COGNITIVE APPROACH IN SAFETY DRIVING ACTIVITIES
Jean-Marc MERCANTINI.
Laboratoire des Sciences de l’Information et des Systèmes (LSIS)
University of Aix-Marseille 3 (Paul Cézanne)
13397 Marseille cedex 20
E-mail: jean-marc.mercantini@lsis.org
KEY WORDS
Cognitive Engineering, Human Errors, Road Safety Analysis,
Generic Model of Accident, Causes Search Process.
ABSTRACT
The driving activities of transport vehicles (cars, aircrafts, trains,
etc.) as well as industrial plants, are constraining the drivers to
cognitive overload that can lead to human driving errors.
Depending on the context, the human driving errors may cause
accidents generating extensive consequences on human lives as
well as on economical activities. To deal with this important
problem a lot of research works have been investigated: to
improve knowledge in accident mechanisms, to formalise the
human error concept, to improve the conception of Human-
Machine interfaces ergonomically adapted to Safety Critical
Systems, to develop specific tools to assist experts and drivers
when solving safety problems, etc. The purpose of this paper is
to present the contributions of a cognitive approach to the study
of safety driving activities.
I. INTRODUCTION
According to Amalberti (Amalberti 1999) safety in
transport or industrial activities is approaching to an
asymptotic value and the human actors could be one of the
principle reason for this limit. The human errors have
almost always been considered the main cause of
accidents. This is the result of differences in work pace
and in representation languages which lead into
misunderstandings, responsible for the majority of
accidents reported (Ganascia 1999). On the other hand,
when attempting to manage their own abilities and error
rate, the operators risk to increase their fatigue and stress.
This problem reaches higher proportions in industrial or
transport applications supported by complex systems
considered safety critical from the viewpoint of the
consequences of errors and faults, whether in financial
terms or in terms of their catastrophic consequences. For
those systems, beyond precision and functionality, it is
imperative to offer their users: built in safety, adaptability
to different degrees of expertise and work situations and
support to easy the learning.
The paper is organized as follows. We first present
three examples of the application of a cognitive approach
to the study of safety driving activities. These examples
highlight the need for this approach in order to improve
the understanding of the mechanisms implicated in
accidents. Next, we present an application of a cognitive
engineering method to the Road Safety Analysis domain.
II. THE NEED FOR A COGNITIVE APPROACH
The study of safety in the organisations whether in the
context of transport activities or that of industrial
production, need to seize complex situations and make
evident the relationship with one or multiple actors, the
tasks associated to their missions, their artefacts as well as
their environment. The growing evolution of equipment
technology as well as the raise of professional competence
at the heart of the organisations has lead into an evolution
of the relations man-machine towards an exchange of
information and knowledge in the context where they
occur (Chouraqui et al. 1998). In a man-machine
relationship where the equipments are more and more
complex and reliable shows that safety depends even more
on the humans and their ability to master their machines,
particularly during emergency (malfunctioning,
breakdown, etc.). The understanding of the phenomenon
which happens during these situations demands models
which account for all the factors (actors, tasks, artefacts
and environment). Thus, the cognition plays a role of
growing importance in the studies of human work because
his intellectual, psychic and cognitive abilities are under
growing demands to solve the problems with which he is
confronted. In this context it seems clear that the solution
to safety problems, inherent to the human activity of
interacting with sophisticated machinery must account for
the human cognitive behaviour and hence contribute to its
problem solving borrowing from the cognitive sciences
their approaches and methodologies. We rejoin in this
sense W. D. Gray and E. M Altmann in (Gray and
Altmann 2001), to whom the process of cognitive
modelling constitute an appropriate methodology in order
to seize the complexity of these situations.
The malfunctioning of such systems may originate from
one of their components (the operator, the task, the.....