Uncertainty handling in cyber–physical systems: State‐of‐the‐art approaches, tools, causes, and future directions

Abstract

Cyber-Physical System (CPS) is the set of heterogeneous physical units linked to a network and performs complex operations to achieve a goal. Uncertainty increases with the increase in complexity of CPS. Thus, uncertainty needs to be mitigated to assure the quality and reliability of a CPS. This study aims to identify current stateof-the-art approaches, tools, root causes, and metrics for uncertainty in the domain of CPS. We performed a systematic literature review and employed keyword-based search on publisher sites to find potential studies. After applying the devised inclusion and exclusion criteria on identified potentially relevant studies, selection of studies is validated using an index engine. The core contributions of this study are (i) to categorize the tools used for uncertainty mitigation and existing root causes of uncertainty in CPS domain, (ii) to categorize the tools used for uncertainty mitigation and existing root causes of uncertainty in CPS domain, and (iii) to identify the stateof-the-art methods that lack the ability to elaborate the metrics to measure the uncertainty in CPS. The results of the proposed study are beneficial in guiding future research on devising new approaches or tools to mitigate the causes of uncertainty in CPS.

Towards the portability of knowledge in reinforcement learning-based systems for automatic drone navigation

“…On one hand, the physical elements that are in direct contact with the environment such as sensors or actuators. On the other hand, we have at the computational level we have software elements (such as intelligent agents) in charge of different tasks such as managing user queries, management of incidents arising from the presence of uncertainty, real-time control, knowledge management, and more importantly, uncertainty handling, with several examples of uncertainty drivers given in Asmat, Khan & Hussain (2022) .…”