Interactive Safety

Acronym: 
ActSafe
Term: 
2009-05 till 2012-10
Research Areas: 
C
Abstract: 

ActSafe is working to ensure the safety of autonomous interactive systems, among them mobile robots, highly-automated aircraft, and other systems. The work involves developing and applying design principles for human-machine interfaces, analyzing accidents and incidents with complex socio-technical systems. Further study concerns the creation and analysis of large accident databases, the publication of a textbook on computer safety and the organization of an engineering workshop "Interacting with Extreme Events.

Methods and Research Questions: 

The increasing use of computers and autonomous systems in all areas of life, as well as the continuing growth of automation in transportation means that the operation of such systems affects ever more people. Ensuring the safety of such systems becomes an important task with aspects such as the design of human interfaces, development of dependable software, and the analysis of incidents to prevent a recurrence. Ensuring the safety of automated interactive systems requires multiple strategies to cover all aspects of the development and operation of such systems. These systems can be as diverse as autonomous mobile robots, modern computerized airliners, or nuclear power stations. ActSafe explores design principles for the human interfaces for and human interaction with such systems, which we view as socio-technical systems, i. e. systems that comprise the technical hardware and software, as well as human participants, either dedicated operators (in the case of airplanes or power stations), or counterparts in other interactions, for instance when interacting with an assistive robot. A further aspect is the dependability of the hardware and software. Here, ActSafe is involved in developing and improving international standards (IEC 61508) for the functional safety of electric/electronic/programmable electronic (E/E/PE) systems; (a) by developing new, world leading qualitative methods for hazard analysis (OHA), which are unique amongst hazard analysis methods in that they assess and ensure relative completeness, (b) by gaining insight into interactive systems through causal analysis of incidents with such systems. The method Why-Because-Analysis identifies causal factors for an accident formally or semi-formally way, to find similarities to other occurrences and effectively look for possible countermeasures to prevent a recurrence. It is extended for analysis of incident databases. ActSafe also promotes international co-operation with the Universitas Gunadarma in Jakarta, Indonesia with regular exchange of students. ActSafe has initiated the international workshop “Interacting with Extreme Events” that will work on aspects of technology with high-severity/low-probability events such as nuclear accidents, focusing on the Fukushima Dai-ichi nuclear power plant accident.

Outcomes: 

ActSafe has developed guiding principles for the design of interfaces of interactive automated systems. These involve Rational Cognitive Models of all involved computer and human agents and stipulates that all must possess coherent knowledge about the world and be cognizant of all parameters relevant for the current task.
ActSafe is preparing publication of one of the few textbooks on computer safety, covering engineering matters such as system analysis, accident analysis, system definition, hazard analysis, risk analysis, functional safety, secure interfaces and limits of testing. We are actively involved in the German standardization committee DKE/GK914 for the functional safety of electric/electronic/programmable electronic systems. We have also performed analyses of aspects of aviations accidents involving ultralight aircraft, single-aisle and wide-body airliners.