YU LIU et al: A STUDY ON THE APPLICATION OF TRAIN TO TRAIN COMMUNICATION IN TRAIN CONTROL ...

A Study on the Application of Train to Train Communication

in Train Control Systems

Yu LIU1, Lei YUAN2 and Peng WANG1

1
National Engineering Research Center Of Rail Transportation Operation And Control System,
Beijing Jiaotong University, Beijing, China
2
State Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University, Beijing, China

Abstract — With the rapid development of GPS and wireless communication, train control technology is becoming the focus of
system-wide optimization and intelligent control. Active train safety protection has become an important development direction.
This paper introduces the background and significance of applying the train to train communication (T2T) into train control
system, and suggests a design scheme and key working procedure of the system. The two trains interval tracing scenario was
chosen to model and simulate the system by colored Petri Nets. For the fail-safe requirement, the fault model was specially
introduced into system modeling to verify the functional correctness of the T2T based train control system. We also analyzed the
packet loss rate with different channel qualities and different information transmission intervals. Our work provides the
theoretical basis and suggestions for real system applications.

Keywords- train- train communication; train control system; CPN modeling

control system framework and working process, described

I. INTRODUCTION
Train control technology has been used to solve the
safety and automation of train operation over the past
decades. However, train control system in widely used is
almost based on train-ground communication, ground
equipment generates movement authority for each train
according to the front track condition, and sends it to
onboard, the onboard equipment calculate the permission
speed and prevent the train over speeding or rear-end
collisions. It is formed a linear information flow of “front
train” ground ”behind train”. With the rapid development
of GPS, sensor and wireless communication technology,
train control technology in USA, Europe and Japan has been
started to focus on system-wide optimization and
intelligentialize[1]-[2]. The new vehicle-vehicle
communication (V2V) technology has provided the new
condition for the onboard active safety protection[3], train
control system based train to train (T2T) communication can
break through the traditional limit that ground transmits the
onboard information, improve the system safety, and
optimize the system structure. Therefore, introducing T2T
communication into train control system is of great
significance and development prospect.
However, the study of T2T communication in train
control area are still at an early schematic studies and testing
system development. JR East in Japan suggested translating
groud controlling into onboard active controlling based on
T2T communication in the next generation train control
system[4]-[5]. Alstom group in France developed a simplified
CBTC train control system based on T2T communication
which has been tested run in Lille[6]-[7].
This paper analyzed the characteristics of T2T
communication in railway, and suggested a T2T based train
the fault-security policy when communication failure occurs.
In this paper, it chose a multi-train movement scenario as an
example, and a built colored Petri nets model of train to train
communication train control system to simulate and analyze
the communication performance and system functions.
Meanwhile, this paper proposed several suggestions for the
establishment of real system.
II. THE DESIGN OF A TRAIN CONTROL SYSTEM
BASED ON TRAIN-TRAIN COMMUNICATION
The meaning of train control system based on T2T
communication is that extending the train-ground
communication to train-train communication, the trains can
directly transmit data without ground equipment. The new
system achieves the redistribution of functionality by
transforming some ground functions into onboard, thus it can
reduce the number of ground equipment and optimize the
system structure. Through T2T communication, the train can
directly identify the front train’s position and determine the
movement authority by itself.
Because of the safety-critical and the specificity of
railway operation, there are some new features of T2T
applying in train control system which include some
challenges as well as convenience. Specific information as
follows:
 The speed of moving communication nodes is more
than 80km/h, and the high speed mobile nodes has higher
requirement of the channel quality.
 The train nodes move regularly, so that we can
predict its movement locus easily. After all the train can only
be one dimensional moving and there won’t be any
crossroad.

