Its report on "Traffic Prediction for Intelligent Transportation System using Machine Learning" download: https://www.entireweb.com/?a=62d7ce5d4d964

1. A SEMINAR REPORT ON
TRAFFIC PREDICTION FOR INTELLIGENT TRANSPORTATION
SYSTEM USING MACHINE LEARNING
SUBMITTED TO THE SAVITRIBAI PHULE PUNE UNIVERSITY, PUNE
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE AWARD OF
THIRD YEAR COMPUTER
ENGINEERING SUBMITTED BY
OM DNYANOBA SURYAWANSHI
UNDER THE GUIDANCE OF
Prof. Grishma Bobhate
DEPARTMENT OF COMPUTER ENGINEERING
RMD SINHGAD SCHOOL OF ENGINEERING, WARJE, PUNE
2021-22
AFFILIATED TO

2. CERTIFICATE
This is to certify that the seminar report entitles
” Traffic Prediction for Intelligent Transportation System using
Machine Learning”
submitted by
Mr. Om Dnyanoba
Suryawanshi Roll No :- 56
is a bonafide work carried out by above student under the supervision of Prof. Grishma Bobhate
and it is approved for the partial fulfillment of the requirement of Savitribai Phule Pune
University, Pune for the award of the degree of Bachelor of Engineering(Computer Engineering).
Prof. Grishma Bobhate Ms. Vina M. Lomate
Seminar Guide Head of Department
Computer Engineering Computer Engineering
Dr. V.V. Dixit
RMD SINHGAD SCHOOL OF ENGINEERING, WARJE, PUNE
Examiner Name & Sign :
Place: Pune

3. ii
ACKNOWLEDGEMENT
With due respect and gratitude I take the opportunity to thank those who have helped me directly
and indirectly. I convey my sincere thanks to Ms. Vina M. Lomate, HOD Computer Dept. and
Prof. Grishma Bobhate for their help in selecting the seminar topic and support.
I thank to my seminar guide Prof. Grishma Bobhate for her guidance, timely help andvaluable
suggestions without which this seminar would not have been possible. Her direction has always
been encouraging as well as inspiring for me. Attempts have been made to minimize the errors in
the report.
I would also like to express my appreciation and thanks to all my friends who knowingly or
unknowingly have assisted and encourage me throughout my hard work.
MR.OM DNYANOBA SURYAWANSHI (ROLL.NO : 56)
T.E COMPUTER(III year) 2021

4. iii
ABSTRACT
The aim is to develop a tool for predicting accurate and timely traffic flow Information. Traffic
Environment involves everything that can affect the traffic flowing on the road, whether it’s
traffic signals, accidents, rallies, even repairing of roads that can cause a jam. If we have prior
information which is very near approximate about all the above and many more daily life
situations which can affect traffic then, a driver or rider can make an informed decision. Also, it
helps in the future of autonomous vehicles. In the current decades, traffic data have been
generating exponentially, and we have moved towards the big data concepts for transportation.
Available prediction methods for traffic flow use some traffic prediction models and are still
unsatisfactory to handle real-world applications. This fact inspired us to work on the traffic flow
forecast problem build on the traffic data and models.It is cumbersome to forecast the traffic
flow accurately because the data available for the transportation system is insanely huge. In this
work, we planned to use machine learning, genetic, soft computing, and deep learning algorithms
to analyse the big-data for the transportation system with much-reduced complexity. Also, Image
Processing algorithms are involved in traffic sign recognition, which eventually helps for the
right training of autonomous vehicles.

5. iv
List of Figures
Fig 1.1: Machine Learning
Fig 1.2: ITS Depiction
Fig 3: Traffic Jam
Fig 5.1.1: System Architechture
Fig 5.1.2: Process
Fig 5.2.1: Support Vector Machine
Fig 5.2.2: Random Forest
Fig 5.2.3: Support Vector Regression
Fig 5.2.4:Decision Tree
Fig 5.2.5 :Evaluation

6. v
Contents
ACKNOWLEDGEMENT
………………………………………………………. ii
ABSTRACT
………………………………………………………………………... iii
List of
Figures…………………………………………………………………….
. iv
1. INTRODUCTION 1
1.1 MACHINE
LEARNING………………………………………………… 1
1.2 INTELLIGENT TRANSPORTATION SYSTEM (ITS)…………2
1.3 NEED FOR ITS !
…………………………………………………………. 3
2. OBJECTIVES 4
3. MOTIVATION 4
4. LITERATURE SURVEY 5
5. METHODOLOGY 7
5.1 SYSTEM DESIGN
………………………………………………………. 7
5.2 PREDICTION ALGORITHS
……………………………………….. 9
5.3 Proposed algorithm for predicting the traffic congestion........... 13

