Design of Wireless Sensor Network for Intra-vehicular Communications

Full text

Design of Wireless Sensor Network for
Intra-vehicular Communications
Md. Arafatur Rahman
Department of Electrical Engineering and Information Technologies (DIETI)
University of Naples Federico II, Naples, Italy
Laboratorio Nazionale di Comunicazioni Multimediali (CNIT), Naples, Italy
E-mail: arafatur.rahman@unina.it
Abstract. The number of sensor nodes in the vehicle has increased sig-
nificantly due to the increasing of different vehicular applications. Since,
the wired architecture is not scalable and flexible because of the internal
structure of the vehicle, therefore, there is an increasing level of appeal to
design a system in which the wired connections to the sensor nodes are
replaced with wireless links. Design a wireless sensor network inside the
vehicle is more challenging to other networks, e.g., wireless, sensor and
computer networks, because of the complex environment inside the vehi-
cle. In this paper, we design a wireless sensor network for intra-vehicular
communications. Firstly, we discuss about the link design between a base
station and a sensor node and then we design a network scenario inside
the vehicle for reliable communication. Finally, the performance is eval-
uated in terms of network reliability. The simulation results assist to
design a robust system for intra-vehicular communications.
Keywords: Controller Area Network, ZigBee, Intra-vehicular Communications.
1 Introduction
The Controller Area Network (CAN) is a serial communication protocol capable
of managing high efficiency distributed realtime control with a high level of
security. A CAN network is composed of a linear bus made with a twisted pair
of wires and number of nodes connected to each other via the transmission
medium. It is the most widespread system of communication between Sensor
Nodes (SNs) inside a vehicle with wired connections. Fig. 1 is the example of a
controller area networks and Fig. 2 depicts the frame format of CAN. For more
information about CAN, we refer to [1, 2].
The number of sensors in the vehicle has increased significantly due to the
various safety and convenience applications. Since, the wired architecture is not
scalable and flexible because of the internal structure of the vehicle [3]. Therefore,
there is an increasing level of appeal to design a system in which the wired
connections to the SNs are replaced with wireless links. To this end, several

Fig. 1. Example of a Controller Area Network.
Fig. 2. CAN Frame Format.
technologies, such as Radio Frequency IDentification (RFID) and Zigbee, have
been investigated in literature [4–6].
Wireless channels are by nature extremely complex and unpredictable sys-
tems. Several models and parameters are used to characterize wireless channel
[7, 8], whereas many of them are backed up by intuition and physical theory [9,
10]. However, no single concrete method for determining the characteristics of
a wireless channel has been established. Therefore, the most accurate method
of characterizing a particular wireless channel is experimental measurements,
particularly for practical purposes.
The design of an Intra-Vehicle Wireless Sensor Networks (IVWSNs) can not
be separated from the study on the link between the different sensor nodes
distributed in the vehicle. Therefore, link designing between Base Station (BS)
an SN is an important issue in IVWSNs. The level of network performance
varies with different communication parameters such as distance between BS
and SN, transmission power and channel fading. From Fig. 3, we can see that
the transmission power of the BS is set P
t1
if the distance between BS and SN
is d
1
. When the distance is increased from d
1
to d
2
then the transmit power
needs to be increased from P
t1
to P
t2
for receiving the same level of received
signal by the SN. We also notice that due to the increasing of distance between
BS and SN, the obstacles may come in the propagation path that changes the
line-of-sight (LOS) to non LOS (NLOS). As a result, the fading distribution of a
channel will be changed. For achieving the better performance in IVWSNs, the
above parameters need to be adjusted. In fact, design a wireless sensor network

