Motor imagery EEG-based online control system for upper artificial limb

Abstract

For the characters of EEG signal due to motor imagery, six motor imagery tasks are planned and the methods of common spatial pattern (CSP) feature extraction and probabilistic neural network (PNN) classification of EEG signal are studied. And an online control system for upper limb prosthesis based on motor imagery EEG is realized, which consists of the brain computer interface (BCI) unit, the motion controller and the objective orientation module. Online tests show that six assigned movements can be finished successfully by the artificial limb and the operation of moving to the random position by synchronized motion of each joint can be controlled. As a result, a new idea is provided for designing the online robot control system based on EEG, which can further promote the development and application of EEG control technology in the field of rehabilitative and service robot.