7. vi
6. FUTURE SCOPE 15
7. CONCLUSION 15
8. REFERENCES 16

8. 1
INTRODUCTION
1.1 MACHINE LEARNING
Machine learning (ML) is the study of computer algorithms that can improve automatically
through experience and by the use of data. It is seen as a part of artificial intelligence. Machine
learning algorithms build a model based on sample data, known as training data, in order to
make predictions or decisions without being explicitly programmed to do so. Machine learning
algorithms are used in a wide variety of applications, such as in medicine, email filtering, speech
recognition, and computer vision, where it is difficult or unfeasible to develop conventional
algorithms to perform the needed tasks.
A subset of machine learning is closely related to computational statistics, which focuses on
making predictions using computers; but not all machine learning is statistical learning. The
study of mathematical optimization delivers methods, theory and application domains to the
field of machine learning. Data mining is a related field of study, focusing on
exploratory data analysis through unsupervised learning.Some implementations of machine
learning use data and neural networks in a way that mimics the working of a biological brain.In
its application across business problems, machine learning is also referred to as predictive
analytics.
Fig 1.1: Machine Learning

9. 2
1.2 INTELLIGENT TRANSPORTATION SYSTEM
An intelligent transportation system (ITS) is an advanced application which aims to provide
innovative services relating to different modes of transport and traffic management and enable
users to be better informed and make safer, more coordinated, and 'smarter' use of transport
networks.
Some of these technologies include calling for emergency services when an accident occurs,
using cameras to enforce traffic laws or signs that mark speed limit changes depending on
conditions.
Although ITS may refer to all modes of transport, the directive of the European Union
2010/40/EU, made on July 7, 2010, defined ITS as systems in which information and
communication technologies are applied in the field of road transport, including infrastructure,
vehicles and users, and in traffic management and mobility management, as well as for interfaces
with other modes of transport.[1] ITS may improve the efficiency and safety of transport in a
number of situations, i.e. road transport, traffic management, mobility, etc.[2] ITS technology is
being adopted across the world to increase capacity of busy roads and reduce journey times
Fig 1.2: ITS Depiction

10. 3
1.3 NEED FOR ITS
Various Business sectors and government agencies and individual travellers require precise and
appropriately traffic flow information. It helps the riders and drivers to make better travel
judgement to alleviate traffic congestion, improve traffic operation efficiency, and reduce carbon
emissions. The development and deployment of Intelligent Transportation System (ITSs) provide
better accuracy for Traffic flow prediction. It is deal with as a crucial element for the success of
advanced traffic management systems, advanced public transportation systems, and traveller
information systems. [1]. The dependency of traffic flow is dependent on real-time traffic and
historical data collected from various sensor sources, including inductive loops, radars, cameras,
mobile Global Positioning System, crowd sourcing, social media. Traffic data is exploding due
to the vast use of traditional sensors and new technologies, and we have entered the era of a large
volume of data transportation. Transportation control and management are now becoming more
data-driven. [2], [3].However, there are already lots of traffic flow prediction systems and
models; most of them use shallow traffic models and are still somewhat failing due to the
enormous dataset dimension. Recently, deep learning concepts attract many persons involving
academicians and industrialist due to their ability to deal with classification problems,
understanding of natural language, dimensionality reduction, detection of objects, motion
modelling. DL uses multi-layer concepts of neural networks to mining the inherent properties in
data from the lowest level to the highest level [4]. They can identify massive volumes of
structure in the data, which eventually helps us to visualize and make meaningful inferences
from the data. Most of the ITS departments and researches in this area are also concerned about
developing an autonomous vehicle, which can make transportation systems much economical
and reduce the risk of lives. Also, saving time is the integrative benefit of this idea. In current
decades the lots of attention have made towards the safe automatic driving. It is necessary that
the information will be provided in time through driver assistance system (DAS), autonomous
vehicles (AV)and Traffic Sign Recognition (TSR) [5].

