Modified Coverage Hole Detection Algorithm for Distributed WSNs

Full Paper
Proc. of Int. Conf. on Advances in Recent Technologies in Electrical and Electronics

Modified Coverage Hole Detection Algorithm for
Distributed WSNs
K. Lakshmi Joshitha1 and Dr.S.Jayashri2
1

Research Scholar, Anna University, Chennai, India
Email: lakshmijoshitha@yahoo.com
2
Director, Adhiparasakthi Engineering College, Chennai, India
Email: jayaravi2010@gmail.com
performance of the sensor network. Number of holes is
inversely proportional to efficient communication.

Abstract— Wireless sensor networks (WSNs) are spatially
distributed sensors that find wide applications in various fields
such as environmental control, Medicine and Health care,
Military surveillance etc. The sensing and communication
within the network should be effective for such applications.
Holes are the voids created in the network when accidental
death of nodes is caused due to technical
or improper
coverage. The detection of the holes becomes essential after
the random deployment. The main objective of the work is to
detect the coverage holes using computational geometry
approach which uses co-ordinates of the sensors and to
implement it in the hardware. The communication range of a
node is considered to be equal to its sensing range. The protocol
is designed for irregular domain which is a real time scenario
and takes the help of two-hop neighbors’ of a node to detect
the hole around it. The proposed system also allows only few
nodes to initiate the detection algorithm so that the energy
and time is conserved.

II. RELATED WORK
In wireless sensor networks, determination of hole is of
prime importance and several authors have proposed different
types of algorithms for different networks like static ,hybrid
and mobile sensor networks[1].The discussion in [2] includes
the detection of coverage holes based on the Voronoi graph,
and development of algorithms to make the sensor networks
as uniform as possible. The basic triangulation method of
sensor networking of 2D and 3D surface is dealt in [3]. A gridbased system model [4] is proposed to detect the coverage
hole. A two phase sensor relocation is also proposed.. The
hole detection methods in [5] proposes optimal polynomial
time worst and average case algorithm for coverage
calculation. Reference [6] deals with the method to identify
one, two or a larger single coverage hole and suggests the
mobility scheme to provide coverage. Localized voronoi
polygon and neighbor embracing polygon method is used
for identifying hole in [7].Deterministic method for boundary
node detection based on localized Voronoi polygons is
proposed in [8]. The algorithm uses only the one-hop
information with fewer computations.
The proposed algorithm MHDA (Modified Hole
Detection Algorithm) is localized one and requires one-hop
neighbors’ information to detect the holes. A distributed so
called path density (PS) algorithm [9] is developed to detect
the coverage holes in WSNs. It uses the path density to
detect the holes by the neighbors of a dead sensor. The PS
algorithm can detect coverage holes remotely, but requires
more time and power consumption for detecting holes in
practice. Reference [10,11] gives details about the statistical,
topological and the geometrical approaches of the sensor
network and gives the Hop based hole detection process.
A k-coverage verification scheme for a target field is
proposed, which requires a predefined value of k. According
to this scheme, each node has only localized distance
information of the distance between adjacent nodes in its
vicinity and their sensing radius. Besides, an upper-bounded
sensing radius and lower-bounded transmission radius of
each node is considered. As per the proposed scheme, this
could be achieved by configuring the nodes before placing
them or by walking along the target field boundary with a
handheld device. However, it is quite difficult due to random

Index Terms— Coverage holes, computational geometry
approach, communication range, sensing range, two-hop
neighbors.

I. INTRODUCTION
In cases like volcanic regions, forests areas or monitoring
physical conditions of a region, wireless sensor networks
plays an important role which collects the physical parameters
and converts it into electrical parameters with the help of
sensors. In wireless sensor networks (WSNs), sensors are
distributed randomly to form the network dynamically
without help of any infrastructure. The distributions of
sensors are not usually uniform due to random aerial
deployment, presence of obstructions, and node failures
caused by power depletion. In the post deployment
scenarios, nodes deployed over certain region may be
destroyed due to intrusion, explosion or environmental
factors like heat, vibration, and failure of electronic
components or software bugs. In another scenario, power
sources of the nodes may lead death of the nodes, thereby
affecting the coverage of the original network giving rise to
holes.
Holes are ‘communication voids’ as they act as an
obstacle for communication. Hence, holes can hardly be
avoided in wireless sensor networks. On the other hand, holes
are important indicators of the general health of a sensor
network. The presence of holes in the underlying geometric
environment could have important consequences on the
© 2013 ACEEE
DOI: 03.LSCS.2013.5.32

