Design of an Asynchronous BCI Based on a Facial Expression Paradigm for the Remote Control of Robotic Systems

In this paper, an asynchronous electroencephalogram (EEG) Brain-Computer Interface (BCI) based on the facial expression paradigm was designed for the control of the Smart Video Car robot. The proposed BCI was designed to address the problem of lengthy and tedious BCI experiments, which is one of the limitations to their general public use. To minimize the duration of the Subjects' training session while ensuring a good classification of the EEG data, an approach based on the “curse of dimensionality” theory was used. The designed BCI achieved offline accuracy in the range of 80.30-99.76% and performed asynchronous online mode without false detection, which means a 100% accuracy during the robot control thanks to the specified thresholds. The experimental results show that a good trade-off can be found between the comfortability of the experiment for the Subjects and the need to have a sufficient amount of training data for satisfactory classification.