IRJET- Traffic Prediction Techniques: Comprehensive analysis

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 03 | Mar-2018 www.irjet.net p-ISSN: 2395-0072
© 2018, IRJET | Impact Factor value: 6.171 | ISO 9001:2008 Certified Journal | Page 2494
Traffic Prediction Techniques: Comprehensive analysis
Kawaljit Kaur1, Samandeep Singh2, Megha3
1Student, Dept of Computer Science Engineering, GIMET Amritsar, Punjab, India
2,3Assistant Professor, Dept of Computer Science Engineering, GIMET Amritsar, Punjab ,India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - In the modern era, the road infrastructure failed
to cope up with the exponential increase of road traffic. There
is a thrust to find a smarter ways to deal with such
transportation system. Intelligent Transport System is at the
forefront edge of this, one of the points is exact and hassle free
forecasts that guarantee smooth and bother free driving and
authoritative experience. In such manner, ITS being looked
into for quite a few years and furthermore a field of consistent
growth of works and advancement after some time, there is a
wealth of writing on traffic expectation. Traffic datasets
generated through the application of IoT are operated upon
by the existing techniques. Traffic flow analysisisconductedto
tackle the issues of traffic forecasting. Paper presents a
systematic analysis of previous aggregate work on traffic
prediction, highlight the marked changes and presents future
directions for research work.
Key Words: Traffic prediction, Traffic Dataset, IoT,
Traffic flow, traffic forecasting.
1. INTRODUCTION
Intelligent transportation system technique in electronicor
non electronic forms for producing information through
advanced sensors, computers and communication
technology that improve processof traffic forecasting.ITSis
wide field providing assistance in the field of driver
assistance, inter vehicle communication, air traffic control,
road sign prediction, number plate detection, congestion
control, dynamic routing etc. ITS caters to the
multidimensional needs of traffic management overlapped
with number plate detection and road traffic signal
prediction[1].Most of issues of traffic prediction are caused
due to existing infrastructure however someoftheissuesare
also caused by poor management of traffic flow and
congestion control[2].ITS tackles the issue of poor
management of traffic flow by the use of accurate traffic
monitoring and control strategies. The distributed and
shared judgment and care management has be remolded an
open issue at all levels of traffic forecasting systems. For the
estimation of traffic prediction it requires the information
that is simple and diverse from the sensors and skills[3].To
work efficiently there should be a ITS softwaresysteminthis
environment. But this system also requires credible and
timely information to ensure that software can work
securely and produce results within specified time.
Computer systems make the interactionbetweenhumanand
computational devices very natural so that users can get
desired data in a transparent manner. The newly introduced
gadgets like mobiles, PDAs, laptops etc. make every
information available anywhere at any time. By using ITS,
interactive feedback loops and video games, we can analyze
the traffic related behavior changes that may occur. ITS is
associated with many applications and in long term it is
viable to get feasible into larger frameworks in health
care.[4]According to researchers it is suggested that use of
ITS and emergence in technology is efficient enough to
aware usersabout the current traffic and providepreventive
measures. The ITS also enable user for behavior change.
Distinct elementsof ITS are enhancementindecisionmaking
anw objective oriented. Diverting the traffic greatly depend
upon the awarenessof driver which will be accomplishedby
the use of ITS. Routing adherence is greatly impacted by this
mechanism. with the help of transportation system drivers
can analyze his behavior and prepare himself for taking
appropriate action[4].
2 .Related work
To tackle the requirements of systematic review,
background analysis is conducted. The background analysis
present the existing techniques that are comprehensively
used to predict on road traffic. O.U.Chinyere et al. discussed
examination encounters of building a keen framework to
screen and control street traffic in a Nigerian city. A half and
half approach got by the intersection of the Structured
SystemsAnalysis and Design Methodology (SSADM) andthe
Fuzzy-Logic based Design Methodology was conveyed to
create and actualize the framework. Issues were related to
present traffic control framework at the '+' intersectionsand
this required the plan and usage of another framework to
take care of the issues. The subsequent fluffy logic based
framework for activity control was recreated and tried
utilizing a prominent crossing point in a Nigerian city;
infamous for extreme activity logjam. The new framework
dispensed with a portion of the issues distinguished in the
current activity checking and control frameworks. Traffic
flag controller is playing increasingly and more criticalparts
in present day administration and control of urban traffic.
C.Xiao-feng et al. introduces a shrewd traffic flag controller
in light of multi-microcomputer innovation.Thearchitecture
and crucial elements of the clever traffic flag controller U
initially presented in detail, at that point the human-PC
interface in light of visual innovation intended for the
controller is figured, and lastly an application case by andby
is talked about[6].L.Kdqj discussed propelled activity data
benefit framework not just give opportune and precise
traffic data for activity administration work force who can
adequately adjust the traffic administration control
framework to an assortment of traffic conditions and street

