Automatic Train Control System using Wireless Sensor Networks

Outline Motivation Present Signalling Field Study Thesis Contributions Simulation Conclusion References
Automatic Train Control System for Railways
using Wireless Sensor Network
Prakhar Bansal
2011CS29
under the guidance of
Prof. M.M. Gore
Computer Science and Engineering Department
Motilal Nehru National Institute of Technology Allahabad,
Allahabad, India
June 11, 2013
Prakhar Bansal, MNNIT Allahabad 1 / 66
Automatic Train Control System

Table of Contents
1 Motivation
2 Introduction to Present Railway Signalling Architecture
3 Field Study
4 Thesis Contributions
Proposed Architecture
Algorithms
5 Simulation Implementation
TinyOS, nesC and TOSSIM
Sensor Motes and Sensor Boards
Routing Protocols
Simulation Experiences and Results
6 Conclusion and Future Work
7 References

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

Motivation
Why Railways?
Figure: i.) Railways as a Transportation ii.) Frequency of Rail Accidents
iii.) Need of Sustainable Transport Solution

Motivation
Railways as a Transport
Railways as a Passenger Solution
44446 million passengers travel globally/year via railways [1].
6.8% people travel via rail all over the world [1].
24 million people/day travel via rail in India [2].
Japan, China and Russia has high passenger modal split of
29%, 31.7% and 41.1% respectively [3].
Railways as a Carriage Solution
5439 mtk goods is carried via rail globally in 2011 [1].
IR carries 2.8 million tons of freight/day [4].
USA and Russia has rail freight modal share of 88.8% and
67.9% respectively [5].

Motivation
Accidents Rate
Old technology and manual signalling in most countries.
More than 1000 people die per year globally [6].
715 people die and 1118 injured in last 3 years in 49 accidents
in India [2].
Mostly accidents happen due to manual errors in signalling,
lack of visibility, communication faults and derailments [5].
Trains usually run out of schedule and even get canceled in
winter season due to low visibility.

Motivation
Need for Sustainable Transport
Figure: (a) CO2 Emissions [7] Figure: (b) Energy Consumption [7]

Motivation
Need for Sustainable Transport
Table: CO2 Emissions [7]
Rail 48.8 gram/passenger/km
Roads 418 gram/passenger/km
Navigation 200 gram/passenger/km
Aviation 316 gram/passenger/km
Table: CO2 Emissions in India in the period 1998-2009 [8]
Rail 8 million tons
Roads 128 million tons
Navigation 18 million tons
Aviation 4 million tons

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

Present Railway Signalling Architecture
Figure: General Signalling Boards

Figure: Color Light Signals

Figure: Semaphore Signals

Figure: Convergence and Divergence Signals

Figure: Shunting and Repeater Signals

Successfully Deployed WSN Projects
Smart-Grid Project [9]: entire process from generation,
transmission, distribution of electricity to integration of
renewable and alternative energy sources, is handled by
wireless sensors.
Microsoft SensorMap [10]:
100s of mini weather stations deployed in schools throughout
Singapore.
sensor grid, to automatically collect and aggregate the weather
data in real time.
studies correlation between the weather patterns and dengue
fever.
CodeBlue [11]: wireless sensors for medical care.
Ultra-wideband sensing and communication for biomedical
applications [12].

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

Field Study
Interaction with Senior Section Engineer, North Central Railways
Figure: Ghaziabad Train Control
Room c Indian Railways [13]
Figure: Typical Train Control Room
c Indian Railways [13]

Study Findings
Lots of mechanical equipments used [14].
Completely depends on manual expertise.
As traﬃc increasing, needs good computerized managing
solutions.
Route relay interlocking installed only on busy stations.

WSN Suitability to Railways
Huge scope of wsn in railways [15].
Think about station master itself getting incoming train
readings via sensors.
No need for manual signalling.
Train itself asks for clearance to next block head, no need to
stop and wait.
Accurate location information without GPS.

