A SLAM Integrated Hybrid Brain-Computer Interface for Accurate and Concise Control

In this paper we present a hybrid brain-computer interface (BCI) system that manipulates simultaneous localization and mapping (SLAM) for convenient control of a robot. Due to the low accuracy of classifying multi-class neural signals, using brain signals alone has been considered inadequate for precise control of a robotic systems. To overcome the negative aspects of BCI systems, we introduce a hybrid system where the BCI control of a robot is aided by SLAM. Subjects used electroencephalography (EEG) and electrooculography (EOG) to remotely control a turtle robot that is running SLAM in a maze environment. With the supplementary information on the surroundings provided by SLAM, the robot could calculate potential paths and rotate at precise angles while subjects give only high-level commands. Subjects could successfully navigate the robot to the destination showing the potential of utilizing SLAM along with BCIs.