On the use of artificial intelligence techniques in intelligent transportation systems

Full text

On the use of Artificial Intelligence techniques in
Intelligent Transportation Systems
Mirialys Machin, Julio A. Sanguesa, Piedad Garrido, Francisco J. Martinez
iNiT Research Group
University of Zaragoza, Spain
{mmachin, jsanguesa, piedad, f.martinez}@unizar.es
Abstract—Due to the progressive increase in the population
and the complexity of their mobility needs, the evolution of
transportation systems to solve advanced mobility problems
has been necessary. Additionally, there are many situations
where the application of traditional solutions is not entirely
effective, e.g., when the processing of large amounts of data
collected from in-vehicle sensors and network devices is required.
To overcome these issues, several Artificial Intelligence-based
techniques have been applied to different areas related to the
transportation environment. In this paper, we present a study
of the diverse Artificial Intelligence (AI) techniques which have
been implemented to improve Intelligent Transportation Systems
(ITS). In particular, we grouped them into three main areas
depending on the main field where they were applied: (i) Vehicle
control, (ii) Traffic control and prediction, as well as (iii) Road
safety and accident prediction. The results of this study reveal
that the combination of different AI techniques seems to be very
promising, especially to manage and analyze the massive amount
of data generated in transportation.
Index Terms—Intelligent Transportation Systems, Artificial
Intelligence, Artificial Neural Networks, Genetic Algorithms,
Fuzzy Logic, Expert Systems.
I. INTRODUCTION
Transportation systems have become an essential part of
daily human life. It is estimated that an average of 40% of the
world population spends at least one hour on the road every
day [1].
Intelligent Transportation Systems (ITS) are those systems
which integrate both advanced control systems and wireless
communication technologies to provide innovative solutions in
transport and traffic management [2]. In the last years, mainly
due to the progressive increase of the population and the
complexity of their mobility needs [3], ITS have incorporated
different Artificial Intelligence (AI) [4] techniques to provide
new services. These services usually involve managing a
significant amount of data generated by vehicles and drivers
[5], [6]. In particular, modern ITS seek to overall improving
traffic safety and sustainability, while additionally introducing
a positive impact on people living [7], [8]. In fact, the use of
AI has extended to many areas and disciplines in real life,
and transportation systems are no exception since multiple
applications using AI techniques and methodologies, with the
aim of improving ITS, can be carried out.
In this paper, we perform a thorough study of existing
literature related to the application of AI techniques in ITS.
More specifically, we determine the most used techniques and
analyze their advantages and drawbacks when they are applied
to different areas of transportation.
Particularly, we grouped all the proposals into four tech-
niques: (i) Artificial Neural Networks (ANNs), (ii) Genetic
Algorithms (GAs), (iii) Fuzzy Logic (FL), and (iv) Expert
Systems (ESs) (see Figure 1). Next, we present a brief
description of the selected AI techniques.
•Artificial Neural Networks are based on the mathematical
model of neuron proposed by McCulloch and Pitts in
1943 [9], and they can be considered as a processing
paradigm that works similarly to the human brain. ANNs
present some features, such as adaptive learning, self-
organization, and fault tolerance, and they are usually
implemented to solve problems related to pattern recog-
nition, data coding, optimization, and data mining
•Genetic Algorithms are heuristic search methods applied
to find solutions to problems using evolutionary biology
and natural selection strategies [10]. GAs are often used
for searching large and complex datasets, and they are
highly capable of solving optimization problems in many
fields such as robotics, telecommunications, medicine,
and transportation
•Fuzzy Logic techniques emulate the human reasoning
capacity to make rational decisions in an environment of
uncertainty and imprecision [11]. FL is usually applied
to solve many problems, such as complex industrial
processes, recognition of handwritten symbols, driving
comfort, and prediction systems [12].
•Expert Systems are rule-based approaches based on pre-
defined knowledge [13]. They are commonly applied to
decision-making through logical deduction. ESs emulate
the behavior of human experts solving problems. More
Fig. 1. Artificial Intelligence-based techniques mainly used in ITS.

