生活支援ロボットのための SSVEP を用いた操作インタフェースの開発

抄録

This research aims to design an user interface for controlling a tele-operation robot by brain machine interface (BMI). In this research,we use a BMI based on Steady-State Visual Evoked Potentials (SSVEP). We carried out experiments to compare Support Vector Machine (SVM) and Multivariate Synchronization Index (MSI) for analysing SSVEP. Consequently, SVM achieves higher discrimination accuracy than MSI, and thus we decided to use SVM to analyze SSVEP. Since a BMI can transfer only a small amount of information at a time, we need to design an interface which can get the user’s intention efficiently. For this purpose, we have designed an interface which combines the robot’s scene recognition abilities with a BMI. We carried out tele-operation experiments to validate the effectiveness of the interface.