arrange limit, yet in addition help street clients, viably
staying away from roads turned parking lots, diminishing
auto collisions. Notwithstanding, the existing dynamic
activity data is discharged for generally group of onlookers.
On the off chance that the majority of the drivers utilize the
dynamic traffic data to design ongoing travel courses, atthat
point the in general activity framework might be bothered
generally, and another road turned parking lot appear inthe
meantime maintaining a strategic distance from the current
activity stick. In light of the GIS spatial information
demonstrate and the hypothesis of multi-operator, we
ponder a dynamic activity data administrations innovation
in view of collective multi-specialist techniques all together
to show signs of improvement travel waythroughupgrading
the communication what'smore, coordinatedeffortbetween
the data suppliers and voyagers. At that point the test model
framework is outlined what's more, created in view of the
swarm stage and java language, and some analysis data is
produced by the prototype system[7].B.Singh and A.Gupta
dealswith the expanding activity i.e majorissueeverywhere
throughout the world. Wise Transportation System (ITS)
answers these issues with the assistance of new
advancements. ITS is an incorporated framework that
executes an expansive scope of correspondence, control,
vehicle detecting and hardware advances to take care ofand
deal with the traffic issues. ITS is being utilized as a part of
the created nations since past two decades, however it is as
yet another idea when creating nations like India, Brazil,
China, South Africa and so on is concerned. In the present
examination we have considered four noteworthy parts of
the ITS i.e., Advanced Traveller Data System (ATIS),
Advanced Traffic Management System (ATMS), Advanced
Public Transportation System (APTS), and Emergency
Management System (EMS). Target of the paper isto ponder
different ITS engineering and model and audit such models
to get top to bottom of their design. Subsequently
engineering and created modelsthroughoutthetimesoffour
noteworthy branches of ITS have been inspected here to
make an examination investigation of various models that
have been produced by the scientists in their examinations.
It will prompt the holes in the information which can be
additionally considered. The paper featuresthe conclusions
extricated from the investigations of various frameworks
and furthermore gives what's to come scope in the field of
ITS to make it more easy to use and open.[8] H.O.Al-sakran
suggested As of late notoriety of private autos is getting
urban activity more swarmed. As result traffic is getting to
be plainly one of vital issues in huge urban areas in
everywhere throughout the world. A portion of the activity
concerns are clogs and mischance which have caused a
colossal exercise in futility, property harm and ecological
contamination. This exploration paper introduces a novel
smart activity organization framework, in viewofInternetof
Things, which is included by ease, high adaptability, high
similarity, simple to redesign, to supplant conventional
traffic administration framework and the proposed
framework can enhance street activity hugely. The Internet
of Things depends on the Internet, organize remote
detecting and discovery advances to understand the canny
acknowledgment on the labelled activity protest, following,
observing, overseeing and handlednaturally.Paperproposes
a design that coordinates web of things with operator
innovation into a solitary stage where the specialist
innovation handles successful correspondence and
interfaces among countless exceptionally dispersed and
decentralized gadgets inside the IoT. Thedesignpresentsthe
utilization of a dynamic radio-recurrence distinguishing
proof (RFID), remote sensor advances, question specially
appointed systems administration, and Internet-based data
frameworks in which labelled activity items can be
consequently spoken to, followed, and questioned over a
system. This examination shows a review of a structure
conveyed traffic reproduction display inside NetLogo, an
operator based condition, for IoT activity checking
framework utilizing versatile specialist
innovation.[9]T.Osman et al. discussed incorporatestheplan
and usage of a clever and robotized activity control
framework which takes points of interest of PC vision and
picture handling systems. Alongside regular PC vision
strategies; this paper presents two new techniques which
has low preparing cost. One of the techniques has been
developed with the assistance of equipment what's more;
the other one is outlined without equipmentbolster.Thisisa
finish activity administration framework which has
possessed the capacity to decrease roadsturnedparkinglots
and clog on re-enacted condition. It distinguishes the
quantity of vehicles on every street and relying upon the
vehicles stack on every street, this framework allots
improved sum of holding up time (red flaglight)andrunning
time (green flag light). This framework is completely
robotized framework that can supplant the regular pre-
decided settled time based activity framework with a
progressively oversaw activity framework. It can likewise
distinguish vehicle condition on street and auto-change the
framework as indicated by the changing street conditions
which makes the framework insightful. The composed
framework can help tackling traffic issuesin occupiedurban
communities to an awesome degree by sparing a lot of
worker hours that get lost attending to stuck streets. This
examination concentrateson factors, ease picturepreparing
and activity stack adjusting.[10]Y.Wang etal. Asindicatedby
city open travel issue trademark, the fundamental body of a
paper has been submitted and has worked out one sort of in
view of the Internet of things outline intelligent
transportation framework. That framework gathers
information by vehicle terminal and transfersinformationto
the server through the system and makes information
obviousto the purchaser passing an algorithm inthe server.
One viewpoint, the customer may ask about open travel
vehicle data by Web. On another viewpoint, the shopper can
know open travel vehicle data by station terminal. The
investigations have tried that the intelligent transportation
framework can offer open travel vehicle data to numerous
shoppers with helpful way along these lines this framework
can take care of the city mass travel issue.[11] X.Yu
concentrated on the fundamental structure of canny urban
Traffic Management System Based on Cloud Figuring and
Internet of Things, proposed the design of canny urban