Fig. 3. Scenario is changed with the varying of communication parameters.
inside the vehicle is more challenging to other networks, e.g., wireless, sensor
and computer networks, because of the complex environment created by a large
number of parts inside the vehicle. Therefore, it is an active research area to
design a network for intra-vehicle communications.
In this paper, we design a wireless sensor network for intra-vehicular com-
munications and evaluate its performance in terms of network reliability. More
in details, firstly, we study about the link analysis between BS and SN, then
based on that we design an IVWSNs by utilizing ZigBee standard instead of the
traditional CAN technology. Finally, we define network reliability in terms of
end-to-end delay to measure the performance of the networks. The simulation
results assist to design a robust system for intra vehicular communications.
The rest of the paper is organized as follows. In Section 2, we provide the
related works, while in Section 3, we describe the design of IVWSNs. In Sec-
tion 4, we discuss about the network reliability, while in Section 5, we present
the simulation results. Finally, in Section. 6, we conclude the paper.
2 Related Works
There have been active researches on the design of wireless and sensor networks
[12, 14–16]. For example, In [12], the authors present an empirical study based
reliability estimation in wireless networks. However, there are less numbers of
work have addressed particularly the design of network for intra-vehicular com-
munications [3, 11, 17–20]. In [19], the authors present the viability of the optical
wireless channel for use in intra-vehicular communications applications. In [20],
the authors investigate the coverage area performance of multi band orthogonal
frequency division multiplex ultra wide band intra-vehicular communication in
the presence of plural mobile terminals.
The ZigBee is a key technology to design a wireless sensor network for var-
ious purposes. In [21, 22], the authors design a monitoring and control system

based on ZigBee wireless sensor network. In addition, it plays an important role
in the intra-vehicle networks. However, few works have addressed this technol-
ogy for intra-vehicular communications. In [3], the authors report the statistical
characteristics of 4 representative intra-vehicle wireless channels on the basis
of the results of received power measurements and verify the level of reliability
of the channels. In [11], the authors propose another work to characterize the
wireless channel for intra-vehicle wireless communication. In [17], the authors
design and analysis a robust broad-cast scheme for the safety related services
of the vehicular networks. In [18], the authors study the performance of ZigBee
sensor networks for intra-vehicle communications, in the presence of blue-tooth
interference.
Unlike all the aforementioned works, in this paper we design a ZigBee based
wireless sensor network for intra-vehicular communications and evaluate its per-
formance in terms of network reliability.
3 Design of IVWSNs
The main design of the intra-vehicle wireless sensor networks including two
parts: link design between BS and SN and network scenario design. The link
design presents the suitability of the communication parameters for single link
in IVWSNs, such as transmit power and the distance between BS and SN. The
network scenario design part presents the detailed description about IVWSNs.
We are explaining them in the following.
3.1 Link Design between BS and SN
In this sub-section, we study about the analysis of single link between BS and
SN, since the design of a IVWSNs can not be separated from the study on the
link between the different sensor nodes distributed in the vehicle. In order to
do that, we have carried out a simulation through a discrete event simulation
software, OPNET, with the relative packages for the ZigBee module. A pair of
transmitter (i.e., SN) and receiver (i.e., BS) communicates each other within a
vehicle. The BS collects the packets that are transmitting periodically by the
SN. The BS and the SN are placed at a distance d. The Transmit Power set:
{-10, -15, -20, -25} dBm, which is suitable for ZigBee, such as the Crossbow
MICAz MPR2400 [24]. The Carrier frequency is 2.4 GHz (ISM band). There
are two channels 1 (a, b), which are for NLOS paths with Rayleigh fading. The
path loss exponent γ for channel 1(a) 3 and for channel 1(b) is 4. The values
of shadowing deviation σ[dB] is 8. The suitability of the considered parameters
has been discussed elaborately in our previous work [23].
Fig. 4 shows the behavior of the average throughput with the variation of
distance between BS and SN for Channel 1 (a, b). The figure clearly shows a
decreasing trend of the average throughput with increasing distance. The cause
of this trend is due to the low power level of the packets arriving to the antenna
of BS. We know that the path loss increases with distance and the effect of the