specifically, they store knowledge about a given field and
can solve problems by using logical deduction about that
knowledge. ESs are usually used in telecommunications,
medicine, accounting, and transportation
Once we have introduced the different AI techniques which
have been commonly used in ITS, now we present the main
areas related to transportation in which the AI techniques
described before have been widely applied with promising
results. They are:
1) Vehicle control. In the literature, we find several pro-
posals implemented to develop new control systems for
vehicles. Examples include autonomous driving, anti-
lock braking systems, vehicles’ consumption manage-
ment, and emission control systems
2) Traffic control and prediction, where developed systems
aim at reducing traffic congestion or even predicting
traffic jams in advance
3) Road safety and accident prediction.Wu15 Finally, other
contributions focus on increasing the road safety, avoid-
ing or preventing possible accidents by incorporating
systems that autonomously anticipate accidents, or at
least mitigate their consequences
II. ANA LYSIS OF THE USE OF AI TECHNIQUE S IN ITS
APPLICATIONS
To analyze the most frequently Artificial Intelligence tech-
niques used in Intelligent Transportation Systems, we made
a thorough search of the existing papers in the literature
that combine those fields, (i.e., AI and ITS). As previously
commented, we grouped AI techniques into four different cat-
egories (i.e., Artificial Neural Networks, Genetic Algorithms,
Fuzzy Logic, and Expert Systems).
Table I summarizes all the proposals studied in this work.
In particular, we accounted for the papers that include any
AI technique when proposing new applications and services
in the transportation environment. As shown, the most used
technique in ITS-related approaches is ANNs (twenty-four
times), closely followed by the use of FL (twenty-one times),
and GAs (eighteen times).
It is noteworthy that Expert Systems are the least technique
implemented in ITS (only used in six of the proposals), in spite
of representing one of the most consolidated AI techniques
in other fields, such as medicine or accounting. Due to the
excellent results obtained in other domains [14], it seems
appropriate to extend their application to the transportation
environment, especially in those services that require the
manipulation of large volumes of data.
Additionally, it is necessary to mention that in the majority
of the proposals, the authors combined two AI techniques to
obtain better results.
Figure 2 presents a taxonomy of the analyzed ITS areas
found in the papers studied. More specifically, it depicts the
different services or systems that were presented and the AI
techniques that were applied to improve those systems.
As shown, we found eight systems or services, related to
vehicle control, which were designed relying on the use of AI,
TABLE I
SUM MARY O F TH E PAPER S ST UDI ED : USE O F AI T ECH NI QUE S IN
DIFFERENT ITS AREAS
AI Techniques
ITS Areas ANNs GAs FL ESs
Vehicle control 6 9 6 2
Traffic control and prediction 12 5 8 1
Road safety and accident prediction 6 4 7 3
Total 24 18 21 6
seven systems related to traffic control and prediction, whereas
nine systems related to road safety and accident prediction.
Additionally, we included the techniques used for each of the
systems or services (e.g., Fuzzy Logic was used to improve
the stability of vehicles, the intelligent systems designed
for visual monitoring relied on Fuzzy Logic and Artificial
Neural Networks, or a Genetic Algorithm was implemented
to enhance route planning, and thus avoiding traffic jams).
In the next subsections, we detail the results of our analysis
for each of the ITS areas studied.
A. Vehicle control systems
In the last few years, automotive manufacturers have been
working on adding many advanced technologies to lift existing
vehicle control systems. Basically, they are mainly focused
on promoting safety, fuel efficiency, as well as driver and
passengers’ comfort.
In particular, these new approaches are especially focused
on different control systems, including:
•Anti-lock braking systems
•Lateral and frontal control of vehicles
•Adjusting trajectories in Remotely Operated Vehicles
(ROV)
•Parameters optimization for unmanned vehicles
•Stability improvement in vehicles
•Optimization of vehicle suspension systems
•Minimize consumption and emissions of Plug-in Hybrid
Electric Vehicles (PHEVs) and Electric Vehicles (EVs)
•Human-knowledge integration into Automatic Guided
Vehicles (AGVs)
Figure 3 shows that 39% of the proposals regarding vehicle
control systems relied on GAs. These approaches comprise
different services and applications such as reducing fuel or
energy consumption, improving autonomous driving, or ad-
vanced braking systems. Two of the techniques (i.e., ANNs
and FL) are used in the 26%, and Expert Systems are only
used in the 9% of the proposals.
