A 3D Visual Stimuli Based P300 Brain-computer Interface

Abstract

The brain-computer interface (BCI) is a technology that allows human brain to directly control devices. Among various kinds of BCI based system, the brain actuated robotic arm is a meaningful research for many fields. Due to the multiple freedom of robotic arm, the evoked potential is usually adopted to build the BCI in which a computer displayer is usually adopted as the stimulation interface. However, the introducing of the displayer is usually leading to the distracting for the user. To alleviate this problem, we designed a LED stimulator to mount on the end-effector of the robotic arm so that the stimulator and the controlled devices are integrated together. This stimulator includes six LEDs, and each LED positioned on the corresponding direction of the robotic arm. In this way, the visual stimulus is overlapped on the device. Therefore, the distracting in robotic arm is effectively restrained. Additionally, a slide-window based P300 classifier is designed to fast identify human intent. With this system, human could control the robotic arm in real-time. Online experiment is performed; three subjects participated in the test. They are instructed to control the robotic arm to grasp a bottle on the desk and place it in the target box. Each subject takes 10 trials, and all of them could complete more than seven tasks. The results proved the effectiveness of our system.