Design of an Affordable Brain-Computer Interface for Robot Navigation

Abstract

Brain-computer interface (BCI) devices are designed to bypass neuromuscular pathways enabling an individual to operate an external device using neural activity, as opposed to motor activity. Neural outputs are detected and utilized to drive a command signal to control a device, such as a wheelchair. BCIs provide individuals with severe motor impairment with an alternative mechanism to interact with their environment. Ultimately BCIs help to restore some degree of autonomy to an individual with limited motor control. Increased autonomy will likely lead to an enhanced quality of life of the individual as well as the caregiver(s). In this paper, we developed a BCI-controlled, ground-based robot. A wireless EEG headset was used to obtain the EEG signals from the user and stream the signals wirelessly into a host computer. The computer processed the EEG signals, extracted possible features from the signals to interpret the user's intents as the control commands, and then sent the command signals wirelessly to navigate the robot. This paper discusses the methodology used in the design as well as the response speed and accuracy in robot navigation from the test results. Future applications in assistive technology by using similar design concept are also discussed.