Nonlinear subspace-based extended prediction self-adaptive control for individualized anesthesia care

Throughout reported studies of closed-loop anesthesia control, model-based control strategies have been applied as a series of effective and promising algorithms. However, due to intra- and inter-subject variations, the accuracy of identified model and robustness of control algorithm seem especially crucial. At this point, a combination of subspace-based Wiener system identification method and Extended Prediction Self-Adaptive Control (EPSAC), which can be regarded as a data-driven model predictive control strategy, have been applied to solve the individualized anesthesia care problem and achieved an acceptable performance. The suggested entire algorithm, which can be operated with little prior knowledge, is effective to control nonlinear Wiener system with the advantages of precision and stability. In Simulation Section, 24 diverse virtual patients in the Wang's Anesthesia Simulator have been successfully employed to demonstrate the efficiency and robustness of the proposed method.