Fig. 4. Average throughput versus Distance between BS and SN with the varying of
Transmit power.
log-normal shadowing involves a fluctuation in time of the received power, which
can further degrade the performance of the communication. These fluctuations
may lead the level of received power below the receiver sensitivity (-95 dBm).
Then, the BS evaluates the received packet as noise and consequently, the packet
is lost. From this analysis, we can see that the transmit power -15 dBm (both
γ = 3 and 4) is suitable for IVWSNs. We also notice that when the distance
between BS and SN is less than 4 the performance is very good, consequently,
we can say the BS should be placed in the center of the car for getting good
performance. The detailed analysis for single BS and SN is found in [23].
3.2 Network Scenario Design
There is only on BS that is placed in the center of the vehicle and several number
of SNs are placed around it, as shown in Fig. 5. In this way, the distance between
SN and BS will be less than any other scenarios where BS is set at any palace
in the vehicle, as a result the BS can receive packets from the SNs with better
signal strength. Two different SNs are considered: one is Green (G) and other
is Yellow (Y), whose transmission period is 120 ms and 60 ms, respectively.
We also consider four different cases according to traffic load in the networks.
These considerations will help to measure the performance of the network while
the traffic load is high. In case I, 100% Green, in case II, 70% Green and 30%
Yellow, in case III, 50% Green and 50% Yellow and in case IV, 30% Green and
70% Yellow SNs will be from the total sensor nodes, see Table 1. The more
number of Yellow SNs means the more traffic in the network because of its less
transmission period. The number of sensor node set is {10, 30, 50, 70, 90, 110}.
The communication parameters of the networks are as follows:
– Transmit Power: -15 dBM, as discussed in the previous subsections;

Number Case I Case II Case III Case IV
of NS G Y G Y G Y G Y
10 10 0 7 3 5 5 3 7
30 30 0 21 9 15 15 9 21
50 50 0 35 15 25 25 15 35
70 70 0 49 21 35 35 21 49
90 90 0 63 27 45 45 27 63
110 110 0 77 33 55 55 33 77
Table 1. Considered Cases
– Carrier frequency: 2.4 GHz (ISM band), which is used on ZigBee sensor node
[24]
– Receiver sensitivity: The reception threshold of the BS is set equal to -95
dBm, typical for ZigBee [24];
– Transmission Period: 120 ms and 60 ms for Green and Yellow SN, respec-
tively ;
– Channel: The channel is for NLOS paths with Rayleigh fading. The path
loss exponent γ is 4. The values of shadowing deviation σ[dB] is 8. These
values are suitable for intra-vehicle communication [3, 11].
– Packet size: 210 bits (ZigBee packet header 120 bits + data 90 bits);
Remark 1. 90 bits for data is selected with the reference of CAN message
used in [13].
– Parameters MAC:
• ACK Mechanism:
∗ ACK Wait Duration: 0.05 second;
∗ Number of retransmissions: 5.
• CSMA-CA Parameters:
∗ Minimum Backoff Exponent: 3;
∗ Maximum Number of Backoffs: 4;
∗ Channel Sensig Duration: 0.1 second.
4 Network Reliability
In this section, we define the network reliability, which will be utilized for mea-
suring the performance of the network. The definition is in the following:
Definition 1 (Network Reliability R). Network reliability is defined as the
ability to deliver the packets to the destination (BS) within a certain time limit
(called Deadline). The expression of the reliability can be written:
R = P
r
(D
ete
≤ D) (1)
were R is the network reliability, D
ete
is the end to end delay (i.e., the overall
delay between the time instant when it creates a package from the application
layer and the time instant when it is received) and D is the deadline (i.e., the
limits on the chosen end-to-end delay of the packet).

Fig. 5. Example of SNs distribution inside the vehicle.
Remark 2. In this paper, we consider two deadlines: one is called restrictive
deadline denoted as D
1
and other is called less restrictive deadline denoted as D
2
.
Note that D
1
and D
2
represent 25% and 50% of the maximum SN transmission
period, i.e., 120 ms. This consideration is reasonable because less end to end
delay of the packet increases the reliability of the network due to its less packet
retransmission.
5 Simulation Results
In this section, we analyse the reliability of the network, varying of traffic load,
by taking account of Definition 1. Due to the increasing of traffic load, the intra-
vehicle network becomes congested. The effect of congestion on the network is
also investigated. In order to assess the level of reliability, we have carried out a
series of simulations through a discrete event simulation software, OPNET, with
the relative packages for the ZigBee module. The performance of the network is
measured based on this network reliability.
In Fig. 6, we report the CDF of the end-to-end delay versus the number
of nodes for analyzing the reliability in the case I. From the figure, we notice
that as the number of SN increases in IVWSN, the CDF shift to the right.
The cause of this performance is due to the collisions among the packets, which
increases with the increasing number of SN in the network. In fact, after the
collision SN waits for a certain period of time (Backoff + sensing period) and
then if the channel is free, it retransmits the packet that already caused collision
previously. The new re-transmissions can be subjected to other collision. The
repetition of the procedures is explained under the CSMA-CA protocol. It is
easy to understand at this point that the increasing number of collisions results
the increasing end-to-end delay experienced by the packets.