Traffic Management System BasedonDistributedcomputing
and Internet of Things. The paper made a profound research
on the data observing in light of Internet of things,
estimation and the shrewd displayingsegmentswhat'smore,
learning coordinating segment. Mass estimation was
acknowledged by the utilization of the distributed
computing stage. The framework generally understandsthe
shrewd observing what's more, administration of urban
traffic and understands the reason for keen dig of urban
traffic. Traffic management with the implicationofsensorsis
complex and required accuracy. Techniques devised so far
still requires further enhancements for increasing accuracy
of prediction.
2.1 Gaps in literature
Analysis of literature indicates that dataset used is offline
and is not derived with the application of IoT. sensor data
utilization within traffic related application is the prime
cause of interest. Accurate prediction related to traffic to
drivers involved along with direction sensing is missing in
existing literature. Advanced application framework
construction for traffic prediction is the solution for the
problem.
3. Comparison tables
The comparison of various techniques that can be used to predict traffic is listed as under:-
Title Technique Datasets Parameters Merit Demerit
A Consumer
Transceiver for Long-
Range IoT
Communications in
Emergency
Environments[12]
IEEE802.11ahWi-
Fi protocol, Time
Domain Least
Square(TDLS)
-------- Packet Error
Rate(PER),
MSE
Increased range of
service
Time of
execution is
substantially
high
The advantages of IoT
and Cloud applied to
Smart Cities[13]
ClouT
architecture
which is
combination of
cloud and IoT is
discussed
-------- ------- Sensorisation ,
Actuatorisationlayer
along with IoT have
been added in CIaaS
layer to extract data
out of API's
CSaaS layer is
still not
completely
defined.
Combining KNN
Algorithm and Other
Classifiers [18]
KNN, C4.5, SVM
And Naive Bayes
Classifier(KNC)
20 UCI
Datasets
Accuracy for
classsification
Higher accuracy Execution time
not considered
Short-term traffic flow
prediction using
seasonal ARIMA model
with limited input
data[14]
SARIMA 3-Lane
roadway in
Chennai,
India
Flow of
vehicles'
accuracy
through MAPE
More accurate
results even with
data shortage
More time for
computations
Smart Disease
Surveillance Based on
Internet of Things (IoT)
[15]
IoT in the field of
health care
Central
Health
Ministry
Prediction
accuracy
Fast prediction of
patterns of disease,
help to take
measures on time
Inadequate data
managers, low
budget, lack of
technical
advisory group
Optimising Power
Consumption of Wi-Fi
inbuilt IoT Devices[16]
Reduce power
consumption of
Wi-Fi enabled
devices
-------- Power
consumption
of various
processors
Wi-Fi is better than
other technologiesin
terms of range and
security
No parameters
enhancements
are suggested
Spatial and Temporal
Patterns in Large-Scale
Traffic Speed
Prediction[21]
Unsupervised
methods(k-
means, self
organising maps,
principal
component
analysis ) to find
out global trends
Road
network
from
Outram
park to
Changi in
Singapore.
Prediction
accuracy
MSE
Spatial and temporal
trends found which
was not possible
through use of SVM
Need to
incorporate
these found
patterns into
route guiding
algorithms

