Using finite state machine and a hybrid of EEG signal and EOG artifacts for an asynchronous wheelchair navigation
Introduction
Assisting the physically disabled and aging community to perform daily activities is challenging as their needs vary depending on their physical limitations. Until today, various forms of human machine interface (HMI) systems have been developed to help the physically challenged to regain some mobility and improve the quality of their lives. A typical HMI system normally takes a signal or a combination of signals as inputs and transform them into a command to be executed by an actuator. These signals can come in different forms like sip and puff (Mougharbel, El-Hajj, Ghamlouch, & Monacelli, 2013), eye tracking control (Nguyen & Jo, 2012), tongue control (Kim et al., 2013), head gesture (Halawani, ur Réhman, Li, & Anani, 2012) electromyogram (EMG) (Xu, Zhang, Luo, & Chen, 2013), electrooculogram (EOG) (Ianez, Ubeda, Azorin, & Perez-Vidal, 2012) and electroencephalogram (EEG) (Perdikis et al., 2014).
A brain-actuated wheelchair is an example of an HMI that can be used for assistive purpose. This paper presents an approach to navigate an asynchronous wheelchair using finite state machine and an EEG signal from the alpha band (8–13 Hz) signal and EOG traces extracted from EEG delta band (<3 Hz) signals. The alpha signal is recorded at the occipital region while the delta signals are obtained from the motor cortex area. From these signals, the eyelid positions (closed or open) and the gaze directions (straight, right or left) can be determined by a classifier. Depending on the eyelid positions and gaze directions, a command is issued to move the wheelchair. The wheelchair navigation system is modeled as a finite state machine to allow it to move in three directions forwardly or backwardly, using three gaze directions only. The user can also instruct the wheelchair to stop when needed.
The remainder of this paper is organized as follows. Section 2 describes the related works that use EEG and EOG signals to control an automated wheelchair. In Section 3, the methodology of the proposed system is presented where details of the data acquisition, sliding window, feature classification, finite state machine model and hardware implementation are given. Section 4 describes the experiments performed and the results achieved. Finally in Section 5, conclusions are drawn and future works outlined.
Section snippets
Related work
EEG and EOG signals have been studied by researchers and medical practitioners for various purposes. One of the applications of EEG and EOG signals is as inputs to a wheelchair navigation system. In the early systems of wheelchair navigation using these signals, the wheelchair could only be steered in forward direction (Cao et al., 2014, Choi and Jo, 2013, Choi and Cichocki, 2008, Diez et al., 2013, Galan et al., 2008, Hai et al., 2013, Iturrate et al., 2009, Kaufmann et al., 2014, Khare et
Data acquisition and signal properties
The signals are acquired using g.Mobilab from Guger Technologies at a sampling rate of 256 Hz. The gold electrodes are placed at C3, C4 and O2 with Cz as reference and FPz as ground. The electrode arrangement follows the International 10–20 montage system and the impedance is kept under 10 kΩ. The signals are recorded and analyzed using LabVIEW of National Instrument. The EOG traces are then extracted from the EEG signals using Butterworth filters. For the O2 signal, the EEG signal related to the
Experimental procedure
A total of 20 healthy participants with age ranging from 23 to 27 years old were recruited in the experiment. The participants were briefed on the purpose and nature of the study before signing a consent form for their participation in the experimental session.
The study is conducted in three separate sessions namely the training, online and navigation sessions. The first session involves collecting offline data to determine some threshold values needed for sliding window adjustment and
Evaluation of online session
Table 4 shows the confusion matrix of the results during an online session. Overall, 783 out of 800 instructions were executed correctly and this figure translates to 97.88% execution rate. It is observed that the go straight, stop and null act instructions are perfectly executed. Seven turn left and ten turn right commands are wrongly executed as going straight. This is mainly due to the operator failing to initiate a response within two seconds after the commands are issued. This occurs when
Conclusions
In this study, a hybrid BCI for the asynchronous wheelchair navigation system using an EEG signal and EOG traces embedded in EEG signals is described. A sliding window is used to capture important cues in the signals for a high classification rate. The variance of the EEG signal is used to determine the condition of closed or open eye by thresholding. Then, features extracted from the EOG traces are used as inputs to a pair of minimum distance classifiers whose outputs reveal the gaze shift
Acknowledgements
This work was supported by High Impact Research Grant (www.hir.um.edu.my) Reference No.: UM.C/HIR/MOHE/ENG/16 from Ministry of Higher Education (MOHE), Malaysia.
References (38)
- et al.
A hybrid brain computer interface system based on the neurophysiological protocol and brain-actuated switch for wheelchair control
Journal of Neuroscience Methods
(2014) - et al.
Commanding a robotic wheelchair with a high-frequency steady-state visual evoked potential based brain–computer interface
Medical Engineering & Physics
(2013) - et al.
A brain-actuated wheelchair: Asynchronous and non-invasive brain–computer interfaces for continuous control of robots
Clinical Neurophysiology
(2008) - et al.
The effects of ocular artifacts on (lateralized) broadband power in the EEG
Clinical Neurophysiology
(2001) - et al.
Assistive robot application based on an RFID control architecture and a wireless EOG interface
Robotics and Autonomous Systems
(2012) - et al.
REG-ICA: A hybrid methodology combining blind source separation and regression techniques for the rejection of ocular artifacts
Biomedical Signal Processing and Control
(2011) - et al.
Assisted navigation for a brain-actuated intelligent wheelchair
Robotics and Autonomous Systems
(2013) - et al.
EOG-based visual navigation interface development
Expert Systems with Applications
(2012) - Al-Haddad, A. A., Sudirman, R., & Omar, C. (2011). Guiding wheelchair motion based on EOG Signals using tangent bug...
- Baker, M., Casey, R., Keyes, B., & Yanco, H. A. (2004). Improved interfaces for human-robot interaction in urban search...