Defesa de Dissertação de Mestrado – Rafael Mendes Duarte – 22/05/2017
22/05/2017 16:54
| Defesa de Dissertação de Mestrado | ||
| Aluno | Rafael Mendes Duarte | |
| Orientador | Prof. Alexandre Trofino Neto, Dr. – DAS/UFSC | |
| Data | 22/05/2017 (segunda-feira) 15h30
Sala PPGEAS I (Piso Superior) |
|
| Banca | Prof. Alexandre Trofino Neto, Dr. – Presidente – DAS/UFSC;
Prof. Jefferson Luiz Brum Marques, Dr. – EEL/UFSC; Prof. Odival Cezar Gasparotto, Dr. – CCB/DCF/UFSC; Prof. Hector Bessa Silveira, Dr. – DAS/UFSC. |
|
| Título | Low Cost Brain Computer Interface System for AR. Drone Control | |
| Abstract: This work presents the design, implementation, and testing of a Brain-Computer Interface (BCI) system based on μ-waves to control the navigation of a drone. BCI systems perform the translation of brain signals into commands to communicate with external applications. The μ rhythm is a type of brain signal response to motor activity which can be easily measured by electroencephalography (EEG). For this reason, μ-waves based BCI systems have been extensively explored in the literature as a way of enabling patients with compromised neuromotor functions to interact with the outside world. To implement the signal processing and application interface routines, a software platform was built based on well-established filter and classification techniques, such as the Common Spatial Patterns (CSP) and the Linear Discriminant Analysis (LDA). For interfacing with the drone, an algorithm for translating the classifier outputs into drone commands was proposed. In addition, the acquisition of brain waves was performed by a low-cost and open-hardware EEG amplifier called OpenBCI. The validation of the designed system was performed using public and an acquired motor imagery EEG datasets, which were supplied to the platform to simulate the real-time performance of the system. The tests, conducted in a drone simulator, demonstrated the correct operation of the proposed methodology and the designed system. | ||
