Operating a mobile robot using a BCI is a great exercise for a number of reasons. First, the ability to reliably control a robot demonstrates the real-time effectiveness a BCI system. Controlling a robot is also a good first step to operating larger machines. For example, one should certainly be able to control a small mobile robot before attempting to drive an electric wheelchair. Robots also offer many opportunities for developing hybrid control. If a robot is able to augment the instructions it is given with intelligent decisions, it may function much more reliably. For example, a robot can avoid walls and other obstacles or move toward objects of interest even if it was not specifically instructed to do so. Finally, driving robots is just plain fun and cool!
At the CSU BCI Lab, we are currently developing modules for our CEBL3 BCI system to control the ER1 robotics platform. Our first prototype uses a serial P300 speller interface to control the robot by allowing the operator to select among L, R, F, B, S for "turn left," "turn right," "move forward," "move backward" and "stay" respectively. Although this approach is rather slow, it demonstrates the proof of concept. Soon, we hope to develop modules for controlling the ER1 using spontaneous, asynchronous BCI.
Here is a picture of Elliott Forney and Doug Hains driving our ER1 robot with a P300 speller interface for a course on Assistive Technology in the Department of Occupational Therapy at CSU on April 11, 2013.
