Modeling of Complex Human-Process Interaction as Framework for Assistance and Supervisory Control of Technical Processes

Modeling of Complex Human-Process Interaction as Framework for Assistance and Supervisory Control of Technical Processes

Dirk Söffker (University of Duisburg-Essen, Germany), Xingguang Fu (University of Duisburg-Essen, Germany), Andreas Hasselberg (German Aerospace Center (DLR), Germany) and Marcel Langer (University of Duisburg-Essen, Germany)
DOI: 10.4018/jitwe.2012010104
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Interactive processes are omnipresent in fields dealing with automation of complex technical processes, allowing a wide range of application scenarios of assistance systems to support human operators. However, the definition and description of the interaction strongly depends on the application. The description of Human-Process-Interaction (HPI) in most cases is the key for the development of interaction models leading to a framework for the development of assistance systems. The goal of such assistance systems is to support human operators, or human factors, and thereby improve the considered process in parallel. For the purpose of creating a framework for the development of assistance systems, the Situation-Operator-Model (SOM) approach is applied. The SOM approach is capable of modeling HPI in a net-like structure representing the complex environment (or process) and human interactions. By this formal modeling of HPI for the development of human-centered assistance systems based on the underlying modeled interaction is available. In this contribution illustrative examples of different applications for the proposed approach are given by a cognitive, individualized driver assistance system for lane changing maneuvers, by the assessment of operator decisions in air traffic control as well as the operator supervision in semi-automated production process.
Article Preview

Introduction

The rising level of integration in modern industries and facilities leads to increasing complexity and vulnerability. For such technical processes involving human operators, the reliability of the interaction and cooperation becomes a critical issue. Although automatic control has been widely applied in complex human-machine systems such as power plant and airplanes, the role of human operators is irreplaceable (Sheridan, 1992; Endsley & Kiris, 1995), especially in the time critical emergency cases. The human operator in a nuclear power plant needs to decide whether to turn off the reactor in a few seconds when one of five sensors sends a warning signal; and the pilot has to decide where to make an emergency landing because of a technical malfunction of the air plane. These emergency situations are always challenging to both designers and operators of the Human-Process-Interaction (HPI).

Human-Process-Interaction (HPI) involves the understanding of how a process and the interface works. Although for very complex processes or very rare situations, human operators are prone to errors. Typical human errors have been classified in research realized in the late 80’s by Dörner (1989).

Due to the mentioned reasons, research in the field of HPI concentrates on the improvement of the process reliability, safety, and efficiency among others. Thereby, the proposed approach focusses on the technical side of the Human-Machine-System by enabling the technical systems to supervise and support humans in central roles with a high level of process responsibility. Sample approaches in the fields of transportation, air traffic control, and industrial automation are given.

The realization of assistance depends strictly on the underlying process and interaction model, and thus in the domain of the interaction designer. In an ideal case, available process knowledge is included within the model, in order to be available to the human operator in critical situations.

Obviously, it is difficult to achieve a broad understanding of a process inside a technical device/program, but on the other hand a technical device has several advantages in comparison to humans with respect to understanding and cognition (Wickens, 1984) like assignment of predefined events, information, and relations, fast and secure execution of algorithms reacting to sensor inputs, and pattern recognition of known patterns. As long as a situation is known to the device (or to the designer of the device), the technical system has the potential to act more reliable than human operators. On the other hand, unknown and unpredicted situations are much more suited to the cognitive capabilities of human operators. This is also true for modeling errors that are not modeled sufficiently or detailed enough during the design process. As a result, the tasks of process interaction are carefully to divide among human operators and technical devices.

The main question answered in this contribution is: How can HPI and processes in general be modeled to enable supervision of human operators by technical devices? Certainly, this has to be solved for each process or type of process individually. Considering the action logic of the process and the interaction logic of the interaction, a common, process invariant framework is defined for modeling of action and interaction. Within this work, the previously developed generic approach Situation-Operator-Modeling (SOM) is used as a framework modeling HPI (Söffker, 2008). This approach and its technical realization are repeated and briefly described in the next subsection. In the following, the underlying basic methods and assumptions of assistance and supervision of human operators using SOM are described, followed by three technical application examples.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 13: 4 Issues (2018): 1 Released, 3 Forthcoming
Volume 12: 4 Issues (2017)
Volume 11: 4 Issues (2016)
Volume 10: 4 Issues (2015)
Volume 9: 4 Issues (2014)
Volume 8: 4 Issues (2013)
Volume 7: 4 Issues (2012)
Volume 6: 4 Issues (2011)
Volume 5: 4 Issues (2010)
Volume 4: 4 Issues (2009)
Volume 3: 4 Issues (2008)
Volume 2: 4 Issues (2007)
Volume 1: 4 Issues (2006)
View Complete Journal Contents Listing