Design And Research Of Closed-loop Control System For Intravenous Anesthesia

The stable implementation of anesthesia is the basis of successful clinical operation.With the deepening of the research on anesthesia control,how to achieve accurate anesthesia control has become a challenging problem in neuroengineering and clinical practice.In the current clinical operation,the control of anesthetics still relies mainly on the experience of anesthesiologists.There is a lack of scientific and effective control methods between the monitoring index of anesthesia and the control of anesthetics.In addition,the long working time of anesthesiologists leads to inevitably deviations in the control of anesthesia.Therefore,how to achieve accurate control of anaesthetic has become an urgent scientific problem and practical demand.The objective of this study is to establish a closed-loop control system for intravenous anesthesia,so as to realize the automatic control of anesthesia during operation.On the one hand,it can reduce the working pressure of anesthesiologists and make them pay attention to more important operation nodes.On the other hand,it can improve the accuracy of anesthesia control and realize the individualized and refined control of anesthetics.The main contents of this paper are as follows: the selection of monitoring indicators in intravenous anesthesia control system,the regression prediction of monitoring indicators in the process of anesthesia control,the design of controller based on monitoring indicators,and the realization of intravenous anesthesia control system.Firstly,by comparing and analyzing the performance parameters of Entropy Index,Bispectral Index and Narcotrend Index,such as algorithm complexity,index quantization degree and time delay,Bispectral Index is selected as the monitoring index of the system.Because the Bispectral Index can not be obtained directly in the course of clinical operation and its algorithm is not public,it is difficult to design the control system with Bispectral Index In this paper,the long-term and short-term memory-pre-feedback neural network model is used to realize the regression prediction of bispectral index through EEG data,which solves the problem that the index algorithm is not public,and improves the applicability and operability of the index in the system.Then,according to the characteristics of clinical anesthesia operation,the controller is designed.The control of anesthesia was divided into induction period,maintenance period and recovery period.After comparing and analyzing the performance indexes of model predictive controller and ant colony optimization PID controller,ant colony optimization PID controller is chosen as the controller of the system.Finally,this paper introduces the software design and implementation of the system.