Design of an affordable hybrid brain-computer interface to control robotic devices

Abstract

Brain-computer interfaces (BCIs) are devices that utilize neural activity as the foundation for an alternate method of communication or control of external devices. BCIs establish a connection between natural electrophysiological signals regularly occurring in the brain and desired computerized outputs. Instead of relying on natural neuromuscular outputs, BCIs produce artificial outputs, such as movements or navigation commands, capable of controlling robotic devices. The use of BCIs in robotic control holds significant promise for assisting individuals with motor disabilities in various applications, such as exploration, home assistance, or control of an electric wheelchair, thereby fostering greater independence in their daily activities. This paper presents the design of a hybrid BCI tailored for controlling a variety of robotic devices, including a quadcopter, wheeled robot, and wheelchair, by modifying the control scheme of the BCI. An inexpensive electroencephalogram (EEG) headset was employed to capture occipital alpha waves, frontal muscular artifacts, and gyroscope signals indicating head movements. These signals were wirelessly transmitted to a host computer, where signal processing algorithms were implemented to process the acquired signals and extract possible features to interpret the user’s intents as control commands to navigate robotic devices. The response speed and accuracy of the BCI design for robot navigation were recorded and analyzed to evaluate both the performance of the design and the effectiveness of user training. The results support the idea that this cost-effective BCI design can effectively control various robotic devices with minimal adjustments to the control scheme, thus reducing the need for major design changes and user training.