Resumo (PT):
Abstract (EN):
The aim of this paper is to illustrate how statistical techniques based on Walsh-Fourier spectral analysis (WFA) contributes to the design of an on-line adaptive control system for neuromuscular blockade that could be used successfully in a wide range of populations.
It is found that neuromuscular blockade levels at the average periods indicated by WFA have a high predictive power for the parameters of the controller, providing additional information about the model.
A classification algorithm, based on the observed values of neuromuscular blockade level at significant WFA periods, enables the on-line individual tuning of a time varying adaptive reference profile, hence decreasing the overshoot.
The system provides strong robustness to inter and intra-individual variability of the patient's responses and adaptation to the individual requirements.
Language:
French
Type (Professor's evaluation):
Scientific
Contact:
ims@fe.up.pt
No. of pages:
7