| Thesis summarizes the relevant design work of a PC software system for physiological multi-parameter monitor.The software realizes various common functions of physiological monitoring software such as equipment communication,data visualization and storage,equipment and user management.At the same time,aiming at the functional scalability of the software system,the overall software structure,data exchange protocol,communication interface with decoupling model and data storage structure are developed design and optimization.Based on the software system,thesis,relying on the Department of Nephrology and Hemodialysis Center of a Class III hospital,carried out physiological monitoring application research for patients with advanced uremia,mainly involving ECG monitoring,cardiac beat classification(abnormal detection),etc.The specific work and achievements of thesis are as follows:1)Design and implementation of physiological multi-parameter monitoring software system.It includes equipment communication,multi-user data monitoring,multi-channel waveform display,file persistence,user information and equipment information management.The monitoring objects include five types of signals: electrocardiogram,respiration,body temperature,inertial navigation,and pulse wave.At present,the software system has been applied to a real-time physiological multi-parameter dynamic monitor,and the actual operation effect is good.2)Research on performance optimization and functional scalability of software system.Focusing on the actual task requirements of multi-communication mode and multiparameter acquisition,the communication interface specification is designed and three kinds of implementations are given,including Bluetooth,Wifi and serial port.The message parsing framework based on state machine and two data protocols,fixed length combination and variable length single packet,are proposed; The design mode of combining observer and intermediary is adopted to establish a ”many-to-many” topology model between the upper service and the lower data interaction.The separation of service and data is realized by designing the service deployment object,which effectively reduces the system coupling.3)For the problem of beat classification,thesis has carried out the research on beat by beat classification algorithm based on RR interval and SVM classifier algorithm based on wavelet transform.Among them,the beat by beat classification algorithm is oriented to four beat categories,and is optimized based on the posterior probability scheme.The improved algorithm improves the sensitivity of PVC category by 7%; The SVM classifier algorithm is oriented to six more detailed heart beat categories,and has achieved accuracy of 99.70% and 99.26% in training and testing on two types of datasets,respectively.Finally,an extension module for heart beat classification was implemented based on the classification algorithm.4)Function and performance test of physiological multi-parameter monitoring software system.According to the module decomposition design use case,the function verification of the system is carried out and the concurrent request capability of the system is tested.The final experiment shows that the software system designed in this paper can complete equipment communication,waveform display,parameter monitoring,information maintenance and other functions,and the monitoring effect is good,satisfying the service requirements.5)An application validation was conducted for a group of late stage uremic patients in a tertiary hospital.A long-term electrocardiogram monitoring experiment and heart beat classification experiment were designed for 118 uremic patients during hemodialysis.The472 minute resting electrocardiogram in the experimental data collection was used as an artificial standard to construct a total of 43632 heart beat heart rate dataset for uremic patients.The heart beat classification statistical function classified 1883 heart beats from10 people with an accuracy of 99.7%. |
PDF Full Text Request