Approche méthodologique pour une architecture de monitoring intelligent

Abstract

This paper aims to describe the recent steps of our research work focusing on the design of an intelligent real-time monitoring system intended to be used in emergency contexts (CICUs, operating rooms, etc.). Based on an analytical study of the needs expressed by the medical staff, our approach rests on two goals: 1) the interpretation of intelligence concepts (taking into account the context, the patient's state) in order to enhance the quality of the diagnosis and to reduce the numerous false alarms; and 2) the application of methodologies specific to real-time systems design, borrowed from software engineering, in order to gain robustness, progressiveness, and above all, generalness, guaranteeing its use in different medical contexts. For this purpose, the fundamentals of a new architecture providing intelligence distribution, modelled on SA/RT methodology and Petri nets, is presented. The joint application of these real-time methodologies allows us to monitor the temporal evolution of the system. The feasibility of this approach is demonstrated by the new evolution of our monitoring system prototype, SIMBAD, where the first biomedical signal processing applications are implemented.