Information Technology Reference
In-Depth Information
Nonetheless, resorting to physical measures such as electroencephalograms is
not foreseeable in the context of a commercial application, for obvious practical
reasons, as well as for reasons of acceptability from drivers. The development of
“non intrusive” technologies, particularly based on the analysis of ocular activities,
is therefore favored. Initial technological realizations are already operational in this
field, such as SAFETRAC
14
,
Driver
State Sensor
(DSS)
15
or
PERcent of the time
eyelids are CLOSed
(PERCLOS)
16
devices, even if their performances in real
driving conditions are still far from being optimal. For a concise report regarding
these devices, see [GON 04]. Regarding this, the
Driver Monitoring System
(DMS)
device, based on the video analysis of eyelid
movements (the time it takes them to
close, for example) coupled with an analysis of driving performance in terms of
lateral control of the vehicle [BOV 08], has promising results.
Experiments carried out on the open road [BOV 05] have enabled us to show a
good correlation between:
- physiological measures characteristic of states of drowsiness that are estimated
based on electroencephalograms (Figure 5.5) and electro oculograms;
- the diagnosis of hypovigilance in real time proposed by the DMS device (see
Figure 5.5);
- a diagnosis of a decrease in vigilance estimated based on the average duration
of eyelid blinking (
means blink duration
in Figure 5.5); and
- a diagnosis of decrease in vigilance based on the calculation of the PERCLOS
system [GRA 01].
This supervision device of the state of the driver is not limited to the single
hypovigilance diagnosis; it can also provide relevant elements of a solution with
regards to problems of inattention and/or distraction at the wheel. It is precisely in
view of this that the DMS system has been used in the context of the European
AIDE project, dedicated to the adaptive management of driving assistance
[TAT 05]. In this particular context, the objective was to use the
state of vigilance
of
the driver and couple it with a diagnosis of “availability”
17
of the driver [BEL 07b],
to adapt it to human-machine interfaces (from the specification of intervention
modalities to management of the intensity of warnings).
14 http://www.assistware.com/.
15 http://www.seeingmachines.com/product/dss/.
16 http://www.attentiontechnology.com/pro-per.html [GRA 01].
17 The concept of
availability
,
such as it defined in the context of this project, corresponds to
the ability of the driver to perceive and deal with information delivered by an assistance
device, given the requirements (effective or potential) of the current driving situation
[BEL 07b].
