Myoelectric prosthesis control algorithm based on Motor Imagery recognition by a Spiking Neural Network.

A tétel rövid nézete
dc.contributor.advisorAlmusawi Abdulkareem, Husam
dc.contributor.authorMartins Alonso, Arthur
dc.contributor.departmentDE--Műszaki Kar
dc.date.accessioned2025-09-04T14:51:59Z
dc.date.available2025-09-04T14:51:59Z
dc.date.created2025
dc.description.abstractEnhancing the control of myoelectric prostheses requires accurate and energy-efficient decoding of user intent. This thesis investigated the integration of Spiking Neural Networks (SNNs) for motor imagery based electroencephalography (MI-EEG) pattern recognition to improve both classification accuracy and energy efficiency for prosthetic command generation. Our proposed approach envisions an electromyographic (EMG) and MI-EEG hybrid control model, using SNNs to decode the movement intention from the MI-EEG, and the EMG providing proportional control. We implemented SNNs through ANN-to-SNN conversion, specifically evaluating lightweight deep convolutional network architectures (LENet) on a 3 class MI-EEG dataset from the same limb. Results demonstrate that the SNN model achieved a mean overall accuracy of 77.23%, on par with its LENet counterpart and competitive with state-of-the-art methods, while its theoretical energy estimations indicate that the SNN model could reduce computational energy consumption by 59.6% compared to its CNN equivalent, primarily due to sparse, event-driven processing. These findings underscore the potential of SNNs to enable accurate, low-power processing, paving the way for more natural and practical neuroprosthetic control systems.
dc.description.courseMechatronical Engineeringen
dc.description.degreeBSc/BA
dc.format.extent89
dc.identifier.urihttps://hdl.handle.net/2437/397232
dc.language.isoen
dc.rights.infoHozzáférhető a 2022 decemberi felsőoktatási törvénymódosítás értelmében.
dc.subjectProsthetic control
dc.subjectMotor Imagery
dc.subjectBrain-Computer Interface
dc.subjectElectroencephalography
dc.subjectElectromyography
dc.subjectDeep Learning
dc.subjectSpiking Neural Networks.
dc.subject.dspaceInformatics
dc.subject.dspaceEngineering Sciences
dc.subject.dspaceBiology::Biotechnology
dc.titleMyoelectric prosthesis control algorithm based on Motor Imagery recognition by a Spiking Neural Network.
Fájlok
Eredeti köteg (ORIGINAL bundle)
Megjelenítve 1 - 2 (Összesen 2)
Nincs kép
Név:
Thesis classification request signed.pdf
Méret:
1.96 MB
Formátum:
Adobe Portable Document Format
Leírás:
Megjelenítés
Nincs kép
Név:
Stripped - Alonso Thesis 2025 - Myoelectric prosthesis control algorithm based on Motor Imagery recognition by a Spiking Neural Network.pdf
Méret:
4.22 MB
Formátum:
Adobe Portable Document Format
Leírás:
Megjelenítés
Engedélyek köteg
Megjelenítve 1 - 1 (Összesen 1)
Nincs kép
Név:
license.txt
Méret:
1.69 KB
Formátum:
Item-specific license agreed upon to submission
Leírás:
Letöltés
Gyűjtemények
Theses (Department of Electrical Engineering and Mechatronics)