32

Full Paper
Circumcenter (Z): Center of the circum circle formed by
location of any three sensors as the vertices of a triangle is
called circumcenter [fig.1].
System framework: In our system A
model, it is assumed that there are multiple coverage holes
in the monitoring region and the network is connected.
Therefore, the one-hop and two-hop neighbors of a node
must be connected with it through one and multi-hops,
respectively.

deployment nature of the sensors and geographical condition
of the monitoring region.
III. PROPOSED SYSTEM
The proposed system has few nodes initiating the algorithm for specific region while the existing system includes
all the nodes in the network to initiate the hole detection
algorithm. Clustering of the given nodes is done by choosing the cluster size of 16.Here the clusters are of overlapping
nature which will eliminate the missing of any hole during the
detection process. For each region the reference node is selected randomly. Thus the energy required for detecting holes
is reduced. Further a comparison of time consumed and the
number of hole detected by both existing and proposed algorithm is also given [fig3,4]. The designed algorithm (MHDA)
Modified coverage Hole Detection Algorithm is also implemented in real-time with a network consisting of three nodes.
IV. SIMULATION
Wireless sensor networks are the one where nodes are
deployed randomly over the monitoring region such that some
part of the network has sufûcient coverage due to the
presence of several redundant nodes, whereas other parts
may have coverage holes due to absence of sensor. As
soon as the network is formed, each node knows its location
information and collects its one and two-hop neighbours’
list. The sensing range (Rs) is equal to the communication
range (Rc) and each node knows its location information via
GPS or any location information system. Each node collects
its two-hop neighbours’ location information as soon as the
deployment is over, i.e. each node knows the location of its
neighbours within the range of 2 Rc .

Fig.1 Circum radius (R) and circumcenter (Z)

For example, as shown in [Fig. 2], if A is considered as a
reference node, it is connected with other nodes of the network with help of its one and two-hop neighbors though
coverage hole exists in the network. In fact, B and D are onehop neighbors of A. Hence, A is connected to C and E through
its one- hop neighbors B and D, respectively.

A. Definitions
The important parameters that are to be considered in the
algorithm are
Sensing range (R S): Sensing range of a node is the
circular disk of radius Rs, which is centered at its location. It
is the sensing range of the sensor used.
Communication range(RC): Communication range of a
node is the circular disk of radius RC, which is centered at its
location. It is the antenna range of the communication device
used in the node.
Reference node (RN): A source node that initiates to
execute the hole detection algorithm is called a reference node
(RN) It is to be noted that a reference node ûrst collects its
neighbors information located within 2Rc and executes the
hole detection algorithm.
Neighbor: If A and B are any two nodes such that distance
between them, i.e.( d AB) d” RC , then A and B are one-hop
neighbors to each other. However, if RC <dABd”2 RC, A and B
are two-hop neighbors to each other.
Circumradius (R): Radius of the circum circle formed by
location of any three sensors as the vertices of a triangle is
called circum radius R.
1

Fig.2 Connectivity with holes in the network

Besides, C, E, F and G are two- hop neighbors of A as
they are within its 2Rc. Though, A is not connected with G
and F through its one-hop neighbors B and D, it is connected with them through its two-hop neighbor E, which is
connected to D. Hence, the whole network is connected
though there are coverage holes in it.
B. Collection Of Neighbor Information
In this phase, a reference node is selected randomly from
any part of the deployed region, which undergoes the
neighbor discovery phase. The reference node broadcasts
HELLO1 message that contains its location information. Upon
receiving the hello message, a node calculates the distance
d1 between the reference node and itself. If d1d”Rc, the node
sets itself as one-hop neighbor of the reference node X and
unicasts its location information and ID to the reference node

Asso Prof, Sri Sai Ram Engineering College, Chennai, India

© 2013 ACEEE
DOI: 03.LSCS.2013.5. 32

33

Full Paper
X. In the next step, each one-hop neighbors of the reference
node X, broadcasts theHELLO2 message that contains the
location information of the reference node X. Upon receiving
the HELLO2 message, each receiver calculates its physical
distance d2 from the reference node X. If d2 d” 2Rc, the sensor
sets itself as a two-hop neighbor of reference node Ultimately,
reference node X records the location information and ID of
each of its one and two-hop neighbors. This procedure is
executed by each node of the network in a distributed manner
by unicasting its location information and ID to the reference
node.

one-hop and another node from its two-hop neighbors. As
given in the hole detection algorithm, the presence or absence
of the hole depends on the nature of the triangle. In order to
justify the correctness of the algorithm analytically about
ten lemmas are proposed [12]. These Lemmas are summarized
in [Table I] .If acute triangle is formed by the reference node
with 1 hop neighbors with the maximum acute angle being π/
2 then there are no holes within 3 sensors.
Proof: The circumcentre will be located at most on one
side of triangle & R<Rs. Common sensing range also exists.
So circumcentre must be covered by any of 3 sensors.

