An EEG-based Brain-Robot Interface for Controlling ASIMO by Imagination of Movement

Project Leaders: Helge Ritter, Yaochu Jin
PhD Student: Andrea Finke

A Brain-Machine Interface (BMI) is a human-machine interface that provides a direct connection of the human brain and a machine, enabling the control of a device without any motor actions. An important type of BMIs are non-invasive systems utilizing electroencephalography (EEG) to measure the brain activity, which provides a high temporal resolution of the measurement and easy as well as locally unrestricted application, enabling portable devices that can operate in real time.

We focus on the application of EEG-based BMIs for robot control, so-called Brain-Robot Interfaces, utilizing two complementary EEG components, the P300 event-related potential, a discrete selection mechanism triggered by rare, relevant stimuli in a sequence of background stimuli, and the imagination of movement, based on the principle that the sensorimotorcortex activity is identical whether a movement is actually performed or only imagined.

Brain-Robot Interfacing is currently subject to fundamental research. The main target group for all Brain-Machine Interfaces are severely paralyzed patients who are unable to perform motor actions. In a future scenario, a Brain-Robot Interface could enable these patients to have the robot perform motor actions in their place while providing the commands for these actions merely with the signals from the brain. Hence, the robot could serve as a form of substitution for the user's own body. Also, a basically autonomously acting robot could be given commands through such an interface when its actions do not match the users wishes or a specific behavior or action is demanded.