Use of Self-recharging Batteries from Vibrations
Piezoelectric vibrational energy harvester (PZeh) are used.
Generates 40mW on an average with a peak operation of
0.3W, when shaken gently.
Generates 280mW with a peak operation of 2.0W, when
shaken vigorously.
Micaz mote processor consumes 8mA in Active mode and
<15µ A Sleep mode.
Micaz mote radio consumes 19.7mA in Receiving mode,
17.4mA TX, 0dBm, 20µA Idle mode, voltage regular ON and
1µA Sleep mode, voltage regulator OFF.

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

Figure: Train Running Signalling using WSN

Figure: Convergence and Divergence using WSN

Figure: WSN based Interlocking

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

Train Running Algorithms
Algorithms
Algorithm 1: Conﬁguration Phase - Learning the BHs
Algorithm 2: Conﬁguration Phase - Learning the CHs
Algorithm 3: Train Event Detection: Seeking Clearance by
BHs
Algorithm 4: Data Aggregation and Forwarding: Data to CHs
Algorithm 5: Topology Updation and Maintenance

Algorithm 1
Algorithm 1 Configuration Phase: Learning the BHs
BH/Station broadcasts a ‘configuration message’CM with a
BHdistance = 1
Ru is the set of nodes that receive the CM message
for each u ∈ Ru do
i=0
if BHdistanceu > BHdistanceCM and
firstsendingu[BhIDCM ]==true and isBH==false then
nextbhu[i] ← BhIDCM
nexthopu[i] ← NIDCM
BHdistanceu ← BHdistanceCM + 1
NIDCM ← TOS NODE ID
BHdistanceCM ← BHdistanceu
node u broadcast the modified CM msg
firstsendingu[BhIDCM ] ← false
i++
else
node u discards the received CM message
end if
end for

After Algorithm 1
Figure: After Blockhead Conﬁguration

Algorithm 2
Algorithm 2 Conﬁguration Phase: Learning the CHs
Clusterhead broadcasts a Clusterhead Declaration Message
(CDM) with a TTL value
Ru is the set of nodes that receive the CDM message
if TTL = 0 and u ∈ BH then
if CDM− > ID /∈ CHQueueu then
add(CHQueueCH, CDM− > ID)
TTL ← TTL − 1
node u broadcasts modiﬁed CDM message
end if
else
node u discards the received CDM message
end if
end for

After Algorithm 2
Figure: After Clusterhead Conﬁguration

Algorithm 3
Algorithm 3 Train Event Detection and Seeking for Clearance
Ru is the set of nodes that detect train
if TrainDetectedu == true and DoubleLane==true then
if flagu == 0 then
// critical section
send clearance signal
flagu = 1
end if
else
send wait signal
end if
if TrainDetectedu == true and DoubleLane==false and
stationu == true then
if flagu == 0 and flagNextStationu == 0 then
// critical section

Algorithm 3
Algorithm 3 Train Event Detection and Seeking for Clearance
(cont.)
send clearance signal
flagu = 1
flagNextStationu = 1
end if
else
send wait signal
end if
if TrainLeavesu == true and DoubleLane==true then
send clearance signal to neighboring BH
end if
if TrainLeavesu == true and DoubleLane==false and
stationu == true then
send clearance signal to neighboring station
end if
end for

Algorithm 4
Algorithm 4 Data Aggregation and Forwarding
Ru is the set of BH/Station
Each BH/Station periodically sends the list of train to all CHs
in the queue
Packet.msg ← TrainInfou
add(Packet.ID[ ], TOS NODE ID)
//The BH/Station, when receives the list from other
BHs/Stations, it aggregates the data and then forwards it
if Packet Received then
buffer=buffer∪Packet
Packet ← buffer
forwards Packet
end if
end for

Algorithm 5
Algorithm 5 Topology Updation and Maintenance
if train list not received by CH or partial list is received then
restart algorithm 1 and 2
end if

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

TinyOS
TinyOS
Free, open-source, BSD-licensed OS designed for low-power
embedded distributed wireless sensor devices [16].
Developed by University of California, Berkeley, Intel Research
and Crossbow Technology.
Designed to support the concurrency intensive operations
required by networked sensors with minimal hardware
requirements.
Written in nesC programming language.

nesC and TOSSIM
nesC
Network embedded systems C, C optimized to support
components and concurrency [17].
Component based, event driven programming language used
to build application for TinyOS platform.
Components are wired together to run applications on
TinyOS.
Programs = software components (connected statically via
interfaces).
TOSSIM
Simulates entire TinyOS applications [18].
Replaces components with simulation implementations.
2 interfaces: c++ and python.

