Abstract
By introducing advanced automation, high accuracy, efficiency, and adaptability have been achieved in large scale and high-tech systems such as nuclear power plants and modern aircrafts.On the other hand, when we consider a situation that a human operator supervises such enhanced systems, the complexity of interaction between the human and the machine causes many unpredictable problems. For example, the user can not follow behavior of the machine through a user-interface, which shows partial information of the current mode of the machine.Thus, a current mode of the machine may differ from that anticipated by the user so that breakdown in the interaction between users and machines occurs.This phenomenon is called an automation surprise, which may yield to a serious human error.These problems mainly result from the incompleteness of information between the machine and the user.So it is an important problem to design a user-interface which plays an intermediate role in human-machine systems.In this paper, we analyze the human-machine interaction based on discrete event system theory and address the verification of automation surprises.Moreover, we develop a software which implements the proposed verification method and refer to the specification of this software.When we model the human-machine systems as discrete event systems, automation surprises can be classified into three types; a mode confusion, a blocking state, and a refusal state.We give a mathematical definition to each undesirable situation and propose a systematic procedure of the verification of automation surprises.In our proposed procedure, we construct a composite model in which a machine model and a user model evolve concurrently and implement an algorithm for searching of automation surprises. Finally, we show the usefulness of our proposed method by using our software.
