Our study aims at developing electroencephalography (EEG) based Brain-computer interface (BCI) for controlling home appliances. BCI enables users to communicate with external worlds via brain activity without using muscles and our BCI system especially focuses on controlling home appliances. In order to building a usable system, our research interests include not only figuring out the best EEG features representing the users’ intention and decoding model for improving the performance of the system, but also the effect of environment or user’s state on the performance of BCI, the environmental effect on the user’s state while they using the BCI system, the optimal design of stimulus paradigm, etc.
The stimulus is given to the user to induce specific EEG patterns. The medium of presenting stimulus includes conventional LCD and augmented reality (AR). EEG signal acquired is sent to the system for signal analysis and undergoes 1) Preprocessing, 2) Feature Extraction, and 3) Classification. The parameters obtained during the training is used in online BCI for control home appliances. We covers developing and improving each step of EEG signal analysis, from preprocessing to classification.
For improving the performance, we study the optimized way of presenting stimulus, the application of various decoding models including deep learning algorithms, and resolving issues embedded in the system. We are also interested in the influence of user’s state or environment on the performance of BCI, and the effect of environment on user’s state. Moreover, we conduct researches on how to design stimulus paradigm to induce optimal EEG pattern in various environments including AR.