Modelling motivations, tasks and human errors in a risk-based perspective

摘要

This paper discusses the modelling architectures of human-machine interaction (HMI) and the human error models implemented to support design and safety assessment of modern highly automated systems. In particular, the architecture of a joint cognitive model, which accounts for motivational and descriptive aspects and enables to predict dynamic HMI, is revised in detail, on the basis of a previously formulated paradigm of a causal structure of categories of human model (Carsten 2007). Then, an error generation model is associated with this human model architecture and can be utilised in risk-based design methodologies. These approaches share a common crucial issue, that is, the need of specific data and empirical correlations obtained from field observations and from reports of end-users. Ways and means to overcome this important bottleneck reside primarily in the improvement of the safety culture at society and organisation level.