Fig. 6. CDF of end-to-end delays Vs the number of nodes in the case I.
Fig. 7. CDF of end-to-end delays Vs the number of nodes in the case II.
As in the first case there is only Green SN and the considered deadlines are
D
1
= 30ms and D
2
= 60ms. As it can be seen in Fig. 6, the condition on less
restrictive deadline, D
2
, is fully satisfied, as shown in Table 2. On the other
hand, the restrictive condition, D
1
= 30ms, it is satisfied in the case of 10, 30,
50 and 70 SNs, while in other cases (90, 110 SNs) are satisfied with a probability
not adequate (about 16% in best case) in terms of reliability.
We also report the CDF of the end-to-end delay versus the number of nodes
for analyzing the reliability in other cases, as shown in Fig. 7 - 9. The above
figures show that the introducing of more Yellow SNs in the network, the end-
to-end delay is increasing i.e., the reliability of the network decreases. In fact,
in the case I with 70 NS, the reliability is 100% for less restrictive deadline.
D
2
, whereas in case II, this falls to about 96% and continuously falling while
increasing in the number of Yellow SNs in the network.

Fig. 8. CDF of end-to-end delays Vs the number of nodes in the case III.
In addition, the increasing number of SN, in particular when it increases the
number of Yellow SN, in the intra-vehicle WSN is subjected to the phenomenon
of congestion. In fact, increasing the traffic up to a certain point where the net-
work is no longer able to handle the traffic then it enters into congestion. As a
result a number of transmitted packets (including retransmitted packets) by SN
never reaches its destination. Higher the degree of congestion of the network, the
greater will be the number of packets that never arrives at the destination. The
Table 2 summarizes the results obtained in four cases, the results highlighted in
bold are ”distorted” due to congestion of the network. The phenomenon of con-
gestion decreases the end-to-end delay that causes the distortion of the results,
since in OPNET the end-to-end delay is calculated on the basis of packets that
reach to their destination.
To mitigate the problem of congestion, we introduce two BSs in the network.
Each BS is consisting of 50% SNs from the total SNs. In Fig. 10, there is 90
SNs with half Green and half Yellow SNs in case of both single and double BS.
We note, in case of single BS, initially the result is distorted due to the large
number of packets that do not reach to the destination because of the congestion
in the network. In case of double BS, there is no congestion effects because of
the proper traffic load distribution. However, introducing additional BS increases
the design complexity of the networks. Therefore, a new MAC strategy can be
designed for congestion network that will be the future direction of this work.
We will also investigate the performance by introducing the concept of cognitive
radio in intra-vehicle wireless sensor networks [25]-[30].
6 Conclusion
In this paper, we design a wireless sensor network for intra-vehicular communica-
tion. We define the network reliability in terms of end-to-end delay to measure

