Action prediction with the Jordan model of human intention: A contribution to cooperative control

Human intentions are internal processes that can be deduced by observation of their resulting actions. Hence, an observation-based model of human intention is needed. The Jordan model of human intention in traffic is presented and applied on empirical data for two different intentions: driver’s braking intention and pedestrian’s crossing intention. The analysis shows that the behavior postulated within the theoretically developed Jordan model is observable in both sets of empirical data of drivers and of pedestrians while they perform intended actions. It can be assumed, that the Jordan model is universally applicable on very distinct intentions, not only limited to traffic scenarios. The Jordan model enhances the development of cooperative control among humans and machines. On the way towards a design of cooperative behavior among humans and vehicles in traffic, the Jordan model contributes with a systematic sequence and description of human behavior that is reflecting intention. Consequently an integration of the Jordan model of human intention with the model of cooperative and shared control Flemisch et al. (Ergonomics 57(3):343–360, 2014) is proposed. This integration facilitates the cooperation between human and system on strategic, tactical, and operational level