DOI 10.5013/ IJSSST.a.17.42.29 29.1 ISSN: 1473-804x online, 1473-8031 print

YU LIU et al: A STUDY ON THE APPLICATION OF TRAIN TO TRAIN COMMUNICATION IN TRAIN CONTROL ...
 The fixed nodes is far more than mobile nodes in
T2T based train control system, thus there is no need to
consider the broadcast storm problem.
 With the powerful computing ability and large
storage capacity, mobile nodes is easier to realize the
functions such as accurate positioning.
According to the above characteristics, the structure of a
T2T based train control system can be suggested shown in
Figure 1. T2T based train control system includes mobile
nodes (trains), database station, relay station, control
center(CTC), etc. The main work of T2T network is
following: realizing the information interaction between train
and train, train and CTC, train and interlocking, providing
safe real-time data for the generation of movement authority,
assuring function correctness when train registering,
initiating, cancelling, normally running, shunting, etc.
Here we choose 2 trains interval tracing scenario to
illustrate the work procedure of T2T based train control
system. (1) A train requests to establish a connection with the
front train. First the information will choose the node that is
Data base
Line data and
DSU
parameters
CBI
Train data
Train2
Figure 1. The structure of a T2T based train control system
Figure 2. The sequence diagram of T2T communication process
DOI 10.5013/ IJSSST.a.17.42.29
closest to the behind train. Then the information will be
transmitted through other fixed nodes according to the
direction of the train moving. The mode of transmission is
Multi-hop. (2) When the front train has gotten the request
from the train behind, it sends agreement to the back train,
the information transmitting process is the same as step 1
except the direction of information transmission is opposite.
(3)After the behind train receive the agreement from the
front train, it confirms the message and ask for the position
of the front train. (4) The front train send its real-time
position, which means entering the data transmission
normally. If there is a communication failure, according to
the fault - security principle, the work process is: (1) the train
takes full service braking until the train stops. (2)The train
gets connection with CTC after the train stops for two
minutes. (3) The train turns to Onsight(OS) mode which
driver is responsible for driving. (4) The train check the
communication condition every moment when it runs in OS
mode. The sequence diagram of T2T communication is
shown in Figure 2.
CTC
Delay station
Train data
Train1
29.2 ISSN: 1473-804x online, 1473-8031 print
YU LIU et al: A STUDY ON THE APPLICATION OF TRAIN TO TRAIN COMMUNICATION IN TRAIN CONTROL ...
III. THE MODELING OF T2T BASED TRAIN
CONTROL SYSTEM
We chose the scenario 2 trains interval tracing to build
the CPN model according to the above design scheme of
T2T based train control system. In this scenario, the
modeling of whole communicate process adopted parting-
module way to do the research and the whole model could be
divided into three parts: sending module modeling, codes
transmitting module modeling and message receiving
module modeling.
When the train get connected with the front train, the
message from the front train need to be send in course. The
steps are listed below. (1) The train data should be copied to
the packet sending module in order to finish the
synchronization of complete information. (2) This module
need to confirm the sending context by checking the
timestamp in real system which is replaced by variable ‘n’ in
this modeling and simulation. Every circulation of sending
and receiving, n will be plus one and waiting for next round.
(3) The front train sends message and the train behind
receives the message. (4)This module need keep record
about the sending and receiving. The CPN model of this part
are shown in Figure 3.
Figure 3. The CPN model of packet sending module
Figure 4. CPN model of message transmitting module
DOI 10.5013/ IJSSST.a.17.42.29
After the message is sent, the fixed nodes will receive,
judge and retransmit it. The steps of this part are following:
(1) The fixed node receive and decode the packet and the
front train information. (2) The fixed node compares ‘n’
which is on behalf of timestamp with preset value ‘k’ in
order to check whether the information is effective. (3)
According to the checking results, change the value of ‘n’
and ‘k’ and send the feedback for the next round sending. (4)
If the receiving message is right, the fixed code store and
forward the message. The CPN model are shown in Figure 4.
After the message packet is transmitted from the fixed
code, the message receiving module of the second train will
receive and save the message, this process is similar with
fixed codes receiving packets. The step is: (1) The second
train receives and decodes the packet to get the front train
information. (2) The second train compares ‘n’ which is on
behalf of timestamp with preset value ‘k’ in order to check
whether the information is effective.(3) According to the
checking results, change the value of ‘n’ and ‘k’ and send the
feedback for the next round sending.(4) Stored the
information. The CPN model of this module is shown in
Figure 5.
When the train data or feedback are lost, ‘n’ and ‘k’ will
remain unchanged. Thus, serial numbers that the front train
receives will be same as before. So, it will ask for message
with the last serial number and retransmission function will
be realized. We can figure out the packet loss probability by
counting for the retransmission times. The CPN model of
this module is shown in Figure 6.
Figure 5. Message receiving module
Figure 6. Message feedback module
29.3 ISSN: 1473-804x online, 1473-8031 print
YU LIU et al: A STUDY ON THE APPLICATION OF TRAIN TO TRAIN COMMUNICATION IN TRAIN CONTROL ...

Figure 7. Fail-safe model

Figure 8. T2T communicate simulation

The most basic principle of railway is fail-safe and it is
necessary to consider what the trains will do when there is
communication failure. There are two links between the two
trains, which are “train – relay station – train” and “train –
fixed codes – train”. The two links are corresponding to
‘link1’ and ‘link2’ in Fig. 10. When the train gets its speed
and position, it will set up these two communication links
with the train ahead. If there is one link gets connected, the
flag bit is set 1 and 0 if both of links are failure. When the
two links are both failure, the system must ensure fail-safe.