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

MICAz Mote
Figure: MICAz Sensor Mote c Crossbow Technology, USA [19]

MICAz Mote
2.4 GHz mote for enabling low-power wireless sensor networks.
IEEE 802.15.4 compliant Radio frequency transceiver.
Radio, resistant to RF interference and provides inherent data
security.
Atmel128L, low power microcontroller.
51-pin expansion connector.
High speed (250 Kbps), hardware security (AES-128).

Sensor Boards
MTS400CA
Acceleration: dual-axis acceleration sensor.
Atmospheric pressure: barometric pressure sensor.
Light: ambient light sensor.
Humidity and temperature: relative humidity and temperature
sensor.
MDA100CB
Light: light sensor and photocell.
92 unconnected soldering points.
51-pin connector.

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

Routing Protocols
Collection Tree Protocol
Collecting data from motes.
One or more collection trees is built, each of which is rooted
towards the speciﬁed destination.
When a node has data which needs to be collected, it sends
the data up the tree, and it forwards collection data that
other nodes send to it after aggregating, or suppressing
redundant transmissions.

Routing Protocols
Dissemination
It is used to maintain consistency across the network.
The dissemination service tells nodes when the value changes,
and exchanges packets so it will reach eventual consistency
across the network.
Blip
BLIP, the Berkeley Low-power IP stack, is an implementation
in TinyOS of a number of IP-based protocols.
Internet of things.

Routing Protocols
Tymo
TYMO is the implementation on TinyOS of the DYMO
[Dynamic MANET On-demand] protocol, a point-to-point
routing protocol for MANET.
TYMO, packet format is changed and implemented on top of
the Active Message stack of TinyOS.
Reactive protocol, DYMO does not explicitly store the
network topology.
Nodes compute a unicast route towards the desired
destination only when needed using RREQ and RREP packets.

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

Simulation Experiences
Initial Design I
Figure: Initial Design with 1000 Nodes

Initial Design II
Figure: Design with Clearance Points along Stations but not along
Junctions with 100 Nodes

Design with BHs but not Intermediate Nodes
Figure: Introduction to Block System with 100 Nodes

Design with BHs with Intermediate Nodes
Figure: Revised Block System Architecture with 115 Nodes

Final Architecture: BHs + CHs + Intermediate Motes
Figure: Present Architecture with 125 Nodes

Topology Framework with respect to Allahabad Junction
Figure: Topology Framework with respect to Allahabad Junction

Results with Variable Number of Nodes
Figure: (a) Energy Consumption
with Variable Number of Nodes
Figure: (b) Success Rate with
Variable Number of Nodes

Results with Variable Frequency of Trains across Allahabad Junction
Figure: (a) Energy Consumption
with Variable Frequency of Trains
across Allahabad Junction
Figure: (b) Success Rate with
Variable Frequency of Trains across
Allahabad Junction

Discussions
In our work we used the energy model where the radio
dissipates energy E = 50 nJ/bit to run the transmitter or
receiver circuitry and amp = 100 pJ/bit/m2 for the transmit
ampliﬁer to achieve an acceptable SNR [20].
Simulation maximum duration is 10000 seconds and it runs 8
rounds/set of nodes.
Topology is generated randomly in each run when doing
simulation for variable number of nodes and it is ﬁxed for
simulation across Allahabad junction.
The success rate is currently decreasing as the number of
packets increase in the network. This is due to collisions of
messages. This needs to be improved.

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

Conclusion
This could be a revolution in railway technology.
Trains can run eﬃciently and accurately as any error can be
easily detected.
Trains can run in low visibility as sensors would take care of
this.

