We describe a teamwork framework to design human-automation interactive systems. Human automation interactions have been primarily considered as supervisory systems in which communication is seen as unidirectional. As systems become more complex, sophisticated, and autonomous, the opportunity for bidirectional cooperation between the human and machine becomes feasible and advantageous. To better understand bidirectional cooperation and develop computational tools we draw on concepts from philosophy, social psychology, human factors, and artificial intelligence to define "team" and understand the requirements for efficient team interaction. We define several team properties such as mutual support, mutual commitment, machine transparency (state, behavior, control) and user transparency (state, intent, action) and responsiveness as a way to analyze and guide the design of human-machine interactions. We illustrated the applicability of the approach using a generic model of a full speed range adaptive cruise control and show the potential of this type of cooperative approach to human- machine interaction.
展开▼