We consider that GAs are a suitable solution due to the
capacity of this kind of algorithms to solve problems where
the simultaneous optimization of several competing objectives
is required. For example, Yeh and Tsai [15] presented a genetic
algorithm-based charging control system for Electric Vehicles,
specially designed to achieve greater efficiency during the
charging process. The proposal can account for power con-
straints, thereby efficiently managing the charging process.

Fig. 2. Taxonomy of ITS services and systems. The AI technologies used are placed between square brackets
ESs
FL
GAs
ANNs
9%
26%
39%
26%
Fig. 3. AI techniques used in vehicle control systems
Related to the application of Artificial Neural Networks,
Chu et al. [16] used a local Recurrent Neural Network to
control the trajectories of a Remotely Operated Vehicle (ROV).
As for the use of Fuzzy Logic, Naranjo et al. [17] developed
a Global Positioning System with wireless LAN support and
artificial vision. They used FL for controlling the vehicle and
incorporated human knowledge into the algorithms to control
both its direction and speed.
Finally, regarding Expert Systems, Yang et al. [18] applied
an ES in a distributed Vehicle Driving Simulator (VDS), where
the users can experience the feeling of driving while updating
the knowledge database.
B. Traffic control and prediction systems
Regarding traffic control and prediction systems, many
studies have been performed, so far. Those systems are mainly
devoted to predicting traffic flow thanks to the traffic data
gathered by the infrastructure or the vehicles themselves.
Most of the works were focused on the following topics:
•Traffic flow prediction in urban networks (short and long
terms)
•Vehicles speed and route prediction
•Traffic volume prediction
•Route planning to avoid traffic jams
•Reduction of the time stopped at intersections
•Traffic signals control
•Traffic congestion reduction
Although different AI techniques have been used, the ap-
plication of Artificial Neural Networks has been demonstrated
as one of the most frequently used alternative for traffic
controlling and predicting mobility patterns. More specifically,
46% of the proposals implemented this technique (see Figure
4). An example can be found in [19], where the authors
proposed a system based on neural networks to predict short-
term traffic flow.
Regarding the application of GAs, Li et al. [20] developed
new cooperative control to reduce the time that vehicles
are usually stopped at intersections. The system receives the
time range of arriving at the intersection from the vehicles
thanks to their vehicle-to-infrastructure (V2I) communication
capabilities, and a genetic algorithm determines the time that
each vehicle must be delayed to arrive at the intersection.
Therefore, the system sends this information to each vehicle
individually, and then, each vehicle can plan its speed profile.
As for the utilization of FL in traffic control, Milan´
es
et al. [21] used vehicle-to-vehicle (V2V) communications to
estimate vehicles location and speed at intersections. Then, a
fuzzy controller modifies the speed of the vehicles to optimize
the number of vehicles driving through intersections, and thus
reducing traffic congestion in urban environments.
Finally, regarding the use of ESs, Hossain et al. [22]
designed a Belief Rule-Based Expert System (BRBES) to

ESs
FL
GAs
ANNs
4%
31%
19%
46%
Fig. 4. AI techniques used in traffic control and prediction systems
handle uncertainty in traffic control signals.
C. Road safety and accident prediction systems
The third ITS area includes all the systems related to road
safety and accident prediction, i.e., the proposals that: (i) aim
at preventing traffic accidents, (ii) analyze and process the
circumstances that cause vehicle accidents, or (iii) mitigate
the severity of accidents, thereby increasing the chances of
drivers and passengers’ survival [23].
Accidents can be caused by drivers’ behavior, traffic condi-
tions, environmental factors, or road state. However, to avoid
them, it is crucial that vehicles can process all data gathered
by in-vehicle sensors and extract useful information to avoid,
mitigate, and even predict traffic accidents.
Focusing on the works closely related to road safety and
accident prediction, we found that they comprise several
systems and applications. Particularly, they are:
•Intelligent systems for visual monitoring
•Vehicular accidents modeling
•Accident frequency analysis
•Determining the causes of the accident
•Traffic accidents evaluation
•Driver fatigue detection
•Dangerous driving identification
•Automatic incident detection
•Automated braking systems
Figure 5 shows that, in this area, FL and ANNs are the
most frequently used techniques. In particular, the 35% and
the 30%, respectively, of the proposals included them. Note
that this ITS area is by far the one where the different AI
techniques have been used in a more balanced manner since
the 30% of the approaches relied on ANNs and the 15% on
Expert Systems.