11. 4
OBJECTIVES
 The aim is to research different machine learning algorithms capable of producing
accurate traffic flow prediction.
 Which technique is best ?
 What is proposed method ?
MOTIVATION
India is a country of huge population. The Road traffic in all cities of India is of greater concern.
There is always a long wait for the people on the roades of the cities. India is among the top
countries with large traffic index in the worldand, it is also 4th among the traffic index rankings
of 2019 [3]. With high time index and also the C02 (Carbon di oxide) percent among all the
cities [3]. So it is important to find effective solutions through ML to solve traffic problem.
Fig 3: Traffic Jam

12. 5
LITERATURE SURVEY
PAPER NAME - Traffic Prediction for Intelligent Transportation System using Machine
Learning
AUTHORS - Gaurav Meena,Deepanjali Sharma, Mehul Mahrishi.
PUBLICATION - IEEE(2020)
INFERENCE - ITS provides a smooth and safe movement of road transportation. Decision
Tree ,Random forest and SVM algorithm are used to identify classification and regression .
PAPER NAME- Smart Traffic Analysis using Machine Learning
AUTHORS- Aditya Krishna K.V.S, Abhishek K, Allam Swaraj, Shantala Devi Patil, Gopala
Krishna Shyam
PUBLICATION – IJEAT(2019)
INFERENCE- Analysis using Random Forest Algorithm, predicting the Mean square
error(MSE), calculate Mean Absolute error(MAE) which means the difference between two
continuous variables may be X and Y, also calculating the Root mean squared error(RMSE)
which means the frequently used measure of difference in the values predicted by th machine
learning model.
PAPER NAME- Parallel Control and Management for Intelligent Transportation Systems:
Concepts, Architectures, and Applications
AUTHORS- Fei-Yue Wang
PUBLICATION – IEEE(2011)
INFERENCE- ACP-based parallel control and management systems And use of ATS(Artificial
Transportation System) in it.Studied about 5 components of ATS and also System Architecture
of PTMS.

13. 6
PAPER NAME- A Decentralized Approach for Anticipatory Vehicle Routing Using Delegate
Multiagent Systems
AUTHORS- Rutger Claes, Tom Holvoet, and Danny Weyns
PUBLICATION- IEEE(2011)
INFERENCE – This paper presents a decentralized approach for anticipatory vehicle routing
that is particularly useful in large-scale dynamic environments. The approach is based on
delegate multiagent systems.
PAPER NAME- Index point detection and semantic indexing of videos - a comparative review
AUTHORS- Mehul Mahrishi and Sudha Morwal
PUBLICATION-IEEE(2021)
INFERENCE- To study the existing methods of automatic video indexing and annotation to
analyze the outcomes and gaps by Use of YoloV4.
PAPER NAME- Decision tree methods: applications for classification and prediction
AUTHORS- Yan-yan SONG, Ying LU
PUBLICATION- Shanghai Archives of Psychiatry(2015)
INFERENCE- Decision tree methodology is a commonly used data mining method for
establishing classification systems based on multiple covariates or for developing prediction
algorithms for a target variable.

14. 7
METHODOLOGY
5.1 SYSTEM ARCHITECTURE
Fig 5.1.1: System Architechture
1) Dataset Generation:
The dataset for this project is generated based on available datasets for traffic analysis. The
dataset is created for a particular location in bangalore called Yelahanka For easy understanding.
The dataset will be in the form of a .csv file
1) Dataset Generation
5) Verification
2) Feature
Identification
4)Machine Learning
Algorithm used for
Analysis
3) Feature
Extraction

15. 8
2) Feature Identification:
The neccessary features for the project are to be identified like time, distance, delay, Vehicle
Number etc. The features which are associated with the project are identified for the dataset by
using which the analysis could be easily performed
3) Feature Extraction:
Feature extraction will in general make use of the dimensionality reduction procedure to reduce
and consider only those neccessary attributes neccessary for the project like time
,distances, Nodes between which the traffic in general is identified.
4) Machine Learning Algorithm used for Analysis:
The Machine learning algorithm that is used for the traffic analysis we have used for our TAM
algorithm is the Algorithm.The Algorithm will help in classifying whether the traffic is more or
less in a particular area based on the dataset loaded to the algorithm
5) Verification:
The Verification step will check whether the analysis done on the dataset is proper or not. This
means that the analysis step is giving the proper result or not.
Fig 5.1.2: Process

