Development of Real-Time Brain-Computer Interface System for Robot Control

Wong, KW; Steve Sai Ho, L; Yang, A

Development of Real-Time Brain-Computer Interface System for Robot Control

Wong, KW Steve Sai Ho, L Yang, A

Permalink

Publication Type:: Journal Article
Citation:: SSRN, 2023
Issue Date:: 2023-04-04

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Adobe PDF

Download Published versionAdobe PDF (1.68 MB)

View on publisher's site

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Wong, KW
dc.contributor.author	Steve Sai Ho, L
dc.contributor.author	Yang, A
dc.date.accessioned	2024-01-02T23:49:30Z
dc.date.available	2024-01-02T23:49:30Z
dc.date.issued	2023-04-04
dc.identifier.citation	SSRN, 2023
dc.identifier.uri	http://hdl.handle.net/10453/173934
dc.description.abstract	Electroencephalogram (EEG) -based brain-computer interfaces (BCI) has been considered one of the prevailing non-invasive method to collect human biomedical signals through attaching electrodes to the scalp. However, it is difficult to detect and use these signals for controlling the online BCI robot in a real environment due to the environmental noise. In this study, a novel state recognition model is proposed to determine and improve the EEG signal states. First, a Long Short-Term Memory Convolutional Neural Network (LSTM-CNN) is designed to extract the EEG features along the time sequence. During this process, errors which are caused by mind randomness or external environmental factors may be generated. Thus, an actor-critic based decision-making model is proposed to correct these errors. The model consists of two networks that can be used to predict the final signal state based on both current signal state probability and past signal state probabilities. Then, a hybrid BCI real-time control system application is proposed to control a BCI robot. The Unicorn Hybrid Black EEG device is used to acquire brain signals. The data transmission system is constructed in OpenViBE to transfer data. The EEG classification system is built to classify BCI commands. In the experiment, EEG data from three subjects was collected, to train and test the performance and reliability of the proposed control system. The system records the robot's spending time, moving distance, and the number of objects pushing down. Experimental results are given to show the feasibility of the real-time control system. Compared with similar BCI studies, the proposed hybrid BCI real-time control system can accurately classify seven BCI commands in a more reliable and precise manner. Overall, offline testing accuracy can achieve 85.22%. When we apply the proposed system to control a BCI robot in a real environment, the best controlling time is 187.4 seconds, and the best running distance is 6.8 meters. This shows that the proposed hybrid BCI real-time control system demonstrated a higher reliability, which can be used in practical BCI control applications.
dc.language	en
dc.relation.ispartof	SSRN
dc.relation.isbasedon	10.2139/ssrn.4402771
dc.rights	info:eu-repo/semantics/openAccess
dc.title	Development of Real-Time Brain-Computer Interface System for Robot Control
dc.type	Journal Article
pubs.organisational-group	/University of Technology Sydney
pubs.organisational-group	/University of Technology Sydney/Faculty of Design, Architecture and Building
pubs.organisational-group	/University of Technology Sydney/Faculty of Design, Architecture and Building/School of Built Environment
utslib.copyright.status	open_access	*
pubs.consider-herdc	false
dc.date.updated	2024-01-02T23:49:28Z
pubs.publication-status	Published online

Abstract:

Electroencephalogram (EEG) -based brain-computer interfaces (BCI) has been considered one of the prevailing non-invasive method to collect human biomedical signals through attaching electrodes to the scalp. However, it is difficult to detect and use these signals for controlling the online BCI robot in a real environment due to the environmental noise. In this study, a novel state recognition model is proposed to determine and improve the EEG signal states. First, a Long Short-Term Memory Convolutional Neural Network (LSTM-CNN) is designed to extract the EEG features along the time sequence. During this process, errors which are caused by mind randomness or external environmental factors may be generated. Thus, an actor-critic based decision-making model is proposed to correct these errors. The model consists of two networks that can be used to predict the final signal state based on both current signal state probability and past signal state probabilities. Then, a hybrid BCI real-time control system application is proposed to control a BCI robot. The Unicorn Hybrid Black EEG device is used to acquire brain signals. The data transmission system is constructed in OpenViBE to transfer data. The EEG classification system is built to classify BCI commands. In the experiment, EEG data from three subjects was collected, to train and test the performance and reliability of the proposed control system. The system records the robot's spending time, moving distance, and the number of objects pushing down. Experimental results are given to show the feasibility of the real-time control system. Compared with similar BCI studies, the proposed hybrid BCI real-time control system can accurately classify seven BCI commands in a more reliable and precise manner. Overall, offline testing accuracy can achieve 85.22%. When we apply the proposed system to control a BCI robot in a real environment, the best controlling time is 187.4 seconds, and the best running distance is 6.8 meters. This shows that the proposed hybrid BCI real-time control system demonstrated a higher reliability, which can be used in practical BCI control applications.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/173934