Wittenboer, Lüke van den (2025) Brain Computer Interfaces for Robust Continuous Control. Master's Thesis / Essay, Artificial Intelligence.
|
Text
MScAIThesisLLvdWittenboer.pdf Download (2MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
Toestemming.pdf Restricted to Registered users only Download (211kB) |
Abstract
This research investigates the feasibility of robust, continuous Brain-Computer Interface (BCI) control using motor execution signals in dynamic, cue-free environments. To bridge the gap between traditional trial-based BCI paradigms and real-world applications, a novel experimental framework involving a simulated driving task controlled via an electromyography (EMG)-based classifier was developed. Simultaneously recorded electroencephalography (EEG) data was analysed offline using two classification approaches: logistic regression with common spatial patterns and a shallow convolutional neural network, to decode discrete motor states. In contrast, a separate regression approach using Ridge Regression was employed to predict continuous muscle activation directly from EEG features. Results showed that while the SCNN achieved high performance in matched conditions, it suffered in transfer settings, highlighting challenges in model generalisability. Conversely, the CSP-based model was more robust across contexts. The Ridge Regression model failed to accurately predict EMG envelopes from EEG, suggesting limitations in current feature extraction under continuous control conditions. This study introduces an EMG-guided approach to BCI development that enables low-frustration data collection and lays the groundwork for future motor-based BCI systems suited for real-time, continuous applications.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Sburlea, A.I. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 08 May 2025 06:37 |
| Last Modified: | 08 May 2025 06:37 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/35151 |
Actions (login required)
 |
View Item |