Detecting Inconsistencies in Rule-Based Reasoning for Ambient Intelligence
Hamdi Aloulou, Romain Endelin, Mounir Mokhtari, Bessam Abdulrazak, Firas Kaddachi, Joaquim Bellmunt · 2016
Rule-based reasoning engines have proven their value for Context-Awareness in Ambient Intelligence. However, the definition of rules is often prone to human-caused inconsistencies or sensor failures. In this paper, we propose to assist the definition of rules for context-awareness by automatically validating the consistency of the rules. Using our approach, we are able to detect conflicts between rules for context-awareness, as well as recommend adding sensors to improve the reasoner outcomes. We propose to define rules as a finite set of conditions to be verified, and provide consistency-checking capabilities to these conditions. We discuss in this paper our approach as well as results and validation.