Information Technology Reference
In-Depth Information
At first glance, the task can seem disproportional. Nonetheless much scientific
knowledge and many technological components likely to have such a use already
exist, whether in terms of embedded sensors and perception technologies,
algorithms for processing or fusion of data, human-machine cooperation, control-
command functions or interfaces with the driver.
In addition, it is not necessarily a matter of designing an “omniscient” copilot,
which has a “universal” intelligence. Much more pragmatic and no doubt more
realistic from an industrial approach, would be to proceed according to the logic of
distributed intelligence
,
i.e. maintaining a “local” intelligence that is specific to the
level of each of the assistance functions (such functions include mapping knowledge
and itinerary management skills for the navigational aid devices; estimation of the
criticality of the situation and regulation of the speed via collision avoidance devices
and ACC; lateral control abilities via functions to maintain a secure trajectory; or
diagnosis of falling asleep and/or inattention via a driver supervision device, etc.). It
is necessary to restrain the skills of the copilot to the fundamentals of its integrative
function: the
knowledge of how to cooperate
18
[MIL 98] and the
adaptive nature
of
assistance [BEL 03] required in order to guarantee the proper functioning of a
human-machine as a whole.
Thus defined, the automobile copilot is no longer just a pipe dream. It is a
conceptual framework for
human-machine coupling
[AMA 92], assorted with
ergonomic specifications enabling us to define an
integrated and coordinated
design
strategy of future aid devices. This is achieved by favoring the optimization of
assistance functions by the mutualization of technical means and the coupling of
human and machine competencies in view of improving the reliability of the human-
machine system tasked with piloting the vehicles of tomorrow.
5.6. Acknowledgements
This work has been carried out with the support of the GDR CNRS 3169 Psycho
Ergo and the ANR-PREDIT projects PARTAGE, PREVENSOR and ARCOS.
Some of the results presented in this chapter are from projects that have also
received funding from the European Commission's Framework Programs (FP6 and
FP7), more particularly under the grant agreements of CEMVOCAS, AWAKE,
SENSATION, SAVE and AIDE projects.
18 The knowledge of how to cooperate between two agents refers - as a minimum - to: (i)
the ability of these agents to manage interferences between their respective goals; and (ii) to
the ability of each agent to facilitate the other agent reaching its goals.
