摘要

In this paper, we develop a theoretical framework for abstraction-based failure prognosis of partially observed discrete event systems. The purpose of using abstraction is to verify prognosability and synthesize a prognoser based on the abstracted models with fewer states and transitions. We present conditions under which prognosability can be verified and a prognoser can be synthesized based on the abstracted models of the system and the specification.

  • 出版日期2016-7