Information Technology Reference
In-Depth Information
evaluations, in real conditions of delivery activity and in real driving situations,
which does not enter into the scope of this project.
7.7. Conclusions and viewpoints
From the point of view of design, the approach implemented was UCD. Its focal
development point was the needs of users and this enabled the design of a system
that was well accepted by the latter. The iterative UCD cycle meets the needs of
regularly re-examining the specifications. However, though the traditional approach
involved sometimes very costly developments, it turned out to be necessary to find
solutions enabling developments carried out in parallel by different teams working
on the project. We therefore proposed a methodology integrating the new validation
cycles to reduce the costs of tests. These cycles allowed the results of the analysis of
driver activities carried out in parallel, and the different stages of the process such as
they are described by [MAG 01] to be taken into account.
The phases of specification of the user needs led to the definition of typical
situations that required specific assistance: two reversing situations (in movement
and when starting to move), two situations of moving forward (in movement and
when starting to move), a situation of turning right after starting at a light, and an
situation of overtaking a vulnerable user on the road going in the same direction.
The assistance system retained consisted of an immobilizer, emergency brake and a
speed regulator, completed by visual informative feedback by rear and side cameras
and by flashing lights and sound warnings in the cab. The ergonomic evaluation
phase on a simulator was necessary to lead to the proposal of this adaptive system.
Designed to diminish the number of accidents between vulnerable users and
industrial vehicles in an urban environment, it is based on dynamic assistance
strategies according to driver behaviors and the behaviors of vulnerable user(s) in a
given situation. Evaluated in a virtual world by a panel of truck drivers, the VIVRE2
system has proven its efficiency on a driving simulator: thanks to the system, 89%
of accidents were avoided over the course of the experiments. Moreover, the results
of the ergonomic evaluation attest that the system should be well accepted by drivers
and that its use should not lead to a cognitive and informational overload.
In conclusion, the generic methodological framework presented in this chapter
proved its efficiency and met both objectives of the project and will be used in
future projects to design an aid system on a simulator that based on the needs of the
driver that uses perception technology.
This type of approach can be extended to the design of demonstrators in real
driving situations. Thus, a follow up of this work proposes deploying the results
broadening the field of application to urban buses and to all the users involved