16. 9
5.2 PREDICTION ALGORITHS :-
1) Support Vector Machine (SVM)
Fig 5.2.1: Support Vector Machine
Support Vector Machine or SVM is one of the most popular Supervised Learning algorithms,
which is used for Classification as well as Regression problems. However, primarily, it is used
for Classification problems in Machine Learning.
The goal of the SVM algorithm is to create the best line or decision boundary that can segregate
n-dimensional space into classes so that we can easily put the new data point in the correct
category in the future. This best decision boundary is called a hyperplane.
SVM chooses the extreme points/vectors that help in creating the hyperplane. These extreme
cases are called as support vectors, and hence algorithm is termed as Support Vector Machine.
Consider the below diagram in which there are two different categories that are classified using a
decision boundary or hyperplane:

17. 10
2) Random Forest
Fig 5.2.2: Random Forest
Random Forest is a popular machine learning algorithm that belongs to the supervised learning
technique. It can be used for both Classification and Regression problems in ML. It is based on
the concept of ensemble learning, which is a process of combining multiple classifiers to solve a
complex problem and to improve the performance of the model.
As the name suggests, "Random Forest is a classifier that contains a number of decision
trees on various subsetsof thegiven dataset andtakestheaverage to improve the
predictive accuracy of that dataset." Instead of relying on one decision tree, the random
forest takes the prediction from each tree and based on the majority votes of predictions, and it
predicts the final output.
The greater number of trees in the forest leads to higher accuracy and prevents the
problem of overfitting.

18. 11
3) Support Vector Regression (SVR)
Fig 5.2.3: Support Vector Regression
Support Vector Regression is a supervised learning algorithm that is used to predict discrete values.
Support Vector Regression uses the same principle as the SVMs. The basic idea behind SVR is to
find the best fit line. In SVR, the best fit line is the hyperplane that has the maximum number of
points.
Unlike other Regression models that try to minimize the error between the real and predicted
value, the SVR tries to fit the best line within a threshold value. The threshold value is the
distance between the hyperplane and boundary line. The fit time complexity of SVR is more
than quadratic with the number of samples which makes it hard to scale to datasets with more
than a couple of 10000 samples.

19. 12
4) Decision Tree
Fig 5.2.4:Decision Tree
A decision tree is a decision support tool that uses a tree-like model of decisions and their
possible consequences, including event change outcomes, resource costs, and utility. It is one
way to display an algorithm that only contains conditional control statements.
Decision trees are commonly used in operations research, specifically in decision analysis, to
help identify a strategy most likely to reach a goal, but are also a popular tool in Machine
Learning.
Decision tree methodology is a commonly used data mining method for establishing
classification systems based on multiple covariates or for developing prediction algorithms for a
target variable. This method classifies a population into branch-like segments that construct an
inverted tree with a root node, internal nodes, and leaf nodes. The algorithmis non-parametric
and can efficiently deal with large, complicated datasets without imposing a complicated
parametric structure. When the sample size is large enough, study data can be divided into
training and validation datasets. Using the training dataset to build a decision tree model and a
validation dataset to decide on the appropriate tree size needed to achieve the optimal final
model.

20. 13
Fig 5.2.5 :Evaluation
5.3 Proposed algorithm for predicting the traffic
congestion which can be seen below:
Step 1: For identifying the congested situation :
 Collect the traffic data in every 5 min with various features
 Group every 5 min interval with their corresponding data
 Calculate the distance between each vehicle with all another vehicles within
specified junction.
 if the distance is less than the specific threshold between two vehicles then
those vehicles are considered to be the neighbourhood vehicles
else
Not considered as neighbour vehicles.
end if

21. 14
Step 2: For classifying the congested situation
 This will eventually give us the matrix A.
 Now assign 1 to A[i, j]
 if A[i, j] < threshold then
A[i, j] = 1
else
A[i, j] = 0
endif
 Count A[i, j]=1 and label i, j as neighbourhood vehicles
 Repeat above steps in every 5 min for 45 min
 Plot the graph between neighbourhood vehicles and time interval
Step 3 : Evaluation
 if the neighbourhood vehicles shows an increasing graph
then
else
the traffic congestion is identified
No traffic
end if

