Development of a brain-controlled robotic arm system for motor imagery based on MR visual guidance

摘要

Brain-computer interface (BCI) aims to communicate with external devices through EEG signals to realize the control of external devices. In order to improve the training efficiency of visually guided motor imagery (MI) and the classification accuracy of electroencephalography (EEG) signals, this paper investigates MR-based visually guided motor imagery training for a lightweight robotic arm control system. The system integrates two kinds of biological signals, EEG and EEG, in stages, and uses dual EEG as the task switch, motor imagery EEG signal to control the movement of the robotic arm, and single EEG to control the stage switching, which realizes the dichotomous motor imagery to generate multiple control commands, and completes the continuous control of the robotic arm. In which the motor imagery EEG signals were extracted using the method of combining common spatial patterns (commonspatialpattern,CSP) and classified using supportvectormachines (SVM). The results show that there is a significant improvement in the classification accuracy of EEG signals after training with motor imagery. In order to verify the feasibility of the system, a brain-controlled robotic arm grasping experiment was designed, and the subjects were tested through online experiments to realize the robotic arm control through the use of EEG signals, which is conducive to the further promotion of the practical application of brain-computer interface technology.Keywords: electroencephalography; motor imagery; mixed reality environments; visual guidance.