Emotion recognition using EEG and physiological data for a robot-assisted rehabilitation system rehabroby

Robot-assisted rehabilitation systems have been developed to monitor the performance of the patients and adapt the rehabilitation task intensity and difficulty level at each session accordingly to meet the needs of the patients. Rehabilitation robotic systems can be more prosperous if they are able to recognize the emotional states of the patients, and modify the difficulty level of task considering these emotions to increase the engagement of these subjects during the task performance with robot-assisted rehabilitation systems. We aim to develop an emotion recognition model using electroencephalography (EEG) and physiological signals such as blood volume pulse (BVP), skin temperature (ST) and skin conductance (SC) to be used for a robot-assisted rehabilitation system called RehabRoby in this thesis. We group emotions into three categories, which are positive (pleasant), negative (unpleasant) or neutral. We use a machine-learning algorithm called Gradient Boosting Machines (GBM) and a deep learning algorithm called Convolutional Neural Networks (CNN) to classify pleasant, unpleasant and neutral emotions from the recorded EEG and physiological signals. We ask the subjects to look at pleasant, unpleasant and neutral images from IAPS database. We collect EEG and physiological signals during the experiments. We compare the classification accuracies for both GBM and CNN methods when only one sensor information (EEG, BVP, SC and ST) or the combination of the sensor information from both EEG and physiological signals are used.