A fuzzy based congestion controller for control and balance congestion in gried-based wsn

Natarajan Meghanathan et al. (Eds) : WiMONe, NCS, SPM, CSEIT - 2014
pp. 27–36, 2014. © CS & IT-CSCP 2014 DOI : 10.5121/csit.2014.41203
A FUZZY-BASED CONGESTION
CONTROLLER FOR CONTROL AND
BALANCE CONGESTION IN GRIED-BASED
WSN
Ali Dorri and Seyed Reza Kamel
Department of Computer Engineering, Mashhad Branch,
Islamic Azad University, Mashhad, Iran
Alidorri@mshdiau.ac.ir
Rezakamel@computer.org
ABSTRACT
A Wireless Sensor Network (WSN) is deployed with a large number of sensors with limited
power supply in a wide geographically area. These sensors collect information depending on
application. The sensors transmit the data towards a base station called sink. Due to the
relatively high node density and source-to-sink communication pattern, congestion is a critical
issue in WSN. Congestion not only causes packet loss, but also leads to excessive energy
consumption as well as delay. To address this problem, in this paper we propose a new fuzzy
logic based mechanism to detect and control congestion in WSN. In the proposed approach, a
Monitor Node for each grid in congestion candidate region performs a fuzzy control to avoid
increasing congestion. Fuzzy controller’s inputs are continually fetched from the network by the
Monitor Node. Simulation results show that our approach has higher packet delivery ratio and
lower packet loss than existing approaches.
KEYWORDS
Fuzzy Logic, Wireless Sensor Network, Congestion Control, Packet Lost.
1. INTRODUCTION
A Wireless Sensor Network (WSN) consists of spatially distributed autonomous wireless sensor
nodes to cooperatively monitor physical or environmental conditions, such as temperature, sound
and pressure. In addition, WSN is a network made of hundreds or thousands of sensor nodes
which are densely deployed in hazardous/unattended environment with capability of sensing,
computing and sending information wirelessly to the base station (also called sink) via neighbour
nodes. Figure 1 shows a WSN that collect information and send it to the sink.

28 Computer Science & Information Technology (CS & IT)
Figure 1. Wireless Sensor Network [1]
In WSN special applications, once an event occurs, a sudden surge of data traffic will be triggered
by all sensor nodes in the event area, which may easily lead to congestion. Congestion in WSN
has a negative impact on network performance. It increases the packet loss, end-to-end delay and
wastes nodes energy. When a packet is lost, source node must retransmit it again. Therefore,
node’s energy is wasted and network lifetime will be decreased [2,3]. However, some
characteristics of WSN, such as constrained resources and leak of centralized coordination, make
the congestion problem in WSN more challengeable than any other networks. In WSN all nodes
send their sensed data to the sink. This flow of packet (also called source-to-sink traffic), increase
congestion probability and energy consumption in nodes near the sink. The reason is that, the
neighbouring nodes of the sink should forward other nodes packets in addition to their own traffic
[4,5].
To address these challenges, we present a fuzzy congestion control. In the proposed approach, the
network is divided into grids by the sink. Then, the sink specifies congestion candidate areas by
use of a calculated threshold. In each congestion candidate area, a Monitor Node (MN) uses a
fuzzy controller to detect and avoid congestion in its grid. Congestion level of each grid maybe
different from other grids and if the congestion level of any congestion candidate area reaches to
the acting level, the Border Node forwards packets out of the congested area. Therefore,
congestion and packet loss will be decreased. Simulation results show that the presented fuzzy-
based system decreases the packet loss in congested area and increases packet delivery ratio to the
sink. The remainder of this paper is organized as follows: section 2 presents a literature review of
the congestion detection and control mechanisms. Section 3 provides detailed description of the
proposed fuzzy controller approach. Performance evaluation of the proposed approach is
presented in section 4. Finally section 5 concludes and discusses the future directions of this
research.
2. RELATED WORKS
In literature many congestion control schemes have been proposed for WSN. Congestion control
schemes for WSN either focus on MAC layer or on both MAC and network layer [6]. Authors in
[7] presented an approach based on a threshold. This threshold refers to ratio of received packets
to serviced packets. In this approach each node has a priority. To detect congestion, both
threshold and priority of nodes are influenced. Proposed approach is efficient in term of Quality
of Service (QOS) as it sends data through multipath. Authors in [8] presented a Medium Access
Control (MAC) technique to coordinate the access of nodes to the shared medium. It uses the
queue buffer length of the sensor nodes to estimate the congestion. Then the traffic dynamically
disseminates along with classifying nodes into different priority classes to provide a congestion-
free routing path to the destination with improved QOS. In addition, it uses multiple forwarder
traffic diffusion, which has advantages like increasing network reliability and reducing
congestion. In [9] an optimal routing algorithm that allows optimizing transmission between the
peripheral nodes and central node is presented, in order to increase the residual energy of the
network. This protocol only aims to provide routing fidelity and does not address time
transmission requirements. Authors in [10] presented a cross-layer congestion controller that has
three parts: 1) multipath routing 2) adjusted ratio 3) application oriented design. In this algorithm,
each node has multiple downstream nodes to be transmitted. The probability of forwarding nodes