Data Mining for the
Internet of Things:
Literature Review and
Challenges[17]
Review of various
data mining
techniques and
its applications is
performed
-------- 3 views of
data mining-->
knowledge,
technique,
application
view.
Big data, data mining
are hot topics to
discover deep.
Parameter
optimization is
not considered
An Aggregation
Approach to Short-
Term Traffic Flow
Prediction[24]
Integration of MA,
ES and ARIMA
using NN
National
Highway
107,
Guangzhou,
Guangdong,
China
RMSE, PAE
and MAPE
Accuracy is high Situation
involving
multiple
detectors is
missing
Internet of things:
Vision, applicationsand
research challenges[19]
Review of IoT
along with the
challenges is
discussed.
-------- ------- IoT applications are
described ensuring
its efficient use in
future work
No parameter
enhancement
mechanism is
considered
Utilizing Real-World
Transportation Datafor
Accurate Traffic
Prediction[23]
H-
ARIMA+(Hybrid
model of HAM
and ARIMA)
Los Angeles
County
Transport
Network
MAPE and
RMSE
Short term and Long
term prediction
accuracy better than
ARIMA, ES, NNet
Data from each
sensor is studied
individually.
need for spatial
correlations
between sensors
Smartphone Based
Automatic Abnormality
Detection of Kidney in
Ultrasound Images[20]
Viola Jones
algorithm, SVM,
Genetic algorithm
Ultrasound
imagesfrom
ultrasound
scanner
Prediction
accuracy
Benefits rural
people, can be used
for emergency
Only cyst and
kidney stone is
considered
Smart video
surveillance system for
vehicle Detection and
traffic flow control[22]
Image Processing-
->Background
Subtraction using
Threshhold
Adjusting process
Video
Database
False
Rejection
Rate(FRR),
False
Acceptance
Rate(FAR),
Total Success
Rate(TSR)
Prediction accuracy
is increased by the
use of video
surveillance
Cameras not for
night vision,
situations to
suspect danger
not covered.
Traffic Flow
Forecasting Using
aSpatio-temporal
Bayesian Network
Predictor[25]
BayesianNetwork -------- Accuracy
throughMMSE
Prediction accuracy
is improved since
pre-processing
reduces the impact
of error
No real time
dataset is
considered
Table1: comparison of traffic prediction techniques
4. CONCLUSION AND FUTURE SCOPE
Traffic prediction using the application of fog computing is
critical that can be used to monitor time critical applications
such as preventing road accidents. The relevant information
is required to be transferred to the source so that user who
can be a driver can take appropriate action regarding route
towards the destination is the prime objective of this study.
Dataset derived from sensor will be used to construct real
time traffic prediction framework. Accuracy will be the key
parameter that could be enhanced by the application of
proposed methodology.
REFERENCES
[1] W. Min and L. Wynter, “Real-time road traffic
prediction with spatio-temporal correlations,”
Transp. Res. Part C, vol. 19, no. 4, pp. 606–616, 2011.
[2] S. V. Kumar and L. Vanajakshi, “Short-term traffic
flow prediction using seasonal ARIMA model with
limited input data,” Eur. Transp. Res. Rev., vol. 7, no.
3, pp. 1–9, 2015.
[3] X. Yu, “Intelligent Urban Traffic Management System
Based on Cloud Computing and Internet of Things,”
pp. 2169–2172, 2012.