Fig. 9. CDF of end-to-end delays Vs the number of nodes in the case IV.
Number R in case I R in case II R in case III R in case IV
of NS D
1
D
2
D
1
D
2
D
1
D
2
D
1
D
2
10 100% 100% 100% 100% 100% 100% 100% 100%
30 100% 100% 100% 100% 100% 100% 100% 100%
50 100% 100% 72% 97% 52% 90% 45% 87%
70 92% 100% 40% 96% 25% 88% 66% 86%
90 16% 100% 30% 95% 62% 89% 63% 87%
110 6% 100% 65% 75% 77% 92% 82% 89%
Table 2. Reliability in Different Cases
the performance of the networks. After the analysis, we note that, the phe-
nomenon of congestion plays an important role in the network while the traffic
load is high. To mitigate the congestion problem, we could increase the number
of BS in the network. In fact, introducing additional BS increases the design
complexity of the networks. Therefore, a new MAC strategy can be designed for
congestion network that will be the future direction of this work. The simulation
results assist to design a robust system for intra vehicular communications.
Acknowledgments
This work is partially supported by the project ”Mobile Continuos Connected
Comprehensive Care (MC3CARE), ”DRIVEr monitoring: technologies, method-
ologies, and IN-vehicle INnovative systems for a safe and ecocompatible driv-
ing (DRIVE IN
2
)” founded by the Italian national program Piano Operativo
Nazionale Ricerca e Competitivit 2007-2013 and the project, ”Sviluppo di Tec-
niche di Comunicazione di Sistemi Embedded Distribuiti” founded by POR Cam-
pania FSE 2007/2013.

Fig. 10. Comparison between CDF in the case of single BS and double BSs.
References
1. K.W. Tindell, H. Hansson, and A.J. Wellings, Analysing real-time communications:
Controller Area Network (CAN). in Proc. of 15th Real-Time Systems Symposium,
Pages 259-263, IEEE Computer Society Press, 1994.
2. K. Tindell, A. Burns, Guaranteeing Message Latencies on Controller Area Network
(CAN). in Proc. of 1st International CAN Conference, Pages 1-11, 1994.
3. H.M. Tsai, W. Viriyasitavat, O. K. Tonguz, C. Saraydar, T. Talty and A. Mac-
donald, Feasibility of In-car Wireless Sensor Networks: A Statistical Evaluation. in
Proc. IEEE SECON, Pages 101-111, 2007.
4. O. K. Tonguz, H.M. Tsai, C. Saraydar, T. Talty, and A. Macdonald, Intra-car wire-
less sensor networks using RFID: Opportunities and challenges. in Proc. INFOCOM
MOVE Workshop, Pages 43-48, 2007.
5. H.M. Tsai, O. K. Tonguz, C. Saraydar, T. Talty, M. Ames, and A. Macdonald,
Zigbee-based intra-car wireless sensor networks: A case study. IEEE Wireless Com-
mun, vol. 14, Pages 67-77, 2007.
6. W. Niu, J. Li, S. Liu, and T. Talty, Intra-vehicle ultra-wideband communication
testbed. in Proc. MILCOM, Pages 1-6, 2007.
7. A.S. Cacciapuoti, F. Calabrese, M. Caleffi, G. Di Lorenzo, L. Paura, Human-mobility
enabled wireless networks for emergency communications during special events. El-
sevier Pervasive and Mobile Computing, Vol. 9, Pages 472-483, 2013.
8. A.S. Cacciapuoti, F. Calabrese, M. Caleffi, G. Di Lorenzo, L. Paura, Human-mobility
enabled networks in urban environments: Is there any (mobile wireless) small world
out there?. Elsevier Ad Hoc Networks, Vol. 10, Pages 1520-1531, 2012.
9. H. Hashemi, Impulse response modeling of indoor radio propagation channels. IEEE
J. Sel. Areas Commun., vol. 11, Pages 967-978, 1993.
10. A. Saleh and R. Valenzuela, A statistical model for indoor multipath propagation.
IEEE J. Sel. Areas Commun., vol. SAC-5, Pages 128-137, 1987.
11. A. R. Moghimi, H.M Tsai, C. U. Saraydar and O. K. Tonguz, Characterizing
IntraCar Wireless Channels. IEEE Transactions on Vehicular Technology, vol. 58,
Pages 5299-5305, 2009.

