Cooperating for assisting intelligently operators

We are in the process of design of an intelligent cooperative system in a nuclear plant application. The system must cooperate with an operator who accomplishes a task of supervision of a real-world process. We point out in the paper that a cooperation between a cooperative system and an operator has two modes: a waking state and a participating state. During the waking state, the system observes the operator's behavior and the consequences on the process. During the participating state, the cooperative system builds jointly with the user a solution to a problem. In our approach, the cooperation depends of the system capabilities to explain, to incrementally acquire knowledge and to make explicit the context of the cooperation. We develop these ideas in the framework of the design of the cooperative system in the nuclear plant application.