C. Algorithm

E. Simulation Results
The simulation using mat Lab provides the distribution
of nodes is an area and the presence of hole in the region
based on the different lemmas provided in [Table I]

In the algorithm discussed Rs refers to the sensing range.
X is the Reference node (RN) initiating the hole detection
algorithm. N is the set of neighbors of X. (α ,β) is the
Location of X. Nu refers to set of neighbors whose ycoordinate e” β.Nd is the set of neighbors whose y- coordinate
< β. Area of the monitoring region is given as the input.
Values to be found are Length of sides of triangle a, b& c,
Circum radius R, Area of triangle “, Circum center Z, Angles
of the triangle.
1. Phase I: Divide the entire network into16 clusters each
having different size.
2. Select any node X randomly as RN
3. Find 1 & 2 hop neighbors of X & assign to set N.
4. Select nodes of y-coordinate e”β from set N &assign to set
Nu.
5. Arrange nodes of x-coordinate of Nu in ascending order
& assign to a new set N UX.
6. Select nodes of y-coordinate < β from set N and assign to
set Nd.
7. Arrange nodes of x-coordinate of Nd in descending order
& assign to a new set ND X.

TABLE I. LEMMAS FOR HOLE D ETECTION

Hop Neighbor

1

2

3

R=Rs

No
Hole

No
Hole

No
Hole

R>Rs

-

Hole exist

Hole
Exists

Acute Δ

R=Rs
R>Rs
Obtuse
Δ

Acute
angle

R=Rs

R>Rs

-

No Hole

R>Rs
R=Rs

Obtuse
angle

No
Hole
Hole
exist

Hole exist
-

Hole exist

No Hole
Hole
Exists
Hole
Exists

1. Phase 2: Select 1st 2 nodes of Ai & A j of N UX.
2. The x-coordinate of Ai must be less than A j.
3. Compute R & Z of triangle
X Ai A j
4. Check
X Ai A j is acute or obtuse.
5. If
X Ai A j is acute & Rd” Rs, no hole exists, else hole
exists around R N X.
6. If X Ai A j is obtuse & Rd” Rs, no hole exists.
7. Else check if Z is covered by sensor.
8. If Z is covered no hole exists, else hole exists.
9. Update N UX < N UX - {Ai}.
10. The phase 2 steps are repeated in loop with the
overlapping cluster until NUX is not equal to 1 for each node.
1. Phase 3: Choose 1st node Ai of N DX & balance node A j of
N UX.
2. Execute procedure from step 3 to step 8 of phase 2 for set
ND X
3. Update d Nx < ND X - {Ai}.
4. Continue procedure until ND X = null.

Fig.3. Comparison between the MHDA and the Existing Algorithm
for Hole Detection

D. Analytical Proof
Each sensor forms a triangle either with one pair of its
one-hop or two-hop neighbors or with one node from its
© 2013 ACEEE
DOI: 03.LSCS.2013.5. 32

34

Full Paper
Each mote consists of GPS, Zigbee and LPC2103
ARM7TDMI Microcontroller. In GUI the coordinates for the
3 motes are obtained by connecting UART cable to the laptop.
By varying the sensing range we can know whether hole
exists or not.

Fig.4. Comparison between the Time of Detection in MHDA and
the Existing Algorithm

V. REAL-TIME IMPLEMENTATION
The hardware implementation consists of 3 motes. Any
one mote is considered as the reference node. The reference
node collects the location of the other two neighbors. The
location of each mote is known using GPS. The neighbor
locations are transmitted to reference node using Zigbee with
the location information the reference mote executes the hole
detection process. The GPS &Zigbee are interfaced to the
mote using UART as shown in fig [5].To initiate Zigbee
transmission AT commands are used. The GPS updates the
position of each node for every finite interval of time. The
hole detection algorithm is loaded into the microcontroller
using KEIL software & is dumped in the microcontroller using
FLASH NXP software. Thus the hole detection process is
implemented & real-time results are obtained.