[4] X. Pang, C. Wang, and G. Huang, “A Short-TermTraffic
Flow Forecasting Method Based on a Three-LayerK-
Nearest Neighbor Non-Parametric Regression
Algorithm,” no. July, pp. 200–206, 2016.
[5] O. U. Chinyere, O. O. Francisca, and O. E. Amano, “D
ESIGN AND S IMULATION OF AN I NTELLIGENT T
RAFFIC,” vol. 1, no. 5, pp. 47–57, 2011.
[6] C. Xiao-feng, S. Zhong-ke, and Z. Kai, “Research on an
Intelligent Traffic Signal Controller,” pp. 884–887,
2003.
[7] L. Kdqj, “An Intelligent Traffic Information Service
System based on Agent and GIS-T,” 2010.
[8] B. Singh and A. Gupta, “Recent trends in intelligent
transportation systems : a review,” vol. 9, no. 2, pp.
30–34, 2015.
[9] H. O. Al-sakran, “Intelligent Traffic Information
System Based on Integration of Internet of Things
and Agent Technology,” vol. 6, no. 2, pp.37–43,2015.
[10] T. Osman, S. S. Psyche, J. M. S. Ferdous, and H. U.
Zaman, “Intelligent Traffic Management System for
Cross Section of Roads Using Computer Vision,”
2017.
[11] Y. Wang and H. Qi, “Research of Intelligent
Transportation System Based on the Internet of
Things Frame,” vol. 2012, no. July, pp. 160–166,
2012.
[12] M. Kim and S. Chang,”A Consumer Transceiver for
Long-Range IoT
Communications in Emergency Environments[ ,” vol.
62, no. 3, pp. 226–234, 2016.
[13] C. U. Scenarios and R. Architecture, “The advantages
of IoT and Cloud applied to Smart Cities,” pp. 325–
332, 2015.
[14] S. Vasantha Kumar and Lelitha Vanajakshi, “Short-
term traffic flow prediction using seasonal ARIMA
model with limited input data,” pp. 1–9, 2016.
[15] A. Mathew, F. A. S. A, H. N. Pooja, and A. Verma,
“Smart Disease Surveillance Based on Internet of
Things ( IoT ),” vol. 4, no. 5, pp. 180–183, 2015.
[16] B. D. Thomas, R. Mcpherson, G. Paul, and J. Irvine,
“Consumption of Wi-Fi for IoT Devices,” no.
September, pp. 92–100, 2016.
[17] F. Chen, P. Deng, J. Wan, D. Zhang, A. V Vasilakos, and
X. Rong, “Data Mining for the Internet of Things :
Literature Review and Challenges,” vol. 2015, no. i,
2015.
[18] Z. Zhou, C. Du, L. Shu, G. Hancke, J. Niu, and H. Ning,
“Combining KNN Algorithm and Other Classifiers ,”
in 2010 IEEE International Conference on Cognitive
Informatics, 2010,pp.800-805.
[19] D. Miorandi, S. Sicari, F. De Pellegrini, and I.
Chlamtac, “Ad Hoc Networks Internet of things :
Vision , applications and research challenges,” Ad
Hoc Networks, vol. 10, no. 7, pp. 1497–1516, 2012.
[20] P. Vaish, R. Bharath, P. Rajalakshmi, and U. B. Desai,
“Smartphone Based Automatic Abnormality
Detection of Kidney in Ultrasound Images,” 2016.
[21] M. T. Asif, J. Dauwels, C. Y. Goh, A. Oran, E. Fathi, M.
Xu, M. M. Dhanya, N. Mitrovic, and P. Jaillet, “Spatial
and Temporal Patterns in Large-Scale Traffic Speed
Prediction.”
[22] A. A. Shafie, M. H. Ali, F. Hafiz, and R. M. Ali, “SMART
VIDEO SURVEILLANCE SYSTEM FOR VEHICLE
DETECTION AND TRAFFIC FLOW CONTROL,” vol. 6,
no. 4, pp. 469–480, 2011.
[23] B. Pan, U. Demiryurek, and C. Shahabi, “Utilizing
Real-World Transportation Data for AccurateTraffic
Prediction.”
[24] M. Tan, S. C. Wong, J. Xu, Z. Guan, and P. Zhang, “An
Aggregation Approach to Short-Term Traffic Flow
Prediction,” vol. 10, no. 1, pp. 60–69, 2009.
[25] S. Sun, C. Zhang, and Y. Zhang, “Traffic Flow
Forecasting Using a Spatio-temporal Bayesian
Network Predictor,” pp. 273–278, 2005.

IRJET- Traffic Prediction Techniques: Comprehensive analysis

More Related Content

What's hot (20)

Similar to IRJET- Traffic Prediction Techniques: Comprehensive analysis (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Traffic Prediction Techniques: Comprehensive analysis