IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, VOL. 27, NO. 2, FEBRUARY 2016 347
Brain Dynamics in Predicting Driving Fatigue Using
a Recurrent Self-Evolving Fuzzy Neural Network
Yu-Ting Liu, Yang-Yin Lin, Member, IEEE, Shang-Lin Wu, Member, IEEE,
Chun-Hsiang Chuang, and Chin-Teng Lin, Fellow, IEEE
Abstract—This paper proposes a generalized prediction system
called a recurrent self-evolving fuzzy neural network (RSEFNN)
that employs an on-line gradient descent learning rule to address
the electroencephalography (EEG) regression problem in brain
dynamics for driving fatigue. The cognitive states of drivers
significantly affect driving safety; in particular, fatigue driving,
or drowsy driving, endangers both the individual and the
public. For this reason, the development of brain–computer
interfaces (BCIs) that can identify drowsy driving states is a
crucial and urgent topic of study. Many EEG-based BCIs have
been developed as artificial auxiliary systems for use in various
practical applications because of the benefits of measuring EEG
signals. In the literature, the efficacy of EEG-based BCIs in
recognition tasks has been limited by low resolutions. The
system proposed in this paper represents the first attempt to
use the recurrent fuzzy neural network (RFNN) architecture to
increase adaptability in realistic EEG applications to overcome
this bottleneck. This paper further analyzes brain dynamics in
a simulated car driving task in a virtual-reality environment.
The proposed RSEFNN model is evaluated using the generalized
cross-subject approach, and the results indicate that the RSEFNN
is superior to competing models regardless of the use of recurrent
or nonrecurrent structures.
Index Terms—Brain–computer interface (BCI), driving
fatigue, electroencephalography (EEG), recurrent fuzzy neural
network (RFNN).
Manuscript received August 15, 2014; revised October 14, 2015; accepted
October 22, 2015. Date of publication November 18, 2015; date of cur-
rent version January 18, 2016. This work was supported in part by the
Aiming for the Top University Plan within National Chiao Tung University
through the Ministry of Education, Taiwan, under Grant 104W963, in part
by the University System of Taiwan–UC San Diego International Center of
Excellence in Advanced Bio-Engineering within the Ministry of Science and
Technology through the I-RiCE Program under Grant MOST 103-2911-I-
009-101 and Grant MOST 104-2627-E-009-001, in part by the Cognition
and Neuroergonomics Collaborative Technology Alliance Annual Program
Plan through the Army Research Laboratory under Cooperative Agreement
under Grant W911NF-10-2-0022, and in part by the Tung Thih Electronic
Fellowship.
Y.-T. Liu and S.-L. Wu are with the Institute of Electrical Control Engi-
neering, National Chiao Tung University, Hsinchu 30010, Taiwan (e-mail:
tingting76319@gmail.com; slwu19870511@gmail.com).
Y.-Y. Lin is with the Electronic Systems Research Division, National
Chung-Shan Institute of Science and Technology, Taoyuan 32546, Taiwan
(e-mail: oliver.yylin@gmail.com).
C.-H. Chuang is with the Brain Research Center, National Chiao Tung
University, Hsinchu 30010, Taiwan, and also with the Faculty of Engineering
and Information Technology, University of Technology Sydney, Sydney,
NSW 2007, Australia (e-mail: chchuang@ieee.org).
C.-T. Lin is with the Brain Research Center, Department of Electrical
Engineering and Computer Science, National Chiao Tung University,
Hsinchu 30010, Taiwan, and also with the Faculty of Engineering
and Information Technology, University of Technology Sydney, Sydney,
NSW 2007, Australia (e-mail: ctlin@mail.nctu.edu.tw).
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TNNLS.2015.2496330
I. INTRODUCTION
I
NDIVIDUALS typically possess varying degrees of driving
experience, and driving safety has proved to be one of
the most important issues to address in the realm of public
safety. Driving safety is influenced by two main factors:
1) the external environment and 2) the mental states of the
drivers [1]. External factors are usually unpredictable and
difficult to address, and doing so often incurs substantial costs.
However, concerning the mental states of drivers, although the
investigations of the human mind remain difficult, the plenti-
ful and substantial results obtained in the field of cognitive
neuroscience have recently provided an opportunity to resolve
this problem [2].
Various physiological/psychological conditions, e.g.,
fatigue [3], distraction [4], and motion sickness [5], can affect
driving safety. Fatigue driving, or drowsy driving, exposes
a personal and public property to dangerous situations.
Individuals may find themselves in a drowsy or fatigue state
as a result of sleep deficits, long-term driving, midnight
driving, monotonous driving, taking sleeping drugs, or
sleep disorders [3], [6]. Long-term driving is a common
cause of accidents, and has been experienced by numerous
drivers. Therefore, the development of artificial auxiliary
systems for detecting drowsy driving states is of the utmost
importance. In contrast to various other types of systems
[7], [8], brain–computer interfaces (BCIs) are effective,
because they can directly evaluate the cognitive states of
human beings [8], [9]. Many measurement methodologies
have been proposed for the estimation of brain dynamics
[e.g., computer tomography, positron emission tomography,
electroencephalography (EEG), magnetoencephalography, and
magnetic resonance imaging]. As one of the most important of
these methodologies, EEG has attracted substantial amounts
of attention over many years. A significant advantage of
EEG over other extraction methodologies is that it provides
convenient real-time measurements. Therefore, EEG signals
are commonly used in real-world applications [10]–[12].
A considerable body of literature in cognitive science indi-
cates that we have grasped the connections between brain
activation and a variety of tasks through the use of EEG
topography. In the literature related to driving tasks [13]–[16],
EEG spectra have been widely exploited and are commonly
used to identify different cognitive states. The magnitude of
EEG power in the alpha band in the occipital cortex has been
found to be one of the significant features that accompanies the
onset of drowsiness (DS) [15]. Azarnoosh et al. [13] illustrated
2162-237X © 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.