Performance Improvement of WSNs using Joint Reed Solomon and Network Coding

Full text

International Journal of Computer Applications (0975 – 8887)
Volume 177 – No. 41, March 2020
22
Performance Improvement of WSNs using Joint Reed
Solomon and Network Coding
Khalifa A. Salim, PhD
Asst. Prof.
ICE Engineering Department
Baghdad University
Baghdad, Iraq
Maryam K. Abboud
Student Member, IEEE
Computer Engineering Department
Al-Farabi University
Baghdad, Iraq
ABSTRACT
In this paper, a combined error control coding and network
coding (NC) scheme is proposed for performance
improvement of wireless sensor networks (WSNs) in terms of
throughput and Bit error rate (BER). A suggested network
model is conserved in simulations, which integrates the
characteristics of; mesh, stare, and tree network models with
eleven nodes. Reed-Solomon (RS) code is considered as a
Forward Error Control (FEC) coding scheme in simulations.
The simulation results shows that BER may slightly increases
using NC, while at high Signal to Noise power Ratios (SNRs),
throughput performance is improved by about 64%. This
improvement is obtained even when the transmission
experience a frequency selective fading environments.
Keywords
NC, RS code, WSN, Channel Coding.
1. INTRODUCTION
WSN consists of a number of small, autonomous devices
called Wireless Sensor nodes, which formed with number of
sensors integrated with tiny microprocessors. Many
applications where WSN nodes adopted and working in a
cooperative manner [1]. The fundamental characteristic of
WSNs is the multi-hop transmission which affected by the
noisy links and hence effects the throughput performance of
the whole network [2]. Hence, to cover this problem, FEC is
adopted to reduce transmission errors, while NC is used for
required transmissions reduction. FEC and NC was covered
by many researchers in this field with different conditions. A
NC scheme was introduced by [3], where a relay channel
mode used to perform NC between two users and the users
perform channel coding operations. NC was combined with
FEC coding in [4], where channel decoding is performed after
NC at the destination by combined soft/hard decisions. In [5-
8], system robustness increased by applying RS and NC
schemes at intermediate nodes of a WSN, where system
robustness enhanced and the require resources was reduced.
The same combination was proposed in [9], where it adopted
for star network model. In the proposed system, Automatic
Repeat Request (ARQ) has been adopted as a FEC code
combined with NC which integrates multiple packets to form
one packet to be transmitted. The QoS for WSNs can be
achieved by the optimal selection for the relay nodes which
combines NC with FEC code after parameters adjustment as
proposed by [10-13]. A generalized WSN model combines
star, mesh and tree structures is considered in this paper, WSN
nodes are classified into; source nodes, coding nodes, and
destination nodes. A node to node error correction
encoding/decoding scenario with NC was performed in an end
to end manner (encoding/decoding) by the suggested system.
At the source nodes, before data sending through the network,
packets are FEC encoded with RS code, where they
reprocessed at intermediate nodes and RS decoded to be
combined using NC. After NC process, the network coded
packets are recoded using RS code. At the destinations side, to
recover the original data, RS decoding is implemented with
NC decoding for packets received from different links. In this
paper; system model and parameters are proposed in sections
2 and 3. Performance evaluation, in sections 4. Finally, main
conclusions and suggestions for future work is proposed in
section 5.
2. SYSTEM MODEL
The proposed system is shown in Figure 1, where the
combination takes place. Binary Phase Shift Keying (BPSK)
is adopted for modulation and detection. In FEC coding stage,
RS    code is considered, which provides maximum
separable property [14]. In the presence work, the code word
length  and the data block length are the two important RS
code parameters needed to be defined. The correction
capability t of RS code used is      [14].
Berlekamp and Chien’s algorithms are used for RS decoding
[15]. (255,247, 4) representing RS coding parameters
   considered in this work.

International Journal of Computer Applications (0975 – 8887)
Volume 177 – No. 41, March 2020
23
Fig. 1: Proposed System Model
Based on generating independent vectors [16], a Random
Linear Network Coding (RLNC) was adopted at NC block of
Figure 1 at each coding node. Network encoding in performed
at coding nodes, while decoding is performed at sink node. A
First in First out (FIFO) queue is used to handle the received
packets to be coded using NC, where system throughput
improved. For network coding, packets are linearly combined
using independent vectors generated with elements related to
a Galois Field (GF), where NC process follows equation (1)
[17];




 (1)
: Source packets (   
 : Coding vectors (   ) of 
elements
: Coded packet.

