BCI for Robot Control
Human-robot interaction can occur through different types of interfaces. Recently, our lab has begun exploring methods to communicate with and control robotic systems through measures of brain activity, called Brain-Computer Interfaces (BCI). Using electroencephalogram (EEG), we place sensors on the surface of the scalp and record event-related electrical potentials and oscillatory activity which are decoded for use with assistive robots and in rehabilitation.
EEG sensors placed on the scalp measure brain activity that can be decoded to control assistive robotic systems and support rehabilitation.
Robot Control for Activities of Daily Living
For individuals with motor impairment, in-home mobile and manipulator robots can be used to perform activities of daily living (ADL). Depending on the type of impairment, a person may have limited ways to interact with and send commands to the robot. When that is the case, brain activity can provide a mechanism for robot control. Certain patterns of activity can be recorded using EEG and converted into robot action commands. We are exploring interactions through brain-computer interfaces while directing robots in executing ADL tasks.
Sensing Deviations from Desired Robot Motion
Real-time detection of discrepancies between desired robot action and the action which is performed could streamline online control and reduce operator cognitive load. If these deviations could be determined passively from brain activity, robot motion discrepancies could be identified and resolved much earlier into their commission. To build brain-computer interfaces for this purpose, there are empirical questions that must first be answered regarding the latency and localization of the resulting neural activity. We are investigating these questions in collaboration with cognitive neuroscience colleagues.