EEG BASED BRAIN COMPUTER INTERFACE SYSTEMS FOR EXOSKELETON CONTROL A REVIEW WİTH CONCEPTUAL INSİGHTS

Creative Commons License

Akdeniz A. H., Fidan C. B.

7th INTERNATIONAL BLACK SEA MODERN SCIENTIFIC RESEARCH CONGRESS , Artvin, Türkiye, 24 - 26 Haziran 2025, cilt.7, ss.261-271, (Tam Metin Bildiri)

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası: 7
  • Basıldığı Şehir: Artvin
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.261-271
  • Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
  • Erciyes Üniversitesi Adresli: Hayır

Özet

This paper presents a comprehensive review of EEG-based Brain–Computer Interface (BCI)

systems, with a focus on mental command recognition and its integration into human–machine

interaction platforms. Recent advances in machine learning and deep learning techniques have

enabled the translation of raw EEG signals into meaningful commands, facilitating control over

robotic systems, assistive technologies, and exoskeletons. Key EEG paradigms such as motor

imagery (MI), steady-state visual evoked potentials (SSVEP), and P300 are explored in detail

with respect to their signal characteristics, classification challenges, and application

performance.

We further propose a conceptual framework for a BCI-assisted exoskeleton system aimed at

assisting individuals with lower-limb motor impairments. The proposed architecture includes

signal preprocessing, feature extraction via discrete wavelet transform and autoregressive

modeling, and classification using CNN, LSTM, or SVM algorithms. Wireless communication

platforms like OpenViBE and Raspberry are recommended for real-time signal processing and

device control. Although the proposed system has not yet been implemented experimentally,

referenced studies demonstrate the feasibility of such applications with reported classification

accuracies exceeding 90%.

The study also compares deep learning architectures—including CNN, RNN, LSTM,

Autoencoder, and Transformer networks—highlighting their strengths and limitations for

mental command recognition. Practical examples from the literature, such as EEG-based

control of wheelchairs, robotic arms, and drones, are summarized to showcase the real-world

potential of BCI systems. This paper aims to guide future work in designing personalized, low-

latency, and high-accuracy EEG-based interaction systems, with a particular focus on

rehabilitation technologies and cognitive augmentation.