 and 
:  symbols of Y and respectively.
A simple Gaussian elimination applied for the purpose of NC
decoding using encoding vectors attached in packet's header
[7], [8], [10], [18]. AWGN is the Additive White Gaussian
Noise considered in simulations along with; flat fading and a
multipath selective channels with moderate level of distortion,
where its parameters are shown in Table 1 [19].
Table 1. Multipath Channel Model [19].
Taps No.
Relative Delay (µs)
Average Power (dB)
1
0
0
2
0.4
-5
3
0.9
-10
3. PROPOSED NETWORK MODEL
In this work, a network model is proposed which enables
cooperative communication among its nodes in order to
achieve the advantage of combining NC with channel coding.
The vague nature of WSN nodes distribution makes
performance evaluation to be a critical point especially for
large networks. Thus, the proposed network model in this
paper can be assumed to be a more generalized model which
integrates mesh, star, and tree network model’s properties in
one model as shown in Figure 2.
Fig. 2: Proposed Network Model.
The network of Figure 2 consists of a few number of source,
intermediate, and a base station nodes, where it forms an
elementary sub-network belonging to a large WSN. The
intermediate nodes acts as the coding nodes in the network,
while the base station is the sink node. Data packets are
generated at the source nodes where FEC encoding applied.
NC and FEC coding/decoding are performed at the
intermediate nodes. At the other hand, FEC and NC decoding
processes are carried out at the final destination node.
Figure 3 represents the processing block of all node types;
source, intermediate and sink nodes. In Figure 3, there is no
need to perform NC decoding process at the coding nodes
since the original data is requested by the final recipient.

International Journal of Computer Applications (0975 – 8887)
Volume 177 – No. 41, March 2020
24
Fig. 3: Processing Blocks of Different Nodes.
4. SIMULATION TESTS AND RESULTS
In the proposed work, two test conditions are used to measure
Bit Error Rate (BER) and Throughput (Thru); RS coding
without NC process and RS coding in the presence of NC
process. BER and Throughput are calculated by the following
equations (2 and 3);
  
 (2)
To measure throughput performance, two cases must be taken
into account; first, throughput measuring without NC process
and the second is throughput with NC process to be compared
with the first one.
  
 (3)
In the presence of NC, the most effective parameter in
calculating the throughput of the network is defined as NC
gain, which represents the gain achieved in transmissions
reduction as the basic NC advantages. Hence;
 
 (4)
Where, NC gain is defined by;



(5)
Another important parameter used to measure the
performance gain in system throughput is given by equation
(6);
  
   
(6)
In this work, four main scenarios are considered:
1- Transmission without coding.
2- Transmission without NC, with RS coding.
3- Transmission with NC, without RS coding.
4- Transmission with NC and RS coding.
In simulations, link’s capacity is assumed to be 
.
The transmission considers 4Kbps transmission rate, where all
intermediate nodes performs NC. The BER performance of
the proposed network is shown in Figure 4, assuming the
combination of RS of (255, 247, 4) specifications with NC
over various channels, and throughput performance is
presented in Figure 5.
(a) (b)

International Journal of Computer Applications (0975 – 8887)
Volume 177 – No. 41, March 2020
25
(c)
Fig. 4: BER performance of the proposed network considering various channel models.
a- AWGN b- Flat Fading c- Multipath Fading
(a) (b)
(c)
Fig. 5: Throughput performance of the proposed network considering various channel models.
a- AWGN b- Flat Fading c- Multipath Fading
As could be noticed, the proposed combination results a
transmissions reduction, where 10 transmissions are needed
for packet delivering to the end recipients, while 28
transmissions are required in the case where there is no NC
applied. This results in 64% transmission throughput gain. At
the other hand, a low SNR values are required in order to
obtain high throughputs. Table 2 summarizes this benefit
over different wireless channel environments, where it shows
that as transmission throughput improved from 40 Kbps to
117 Kbps, NC addition reduces the required SNR from 20 dB
over AWGN channel to 11 dB over multipath fading channel.

