부산대학교 도서관

A self-tuning algorithm is presented for the online control of muscle relaxation. Although some patient parameters were fixed at a mean value determined by offline analysis of patient data, certain parameters have been explicitly identified online. In addition the patient states are estimated and are used by a novel controller to bring an unrelaxed patient to some desired level of paralysis and maintain it. Simulations show that a twitch depression of 80% can be achieved in less than 20 min using d-tubocurarine without excessive overdosage.