A Wireless Sensor Network (WSN) is a self-organized network that has a collection
of numerous sensor nodes which are haphazardly distributed in a sensor field.
WSN is recently coming up as a premier technology due to its usefulness in wild
life monitoring, military and health applications (Ruan
and Sun, 2011). Various characteristic issues which influence the performance
of a sensor network are fault-tolerance, scalability, operating environment
and topology (Chu et al., 2011). Wireless sensor
network are resource scavenging in nature (Liang et al.,
2011). The communication architectures in WSN in general based on layering
and clustering concepts. Clustering provides a mean of achieving enhanced lifetime
(Akyildiz et al., 2002). Past literature in fault
tolerant and energy aware frameworks (Saleh et al.,
2006) produced good results in terms of throughput and end to end delay
measurement. But lack in prolonging the lifetime. This may be due to Loss of
Data Avoidance (LoDA) and Lack of Power Avoidance (LoPA) (Dhulipala
et al., 2010). When the lifetime enhancement of the network of sensor
nodes is concerned, it is important to develop heuristic approaches for power
and energy calculations. Heuristic approach in this regard could be reasonably
a good solution.
Many clustering protocols or algorithms have been proposed for WSNs in recent
years. The Low-Energy Adaptive Clustering Hierarchy (LEACH) algorithm, which
is based on gradient cluster, plays a great role in reducing the energy consumption
of the nodes and enhancing the network lifetime (Yang and
Zhang, 2009; Xin et al., 2008; Wang
et al., 2011). However, owing to selecting the cluster heads stochastically
in disregard of the size of clusters and the residual energy of nodes and adopting
the one-hop forwarding model in inter-cluster communication, it cannot guarantee
the even energy dissipation of the network (Koushanfar et
al., 2004; Paradis and Han, 2007). A further
improvement of this algorithm known as LEACH-C is proposed (Saleh
et al., 2006), in which the cluster formation is done using a centralized
mechanism by the BS. Results analyzed (Saleh et al.,
2006) have shown that the overall performance of LEACH-C is better than
LEACH due to improved cluster formation by the BS. A new application of the
Particle Swarm Optimization (PSO) algorithm to the problem of clustering nodes
(Gao et al., 2009). The number of nodes and candidate
cluster-heads in each cluster are equalized in the algorithm, with the objective
of minimizing the energy expended by the nodes while maximizing the total data
gathered. However, no comparison with other benchmark clustering protocols in
terms of energy efficiency has been addressed. An energy-aware clustering for
WSNs using PSO algorithm, which is implemented at the BS, is proposed in energy
aware fault tolerant management framework (Dhulipala et
|| Fault classification based Fault Tolerant Framework for WSN
The cost function is defined with the objective of simultaneously minimizing
the intra-cluster distance and optimizing the energy consumption of the network.
Compared with LEACH and LEACH-C, the protocol can achieve better network lifetime
and data delivery at the BS. However, owing to adopting the one-hop routing
in inter-cluster communication, the protocol may cause unbalanced energy dissipation
of the network, especially when the BS is outside of the network. That is, the
cluster heads farther away from the BS are prone to consume more energy to transmit
data to the BS and die quickly (Guo et al., 2011).
The Unequal Clustering Size (UCS) clustering algorithm balances the energy dissipation
through controlling the size of the cluster (Jiang et
al., 2010) and this scheme can prolong the network lifetime.
Energy-balanced unequal clustering (EBUC) routing protocol for periodical environment
monitoring applications in WSNs (Jiang et al., 2010)
concentrating on the lifetime enhancement of the network of nodes. It organizes
the network via unequal clustering and multihop routing. By using Particle Swarm
Optimization (PSO-C) algorithm. The WSN should also be Fault tolerant and reliable
in case of environmental monitoring. Thus the cluster heads of the particular
cluster will consume lower energy during the intra-cluster data processing and
can preserve some more energy for the inter-cluster relay traffic. It also acts
as a Fault tolerant network as proposed in Fault classification based Fault
Tolerant Framework beneath. Simulation results show that LoPA saves the energy
awareness of the individual nodes and balances the energy consumption over the
entire network and enhances the network lifetime.
