Profound correlation of human and NAO-robot interaction through facial expression controlled by EEG sensor

Abstract

Emotion recognition from brain computer interface (EEG) has been studied extensively for the past few years. Time-frequency analysis is widely used in the past research; however, a variation of case study determines the brain signal analysis. In this paper, human emotion from brain waves is recognized in simple ways by calculating a frequency of signal variation. Entirely 35 healthy subjects from students with age 18-25 years old. The students are divided into 3 groups; the first group consists of 15 students; the second group consists of 10 students and the third group consists of 10 students. Each student takes 4 seconds to test his or her internal emotions. The signal speed is recorded during those 4 seconds. Based on stimulus time, various knocks for Z1 and Z2 is observed during a particular time. The experiment can be reproduced for in upcoming future by following the procedure. There are two main elements to measure signal speed which are Delta T and gap. Delta T subject to time differentiation of the changes in time-frequency of Alpha signals. For an evaluation of this work, there is an available benchmark database of EEG labeled with emotions; it mentions that emotional strength can be used as a factor to differentiate between human emotions. The results of this paper can be compared with previous researches which use the same device to differentiate between happy and sad emotions in terms of emotional strength. There is a strong correlation between emotional strength and frequency, we proved that sad feeling is speedier and beyond steady compared to happy since the number of Delta V to Z1 which represents sad emotion of Alpha signals is greater than Delta V to Z2 that represents a happy feeling in the same time period of the interaction process. (C) 2018 The Authors. Published by IASE.