion of information to the implementation of a
mode of delegation of the
function
aiming to entrust lateral control to the automaton via a
mutual control
mode
. These different modes of cooperation are presented and then discussed with
regards to recent experimental results.
The last device (section 5.3.3) adheres to the logic of driver
monitoring
and
deals more specifically with accident risks induced by the driver falling asleep at the
wheel. The objective is to design a warning system able to diagnose a critical
decrease in vigilance and to compensate, as the case may be, for this type of human
failure. It is also a matter of enabling adaptation of interaction modalities and/or
human-machine interfaces, according to the driver's vigilance level.
Finally, beyond these three specific aid functions − by nature limited in their
respective objectives − we ask ourselves about the integrative aspect of what a true
automobile “copilot” should be, while at the same time trying to guarantee
centralized and harmonious management of human cooperation, whatever the
subjacent aid functions on which it is based. This point is more specifically the
focus of a discussion in section 5.4, and will enable us to propose a first hint of
generic architecture for such a copiloting device adapted to the specificities of
automobile driving.
