P300 Wave Detection Using a Commercial Non-invasive Eeg Sensor: Reliability and Performance in Control Applications
Khemri, Nizar Ali
MetadataShow full item record
This thesis proposes a P300-based Brain Computer Interface (BCI) system that uses Emotiv Epoc headset (one of the recently commercially developed headsets for consumers to be used for entertainment purposes) as an EEG sensor to control simple real-time applications. In the recent years, many developers have proposed commercial EEG sensors that can be used in different gaming and entertainment applications. The low cost of these devices compared to the expensive professional devices motivated several individual and group hobbyists to use these sensors in the research field. A decade back in the time, it was impossible for individuals to participate in the research field of BCI and develop their own purposed applications. Nowadays, inexpensive EEG sensors such as Emotiv Epoc headset make it possible for those hobbyists to develop different real life applications. However, the use of such commercial noisy EEG sensors raises the question about the appropriate approaches and algorithms that can process EEG signals acquired by theses sensors and detect the P300 wave among them. Consequently, the work in this thesis first, extensively reviews general BCI systems, different aspects and approaches of P300-based BCI systems proposed in the literature and available low-cost EEG acquisition systems and selected examples of their use in BCI related research. Second, specific P300-based BCI systems data processing and classification algorithms are tested and compared with well-known benchmark P300 datasets provided from the BCI Competition III, 2004. Finally, the tested algorithms and approaches are applied in the proposed P300-based BCI and real-time experiments are performed to test the feasibility of such an inexpensive BCI system in simple control applications such as cursor movement. In the future, more complex applications, such as mobile robot control, can be tested with this system. The overall accuracy of the system is comparable to similar studies and could reach to high fifties of an overall accuracy. To sum up, the results show that using commercial EEG sensors in the research field of BCIs is feasible and promising. Future developments can improve the efficiency of the commercial EEG sensors and that will lead to an affordable more accurate P300-based BCI system.
- OSU Theses