Related to ANNs, Delen et al. [24] applied ANNs to
represent the possible relationships between the severity of
the drivers and passengers’ injuries with the external factors
related to the accident.
ESs
FL
GAs
ANNs
15%
35%
20%
30%
Fig. 5. AI techniques used in road safety and accident prediction systems
As for the use of GAs, Fogue et al. [25] proposed and
analyzed several multi-objective genetic algorithms to deter-
mine the optimal sanitary resource sets once the accident has
occurred. More specifically, the main objective was to adapt
existing resources to the specific needs, as well as to reduce
the severity of the injuries.
Regarding Fuzzy Logic, Mamat and Ghani [26] presented an
automated braking system based on a Fuzzy Logic Controller
(FLC) to automatically stop the vehicle when approaching an
obstacle and thus avoiding collisions.
Finally, regarding the use of ESs, Falamarzy et al. [27]
developed a web-based advisory Expert System to minimize
the risk of collisions between vehicles and pedestrians.
III. HYB RI D PRO PO SA LS I N ITS
In this section, we present those works which used a
combination of different AI techniques. In particular, they
commonly combine them to exploit the benefits of each one,
thereby providing a better performance (e.g., in terms of delay
or accuracy).
Table II includes the works studied where the authors
applied more than one AI technique in their approaches.
Particularly, it can be observed how FL and ANNs are more
often combined with each other, and also among the rest of
techniques.
Regarding vehicle control systems, we found how authors
blended GAs and FL, for example, to implement a control
strategy for HEVs aimed to minimize fuel consumption and
emissions, without affecting the vehicle performance [28], or
the efficient power transmission control in electric vehicles
[29].
Related to traffic control and prediction, we found that the
majority of works combined ANNs and FL to predict traffic
flow and vehicles speed. It seems that this combination is the
most suitable to obtain good results in this field ([30], [31],
and [32]).
More specifically, Yin et al. [30] presented a fuzzy-neural
model with the aim to predict traffic flow in urban networks.

TABLE II
WOR KS TH AT COM BIN E IA T ECH NI QUE S TO IM PROV E ITS
ITS areas Ref. ANNs GAs FL ESs Proposal
Vehicle control [28] A fuzzy logic controller, tuned by a genetic algorithm, specially designed to minimize fuel
consumption and emissions of HEVs, without affecting vehicles’ performance
[29] A hybrid genetic algorithm along with a fuzzy logic controller for the high-efficiency power
control of a wireless power transmission system for electric vehicles. The genetic algorithm
is used for optimizing the parameters of the fuzzy controller
Traffic Control
and prediction
[30] A Fuzzy Neural Model (FNM) to predict the traffic flow in urban scenarios. Firstly, traffic
situations are classified into a set of fuzzy clusters, and then, the ANN is applied
[31] A Fuzzy Neural Network specially designed for short-term traffic flow prediction
[32] A Fuzzy Neural Network to forecast vehicles speed based on data previously gathered from
remote traffic microwave sensors
[33] A Genetic Algorithm-based optimization strategy to reinforce the selection of the appropriate
Neural Network structure for short-term traffic flow prediction
[34] A Hierarchical Fuzzy Rule-Based System optimized by Genetic Algorithms to predict traffic
congestion
Road safety
and accident
prediction
[35] An Artificial Neural Network designed to develop an accident appraisal Expert System. This
system attempts to mimic the judging behavior of accident appraisal committees
[36] An Expert System which uses a radial basis function (RBF) Neural Network to estimate the
relationship between vehicles pre-braking speed and the length of their skid marks
[37] A non-intrusive fatigue detection system, which relies on a Fuzzy Expert System, to
characterize the level of alertness of drivers
Quek et al. [31] proposed a diffuse pseudo-external product
Neural Network using the truth-value restriction method to
predict short-term traffic flow. More recently, Tang et al. [32]
presented a new method in the construction of a Fuzzy Neural
Network to predict the vehicles speed.
