A Proficient and Low Intricacy EEG Motor Imagery Classification Algorithm Using Boosted Decision Subspace Ensemble Learning

Exoskeleton or brain computer interface design is an complicated and challenging effort as it involves many complicated subtasks. Building an synchronized and energy optimized exoskeleton algorithm is the primary objective of this research. In this study the extraction of the frequency band signals from the brain signals are done by the enhanced Morlet Discrete wavelet Algorithm. Further classification of the obtained signals are done through the Boosted Decision Subspace Ensemble Learning (BDSEL) Algorithm. The classified signals are differentiated in to different frequency regions namely alpha, beta, gamma, delta and theta respectively and the accuracy of the acquired signals are checked against the false identification. An enhanced differentiation with high accuracy rate is obtained in this method without trading off with the resolution of the signals. The energy efficiency of the algorithm was proved to be enhanced when compared to the existing Naive Baye's technique. Thus an highly synchronized energy optimized algorithm was designed for a knee exoskeleton system