Computer Science & Information Technology (CS & IT) 29
and the rate of sending packets can be dynamically adjusted according to the congestion state.
Authors in [11] presented a novel approach based on bird’s behaviour. The proposed approach is
simple to implement at the individual node, involving minimal information exchange. In addition,
it displays global self-properties and emergent behaviour, achieved collectively without explicitly
programming these properties into individual packets. Performance evaluations show the
effectiveness of the proposed Flock-based Congestion Control (Flock-CC) mechanism in
dynamically balancing the offered load by effectively exploiting available network resources and
moving packets to the sink. Furthermore, Flock-CC provides graceful performance degradation in
terms of packet delivery ratio, packet loss, delay and energy consumption under low, high and
extreme traffic loads. In addition, the proposed approach achieves robustness against failure and
also has scalability in different network sizes and outperforms typical conventional approaches.
3. THE PROPOSED APPROACH
In previous section, we presented a literature review on congestion controller mechanisms in
WSN. In this section, we briefly discuss our fuzzy-based congestion controller. Fuzzy system is
used in order to increase accuracy of the congestion controller system. Fuzzy controllers convert
crisp inputs to fuzzy based inputs, then by using a rule base, fuzzy system determines an action as
the main output. The basic idea of a fuzzy controller is presented in Figure 2.
Figure 2. Fuzzy Controller System Functionality
In presented approach, fuzzy logic controller is considered as the kernel of the algorithm. It is
associated with the Monitor Nodes (MNs) to detect congestion, based on information which
comes from the network. In the propose approach, MN continuously monitors its grid and fetches
fuzzy controller’s metrics. Therefore, the fuzzy system calculates congestion level in each grid
continuously and dynamically. We describe our fuzzy-based congestion controller in three phases
that are as follows:
1) Congestion candidate generation
2) Congestion identification
3) Generating new phase
We discuss each phase briefly in the rest of this section.
3.1. Congestion Candidate Generation
After the sensors establishment, the sink uses nodes geographically information in order to divide
network to some equal grids. In some applications of WSN, sensors deployed randomly in order
to monitor environment and as the grids are equal in area, number of sensors in each grid maybe
different from other grids. After dividing the network into grids, sink allocates an ID to each grid
called Grid_ID. Then, it sends the Grid_ID of each grid to its members. The next step is choosing
the Monitor Node (MN). Responsibility of the MN is to monitor its own grid continuously in
order to detect congestion. Sink choose the sensor with highest reminded energy as the MN. In

