Human-machine cooperative judgments in non-homogeneous safety monitoring systems

摘要

In safety monitoring systems, uncertain situations exist in which sensors cannot determine whether safety or danger exists. We have proposed a new type of non-homogeneous monitoring system in order to make this determination. In this article, we consider three strategies for dealing with an uncertainty zone. One is an automatic method of deciding deterministically whether the situation is safe or dangerous; the second is an automatic method of deciding stochastically; and the last is a human-machine cooperative method for judgements. We select the optimal strategy and show the conditions under which the third method is effective. Our results are also applicable to a problem of how to decide the optimal balance between human activities and machines in order to perform adaptive operation by humans and to prevent operator omission and commission errors in supervisory systems.