Future Work
Integration with Internet of things evolution using Blip
eﬀectively.
Security as false messages can be spread by attackers;
authenticity and conﬁdentiality could be introduced using
cryptographic solutions.

Table of Contents
1 Motivation
3 Field Study
Algorithms
Routing Protocols
7 References

References I
World Bank Data. http://data.worldbank.org/indicator/IS.RRS.TOTL.KM.
[Online; last accessed June 10, 2013].
Anil Kakodkar, E. Sreedharan, N.Vedachalam, “Report of High Level Safety Review Committee,” Ministry
of Railways, Government of India, February, 2012.
Hiroumi Soejima, “Railway Technology in Japan Challenges and Strategies,” Japan Railway and Transport
Review, September, 2003.
Pawan Bansal, “Speech by Railway Minister,” Ministry of Railways, Government of India, February, 2012.
Sam Pitroda, Deepak Parekh and M.S. Verma, “Report of the Expert Group for Modernizaion of Indian
Railways,” Ministry of Railways, Government of India, February 2012.
Amitabh Agarwal, “Human Interface in Railway Safety? A New Dimension,” Ministry of Railways,
Government of India, 2007.
Jean-Pierre Loubinoux, “Keeping Climate Change Solutions on Track The Role of Rail,” International Union
of Railways, March 2012.
Tan Yigitcanlar, Lawrence Fabian and Eddo Coiacetto, “Challenges to Urban Transport Sustainability and
Smart Transport in a Tourist City: The Gold Coast, Australia,” The Open Transportation Journal, 2008.
Murtala Aminu Bamanga, “Wireless Sensor Network Applications in Smart Grid,” University of sussex, 2012.

References II
Suman Nath, Jie Liu, and Feng Zhao, “SensorMap for Wide-Area Sensor Webs.”
http://atom.research.microsoft.com/sensewebv3/sensormap/.
[Online; last accessed on June 10, 2013].
Matt Welsh, “CodeBlue: A Wireless Sensor Network for Medical Care and Disaster Response,” Harvard
University, 2005.
“Ultra-wideband Sensing and Communication for Biomedical Applications.”
http://www.sussex.ac.uk/crg/projects/wsn/uwbsens/, University of Sussex.
Karan Desai, “Akamai Technologies, Inc..” http://www.quora.com/India/
How-does-the-inside-of-control-room-of-Indian-railways-look-like-How-do-the-guys-in-the-control-r
[Online; last accessed on June 10, 2013].
Mr. Alok Sehgal. Senior section engineer, North Central Railways, Allahabad railway station.
Yamato Fukuta, “Possibility of Sensor Network Applying for Railway Signal System,” IEEE conference, 2008.
P. Levis and D. Gay, TinyOS Programming.
New York, NY, USA: Cambridge University Press, 1st ed., 2009.
David Gay, Philip Levis, Matt Welsh and David Culler, “The nesC Language: A Holistic Approach to
Networked Embedded Systems,” 2002.

References III
P. Levis, N. Lee, M. Welsh, and D. Culler, “Tossim: accurate and scalable simulation of entire tinyos
applications,” in Proceedings of the 1st international conference on Embedded networked sensor systems,
SenSys ’03, (New York, NY, USA), pp. 126–137, ACM, 2003.
Crossbow Technology, “Micaz Speciﬁcation.”
www.openautomation.net/uploadsproductos/micaz_datasheet.pdf.
Leandro Aparecido Villas, Azzedine Boukerche and Heitor Soares Ramos,, “DRINA: A Lightweight and
Reliable Routing Approach for In-Network Aggregation in Wireless Sensor Networks,” vol. 62, IEEE
Transactions on Computers, April 2013.

Thankyou
Questions Please.

Automatic Train Control System using Wireless Sensor Networks

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Automatic Train Control System using Wireless Sensor Networks

Similar to Automatic Train Control System using Wireless Sensor Networks (20)

Recently uploaded

Recently uploaded (20)

Automatic Train Control System using Wireless Sensor Networks