International Journal of Computer Applications (0975 – 8887)
Volume 177 – No. 41, March 2020
26
Table 2. Required SNR for Throughput Improvement
with NC over Different Channel Environments
AWGN
Flat Fading
Multipath Fading
20 dB
14 dB
11 dB
NC
No NC
NC
No NC
NC
No NC
117
Kbps
40
Kbps
117
Kbps
40
Kbps
117
Kbps
40 Kbps
Regarding BER performance over AWGN channel, an
improvement achieved by about 1.7dB in SNR by applying
RS coding in the presence of NC. And by testing the network
performance when there is no NC applied, the addition of RS
code improves the performance by about 0.7dB. Regarding
multipath and flat fading channels, the BER performance is
decayed clearly at low SRN values of a fading environments
when adopting NC process. At the other hand, its addition
improves BER at high SNR by about 2dB over multipath
fading channel at a BER of as compared to the BER
degradation appears at low SRN values.
5. CONCLUSIONS
In this paper, FEC and NC combinations are tested
considering a generalized network model. An improvement
in BER performance by combining RS coding with NC was
achieved with a considerable reduction in number of
transmissions which improves system throughput. For BER
performance tests, the SNR gain over AWGN channel is
relatively grater as compared to that over fading channels,
while throughput improvement tests shows that the required
SNR is reduced especially in multipath fading channels,
where 117 Kbps was achieved using NC as compared to 40
Kbps for the case where no NC used. Finally, NC with RS
combination can provide an increasing in throughput on a
cost of slight BER degradation at low SNR values.
6. REFERENCES
[1] H. Karl and A. Willig, "Protocols and Architectures for
Wireless Sensor Networks", John Wiley and Sons, Ltd,
Sussex, England, 2005.
[2] Z. Hu, B. Li, "Fundamental performance limits of
wireless sensor networks", Journals of ad Hoc and
sensor networks AHSWN, 2004.
[3] C. Hausl and J. Hagenauer, "Iterative Network and
Channel Decoding for the Two-Way Relay Channel", in
Proceedings IEEE International Conference on
Communications, Vol. 4, Istanbul, June 2006.
[4] C. Hausl and P. Dupraz, "Joint Network-Channel
Coding for the Multiple-Access Relay Channel", IEEE
Communications Society on Sensors and Ad Hoc
Communications and Networks, Vol. 3, Reston,
Virginia, Sept. 2006.
[5] C. Hausl, "Joint Network-Channel Coding for Wireless
Relay Networks", College of Electrical Engineering and
Information Technology, Technical University of
Munich, Munich, Nov. 2008.
[6] Z. Guo, J. Huang, B. Wang, J. Cui, S. Zhou, and W. P.,
“Non-binary joint network and channel coding for
underwater sensor networks”, The International
Workshop on Under Water Networks, September 2008.
[7] Z. Guo, J. Huang, B. Wang, S. Zhou, J.-H. Cui and P.
Willett "A Practical Joint Network-Channel Coding
Scheme for Reliable Communication in Wireless
Networks", in Proceedings of IEEE Transactions on
wireless communications, Vol.11, Issue 6, June 2012.
[8] T. Matsuda and T. Takine. "Multicast Communications
with Reed Solomon/Network Joint Coding in Wireless
Multi-hop Networks", Journal of Communications, Vol.
4, No 11, Dec. 2009.
[9] S. J. Johnson, L. Ong, and C. M. Kellett, "Joint
Channel-Network Coding Strategies for Networks with
Low Complexity Relays", European Transactions on
Telecommunications, Vol. 22, Issue 7, Oct. 2011.
[10] C. Koller, M. Haenggi, J. Kliewer, D. J. Costello, J. L.
Fellow, "Joint Design of Channel and Network Coding
for Star Networks connected by Binary Symmetric
Channels", IEEE Transactions on Communications, Vol.
62, Issue 1, Nov. 2013.
[11] G. Angelopoulos, A. Paidimarri, A. P. Chandrakasan
and M. Medard, "Experimental Study of the Interplay of
Channel and Network Coding in Low Power Sensor
Applications", IEEE International Conference on
Communications (ICC), Budapest, June 2013.
[12] Y. Jin and P. Ruan, "Adaptive Cooperative FEC Based
on Combination of Network Coding and Channel
Coding for Wireless Sensor Networks", Journal of
networks, Vol. 9, No. 2, Feb. 2014.
[13] D. C. Adams, J. Du, M. Medard and C. C. Yu, "Delay
Constrained Throughput-Reliability Tradeoff in
Network-Coded Wireless Systems", IEEE Global
Communications Conference (GLOBCOM), Austin,
Texas, Dec. 2014.
[14] J. Moreira and P. Farrell, “Essentials of Error Control
Coding”, John Wiley and Sons Ltd, West Sussex,
England, 2006.
[15] X. T. Vu, P. Duhamel, and M. D. Renzo, "Performance
Analysis of Network Coded Cooperation with Channel
Coding and Adaptive DF-Based Relaying in Rayleigh
Fading Channels", IEEE Signal Processing Letters,
Institute of Electrical and Electronics Engineers (IEEE),
Vol. 22, No. 9, Mar. 2015.
[16] P. Chou and Y. Wu, " Network Coding for the Internet
and Wireless Networks", IEEE Signal Processing
Magazine, Volume 24, Issue 5, June 2007.
[17] S. Katti, "Network Coded Wireless Architecture", Ph.D.
Thesis, Department of Electrical Engineering and
Computer Science, Massachusetts Institute of
Technology, Aug. 29, 2008.
[18] S. Zhang, Y. Zhu, S.-C. Liew, and K. B. Letaief, "Joint
Design of Network Coding and Channel Decoding for
Wireless Networks", IEEE on Wireless communications
and networking Conference WCNC, Kowloon, 2007.
[19] IEEE 802.16 Broadband Wireless Access Working
Group, “Channel Models for Fixed Wireless
Applications”, IEEE 802.16a-03/01, 27, May, 2003.
IJCATM : www.ijcaonline.org