The autonomous monitoring system via modelling depth of anesthesia using ECG signals

• Main Authors: Tzu-Li Chen, 陳子秝
• Other Authors: Jiann-Shing Shieh
• Format: Others
• Language: en_US
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/yz5qsb

碩士 === 元智大學 === 機械工程學系 === 107 === According to the concept and development of the autonomous system, the practical control system architecture is proposed and applied to the depth evaluation of anesthesia during medical surgery. The design of the autonomous system is to imitate the characteristics of human on problem solving and division. In an unknown environment, the autonomous system must have a high degree of independence to implement self-management and division of labor in order to achieve the given tasks. The autonomous system is an extension of the control system, which consists of five functions (process, model, critic, fault detection, and specification) from low-level to high-level. These five functions enabling the autonomous system to be self-regulating, self-adapting, self-organizing, self-repairing and self-governing. This thesis applies the autonomous system to the field of biomedical science, and the depth of anesthesia is the subject of this research. First applied Continuous wavelet transform to the ECG signal collected during the surgery for pre-processing, then it is trained with the artificial neural architecture (i.e., convolutional neural network) to generate the model for depth evaluation of anesthesia. In order to verify the accuracy of the anesthesia depth prediction, this study used a commercially available anesthesia depth monitor as a reference index, and divided the anesthesia depth value into light, moderate, and deep anesthesia for training comparison.