DOI 10.5013/ IJSSST.a.17.42.29 29.4 ISSN: 1473-804x online, 1473-8031 print

YU LIU et al: A STUDY ON THE APPLICATION OF TRAIN TO TRAIN COMMUNICATION IN TRAIN CONTROL ...
The way and process of ensuring fail-safe are following: (1)
When the train can’t get the message from the train ahead, it
will take the message that the last time received as EOA for
the front train can never run back. (2) After the train stops, it
will report the condition of it to CTC, such as communicate
status and position. (3) CTC will send the broken-down train
information to other train on this line through relay station.
(4) When the train stops for 2 minutes, it can ask CTC for
OS driving if the condition if satisfied. (5) When the train is
running under OS, it constantly check the connection with
the front train. If the communicate condition is satisfied, it
changes to T2T communication and goes into FS mode
driving. The fail-safe model is shown in Figure 7.
IV. SIMULATION AND ANALYSIS
We combined the modules above and built the overview
model with CPN Tools 4.0 as the simulation environment.
The model and simulation results are shown in Figure 8.
‘Train1’ in top left of Figure 8 means data that the front train
need to send and “train 2” in top right of the figure means
data that the train behind receives. “code” in this figure
means fixed nodes of which function is transmitting
message. Because the transmitting media among nodes is
optical fiber and the transmission loss is very low, so it is
simplified as one node here. In this simulation, train 1 has ten
messages to send and each message has one serial number. If
train 2 can’t get one message, it will ask for one more time
according to the number until it receive all of the ten
messages. The simulation result is shown in Figure 8.
In this simulation, it can set different failures to make
train2 couldn’t receive the information from train1. We could
observe the system working process and behavior in different
condition. The simulation results are following: (1) The flag
bit is 1 if one of ‘link1’ and ‘link2’ is normal and train2 could
get all messages. (2) If none of ‘link1’ and ‘link2’ is normal,
train 2 couldn’t get train1 information, but it can stop on time
and turn to OS when it stops two minutes later. And CTC or
other trains can get train1 position, lest affect other trains’
operation.
The design suggestions of actual system: although safe-
failure process can protect the train from crash, it will affect
the operation efficiency. Therefore, we need to consider
using effective retransmission algorithm to ensure the train
can receive information within a certain time limit and avoid
it going into the fault-processing mode.
We can also get the relation between different channel
quality and delivery rate by setting the value of channel
quality. Figure 9 shows the delivery rates in different
conditions and the rates come from 100 times simulation
each channel quality. From the results we can see the
delivery rate will lift with the channel quality rise. The trends
tend to be linear.
DOI 10.5013/ IJSSST.a.17.42.29
Figure 9. The delivery rate in different channel quality
The design suggestions of actual system: It is necessary
to choose optical fiber astransmission media among fixed
codes for there is nearly no loss using optical fiber. But the
channel quality between trains is not easy to control,
therefore it reinforces that effective retransmission algorithm
is needed, so that the message can be received though the
loss rate is high.
V. CONCLUSIONS
With the developing of communication and control
technology, the research on the application of T2T
communication in train control system is the development
trend to increase the train intelligent degree. This paper first
analyzed the current research on T2T based train control
system and the feasibility of applying it in the field of
railway. Then the T2T based train control system general
design scheme and main working procedure are illustrated.
The scenario that trains interval tracing was chosen to build
CPN model and simulate. It is verified the functional
correctness and fail-safety of the T2T based train control
system. Finally, in this passage the simulation of the delivery
rate under different channel quality and sending interval was
analyzed. Meanwhile, it provided a theoretical basis and the