the case that two sensors have the same reminded energy, the nearest sensor to the sink is chosen
as the MN. The MN needs to perform some monitoring works that consumes energy. Therefore,
the sensor with highest reminded energy is chosen as the MN.
The presented fuzzy system uses each grid’s density in order to detect congestion areas.
Therefore, the first duty of each MN is to calculate its grids density. Whoever, at the beginning,
sink knows the number of sensors in each grid, but in the case that a sensor dies as lack of energy,
the MN must recalculate its grids density. When sensors energy reaches to the alarm level, it
sends a packet for MN and aware MN of its death. The alarm levels energy is enough just for
sending a packet to MN.
In order to calculate grids density, the MN put Grid_ID in a packet and broadcast it for its
neighbours. Each sensor that receives the packet compares its own Grid_ID with the received
Grid_ID. If both are the same, the sensor sends a replay to the MN and then rebroadcasts packet.
Sensor drops the packet, either if it received the same packet before or if packet’s Grid_ID is not
the same as node’s Grid_ID.
Each MN sends the density of its grid to the sink. After receiving all grids densities, sink
calculates the average of densities and uses it as a threshold for defining congested areas. Each
grid with higher density than the calculated threshold is marked as the congested candidate area.
If any change in number of sensors happened, the MN sends its new density for the sink. Sink
recalculates the new threshold and updates congested candidate regions. High density regions
have higher congestion probability and higher collision and packet lost rate. A summary of this
phase is shown in Figure 3.
Figure 3. Steps for Congested Region discovery
3.2. Congestion Identification
In each congestion candidate grid, the MN has responsibility for detecting the congestion. In each
MN there is a fuzzy-based system, which controls the congestion in grid and balance traffic in
order to reduce the congestion. Fuzzy system uses three different metrics in order to decrease
congestion and reduce its effects. Fuzzy controller for each MN is shown in Figure 4. MN
continuously calculates three input parameters and then uses fuzzy logic to determine specific

level of congestion. Based on the congestion level, either packets forward through the grid or
through the relay nodes.
Figure 4. The Proposed Fuzzy Logic Controller
Three fuzzy input parameters are as follows:
ETD (Extended Transmission Delay): The ETD metric is the transition time that is required to
transmit a packet to the next hop. The metric is calculated using the convex combination
presented in Formula.1.
= Formula.1
In this formula, Delay(t) determines delay in time t, and α is weight value. α is used to determine
the priority of delay or previous ETD.
Grid’s Density: High density of nodes in each grid can cause more congestion. In addition, grids
density maybe increased or decreased during network lifetime. The reason is that a sensor may
die or new sensors may add to the network. Number of sensors in each grid has a direct effect on
congestion. Therefore, number of sensors is an important metric in the proposed fuzzy system.
DPC (Dropped Packets): Dropped packets, refers to number of lost packets in each grid. Packet
loss is the result of congestion. Therefore, increasing packet loss means increasing congestion.
These parameters have membership functions that are presented in Figure 5. By use of these
membership functions, a rule base is designed for fuzzy system. When MN calculates inputs,
fuzzy system applies them in the rule base. Fuzzy system output determines three different
actions that are as follows:
Action1) relay all packets through the grid
Action2) relay half of the packets through the grid and half through the relay nodes
Action3) relay all packets through the relay nodes
We will discuss the relay nodes and these actions in the next phase. Using presented mechanism
in this phase, fuzzy controller detects congestion and try to reduce the packets in the congested
area.

a) Grids Density b) ETD c) Dropped packets
Figure 5. Membership Functions Diagrams
3.3. Generating New Path
When fuzzy system determines the action, the MN performs the specified action. If there is no
congestion, output will be action1. In this case, there is no need for the MN to perform any action.
But for other two actions, the MN must informs its grid’s Border Node (BN) about the congestion
level. At the first time, the MN sends a packet to the sink and asks for its BN’s ID. The sink
selects the last node in grid as the BN. The detail of selecting BN is presented in [13]. Figure 6
shows the position of BNs and MNs in the network. After selecting the BN, the sink informs the
MN of its grids BN. BN has the responsibility of relaying the packets either through the grid or
through the relay nodes. The BN relay packets based on action level and this will continue until
congestion level in the grid reduce to action1. When the BN gets informed of any changes in
congestion level by the MN, it sends packets through relay path based of congestion level. Since
radio frequency of relay nodes is different from ordinal sensors, sending packet using relay path
has no effect on congestion in the grid. Sink selects the best relay path for BN based on [13].
In this section, we briefly discussed our proposed approach. A summary of the proposed approach
is shown in Figure 7.
Figure 6. Example of BN and MN Assignments

