Design And Research Of Wearable Affective Computing Devicefor Online Learning

The whole world has witnesses the incredible advancement of online learning, and there are increasingly more people acquiring knowledge through this medium. While the lack of affection and interaction have been the pain of online learning for a long period. Affective computing is one of the most cutting-edge research topics these days, which can solve the problems aforementioned of online learning. Physiological signals are widely used in the field of affective computing for recognizing different emotions. Among various signals, galvanic skin response, also known as EDA is a popular choice, while this method is not applicable to the research of online learning in the actual learning scene. For detecting galvanic skin response, many studies prefer to use the bulky equipment in the laboratory. Based on the fact that wearable devices are widely used, we plan to design a new wearable device to obtain physiological signals in real time and do the investigation and research in the actual learning scene. This study has mainly completed following tasks:In terms of hardware, this study designed a wearable device to detect galvanic skin response for the purpose of’design requirement’. This newly designed equipment is light and portable, which can be worn on the wrist. At the same time, this device has abundant ports that can be connected to other sensors, such as heart rate sensor, pulse sensor, body temperature sensor for different research aims. Furthermore, the data collected by the device can be transferred by Bluetooth or Wi-Fi to a remoting host, which makes it possible for investigating emotions in the wild.As for software, this study takes the idea of’Visualization’,’Big data’ and ’Machine learning’ into account, designed a software to store the data from wearable GSR device and implement emotion recognition. This software is built on Matlab and designed in a GUI way, with this platform we can save and observe physiological signals as well as extract features and detect emotions from them. At the same time, this study has done several experiments to measure the performance of the independently designed software and hardware.About system design, with the independently designed hardware and software, features extracted from galvanic skin response, different emotions calculated with machine learning and the emotion regulation mechanism, this study proposes the design framework for adjust learning contents in real-time. Based on the qualitative research of the subjects, we find out the acceptability of this system was rather high.