Case-based situation awareness

Abstract

Situation-aware case-based decision support (SACBDS) systems comprise two distinct parts: situation awareness (SA) and case-based reasoning (CBR). The SA part keeps a finite history of the time space information of the domain and uses rules to interpret cues from the environment with respect to an individual user's context, and then anticipates future situations by performing statistical inference over historical data. The CBR part is the part that seeks to accomplish a particular task with knowledge of the environment from the SA component. This paper discusses the fusion of the CBR model and the SA model into a case-based situation awareness (CBSA) model for situation awareness based on experience rather than rule, similarity assessment and problem solving prediction. The CBSA system perceives the users' context and the environment and uses them to understand the current situation by retrieving similar past situations. Every past situation has a history. The future of a new situation (case) is predicted through knowledge of the history of a similar past situation. The paper evaluates the concept in the flow assurance control domain to predict the formation of hydrate in sub-sea oil and gas pipelines. The results provided the CBSA system with greater number of accurate predictions than the SACBDS system.