Transactions on Industrial Informatics

av
e
rage
 wai
ti
ng
 t
im
e
 (s
e
conds
)
v/c
Average waiting time (uniform pattern)
AdaptiveLight
NoGroup
FuzzyGroup
FuzzyGroupLearning
0
20
40
60
80
100
0.50
0.60
0.65
0.70
0.80
0.85
0.90
av
e
rage
 wai
ti
ng
 t
im
e
 (s
e
conds
)
v/c
Average waiting time (backbone pattern)
AdaptiveLight
NoGroup
FuzzyGroup
FuzzyGroupLearning
Fig. 5. Waiting time variance 
0
10
20
30
40
0.50
0.60
0.65
0.70
0.80
0.85
0.90
wai
ti
ng
 t
im
e
 v
ari
ance
v/c
waiting time variance (uniform pattern)
AdaptiveLight
NoGroup
FuzzyGroup
FuzzyGroupLearning
0
20
40
60
0.50
0.60
0.65
0.70
0.80
0.85
0.90
wai
ti
ng
 t
im
e
 var
iance
v/c
waiting time variance (backbone pattern)
AdaptiveLight
NoGroup
FuzzyGroup
FuzzyGroupLearning

---

1551-3203 (c) 2016 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.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2016.2590302, IEEE
Transactions on Industrial Informatics
8 
the FuzzyGroupLearming algorithm against the other three 
algorithms in Table IV. The percentages in the table are 
calculated based on the average values of correspoinding 
metrics in different traffic levels and traffic patterns. 
VI.
C
ONCLUSION AND 
F
UTURE 
W
ORK
In this paper, we study the intersection control problem for 
smart transportation systems in smart cities. Different from 
existing works, we propose to divide the waiting vehicles in a 
lane into different groups. The permission of passing 
intersection is then granted in terms of vehicle groups, via V2I 
communications. The major challenge lies in determining the 
appropriate groups based on real-time traffic conditions, with 
respect to requirements of average waiting time and fairness. 
We adopt neuro-fuzzy network to do the grouping and also the 
group scheduling. Furthermore, we apply reinforcement 
learning to fine-tuning the parameters of the network and make 
it adaptive to various traffic conditions. Grouping vehicles 
makes our intersection control algorithm efficient and fair, 
especially in high dynamic traffic flows. Such advantages have 
been validated via extensive simulations.
Group based intersection control is a novel approach and 
more efforts need to be made for better solutions. Possible 
directions include improving the grouping fuzzy rules with 
other machine learning techniques, considering more complex 
scenarios of multiple intersections. 
R
EFERENCES
[1]
Berenji H. R., Khedkar P. Learning and tuning fuzzy logic controllers 
through reinforcements. Neural Networks, IEEE Transactions on, 1992, 
3(5): 724-740. 
[2]
Bingham E. Reinforcement learning in neurofuzzy traffic signal control. 
European Journal of Operational Research, 2001, 131(2): 232-241. 
[3]
Cai C., Wang Y., Geers G. Adaptive traffic signal control using vehicle-
to-infrastructure communication: a technical note[C]. Proceedings of the 
Second International Workshop on Computational Transportation 
Science. ACM, 2010: 43-47. 
[4]
Chan K. Y., Khadem S., Dillon, T. S., Palade, V. Selection of Significant 
On-Road Sensor Data for Short-Term Traffic Flow Forecasting Using the 
Taguchi Method. Industrial Informatics, IEEE Transactions on, 8(2): 255 
– 266, May 2012. 
[5]
Dresner K., Stone P. Multiagent traffic management: A reservation-based 
intersection control mechanism, Proceedings of the Third International 
Joint Conference on Autonomous Agents and Multiagent Systems-
Volume 2. IEEE Computer Society, 2004: 530-537. 
[6]
Dresner K. M., Stone P. A. Multiagent Approach to Autonomous 
Intersection Management. J. Artif. Intell. Res.(JAIR), 2008, 31: 591-656. 
[7]
Ferreira M., Fernandes R., Conceição H, et al. Self-organized traffic 
control[C]. Proceedings of the seventh ACM international workshop on 
VehiculAr InterNETworking. ACM, 2010: 85-90. 
[8]
Ferreira M., d'Orey P. M. On the impact of virtual traffic lights on carbon 
emissions mitigation. Intelligent Transportation Systems, IEEE 
Transactions on, 2012, 13(1): 284-295. 
[9]
Gokulan B. P., Srinivasan D. Distributed geometric fuzzy multiagent 
urban traffic signal control. Intelligent Transportation Systems, IEEE 
Transactions on, 2010, 11(3): 714-727. 
[10]
Glaser S., Vanholme B., Mammar S, et al. Maneuver-based trajectory 
planning for highly autonomous vehicles on real road with traffic and 
driver interaction. Intelligent Transportation Systems, IEEE Transactions 
on, 2010, 11(3): 589-606. 
[11]
Gradinescu V., Gorgorin C., Diaconescu R, et al. Adaptive traffic lights 
using car-to-car communication[C]. Vehicular Technology Conference, 
2007. VTC2007-Spring. IEEE 65th. IEEE, 2007: 21-25. 
[12]
