Fully autonomous cars are expected to appear on roads in the near future. Nowadays, the driving task is already supported by automation, which has begun to change the task, especially in terms of attentional processes: the more the car is automated, the more the monitoring sub-task is diversified. Thus, the increasing part of attention involved in monitoring should be considered in the design of fully automated cars. This paper proposes an embodied cognitive architecture, that senses driver attentional states in order to design intelligent interfaces supporting the driving performance. That new module entitled »Driver Cognitive Model of Attention», globally inserted in the autonomous driving car architecture implies to equip the vehicle with sensors that allows to infer driver cognitive states, specifically attention