12. S. Woo and H. Kim, Estimating Link Reliability in Wireless Networks: An Empir-
ical Study and Interference Modeling. in Proc. INFOCOM, Pages 1-5, 2010.
13. M. Ellims, S. Parker, J Zurlo, Design and Analysis of a Robust Real-Time Engine.
IEEE Micro, Vol. 22, Pages 20-27 2002.
14. B. Xing, S. Mehrotra, and N. Venkatasubramanian, RADcast: Enabling reliability
guarantees for content dissemination in ad hoc networks. in Proc. IEEE INFOCOM,
Pages 1998-2006, 2009.
15. A.S. Cacciapuoti, M. Caleffi, L. Paura, A theoretical model for opportunistic routing
in ad hoc networks. in Proc. of International Conference on Ultra Modern Telecom-
munications Workshops (ICUMT ’09), Pages 1-7, 2009.
16. A.S. Cacciapuoti, M. Caleffi, L. Paura, Optimal Constrained Candidate Selection
for Opportunistic Routing. in Proc. of IEEE Global Telecommunications Conference
(GLOBECOM 2010), Pages 1-5, 2010.
17. X. Ma, J. Zhang, X. Yin, K.S. Trivedi, Design and Analysis of a Robust Broad-
cast Scheme for VANET Safety-Related Services. IEEE Transactions on Vehicular
Technology, vol. 61, Pages 46-61, 2012.
18. R. D. Francisco, L. Huang, G. Dolmans, H. D. Groot, Coexistence of ZigBee wireless
sensor networks and Bluetooth inside a vehicle. IEEE 20th International Symposium
on Personal, Indoor and Mobile Radio Communications, Pages 2700-2704, 2009
19. M.D. Higgins, R.J Green, M.S. Leeson, Channel viability of intra-vehicle optical
wireless communications. in Proc. IEEE GLOBECOM Workshops, Pages 813-817,
2011.
20. T. Arai, T. Shirai, Y. Watanabe, F. Maehara, Coverage performance of UWB in-
car wireless communication in the presence of multiple terminals. IEEE Radio and
Wireless Symposium (RWS), Pages 111-114, 2012.
21. Z. Yiming, Y. Xianglong, G. Xishan, Z. Mingang and W. Liren, A Design of Green-
house Monitoring and Control System Based on ZigBee Wireless Sensor Network.
in Proc of International Conference on Wireless Communications, Networking and
Mobile Computing, Pages 2563-2567, 2007.
22. L. Chen S. Yang and Y. Xi, Based on ZigBee wireless sensor network the moni-
toring system design for chemical production process toxic and harmful gas. in Proc
of International Conference on Computer, Mechatronics, Control and Electronic
Engineering, Pages 425-428, 2010.
23. M. A. Rahman, Reliability Analysis of ZigBee based Intra-vehicle Wireless Sensor
Networks. in Proc of Nets4Cars 6th International Workshop on Communication
Technologies for Vehicles, 2014.
24. MPR/MIB mote hardware users manual. http://www.xbow.com.
25. A.S. Cacciapuoti, M. Caleffi, L. Paura, R. Savoia, Decision Maker Approaches for
Cooperative Spectrum Sensing: Participate or Not Participate in Sensing?. IEEE
Transactions on Wireless Communications, Vol. 12, Pages 2445-2457, 2013.
26. A.S. Cacciapuoti, M. Caleffi, L. Paura, Reactive routing for mobile cognitive radio
ad hoc networks. Elsevier Ad Hoc Networks, Vol. 10, Pages 803-805, 2012.
27. M.A. Rahman, M. Caleffi, L. Paura, Joint path and spectrum diversity in cogni-
tive radio ad-hoc networks. EURASIP Journal on Wireless Communications and
Networking, Vol. 2012(1), Pages 1-9, 2012.
28. A.S. Cacciapuoti, C. Calcagno, M. Caleffi, L. Paura, CAODV: Routing in Mobile
Ad-hoc Cognitive Radio Networks. in Proc. of IEEE IFIP Wireless Days 2010, Pages
1-5, 2010.
29. A.S. Cacciapuoti, M. Caleffi, L. Paura, Widely Linear Cooperative Spectrum Sens-
ing for Cognitive Radio Networks. in Proc. of IEEE Global Telecommunications
Conference (GLOBECOM 2010), Pages 1-5, 2010.

30. A.S. Cacciapuoti, M. Caleffi, D. Izzo, L. Paura, Cooperative Spectrum Sensing
Techniques with Temporal Dispersive Reporting Channels. IEEE Transactions on
Wireless Communications, Vol. 10, Pages 3392-3402, 2011.