Predictive Control Strategy with Online Time Delay Estimation Applied in General Anaesthesia

General anaesthesia refers to the state of total unconsciousness resulted from the administration of several drugs. The depth of anaesthesia can be monitored with a device called BIS (Bispectral index) monitor. A manifold of artifacts corrupt the signals measured with this monitor, e.g. coughing, movement of feet and arms, face washing, etc., which challenge the correct estimation of the bispectral index. As such, estimation techniques evaluate the signal to noise ratio and if insufficient information is available, data from the past measurements is used to evaluate the current state of the patient. In this manner, instrumental time delay is introduced in the closed loop regulation of general anaesthesia. This paper evaluates whether or not an incorrect estimation of the time delay has an influence on the stability and robustness of the closed loop control. The performance of the EPSAC (Extended Prediction Self-Adaptive Control) controller was tested using different scenarios. Under- and over-estimations of the real time delay were considered in the prediction. The online time delay estimation was added to the control algorithm and the performance was evaluated.