Vlahogianni et al. [33] used together ANNs and GAs.
Particularly, they presented a genetic algorithm-based opti-
mization strategy to reinforce the selection of the appropriate
Neural Network structure to short-term traffic flow.
In contrast, Zhang et al. [34] presented a Hierarchical Fuzzy
Rule-Based System optimized by a Genetic Algorithm, with
the aim of developing a multi-variable input prediction system
for short-term traffic congestion prediction.
Finally, regarding road safety and accident prediction, we
found some proposals which also blended different AI tech-
niques. In fact, two of them combined ANNs with ESs ([35]
and [36]), and another one used FL and ESs ([37]).
In particular, Chiou [35] used ANNs to develop an accident
appraisal Expert System specially designed for the evaluation
of traffic accidents. In addition, two ANNs models (i.e., party-
based and case-based) were trained and validated using the
cross-validation method. Tseng et al. [36] proposed an Expert
System, which uses a Radial Basis Function (RBF) Neural
Network, to analyze the relationship between vehicles speed
and the length of their skid marks. Azim et al. [37] presented a
Fuzzy Expert System (FES) that classifies the level of alertness
of drivers, as part of a system able to detect drivers’ fatigue,
by analyzing cockpit videos.
IV. CONCLUSIONS
In this paper, we presented a thorough study of existing
works where Artificial Intelligence techniques were applied to
propose new applications and services, or mitigate problems in
Intelligent Transportation Systems. In particular, we focused
our analysis on ITS areas, such as vehicle control, traffic
control and prediction, and road safety.
We observed how some AI techniques have emerged over
the others to address some specific issues.
•Genetic Algorithms have been widely used in vehicle
control systems, especially when optimizing all the pa-
rameters involved. As expected, they are suitable for
multi-objective optimization
•Regarding traffic control and prediction services, most
of the authors implemented Artificial Neural Networks.
More specifically, they are used to make predictions or
classifications
•As for road traffic and accident prediction, both Fuzzy
Logic and ANNs have usually been applied. FL seems to
be more suitable for introducing the human factor to es-
timate accidents frequency, whereas neural networks are
implemented to predict injury severity in traffic accidents
As demonstrated, the use of Expert Systems in ITS-related
approaches is quite marginal, compared to the utilization
of ANNs, FL, or GAs. However, they are combined with
another IA-based technique. This kind of couplings seems very
promising in complex environments such as the transportation
systems. It is quite common to find that Neural Networks
are trained by using Genetic Algorithms, or that they have
been designed through of Fuzzy Logic. Furthermore, Genetic
Algorithms can be used to optimize fuzzy rule-based systems.
Finally, our analysis highlights that the implementation of
ANNs, solely or combined, is very effective, especially to
make accurate predictions based on wireless big data anal-
ysis. Therefore, we consider that they could be applied to
other fields of vehicular environments (e.g., vehicle predictive
maintenance, prediction of driving behavior, etc.).

V. ACK NOWLEDGMENTS
This work has been partially supported by the 2016 Mo-
bility Scholarship for Latin American PhD Students Program
granted by the University of Zaragoza and the Santander Bank,
and by the Government of Arag´
on and the European Social
Fund (T91 Research Group).
REFERENCES
[1] J. Zhang, F.-Y. Wang, K. Wang, W.-H. Lin, X. Xu, and C. Chen, “Data-
driven Intelligent Transportation Systems: A survey,” IEEE Transactions
on Intelligent Transportation Systems, vol. 12, no. 4, pp. 1624–1639,
2011.
[2] A. Sładkowski and W. Pamuła, Intelligent Transportation Systems–
Problems and Perspectives. Springer, 2015, vol. 32.
[3] Y. Pei, X. Li, L. Yu, G. Li, H. H. Ng, J. K. E. Hoe, C. W. Ang, W. S. Ng,
K. Takao, H. Shibata, and K. Okada, “A cloud-based stream processing
platform for traffic monitoring using large-scale probe vehicle data,” in
IEEE Wireless Communications and Networking Conference (WCNC),
March 2017, pp. 1–6.
[4] S. Russell and P. Norvig, Artificial Intelligence: A modern approach,
3rd ed. Pearson, 2016.
