Hierarchical rule-based and self-organizing fuzzy logic control for depth of anaesthesia

Abstract
A multi-stage hierarchical fuzzy control system with a multi-echelon structure for depth of anaesthesia (DOA) is described in this paper. There are four echelons that monitor, control, interpret and assess the whole surgical operation. Echelon 1 is a measurement and control action level that involves: instrument sensing (Dinamap); anaesthetist observations [measurement of sweating (SW), lacrimation (LA) and pupil response (PR)]; and a syringe pump (Graseby pump). Echelon 2 is an interpretation level that involves: interpreting systolic arterial pressure and heart rate to provide the primary DOA; interpreting SW, LA and PR to provide the degree of lightness; and interpreting bolus drug effects to estimate the sensitivity of patients. Echelon 3 is a regulation level that involves: controlling the drug from either a hand-crafted anaesthetists' rule-base or a self-organizing fuzzy logic controller algorithm; planning the drug profile to avoid long recovery; and managing alarm situations. Finally, echelon 4 is an assessment level that assesses the whole surgical procedure according to the patient recovery time. Testing this system in clinical trials as an intelligent adviser has provided a preliminary proof-of-concept of the applicability of this hierarchical structure for DOA management