A Non-User-Based BCI Application for Robot Control

Abstract

Steady State Visual Evoked Potential (SSVEP) based Brain Computer Interfaces (BCI) can be great assistance for people suffering from physical disabilities due to their high accuracy, high speed, an acceptable number of possible targets, etc. Many researchers have managed to design such systems. Most of these BCIs utilize methods for frequency detection which cause the system to need a training phase for each new user, making the system a user-based one. That is why our goal was to design a BCI that not only has accuracy and speed comparable to similar systems, but also does not need any training phase and thus can be used by new users right away. Our final design got a mean accuracy of 94.85% and a mean ITR of 1.5 bits per trial, both of which are comparable to similar systems, and it was successful to avoid the usual training phase for a new user with a BCI device.