[5] J. Barrachina, P. Garrido, M. Fogue, F. J. Martinez, J. C. Cano, C. T.
Calafate, and P. Manzoni, “CAOVA: A Car Accident Ontology for
VANETs,” in IEEE Wireless Communications and Networking Confer-
ence (WCNC), April 2012, pp. 1864–1869.
[6] M. Fogue, J. A. Sanguesa, F. Naranjo, J. Gallardo, P. Garrido, and F. J.
Martinez, “Non-emergency patient transport services planning through
genetic algorithms,” Expert Systems with Applications, vol. 61, pp. 262
– 271, 2016.
[7] K. N. Qureshi and A. H. Abdullah, “A survey on Intelligent Transporta-
tion Systems,” Middle-East Journal of Scientific Research, vol. 15, no. 5,
pp. 629–642, 2013.
[8] F. J. Martinez, C. K. Toh, J.-C. Cano, C. T. Calafate, and P. Manzoni,
“Determining the representative factors affecting warning message dis-
semination in VANETs,” Wireless Personal Communications, vol. 67,
no. 2, pp. 295–314, November 2012.
[9] W. S. McCulloch and W. Pitts, “A logical calculus of the ideas immanent
in nervous activity,” The bulletin of mathematical biophysics, vol. 5,
no. 4, pp. 115–133, 1943.
[10] M. Mitchell, An introduction to Genetic Algorithms. MIT press, 1998.
[11] L. A. Zadeh, “Fuzzy logic,” Computer, vol. 21, no. 4, pp. 83–93, 1988.
[12] C. Wu, X. Chen, Y. Ji, F. Liu, S. Ohzahata, T. Yoshinaga, and T. Kato,
“Packet size-aware broadcasting in VANETs with Fuzzy Logic and RL-
Based parameter adaptation,” IEEE Access, vol. 3, pp. 2481–2491, 2015.
[13] B. P. Mc Cune, R. M. Tong, J. S. Dean, and D. G. Shapiro, “RUBRIC:
A system for rule-based information retrieval,” IEEE Transactions on
Software Engineering, no. 9, pp. 939–945, 1985.
[14] W. P. Wagner, “Trends in expert system development: A longitudinal
content analysis of over thirty years of expert system case studies,”
Expert Systems with Applications, vol. 76, no. Supplement C, pp. 85–
96, 2017.
[15] Y.-C. Yeh and M.-S. Tsai, “Development of a genetic algorithm based
electric vehicle charging coordination on distribution networks,” in IEEE
Congress on Evolutionary Computation (CEC), 2015, pp. 283–290.
[16] Z. Chu, D. Zhu, and S. X. Yang, “Observer-based adaptive neural
network trajectory tracking control for remotely operated vehicle,” IEEE
Transactions on Neural Networks and Learning Systems, 2016.
[17] J. E. Naranjo, M. A. Sotelo, C. Gonzalez, R. Garcia, and T. De Pedro,
“Using fuzzy logic in automated vehicle control,” IEEE Intelligent
Systems, vol. 22, no. 1, pp. 36–45, 2007.
[18] Y. Yang, J. Hu, and D. Chen, “Research on driving knowledge ex-
pert system of distributed vehicle driving simulator,” in 11th Interna-
tional Conference on Computer Supported Cooperative Work in Design
(CSCWD), 2007, pp. 693–697.
[19] J. Z. Zhu, J. X. Cao, and Y. Zhu, “Traffic volume forecasting based
on radial basis function neural network with the consideration of traffic
flows at the adjacent intersections,” Transportation Research Part C:
Emerging Technologies, vol. 47, pp. 139–154, 2014.
[20] J. Li, M. Dridi, and A. El-Moudni, “A cooperative traffic control for
the vehicles in the intersection based on the genetic algorithm,” in 4th
IEEE International Colloquium on Information Science and Technology
(CiSt), 2016, pp. 627–632.
[21] V. Milan´
es, J. P´
erez, E. Onieva, and C. Gonz´
alez, “Controller for urban
intersections based on wireless communications and fuzzy logic,” IEEE
Transactions on Intelligent Transportation Systems, vol. 11, no. 1, pp.