4. SIMULATION RESULTS
To show the advantages of the proposed approach in compare with the Base Work (BW) [12],
both approaches implemented using Opnet Modular 14.5 simulator. The simulation was
performed using a WSN of size 300 m* 300 m. Table 1, lists simulation parameters used in our
study.
Both the BW and our approach were implemented using two different scenarios. For both
scenarios two parameters have been measured and evaluated. These parameters are as follows:
Packet delivery: number of packets reached to the sink
Number of dropped packets: number of packet lost because of congestion in congestion
candidate grids.
Table 1. Simulation Parameters.
Parameter Value
Simulation duration 180 sec
Number of nodes 26
Transmission range 20 m
Traffic type CBR(UDP)
Packet rate 2 packets/sec
Data payload 512 byte/packet
Number of relay nodes 6
The aim of the proposed fuzzy congestion controller is to detect and avoid increasing congestion
in congestion candidate regions. When a packet is lost, it should be retransmitted by the source
node which wastes node’s energy and the network bandwidth.
Figure 8 presents packet delivery rate for both approaches.
Figure 7. The Proposed Approach

Figure 8. Packet Delivery Rate
In this figure horizontal axis refers to time and the vertical axis, to number of packets delivered to
the sink. This figure shows the packet delivery ratio in the sink. Firstly in this figure, both curves
have overlap as our fuzzy mechanism hasn’t detected congestion yet. As fuzzy controller uses
three metrics to detect congestion, after a while, it detects a level of congestion and starts to lead
packets toward relay nodes, based on congestion level. As number of packet lost increased, BW
starts to send all packets through relay nodes. Therefore, after a while both approach send packets
through relay nodes and packet delivery ratio will be the same in the sink for both approaches.
Figure 9 shows the packet lost figure for both approaches.
Figure 9. Packet Loss in Congested Grid rate
Vertical axis refers to packet lost in congested grid, and horizontal axis refers to time. Like Figure
8, in this figure, firstly packet lost in both approaches is the same. After increasing packet lost,
fuzzy system detects congestion and according to congestion level, it sends packets through either
congested grid or relay nodes. As a result, congestion and packet lost in congested candidate grid
will be decreased. Whoever, after a while, as both approaches relay packets through the relay
nodes, packet lost is equal in both approaches.
Referred to simulation results, proposed approach increases packet delivery ratio in the sink and
decreases packet lost in congested grids. In addition, fuzzy system controls the congestion and
avoids growth in congestion. Also, fuzzy system is more flexible as it continuously fetches

