Mind guided motion control of robot manipulator using EEG signals

Most of current brain computer interfaces (BCI) designed for environment control or navigation are usually made using motor imagery and expensive data acquisition equipment. This study presents an alternative way of utilizing Electroencephalogram (EEG) signals acquired through a low-cost headset to steer a virtual object. In this paper, a preliminary framework of controlling the end effector of a 7 degree of freedom (DoF) virtual robot manipulator by EEG signals and the P300 speller BCI are built. The EEG signals emitted by the user brain are collected by the Emotiv EPOC neuro-headset, which is connected to the OpenViBE platform, where a BCI is designed to retrieve and process the EEG recording in real time. P300 is an observed peak evoked 250 to 500 ms after a visual stimulus in an EEG signal. The P300 speller BCI is an OpenViBE scenario that allows the user to spell letters using P300 signals. The main idea is to associate a letter with a command to control the virtual robot, where the six letters are corresponding to 6 robot movements in the Cartesian space: +X, -X, +Y, -Y, +Z, -Z. Letters are sent via a VRPN server on Open Vibe platform and transmitted via a C++ client (designed on Microsoft Visual Studio 2010) to V-REP, a software for creating and testing virtual robot. V-REP remote API is used on the C++ client to control directly the end effector. Online experiments have been performed to test the system performance and the functionality.