Joint visual attention modeling for naturally interacting robotic agents