suggestions for the real system established.
ACKNOWLEDGMENTS
This work is partially supported by “the Fundamental
Research Funds for the Central Universities” (2015JBM014).
The authors also gratefully acknowledge the helpful
comments and suggestions of the reviewers, which have
improved the presentation.
REFERENCES
[1] European Commission within the Seven Framework Programme,
Next generation of train control systems project[R],2013, WEB:
http://www.NGTC.eu.
[2] Brosseau, J., Moore Ede, B. (June 2009) “Development of an
Adaptive Predictive Braking Enforcement Algorithm,” DOT-FRA-
ORD-09/13, U.S. Department of Transportation, Federal Railroad
Administration, Office of Research and Development, Washington,
D.C.
[3] Hartenstein H, Bochow B, Ebner A, et al. Position-aware ad hoc
wireless networks for inter-vehicle communications: the Fleetnet
29.5 ISSN: 1473-804x online, 1473-8031 print
YU LIU et al: A STUDY ON THE APPLICATION OF TRAIN TO TRAIN COMMUNICATION IN TRAIN CONTROL ...
project[C]//Proceedings of the 2nd ACM international symposium on
Mobile ad hoc networking & computing. ACM, 2001: 259-262.
[4] Yasuyuki Nakamura, Overview of the Next-generation Railway
Operation System in the Tokyo Metropolitan Area, JR East Technical
Review No. 20(2011): 03–06.
[5] Mayumi Korekoda, Hiroshi Kawasaki, Yasuhiro Tanaka,
Considerations for a Next-generation Railway System in the Tokyo
Metropolitan Area, JR East Technical Review No. 20(2011): 16–19.
[6] Alstom Transport, Lille metro RS and Urbalis Fluence - Case
Study[R]. WEB: http://www.alstom.com/transport.
[7] Jikang Xu, Train-train communication based CBTC system analysis,
Railway Signalling & Communication, 2014 50(6):78-80.
[8] Su Jing, Wang Dong, Zhang Fei-fei. Review of vehicle networking
technology application [J]. Intelligent Processing and Application,
2014, 06: 69-12.
[9] Wang Zhe. A study on performance analysis of wireless inter- vehicle
communication system in traffic streams [D]. Hefei: USTC, 2009.
[10] García C R, Lehner A, Strang T, et al. Channel model for train to
train communication using the 400 MHz band[C]//Vehicular
Technology Conference, 2008. VTC Spring 2008. IEEE. IEEE, 2008:
3082-3086.
[11] Liu P, Ai B, Zhong Z, et al. A Novel Train-to-Train Communication
Model Design Based on Multihop in High-Speed Railway[J].
International Journal of Antennas and Propagation, 2012.
[12] Chen Qi-xiang. Research on Train to train direct communication
based on ultra-short wave [D]. Lanzhou: Lanzhou Jiao Tong
University. 2014.
[13] Ning Bin, Tang Tao, et al. High-speed train control system [M].
Beijing: Beijing Publishing House,2012.
DOI 10.5013/ IJSSST.a.17.42.29
[14] Wang Zhao-ran, Xie Xian-zhong, Zhao Ding-xin. Key technique of
vehicular Ad Hoc Network [J]. Science Technology. 2011, 27(1): 44-
51.
[15] Yang Qiong, Shen Lian-feng. System architecture and
communication protocols in vehicular Ad Hoc networks [J]. ZTE
Communications. 2011, 03:8-11.
[16] Lv, Z., Li, X., Hu, J., Yin, L., Zhang, B., & Feng, S. (2015, June).
Virtual geographic environment based coach passenger flow
forecasting. In 2015 IEEE International Conference on Computational
Intelligence and Virtual Environments for Measurement Systems and
Applications (CIVEMSA) (pp. 1-6). IEEE.
[17] Lv, Z., & Su, T. (2014, November). 3d seabed modeling and
visualization on ubiquitous context. In SIGGRAPH Asia 2014 Posters
(p. 33). ACM.
[18] Li, X., Lv, Z., Hu, J., Zhang, B., Yin, L., Zhong, C., ... & Feng, S.
(2015, May). Traffic management and forecasting system based on 3d
gis. In Cluster, Cloud and Grid Computing (CCGrid), 2015 15th
IEEE/ACM International Symposium on (pp. 991-998). IEEE.
[19] Lv, Z., Li, X., Zhang, B., Wang, W., Zhu, Y., Hu, J., & Feng, S.
(2016). Managing big city information based on WebVRGIS. IEEE
Access, 4, 407-415.
[20] Li, X., Lv, Z., Hu, J., Zhang, B., Yin, L., Zhong, C., ... & Feng, S.
(2015, May). Traffic management and forecasting system based on 3d
gis. In Cluster, Cloud and Grid Computing (CCGrid), 2015 15th
IEEE/ACM International Symposium on (pp. 991-998). IEEE.
[21] Huang, L., Wu, J., You, F., Lv, Z., & Song, H. (2016). Cyclist Social
Force Model at Unsignalized Intersections with Heterogeneous
Traffic. IEEE Transactions on Industrial Informatics
.
29.6 ISSN: 1473-804x online, 1473-8031 print