基于最小二乘的Mesic呼吸系统模型辨识研究  

A Least Square Based Parameter Identification of the Mesic Respiratory System Model.

作　　者：刘亚杰[1] 丰继华[2] 施心陵[3] 陈建华[3] 朱玲[4] 肖琦[3] 

机构地区：[1]昆明医学院第三附属医院(云南省肿瘤医院)设备科,云南昆明650118 [2]中山大学信息科学与技术学院,广东广州510275 [3]云南大学电子信息工程系,云南昆明650091 [4]云南农业大学电子信息工程系,云南昆明650221

出　　处：《航天医学与医学工程》2009年第1期44-48,共5页Space Medicine & Medical Engineering

基　　金：国家自然科学基金资助项目60661002

摘　　要：目的本文对Mesic呼吸系统仿真模型做系统参数辨识,为人体呼吸系统模型精确度的提高、呼吸疾病的临床诊断和新型呼吸机的设计提供理论和数据支持。方法在Simulink平台下构建集总参数呼吸系统仿真模型,采用最小二乘法,分别用理论数据、实测数据和实测拟合数据对模型进行静态阶段和动态阶段的参数辨识,并由实测数据验证辨识参数。结果由实测拟合数据得到的辨识结果能最好地反映呼吸生理特性,压力、容量和流速曲线最接近实际值。结论辨识结果达到预期目的,此项研究为相关呼吸系统模型的辨识研究提供了一种有用的手段,是Mesic呼吸系统模型的一种完善。Objective To study a Mesie respiratory system parameter identification for providing the useful theory and data support in the improvement of human respiratory model accuracy, respiratory disease diagnosis and design of the new ventilator. Methods The Mesic respiratory system model was established based on Simulink platform. The least-square algorithm was then used to do the static and dynamic parameter identification with theoretical data, clinical data and clinical fitting data. Finally, the validation of the parameter identification was performed by the clinical data. Results The parameters got by clinical fitting data could reach the physiological characteristics well. The pressure, volume and flow curve was the most similar compared with clinical data. Conclusion This method provides an efficient way for the identification research of relative models. It is also a consummation of Mesic respiratory system.

关 键 词：Mesic呼吸系统模型 辨识 最小二乘 

分 类 号：R318.6[医药卫生—生物医学工程]