The initial framework for an electroencephalography (EEG) thought recognition software suite is developed, built, and tested. This suite is designed to recognize human thoughts and pair them to actions for controlling a robotic arm.
Raw EEG brain activity data is collected using an Emotiv EPOC headset. The EEG data is processed through linear discriminant analysis (LDA), where an intended action is identified. The EEG classification suite is being developed to increase the number of distinct actions that can be identified compared to the Emotiv recognition software. The EEG classifier was able to correctly distinguish between two separate physical movements.
Future goals for this research include recognition of more gestures, and enabling of real time processing.
Available at: http://works.bepress.com/vijayan_asari/7/
The document available for download is the author's accepted manuscript, provided with permission. Permission documentation is on file.
Some differences may exist between this version and the published version; as such, researchers wishing to quote directly from this source are advised to consult the version of record, available using the DOI provided.