A Conceptual Model for Analysis of Automation Usability Problems in Control Room Settings

Licentiate thesis, 2010

The focus of this licentiate thesis lies on the problems that can emerge when humans and advanced control systems work together in control room environments. In the nuclear domain, automation is used extensively and it is important to maintain safe and efficient operations. Therefore, problems related to operator-control system interaction is a critical field. In the thesis, these interaction problems are argued to be closely related to usability issues. Therefore a distinction was made between problems related to the applied levels of automation and problems related to usability. The latter have been defined as automation usability problems. The first aim of the thesis was to propose a conceptual model that can capture aspects of human-machine interaction that are relevant for analysing automation usability problems. The second aim was to use the conceptual model for analysis of automation usability related problems and explain the emergence of different types of problems in a single framework. By compiling the means-ends hierarchy, the perception-action cycle, control loop, performance influencing factors, levels of automation and mental models into a single framework, a conceptual model that is suitable for explaining automation usability problems has been proposed. The conceptual model was tested in analysis of the out-of-the-loop, loss of skills and trust problems, by using an analytical approach. The conceptual model was also used for analysing the out-of-the-loop and loss of skills problem by utilizing empirical results from a field study performed in the nuclear power domain with a turbine automation interface as an example. The proposed conceptual model is intended to be used by human factors engineers working in industry. The industrial application is important to contribute to the design of safe and efficient human-machine systems. The conceptual model however needs further development before it is fully complete for industrial use. In further work, two tentative procedures for model use are suggested; a procedure for analysis of existing human-machine systems, and a for aiding the design of new human-machine systems.