Autonomuos robot control based on EEG and cross-correlation

This paper focuses on the real time control of autonomous robot through the detection of multiple types of motor imagery. The EEG data contains when subject performs imagery motor activity. The motor imagery activity includes movements of left hand, right hand, left leg and right leg. For feature extraction, the cross-correlation method is used to calculate six statistical features (mean, median, mode, maximum, minimum, and standard deviation). For classification, Support Vector Machine (SVM) was used. As SVM is binary classifier, one verses rest approach is implemented for multiclass classification.