Moccaldi, Nicola (2021) Measurement instrumentation in Passive Brain-Computer Interfaces. [Tesi di dottorato]
Preview |
Text
Moccaldi_Nicola_34.pdf Download (5MB) | Preview |
| Item Type: | Tesi di dottorato |
|---|---|
| Resource language: | English |
| Title: | Measurement instrumentation in Passive Brain-Computer Interfaces |
| Creators: | Creators Email Moccaldi, Nicola nicola.moccaldi@unina.it |
| Date: | 13 December 2021 |
| Number of Pages: | 132 |
| Institution: | Università degli Studi di Napoli Federico II |
| Department: | Ingegneria Elettrica e delle Tecnologie dell'Informazione |
| Dottorato: | Ingegneria elettronica e delle telecomunicazioni |
| Ciclo di dottorato: | 34 |
| Coordinatore del Corso di dottorato: | nome email Riccio, Daniele daniele.riccio@unina.it |
| Tutor: | nome email Arpaia, Pasquale UNSPECIFIED |
| Date: | 13 December 2021 |
| Number of Pages: | 132 |
| Keywords: | Brain Computer Interface, Health Monitoring, Wearable Device, Emotions Recognition, Attention Measurement, Stress Assessment, Engagement Detection, Machine Learning, Experimental Reproducibility |
| Settori scientifico-disciplinari del MIUR: | Area 09 - Ingegneria industriale e dell'informazione > ING-INF/07 - Misure elettriche e elettroniche |
| Date Deposited: | 31 Jan 2022 09:51 |
| Last Modified: | 28 Feb 2024 11:42 |
| URI: | http://www.fedoa.unina.it/id/eprint/14287 |
Collection description
EEG Passive Brain Computer Interfaces assess cognitive and emotional condition of the user by means of electric signal acquired from the scalp. In the framework of Industry 4.0, Passive BCI represents a promising monitoring channel to improve human-machine interaction and integration. In this thesis, the prototyping and characterization of BCI measurement instrumentation to detect basic and complex mental states are presented. Both off-the-shelf instrumentation and CE-marked devices for medical use are exploited to acquire brain signals. The proposed solutions address the challenge of maximizing hardware wearability (minimizing the number of channels and employing dry electrodes) without penalizing accuracy and latency. To this end, appropriate signal processing strategies based on data-driven approaches are developed. Semi-custom machine learning algorithms are implemented for feature extraction and classification. Emotional valence, rehabilitation distraction, learning engagement, and work-related stress are the case studies proposed to experimentally validate the measurement instrumentation. Databases of EEG signals available online were consulted and experimental campaigns were conducted for a total of more than 200 subjects. Crucial metrological issues in the measurement instrumentation of passive BCIs are explored: e.g., definition of the measurand and its compatibility with the quantitative approach, experimental reproducibility, as well as cross- and within-subject reproducibility. The within-subjects accuracy exceeded 92% and 95% for distraction and emotional valence, respectively. The cross-subject accuracy reached 99% in recognition of a stressful condition.
Downloads
Downloads per month over past year
Actions (login required)
 |
View Item |