metrics from the network. By increasing the number of curves in membership functions of input
metrics, the accuracy of system increases. As a result, number of actions and control ratio
increases. When fuzzy system reaches to acting level, proposed approach controls congestion and
helps the grid to balance traffic by leading some packets through relay paths.
5. CONCLUSION AND FUTURE WORK
Wireless Sensor Network (WSN) is a set of sensor nodes, which are distributed in an area. In
WSN all sensor nodes sends their packets hop-by-hop to the sink, therefore, nodes nearer to the
sink should forward their own packets and other nodes packets. This feature of WSN, increases
congestion and packet lost in nodes, especially in nearer nodes to the sink. Using congestion
controller mechanisms can avoid or control congestion in WSN. As a result of decreasing
congestion, packet loss and energy consumption of the nodes will decrease. In this paper, we
proposed a novel approach based on fuzzy logic to control and balance congestion in grid-based
WSN. Presented approach uses three parameters that dynamically and continuously fetches from
the network in order to detect congestion regions. When fuzzy system detects a level of
congestion it relay some packets out of the grid in order to control and balance congestion. As our
future work, we decided to make our work more energy aware as energy is the most important
parameter in WSN.
REFERENCES
[1] L.A. Villas, A. Boukerche, H.S. Ramos, H.A.B.F. de Oliveira, ”DRINA: A Lightweight and Reliable
Routing Approach for In-Network Aggregation in Wireless Sensor Networks”, IEEE Transactions on
Computers, (Volume:62 , Issue: 4 ), 2013.
[2] R. Annie Uthra, S.V. Kasmir Raja, ” Energy Efficient Congestion Control in Wireless Sensor
Network”, Recent Advances in Intelligent Informatics Advances in Intelligent Systems and
Computing Volume 235, pp 331-341, 2014.
[3] Sh. Borasia, V. Raisinghani,” A Review of Congestion Control Mechanisms for Wireless Sensor
Networks”, Technology Systems and Management Communications in Computer and Information
Science Volume 145, pp 201-206, 2011.
[4] C. Karakus, A.C. Gurbuz, B. Tavli,” Analysis of Energy Efficiency of Compressive Sensing in
Wireless Sensor Networks”, Sensors Journal, IEEE (Volume:13 , Issue: 5 ), 2013.
[5] C. Sergiou, V. Vassiliou, ” Study of lifetime extension in wireless sensor networks through
congestion control algorithms”, IEEE Symposium on Computers and Communications (ISCC), 2011.
[6] J. Zhao, L. Wang, S. Li, X. Liu, ” A Survey of Congestion Control Mechanisms in Wireless Sensor
Networks”, in Sixth International Conference on Intelligent Information Hiding and Multimedia
Signal Processing (IIH-MSP), 2010.
[7] C. Wang, B. Li, K. Sohraby, M. Daneshmand, “Upstream congestion control in wireless sensor
networks through cross-layer optimization". IEEE Journal on Selected Areas in Communications,
(Volume:25 , Issue: 4 ).
[8] U.S. Visweswaraiya, K.S. Gurumurthy, ” A Novel, Dynamic Data Dissemination [D3] Technique for
Congestion Avoidance/Control in High Speed Wireless Multimedia Sensor Networks”, Fifth
International Conference on Computational Intelligence, Modelling and Simulation (CIMSim), 2013.
[9] E. Hajian, K. Jamshidi, A. Bohlooli, “Improve energy efficiency routing in WSN by using
automata”, International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol. 1, No.2,
2010.
[10] Li. Zilong, W. Zou, T. Qi, ” A cross-layer congestion control strategy in wireless sensor network”, in
4th IEEE International Conference on Broadband Network and Multimedia Technology (IC-BNMT),
2011.
[11] P. Antoniou, A. Pitsillides, T. Blackwell, A. Engelbrecht, L. Michael,” Congestion control in wireless
sensor networks based on bird flocking behavior”, Computer Networks Volume 57, Issue 5, 7 April
2013, Pages 1167–1191.

[12] H. Cha, K. Kim, S.Yoo, ” A node placement algorithm for avoiding congestion regions in wireless
sensor networks”, in Third International Conference on Ubiquitous and Future Networks (ICUFN),
2011.
[13] J. Mena, V. Kalogeraki, ” Dynamic Relay Node Placement in Wireless Sensor Networks” in
International Symposium on Applications and the Internet, SAINT ,2008.
AUTHORS
Ali Dorri received his B.S. degree in computer engineering from Bojnord University,
Iran, in 2012, and now is student in M.S in software engineering in Mashhad branch,
Islamic Azad University, Mashhad, Iran. His research interests cover Wireless Sensor
Networks (WSN), Mobile Ad hoc Network (MANET) and specially Security challenges.
Dr. Seyed Reza Kamel Tabbakh is with the Department of Software Engineering,
Faculty of Engineering, Islamic Azad University - Mashhad branch, Mashhad, Iran. He
received his PhD in communication and network engineering from University Putra
Malaysia (UPM) in 2011. He received his BSc and MSc in software engineering from
Islamic Azad University, Mashhad branch and Islamic Azad University, South Tehran
branch, Iran in 1999 and 2001 respectively. His research interests include IPv6 networks,
routing and security. During his studies, he has published several papers in International journals and
conferences.

A fuzzy based congestion controller for control and balance congestion in gried-based wsn

More Related Content

What's hot (18)

Viewers also liked (20)

Similar to A fuzzy based congestion controller for control and balance congestion in gried-based wsn (20)

Recently uploaded (20)

A fuzzy based congestion controller for control and balance congestion in gried-based wsn