Abstract
During the last decades, electro-biological signals have become the focus of several research institutes. The brain signals obtained non-invasively from the scalp through electroencephalography (EEG) have previously been used for tasks such as controlling a cursor and spelling a word. An EEG-based brain-computer interface (BCI) can also be used to command a semi-autonomous robotic arm by means of motor imagery (MI). The BCI detects the intention to move and provides online feedback to the user. At the same time, the feedback can be used as trigger for different pre-programmed robotic motion tasks. For discrimination of MI, the method of common spatial patterns (CSP) turned out to be highly efficient. The aim of this paper is to present a way to optimize the on-line classification accuracy of a MI-based BCI system by choosing an optimal time window for calculating the CSP. The results showed that for the analyzed datasets the best classification accuracy was achieved with a 4 second time length.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
