Brain-Computer Interface Based on Compressive Sensing and Steady-State Visual Evoked Potentials Applied to Command a Robotic Wheelchair
Name: HAMILTON RIVERA FLOR
Publication date: 21/09/2023
Examining board:
Name |
Role |
|---|---|
| TEODIANO FREIRE BASTOS FILHO | Advisor |
Summary: People with severe physical disabilities are unable to use standard robotic wheelchairs, which generally demands some motor skills, and therefore total usage of associate muscles. Robotic wheelchairs commanded by Brain-Computer Interfaces (BCIs) based on Electroencephalography (EEG) have demonstrated to be an alternative for these end-users, as such systems translate brain patterns ongoing EEG signals into control commands. However, BCIs relying on local processing encounter limitations in power, scalability, and real-time. In general, existing robotic wheelchairs commanded by BCIs require powerful hardware for high speed signal processing. On the other hand, end-users need a long training process for safely driving these devices. As a solution, cloud-based BCIs and cloud robotics have emerged, leveraging cloud computing for high-performance data processing, storage, and analysis. This integration empowers advanced and adaptive robotic assistance, transforming tele-rehabilitation and e-health applications for people with disabilities. However, integrating cloud computing with BCIs introduces its own set of challenges. These include an efficient and reliable transmission of large volumes of data and stable communication between the brain signal sensor, cloud infrastructure, and robotic wheelchair. To address these challenges, this thesis proposes a novel cloud-BCI system for conveying wheelchair commands through the use of Steady-State Visual Evoked Potential (SSVEP), Compressive Sensing (CS), and a communication framework. The system enhances Information Transfer Rate (ITR), ensuring stable communication among the BCI, cloud infrastructure, and robotic wheelchair. Leveraging cloud Service-Oriented architecture and Robotic Operating
System (ROS), the system allows for easy integration of diverse robotic platforms, and provides flexibility to integrate various protocols, classifiers, metrics, and command techniques. In conclusion, the cloud-BCI system developed here demonstrates to be an efficient and flexible solution for commanding a robotic wheelchair, making it a valuable tool for researchers and developers in the field of assistive technologies, tele-rehabilitation, and training scenarios.
