BCI Enabled Prosthetic Arm Movement: Signal Processing (Part-1)

Abstract

This paper has two parts where first part focus on the signal processing and second part will focus on the signal classification and arm movement. This work improves SSVEP detection in EEG-based BCI systems using Empirical Mode Decomposition and FFT. EEG signals recorded with the 10–20 system were broken down into IMFs. IMF2 to IMF4 were identified as key contributors. After averaging and segmenting the data into 2-second windows, FFT-based spectral analysis found SSVEP responses. We evaluated detection accuracy by classifying segments as true or false. The results show that the suggested EMD-FFT method outperforms the traditional method (85.38%) by increasing average detection accuracy to 88.00%. The construction of an effective control system for upcoming prosthetic arm applications utilizing EEG-driven BCI technologies is aided by this improved accuracy, which fortifies the dependability of SSVEP-based feature extraction.

Introduction

A Brain–Computer Interface (BCI) enables direct communication between the human brain and external devices without relying on muscle or nerve movement, making it especially valuable for individuals with severe motor disabilities such as ALS or spinal cord injuries. Most modern BCIs use EEG (electroencephalography), a non-invasive method that records electrical activity from the scalp. EEG reflects collective neural activity and contains distinct brainwave patterns (Delta, Theta, Alpha, Beta, Gamma), which correlate with different cognitive and emotional states.

Among EEG-based BCI types, Steady-State Visually Evoked Potentials (SSVEP) are widely used because they produce strong, frequency-specific responses in the visual cortex when users look at flickering visual stimuli. EEG data for such systems is commonly collected using the standardized 10–20 electrode placement method.

The literature shows how BCIs have advanced from early motor imagery and P300 systems to sophisticated deep learning–based methods. Research from 2021–2025 demonstrates improvements in wheelchair navigation, robotic prosthetic control, hybrid EEG–EOG systems, and low-cost EEG devices capable of converting standard wheelchairs into thought-controlled systems. Recent work also highlights models like Bi-LSTM, Bi-GRU, and lightweight embedded ML for real-time prosthetics.

The primary problem addressed is the need for a reliable EEG-based BCI system that can accurately interpret noisy, weak brain signals and convert them into meaningful commands to assist people with paralysis or amputations. The objectives include detecting mental states such as attention, fatigue, and stress, and improving FFT-based feature extraction for higher detection accuracy.

The methods involve:
• Signal acquisition: EEG recorded from the Oz electrode at multiple flicker frequencies using the 10–20 method.
• Pre-processing: Filtering, segmentation, artifact removal (ICA), and normalization to obtain clean signals.
• Feature extraction: Converting time-domain EEG into frequency-domain data using FFT to identify meaningful patterns.
• Empirical Mode Decomposition (EMD): Breaking EEG into Intrinsic Mode Functions (IMFs) for analysing non-linear, non-stationary components.

Conclusion

The integration of Brain-Computer Interface (BCI) technology with prosthetic arms demonstrates significant potential in restoring motor functions and independence for individuals with disabilities. Research on Steady-State Visual Evoked Potentials (SSVEP)-based BCIs highlights advantages such as high signal-to-noise ratio, rapid response time, and minimal training requirements, making them highly suitable for real-time prosthetic control. By decoding neural responses elicited from visual stimuli, SSVEP-based systems can provide reliable and efficient communication pathways between the brain and prosthetic devices. The Detection Accuracy using FFT is important for motor imagery and mental task classification which is increasing as the day pass by. However, challenges remain, including signal variability, user fatigue, and the influence of noise and artifacts, which must be addressed through advanced signal processing, spatial filtering, and adaptive algorithms. Hybrid approaches that combine SSVEP with other modalities, such as motor imagery or deep learning-based classifiers, may further enhance accuracy, robustness, and user comfort.

