Professional Documents
Culture Documents
04/04/2021 2
INTRODUCTION
• Road accidents are increasing day by day all over the world.
• It also has become the leading cause of death all over the world specially in
developing countries.
• Worldwide the total number of road traffic deaths has reached 1.35 million per year.
• For this Road accidents always has been a major issue for each country.
04/04/2021 3
IMPORTANCE
• Traffic is increasing day by day because the numbers of vehicles are increasing.
• To decrease speed-related accidents the traffic calming devices which are commonly
installed are speed-breakers.
• So detecting the speed-breakers and make them attainable to the drivers will provide
a better driving experience.
04/04/2021 4
PROBLEM DESCRIPTION
So the drivers need to be aware of the speed-breakers within the driver's location.
Much research has been done regarding this particular problem.
04/04/2021 5
LITERATURE REVIEWS
Main Theme
• By a smart-phone based application to alert the driver in advance when the
vehicle is approaching a speed breaker.
• The application constantly monitors the smart-phone accelerometer to detect
previously unknown speed breakers.
Drawbacks
The server does not immediatly add a previously unknown speed-breaker to
the confirmed list.
But instead it waits for other smart-phones to report the same speed-breaker.
**Jain, Mohit, Ajeet Pal Singh, Soshant Bali, and Sanjit Kaul. "Speed-Breaker Early Warning System." In NSDR. 2012.**
04/04/2021 6
LITERATURE REVIEWS
Main Theme
Detection of speed bumps is recognized with the help of image
processing.
Drawbacks
Accuracy rate of detecting different speed bumps is around 68.4
percent.
Uses only stored data.
Works only in daylight.
** W.Devapriya, C.Nelson kennedy Babu, T.Srihari .Advance Driver Assistance System(ADAS)-speed bump detection.2015
iIEEE international conference on computational intelligence and computing research DOI: 10.13140/RG.2.1.2412.3285.**
04/04/2021 7
LITERATURE REVIEWS
Main Theme
In this paper the proposed a vibration based approach for automatic detection of
pothholes and speed-breaker along with their co-ordinates and severity using
androids built in accelerometer..
These events along with their saverity and location co-ordinates are stored in local
database of mobile device.
Drawbacks
These local database is syncronised with the global database as soon as the session
ends.
The syncronization is done in such a way that the previous road surface condition
is replaced by the current one.
They did not describe the methodology of how to calculate distance between
running vehicles and upcoming speed-breakers.
** Rishiwal V, Khan H. Automatic pothole and speed breaker detection using android system. In2016 39th International Convention on Information and
Communication Technology, Electronics and Microelectronics (MIPRO) 2016 May 30 (pp. 1270-1273). IEEE..**
8
OBJECTIVES
By using android device we can detect speed breaker of the current location.
Our methodology is to always update the learnings in the server. To develop
this, the concept which we are using is the Federated Learning.
We are using the Haversine Method which will detect speed breaker nearest to
the current location and the learning will be stored in the local storage.
The system will give a voice sms to the driver of coming speed breaker which
is in the 100m of distance.
The system will give feedback to the central server either after detecting
numbers (n) of speed-breakers or after (t) times of starting the journey.
04/04/2021 9
Federated Learning
The goal is to train a high-quality centralized model with training data distributed
over a large number of clients.
04/04/2021 10
Federated Learning Architecture
04/04/2021 11
EVALUATION PLAN
• By using android device we have collected learning data sets, by these learning
datasets we can derive the general learning sample of a speed-breaker.
• After analyzing the threshold value it will generate the confidence of a speed breaker
of a location.
• After anlyzing the learning obtained from data collected over various experimental
drives we can formulated a threshold value to distinguish between smooth roads and
speed-breakers.
• We are developing an app which will detect the speed breaker while driving and send
to the central server.
04/04/2021 12
04/04/2021 12
PROPOSED METHODOLOGY
Architecture
Central Server
User 1 ………………………..…………
User N
GPS and
Acceleration Vehicular speed
GPS and
Acceleration Vehicular speed
Architecture
Receiving
current
Learnings of User N
Speed- User 1 User 2
breakers
04/04/2021 14
ALGORITHM FOR UPDATING THE LEARNINGS OF A SPEED
BREAKER
At a specific location, L (x,y) which is a speed breaker
where x = latitude value of L
y = longitude value of L
Variables: α, β
Here, α = current learning in the central server at location (L)
β = received new learning experience from a user at location (L)
04/04/2021 17
Z-VALUE FOR GENERATING CONFIDENCE VALUE
Z-Value Confidence-Value
14.00m 16.00m .25
04/04/2021 18
SPEED BREAKER DATA COLLECTION AFTER A TEST
DRIVE 1
Sl. Latitude Longitude Vehicular- Z-Value Confidence-
Speed Value
01 23.790436 90.376650 008.6 14.142294 .25
02 23.790265 90.376712 006.0 17.476656 .50
03 23.789798 90.376898 004.5 15.177095 .25
01 02 03 04/04/2021 19
SPEED BREAKER DATA COLLECTION AFTER A TEST
DRIVE 2
Sl. Latitude Longitude Vehicular- Z-Value Confidence-
Speed Value
01 23.790436 90.376650 012.5 18.539569 .50
02 23.790265 90.376712 003.7 14.326548 .50
03 23.789798 90.376898 005.5 21.184752 .75
01 02 03 04/04/2021 20
PERFOMANCE EVALUATION OF OUR
PROPOSED ALGORTIHM
Direct Updating vs Our proposed algorithm’s updating learning of a speed breaker at location 1, 2, 3
0.8
0.7
0.6 0.55
0.5
0.5
Confidence
0.4
0.4
0.30.25 0.25
0.2
0.1
0
1 2 3
Locations
Direct Update Our Method Direct Update2
04/04/2021 21
Any Question
?
04/04/2021 22
04/04/2021 23