243–248, 2010.
[22] M. S. Hossain, H. Sinha, and R. Mustafa, “A belief rule based expert
system to control traffic signals under uncertainty,” in 1st International
Conference on Computer and Information Engineering (ICCIE), 2015,
pp. 83–86.
[23] M. Fogue, P. Garrido, F. J. Martinez, J.-C. Cano, C. T. Calafate, and
P. Manzoni, “A system for automatic notification and severity estimation
of automotive accidents,” IEEE Transactions on Mobile Computing,
vol. 13, no. 5, pp. 948–963, 2014.
[24] D. Delen, R. Sharda, and M. Bessonov, “Identifying significant predic-
tors of injury severity in traffic accidents using a series of artificial neural
networks,” Accident Analysis & Prevention, vol. 38, no. 3, pp. 434–444,
2006.
[25] M. Fogue, P. Garrido, F. J. Martinez, J.-C. Cano, C. T. Calafate, and
P. Manzoni, “A novel approach for traffic accidents sanitary resource
allocation based on multi-objective genetic algorithms,” Expert Systems
with Applications, vol. 40, no. 1, pp. 323–336, 2013.
[26] M. Mamat and N. Ghani, “Fuzzy logic controller on automated car
braking system,” in IEEE International Conference on Control and
Automation (ICCA), 2009, pp. 2371–2375.
[27] A. Falamarzi, M. N. Borhan, and R. A. O. Rahmat, “Developing a
web-based advisory Expert System for implementing traffic calming
strategies,” The Scientific World Journal, vol. 2014, pp. 1–16, 2014.
[28] A. Poursamad and M. Montazeri, “Design of genetic-fuzzy control strat-
egy for parallel hybrid electric vehicles,” Control Engineering Practice,
vol. 16, no. 7, pp. 861–873, 2008.
[29] W.-C. Wang, C.-C. Tai, S.-J. Wu, and Z.-Y. Liu, “A hybrid genetic
algorithm with fuzzy logic controller for wireless power transmission
system of electric vehicles,” in IEEE International Conference on
Industrial Technology (ICIT), 2015, pp. 2622–2627.
[30] H. Yin, S. Wong, J. Xu, and C. Wong, “Urban traffic flow prediction us-
ing a fuzzy-neural approach,” Transportation Research Part C: Emerging
Technologies, vol. 10, no. 2, pp. 85–98, 2002.
[31] C. Quek, M. Pasquier, and B. B. S. Lim, “POP-TRAFFIC: A novel
fuzzy neural approach to road traffic analysis and prediction,” IEEE
Transactions on Intelligent Transportation Systems, vol. 7, no. 2, pp.
133–146, 2006.
[32] J. Tang, F. Liu, Y. Zou, W. Zhang, and Y. Wang, “An improved
Fuzzy Neural Network for traffic speed prediction considering periodic
characteristic,” IEEE Transactions on Intelligent Transportation Systems,
2017.
[33] E. I. Vlahogianni, M. G. Karlaftis, and J. C. Golias, “Optimized and
meta-optimized neural networks for short-term traffic flow prediction:
A genetic approach,” Transportation Research Part C: Emerging Tech-
nologies, vol. 13, no. 3, pp. 211–234, 2005.
[34] X. Zhang, E. Onieva, A. Perallos, E. Osaba, and V. C. Lee, “Hierarchical
fuzzy rule-based system optimized with genetic algorithms for short
term traffic congestion prediction,” Transportation Research Part C:
Emerging Technologies, vol. 43, pp. 127–142, 2014.
[35] Y.-C. Chiou, “An artificial neural network-based expert system for the
appraisal of two-car crash accidents,” Accident Analysis & Prevention,
vol. 38, no. 4, pp. 777–785, 2006.
[36] W.-K. Tseng and S.-S. Liao, “An expert system using RBF neural
network for estimating vehicle speed based on length of skid mark,”
in International Conference on Natural Computation (ICNC), vol. 2,
2011, pp. 631–635.
[37] T. Azim, M. A. Jaffar, and A. M. Mirza, “Fully automated real time
fatigue detection of drivers through fuzzy expert systems,” Applied Soft
Computing, vol. 18, pp. 25–38, 2014.