DESIGNING AN EXPERT SYSTEM FOR
RISK-MANAGEMENT TO SUPPORT OPERATORS’
Sebastian Spundflasch, Johannes Herlemann & Heidi Krömker
Department of Media Production, Ilmenau University of Technology,
The growing trend of centralization in German road-tunnel control results in an
increasing number of tunnels and a higher amount of information that must be
processed by individual operators. In order to support their main tasks – monitoring,
diagnosis and control (VDI/VDE 3699, 2005) and in particular risk management –
an expert system will be developed within the research project ESIMAS. To find
out the requirements that the user interface of the expert systems must comply, we
studied workflow and mental model of tunnel control operators. The results serve
as basis for designing the interaction between operator and expert system.
In a first analysis we wanted to find out, how risk management is being performed
by different tunnel control centers within the 3 phases of process control [Figure 1].
10 tunnel control centers, different in terms such as number of tunnels controlled,
traffic volume, organizational structure and composition of its staff, were analyzed.
Following the Contextual Design (Beyer & Holtzblatt, 1998) process, operators
were observed at work and in-depth interviews were conducted with tunnel man-
agers and operators. The focus was on the interdependences between operators,
their tasks, their control system and their working environment.
The analysis showed, that the main tasks control, monitoring and diagnosis can be
observed in all analyzed control centers, but the extent to which the tasks are being
Figure 1. Phases of Process Control according to (VDI/VDE 3699, 2005)
expended by activities of Risk Management.
448 S. Spundﬂasch, J. Herlemann & H. Krömker
performed within subtask vary. With regard to risk management, a major difference
between the centralized monitoring of several tunnels (multi tunnel control – MTC)
and monitoring of a single tunnel (single tunnel control – STC) was observed. In
STC, the operators mental model (Yin & Laberge, 2010) and thus the workflow
concerning risk management, is much more precise than in MTC. A Hierarchical
Task Analysis (Shepherd, 2001) of each control center yielded the result. In STC
the process of monitoring includes monitoring for risks, carried out by proactive
control behavior, e.g. the continuous monitoring of video images from inside the
tunnel. The monitoring for risks is the prerequisite for activities in risk diagnosis
and risk control. With an increasing number of tunnels to be monitored in MTC,
a more and more reactive control behavior was observed. The operators just react
to manually or automatically detected events. A monitoring for risks does not take
place. Not seeing the risk, the operator is not able to do preventive tasks. This can
cause events which could have been prevented.
The results showed that there is a need for support in monitoring for risks especially
inMTC.The nextstepwillbetheuser interfacedesignoftheexpertsystem. It should
support the operators by providing information they need to perform effective risk
management. The required Information will be extracted from the analysis of STC
operators’ mental model and their performed subtasks in risk management. The
activity of monitoring will be extended by an automatic detection of risks. The
activities of diagnosis and control are to be improved by selective information for
an efficient assessment of risks and decision making. The ultimate goal of the
expert system is to represent selective data that enables the operators to perform
the differentiated subtasks in MTC as well as in STC.
VDI/VDE 3699 Part 2. 2005, Process control using display screens – Prinicples,
(Beuth Verlag, Berlin).
Beyer, H.and Holtzblatt, K. 1998, Contextual Design. Defining Customer-Centered
Systems, (Morgan Kaufmann Publishers, San Francisco).
Yin, S. and Laberge J. 2010, How Process Control Operators Derive, Update,
and Apply Mental Models. Human Factors and Ergonomics Society 54th Annual
Meeting. San Francisco, CA.
Shepherd, A. 2001, Hierarchical Task Analysis, (Taylor & Francis, London).