Adaptive Goal Selection for improving Situation Awareness: the Fleet Management case study

Lack of Situation Awareness (SA) when dealing with complex dynamic environments is recognized as one of the main causes of human errors, leading to serious and critical incidents. One of the main issues is the attentional tunneling manifested, for instance, by human operators (in Decision Support Systems) focusing their attention on a single goal and loosing the awareness of the global picture of the monitored environments. A further issue is represented by stimuli, coming from such environments, which may divert the attention of the operators from the most important aspects and cause erroneous decisions. Thus, the need to define systems helping human operators to improve SA with respect to the two aforementioned drawbacks emerges. These systems should help operators in focusing their attention on active goals and, when really needed, switching it on new goals, in a sort of continuous adaptation. In this work an adaptive goal selection approach exploiting both goal-driven and data-driven information processing is proposed. The approach has been defined and injected in an existing multi-agent framework for Situation Awareness and applied in a Fleet Management System. The approach has been evaluated by means of the SAGAT methodology