Environmental-perception modeling and reference architecture for cyber physical systems

Abstract

In cyber-physical systems, physical and software components are deeply intertwined, blurring the boundaries between the cyber and physical worlds. Perceiving environmental information is a prerequisite for a cyber-physical system to be reliable and adaptive to the environment. However, the intrinsically open and dynamic nature of the environment brings challenges to the systematic realization of environmentalperception. This article regards environmental-perception as a first-class entity with self-maintenance and abstracts it into three levels: data capture, context awareness, and situation identification. Then, a five-tier reference architecture is proposed, which not only explicitly defines the perception ability at different abstraction levels, but also provides for the storage and management of environmental information, thus facilitating information reuse and the adaptation ability of cyber-physical systems. We evaluate the reference architecture with a case study on a widely used smart room environment, demonstrating its broad applicability and the ability to reuse environmental information. We also performed a survey on the proposed reference architecture to reveal its industrial motivations and benefits.