How to achieve accurately and automatically inject anesthetics is a great challenge both in clinic and neural engineering. Employing the computer to simulate the anesthesia process and control performance are valuable for the design of actual control system.The virtual patient library was established to test the ability of the control system,which responds to differences between patients.And Renyi permutation entropy as the depth of anesthesia monitoring index is proposed,according to the changes of Renyi permutation entropy,the control of propofol input speed is changed.For individualized model of patient,the identification is done, and the model predictive controller is designed and used to control the patient based on identification model.The performance of the control system is analyzed by four indexes: error, median performance error, median absolute performance error, wobble.The simulation results show that the system can well track the desired Renyi permutation entropy and maintain the desired depth of anesthesia.%如何实现麻醉药的精确自动给药是神经工程和临床上一个挑战性的难题.采用计算机模拟麻醉过程中的药物效应和控制性能对于实际系统的设计具有重要的价值.本文建立了一个虚拟病人数据库来测试控制系统应对病人间差异性的能力.并提出采用Renyi排序熵作为麻醉深度监控指标,根据Renyi排序熵的变化调控丙泊酚的输入速度.针对病人个人化模型进行辨识问题,在辨识模型的基础上采用模型预测控制设计控制器进行控制.控制系统性能采用误差、中值性能误差、绝对中值性能误差和摆动度4个指标对结果进行分析.仿真结果表明本系统可以很好地跟踪期望的Renyi排序熵,并维持所需麻醉深度.
展开▼