Fig.6.Output Window Of

V. CONCLUSION
The proposed system thus finds the existence of holes in
the network using computational geometrical approach with
less energy by considering the communication range Rc of
each node equal to the sensing range Rs. The proposed
protocol detects the coverage holes irrespective of any shape
or size of the monitoring region [14]. Holes are detected taking
into consideration the overlapping contours so that the
chance of missing a hole is negligible leading to the reliable
implementation of the proposed algorithm in any form of
monitoring region and therefore could be more useful and
beneficial as compared to the similar hole detection protocols.
The hole detection ensures data reliability and efficient
routing. The performance of the proposed system has been
examined through simulations with variable number of nodes
proving that the number of holes detected and the time taken
for the same are better in MHDA than the existing algorithm
especially when number of nodes deployed are more.
The hole detection algorithm can also be applied in the
topological network in future where energy consumption can
be decreased further. Appropriate clustering algorithm can
be used to perform the best clustering with sensor heads
initiating the algorithm in a cyclic or fixed patteren.A dedicated
device can be constructed to monitor the existence of all
types of holes throughout the entire lifetime of the wireless
sensor network. The device should serve the purpose of both
hole detection and hole healing making the communication
between sensor nodes more efficient.

Fig.5. Block Diagram of a Node used in Real time Implementation

© 2013 ACEEE
DOI: 03.LSCS.2013.5.32

Real-Time Implementation Showing
Hole Exists

35

Full Paper
[7] Erdun Zhao Juan Yao, Hao Wang Yating L V, A “Coverage
Hole Detection Method and Improvement Scheme in Wsns”.
Department of Computer Science. 2011 IEEE
conference.
[8] Chi Zhang, Yanchao Zhang, Yuguang Fang
“Localized
algorithms for coverage
boundary Detection in wireless
sensor networks”, Wireless Networks ,Volume 15,Issue
1,Pages 3-20,Jan 2009 .
[9] Zhang C, Zhang Y, Fang Y, “Detecting Coverage Boundary
Nodes In
Wireless Sensor
Networks”, Proceedings of
IEEE international Conference On networking sensing and
control,
2006.
[10] Corke P, Peterson R, Rus D “Finding Holes In Sensor
Networks,
Proceeding
Of
the
Workshop On
Omniscient Space : Robot
Control Architecture
Geared toward Adapting to Dynamic environments” At ICRA,
2007.
[11] I.M. Khan, N. Jabeur, S. Zeadally, Hop- based approach
for
holes and
boundary detection in wireless sensor
networks, IET Wireless Sensor systems, 2, (4), p.328-337,
IEEE 2012.
[12] Hwa-Chun Maa, C. Prasan Kumar Sahoo b,Yen -Wen Chen
“A computational Geometry Based Distributed Coverage Hole
Detection protocol for the Wireless Sensor Network” Journal
of Networks And Computer Application ,Volume 34,Issue
5,September 2011,Pages 1743-1756

REFERENCES
[1] Nadeem Ahmed, Salil S. Kanhere,
Sanjay
Jha,
The Hole Problem in Wireless Sensor Networks,
Communications review, Volume 1, Number 2, 2005.
[2] Wang G, Guohong Cao, Tom La Porta, “Movement-Assisted
Sensor Deployment”,IEEE Transactions on Mobile
Computing ,Volume 5,Issue 6,IEEE computer And
communications societies ,2006.
[3] H. Zhou, H. Wu, S. Xia, M. Jin, N. Ding, A distributed
triangulation algorithm for Sensor networks on 2D and
3D surface ,in: IEEE INFOCOM, Shanghai, China, April,
2011, pp.1053–1061.
[4] G. Wang, G. Cao, T. La Porta, and
W. Zhang, “Sensor
Relocation in Mobile Sensor Networks” of the 24th
[5] International Annual Joint Conference of the IEEE
Computer and
Communications volume 4, pages 23022312, March 2005.
[6] Meguerdichian S, Koushanfar F, Potkonjak M, Srivastava
MB, “Coverage Problems in Wireless Ad-hoc Sensor
Networks”,INFCOM 2001, Proceedings IEEE, Volume 3,p.
1380–1387.

© 2013 ACEEE
DOI: 03.LSCS.2013.5. 32

36

Modified Coverage Hole Detection Algorithm for Distributed WSNs

More Related Content

What's hot (18)

Viewers also liked (8)

Similar to Modified Coverage Hole Detection Algorithm for Distributed WSNs (20)

More from idescitation (20)

Recently uploaded (20)

Modified Coverage Hole Detection Algorithm for Distributed WSNs