Hartenstein H., Laberteaux K. P. A tutorial survey on vehicular ad hoc 
networks. Communications Magazine, IEEE, 2008, 46(6): 164-171. 
[13]
Hunt P. B., Robertson D I, Bretherton R D, et al. The SCOOT on-line 
traffic signal optimisation technique. Traffic Engineering & Control, 1982, 
23(4). 
[14]
Lee J., Park B. Development and evaluation of a cooperative vehicle 
intersection control algorithm under the connected vehicles environment. 
Intelligent Transportation Systems, IEEE Transactions on, 2012, 13(1): 
81-90. 
[15]
Lertworawanich P., Kuwahara M., Miska M. A new multiobjective signal 
optimization for oversaturated networks. Intelligent Transportation 
Systems, IEEE Transactions on, 2011, 12(4): 967-976. 
[16]
Li C., Shimamoto S. An Open Traffic Light Control Model for Reducing 
Vehicles' Emissions Based on ETC Vehicles. Vehicular Technology, 
IEEE Transactions on, 2012, 61(1): 97-110. 
[17]
Lin C. T., Lee C. S. G.. Neural-network-based fuzzy logic control and 
decision system. Computers, IEEE Transactions on, 1991, 40(12): 1320-
1336. 
[18]
Lin C. T., Lee C. S. G. Reinforcement structure/parameter learning for 
neural-network-based fuzzy logic control systems. Fuzzy Systems, IEEE 
Transactions on, 1994, 2(1): 46-63. 
[19]
Lin W. H., Wang C. An enhanced 0-1 mixed-integer LP formulation for 
traffic signal control. Intelligent Transportation Systems, IEEE 
Transactions on, 2004, 5(4): 238-245. 
[20]
Maslekar N., Boussedjra M., Mouzna J., et al. VANET based adaptive 
traffic signal control, Vehicular Technology Conference (VTC Spring), 
2011 IEEE 73rd. IEEE, 2011: 1-5. 
[21]
Milanés V., Pérez J., Onieva E., et al. Controller for urban intersections 
based on wireless communications and fuzzy logic. Intelligent 
Transportation Systems, IEEE Transactions on, 2010, 11(1): 243-248. 
[22]
Milanés V., Villagra J., Godoy J., et al. An intelligent V2I-based traffic 
management system. Intelligent Transportation Systems, IEEE 
Transactions on, 2012, 13(1): 49-58. 
[23]
Prashanth L. A., Bhatnagar S. Reinforcement learning with function 
approximation for traffic signal control. Intelligent Transportation 
Systems, IEEE Transactions on, 2011, 12(2): 412-421. 
[24]
Priemer C., Friedrich B. A decentralized adaptive traffic signal control 
using V2I communication data, Intelligent Transportation Systems, 2009. 
ITSC'09. 12th International IEEE Conference on. IEEE, 2009: 1-6. 
[25]
Qiao J., Yang N., Gao J. Two-stage fuzzy logic controller for signalized 
intersection. Systems, Man and Cybernetics, Part A: Systems and Humans, 
IEEE Transactions on, 2011, 41(1): 178-184. 
[26]
Sims A. G., Dobinson K. W. The Sydney coordinated adaptive traffic 
(SCAT) system philosophy and benefits. Vehicular Technology, IEEE 
Transactions on, 1980, 29(2): 130-137. 
[27]
Srinivasan D., Choy M. C., Cheu R. L. Neural networks for real-time 
traffic signal control. Intelligent Transportation Systems, IEEE 
Transactions on, 2006, 7(3): 261-272. 
[28]
Vasirani M., Ossowski S. A computational market for distributed control 
of urban road traffic systems. Intelligent Transportation Systems, IEEE 
Transactions on, 2011, 12(2): 313-321. 
[29]
Wang F. Y. Parallel control and management for intelligent transportation 
systems: 
Concepts, 
architectures, 
and 
applications. 
Intelligent 
Transportation Systems, IEEE Transactions on, 2010, 11(3): 630-638. 
[30]
Wu H., Yan G., Xu L. D. Developing Vehicular Data Cloud Services in 
the IoT Environment, Industrial Informatics, IEEE Transactions on, 10(2): 
1587 – 1595, May 2014. 
[31]
Wu W., Zhang J., Luo A., Cao J. Distributed Mutual Exclusion 
Algorithms for Intersection Traffic Control, Parallel and Distributed 
Systems, IEEE Transactions on, 2015, Jan., 26(1): 65-74. 
[32]
Wunderlich R., Liu C., Elhanany I., et al. A novel signal-scheduling 
algorithm with quality-of-service provisioning for an isolated intersection. 
Intelligent Transportation Systems, IEEE Transactions on, 2008, 9(3): 
536-547. 
[33]
Zhang G., Wang Y. Optimizing minimum and maximum green time 
settings for traffic actuated control at isolated intersections. Intelligent 
Transportation Systems, IEEE Transactions on, 2011, 12(1): 164-173. 
[34]
Zhao D., Dai Y. Zhang Z. Computational intelligence in urban traffic 
signal control: a survey. Systems, Man, and Cybernetics, Part C: 
Applications and Reviews, IEEE Transactions on, 2012, 42(4): 485-494. 
[35]
Zhou B., Cao J., Zeng X., et al. Adaptive traffic light control in wireless 
sensor network-based intelligent transportation system, Vehicular 
Technology Conference Fall (VTC 2010-Fall), 2010 IEEE 72nd. IEEE, 
2010: 1-5.