A hybrid BCI system combining motor imagery and conceptual imagery in a smart home environment

Abstract

In this study, we combined the advantages of two spontaneous brain-computer interface instruction paradigms, conceptual imagery and motor imagery, to develop a smart home control system with better semantics for device selection and more types of device operations. The BCI system allowed users to control three kinds of household equipment: lamps, water heaters, and electric fans. A Raspberry Pi was used to simulate the usage scenarios, where users issued instructions for home equipment selection through conceptual imagery and issued specific instructions for home equipment control through motor imagery. We used Emotiv Epoc to collect EEG data and sent the data to Raspberry Pi, and we built a deep learning-based model for data processing and classification, converting EEG signals into command signals that could control home equipment. Five subjects were recruited to test the performance of the smart home control system and completed a questionnaire to evaluate their willingness to use the system after the experiments. The average accuracy rate of the system operation was 68.9%, with the highest of 73.3%, which proved that the brain-computer interface control system combining the two instruction paradigms was feasible. Users generally showed acceptance of the ease of the system use, giving an average of 5.4 out of 6 ratings.