22. 15
FUTURE WORK
For future work it would be interesting to attempt the same experiments but exam- ine
more complex versions of the models used. For example by using more neurons and
hidden layers in the neural network architectures. Doing this would however re- quire
better hardware, such that the training phase execution time does not become unfeasible.
The hardware in mind would be some high end Graphics Processing Unit as they are well
optimized for matrix operations, which is a large part of training neural networks. Even
better would be to perhaps utilize a new technology released by Google in 2017 called
Tensor Processing Units (TPU) [57]. TPUs were built specifically for training neural
networks, and are available for usage in the Google Cloud [58]. Also we have planned to
integrate the web server and the application.
Also, considering entirely new models would also be an option. One example is to utilize
both CNN and LSTM networks in the same model. This has been tried in several previous
projects with success [59], [60]. This works by using a CNN network to capture the spatial
correlations, and letting the LSTM deal with the temporal dependencies. Another
interesting idea was proposed by Ma et al. (2017) [61], where they forecast future traffic
patterns based on images. In other words, they interpret the traffic speed at various
locations of some road network as an image. A CNN network is then used to learn the
patterns of the images.
Finally, for a future project, much more data would be needed as this would allow the
model to learn all the traffic patterns over an entire year. This would likely improve the
results because no matter where the test set is put in time, the training set will at some point
have included similar patterns
CONCLUSION
It is clear that machine learning has great potential when it comes to time series
forecasting. This has been shown in this thesis as well as in other referenced liter- ature.
Existing statistical approaches should however not be underestimated. The baseline
methods did in fact achieve decent results and are faster to evaluate com- pared to the ML
techniques. When faced with a forecasting problem, whether its traffic forecasting or
something else, the traditional approaches should always be tried first. If they do not
perform as well as expected, one could try experimenting with machine learning. If this
option is considered, a few things are important to keep in mind. Powerful hardware is
crucial as this allows one to train very large and complex ML models at fast speeds.
Increased performance due to hardware will in turn open up many doors for further
improvement of the ML models. For one, it
will speed up grid search optimizations which helps finding better hyperparameters.

23. 16
REFERENCES
[1] Fei-Yue Wang et al. Parallel control and management for intelligent transportation systems:
Concepts, architectures, and applications. IEEE Transactions on Intelligent Transportation
Systems, 2010.
[2] Yongchang Ma, Mashrur Chowdhury, Mansoureh Jeihani, and Ryan Fries. Accelerated
incident detection across transportation networks using vehicle kinetics and support vector
machine in cooperation with infrastructure agents. IET intelligent transport systems, 4(4):328–
337, 2010.
[3] Rutger Claes, Tom Holvoet, and Danny Weyns. A decentralized approach for anticipatory
vehicle routing using delegate multiagent systems. IEEE Transactions on Intelligent
Transportation Systems, 12(2):364–373, 2011.
[4] Mehul Mahrishi and Sudha Morwal. Index point detection and semantic indexing of videos -
a comparative review. Advances in Intelligent Systems and Computing, Springer, 2020.
[5] Joseph D Crabtree and Nikiforos Stamatiadis. Dedicated short-range communications
technology for freeway incident detection: Performance assessment based on traffic simulation
data. Transportation Research Record, 2000(1):59–69, 2007.
[6] H Qi, RL Cheu, and DH Lee. Freeway incident detection using kinematic data from probe
vehicles. In 9th World Congress on Intelligent Transport SystemsITS America, ITS Japan,
ERTICO (Intelligent Transport Systems and Services-Europe), 2002.
[7] Z. Zhao, W. Chen, X. Wu, P. C. Y. Chen, and J. Liu. Lstm network: a deep learning
approach for short-term traffic forecast. IET Intelligent Transport Systems, 11(2):68–75, 2017.
[8] C. Zhang, P. Patras, and H. Haddadi. Deep learning in mobile and wireless networking: A
survey. IEEE Communications Surveys Tutorials, 21(3):2224–2287, thirdquarter 2019.
[9] Chun-Hsin Wu, Jan-Ming Ho, and D. T. Lee. Travel-time prediction with support vector
regression. IEEE Transactions on Intelligent Transportation Systems, 5(4):276–281, Dec 2004.
[10] Yan-Yan Song and LU Ying. Decision tree methods: applications for classification and
prediction. Shanghai archives of psychiatry, 27(2):130, 2015.
[11] Yiming He, Mashrur Chowdhury, Yongchang Ma, and Pierluigi Pisu. Merging mobility
and energy vision with hybrid electric vehicles and vehicle infrastructure integration. Energy
Policy, 41:599–609, 2012.
[12] Jason Brownlee. Bagging and random forest ensemble algorithms for machine learning.
Machine Learning Algorithms, pages 4–22, 2016.