References

[1] Vaid Swati, Singh Preeti, Kaur Chamandeep, “EEG signal analysis for BCI interface: a review,” In: 2015 fifth international conference on advanced computing & communication technologies. IEEE; 2015. p. 143–7. [2] Dave p. Burke, simon p. Kelly, philip de chazal, richard b. Reilly and ciarán finucane, “parametric feature extraction and classification strategy for brain-computer interfacing,” IEEE transactions on neural systems and rehabilitation engineering, vol. 13, no. 1, march2005 [3] Dr. Souvik Pal, Dr.G.Suseendran, Dr. D.Akila, Dr. R.Jayakarthik, T.Nusrat Jabeen, “Advanced FFT architecture based on Cordic method for Brain signal Encryption system,\" 2nd International Conference on Computation, Automation and Knowledge Management (ICCAKM) at Amity University, 2021 [4] Shengyu Wang, BowenJi, Dian Shao, Wanru Chen and KunpengGao, “A Methodology for Enhancing SSVEP Features Using Adaptive Filtering Based on the Spatial Distribution of EEG Signals,” Micromachines 2023, 14, 976. [5] Swati Aggarwal, Nupur Chugh, “Signal processing techniques for motor imagery brain computer interface: A review,” j.array.2019.100003 [6] Mukesh Kumar Ojha, Manoj Kumar Mukul, “Detection of Target Frequency from SSVEP Signal Using Empirical Mode Decomposition for SSVEP Based BCI Inference System,” Springer Science+Business Media, LLC, part of Springer Nature 2020 [7] Gernot R. Müller-Putz, Gert Pfurtscheller, “Control of an Electrical Prosthesis With an SSVEP-Based BCI,” IEEE transactions on biomedical engineering, vol. 55, no.1, January 2008. [8] Abdel-Ouahab Boudraa, Jean-Christophe Cexus, “EMD-Based Signal Filtering,” IEEE transactions on instrumentation and measurement, vol.56, no.6, December 2007. [9] Li Lin, Ji Hongbing, “Signal feature extraction based on an improved EMD method,” j.measurement.2009.01.001 [10] Cheng Junsheng, Yu Dejie, Yang Yu, “Research on the intrinsic mode function (IMF) criterion in EMD method,” Mechanical Systems and Signal Processing, Volume 20, 2006. [11] Palumbo, A., Gramigna, V., Calabrese, B., & Ielpo, N. — Motor-Imagery EEG-Based BCIs in Wheelchair Movement and Control: A Systematic Literature Review. Sensors, 2021. [12] Theerat Saichoo, Poonpong Boonbrahm & Yunyong Punsawad — Investigating User Proficiency of Motor Imagery for EEG-Based BCI System to Control Simulated Wheelchair. Sensors, 2022. [13] Shyam Diwakar, Sandeep Bodda, Chaitanya Nutakki, Asha Vijayan, Krishnashree Achuthan & Bipin Nair — Neural Control using EEG as a BCI Technique for Low-Cost Prosthetic Arms. SCITEPRESS, 2014 [14] Sidharth Pancholi, Amita Giri, Anant Jain, Lalan Kumar & Sitikantha Roy — Source Aware Deep Learning Framework for Hand Kinematic Reconstruction using EEG Signal. arXiv, 2021. [15] Huang, Z., Zhang, Y., Yu, T., He, H., & Li, X. — An EEG-/EOG-Based Hybrid Brain-Computer Interface: Application on Controlling an Integrated Wheelchair Robotic Arm System. Frontiers / PubMed, 2019 [16] Saadabi, D.M.M., et al. — Non-Invasive Brain-Computer Interface Technology (BCI) Modalities and Implementation: A Review. Annals of Advanced Biomedical Sciences, 2021. [17] Basit, A., Nawaz, S., Rehman, A., & Shafique, M. — CognitiveArm: Real-Time EEG-Controlled Prosthetic Arm Using Embedded Machine Learning. arXiv, 202 [18] Thapa, R., Paneru, R., & Poudyal, S. — EEG-Based AI-BCI Wheelchair Advancement Using Hybrid Deep Learning Models. arXiv, 2025 [19] Zhang, X., Li, J., Zhang, Y., & Liu, H. — A Brain-Controlled and User-Centered Intelligent Wheelchair: A Feasibility Study. Sensors, 2024.

Copyright

Copyright © 2025 Nitin Yadav, Anjali Thakur, Mainuddin , Arjun Singh. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.