Fault classification based Fault Tolerant Framework for WSN: In this study we proposed a fault classification based fault tolerant framework for WSN which is simulated with a unequal clustering of sensor nodes as shown in Fig. 1.
In Wireless sensor networks, fault-tolerant and energy efficiency are two important
factors. Some link failure may happen during data transmission and some threat
can come from compromised nodes, which might relay incorrect information (packet)
to the next node during routing. Energy is the major concern in networks. The
energy may gets wasted due to collisions, unnecessary traffic, long idle time
etc., In WSN, generally we classify the possible fault as power lagging (fail-stop),
congestion, link and hardware failures (Koushanfar et
al., 2004). The failures are totally differ from each other; it wont
affect or interfere with each other. But it will affect the performance and
functions of the nodes. The first failure can occur due to power lagging of
the nodes. Second, congestion may occur due to large number of nodes present
in the network that nodes transition from power saving state to active state
(Paradis and Han, 2007). The collisions are high, when
large amount of packets during transmission from source to sink. Link failure
is the third one, the result is data loss. It breaks the fine communicated environment
or network and causing changes in topology. The final is hardware failure; it
may be predicted or unpredicted one. The solution is replacement. Replace the
node by manual or automatic depend upon the application or the requirement of
Lack of power avoidance (LoPA): We analyse two mechanisms Lack of Power
Avoidance (LoPA) and Loss of Data Avoidance (LoDA) as shown in Fig.
1. These two algorithms the simulation results of these algorithms proved
good in reliable communication. LoPA fully concentrate on power energy and reliability
the nodes (De Cicco et al., 2008). But this gives
the reliable and secured packet routing to the end user. The fault tolerates
mechanisms handles three types of failure in wireless network. When fault occur,
it deviate the network from the normal flow (Jiang et
al., 2010). To tolerate/recover the fault, come back into the communication
to an original one.
|| Pseudo code for LoPA
LoPA, check the battery level of the sensor and act accordingly. The faults
are mainly worsened by the multihop communication nature of the network. Single
node or link failure leads to miss behaviour or missing reports to the destination/nearby
nodes. The fail-stop failure occurred due to power lagging.
Pseudo code for LoPA algorithm is shown in Fig. 2. The main
parameters are power (P), time (t) and energy (E) are to be initialized. The
process of operation starts, vary the time with constant power. We assume that
each sensor nodes knows the percentage of remaining power level of its battery.
Consider the threshold value into three different conditions like 75% (T1),
50% (T2) and 25% (T3) of power is indicated in Fig.
4. Check the battery level of the node under three boundary conditions.
If the energy level is higher or equal to 75% of threshold value (T1),
then the sensor is act like a sink. It directly routing the packets successfully.
Else check the next to this, if the condition is satisfied (E = T2),
the sensor cannot be a sink. It acts an operational sensor that participates
in any communication or sensing activity. While the same condition checking
has continued till the packets sending successfully or reaches to the destination.
But the last condition takes more time for its operation. If the above two conditions
are not satisfied, we consider that the node is having low battery level. First
examine the battery level of all sensor nodes and compare the nodes energy
level with another one. We assume that each cluster having only four nodes.
Let FS, SS and TS be the energy level of the
first sensor, second sensor and third sensor nodes. Finalize the higher energy
node (BIG) that should be greater than the previous. If yes, packets are routed
through this higher node. If not, we assume that all nodes having lower energy
(LE) or otherwise it gives the default output Lacking Strength
Lacking Strength- it loses all power to routing the packets. The
main aim of this LOPA is routing the packets successfully, avoid packet loss,
improve the reliability of communication and increase throughput of the sensor
node (Yen et al., 1998; Campanoni
and Fornaciari, 2007; Polastre et al., 2004).
Preliminaries of the network model and simulation parameters
||All sensors are deployed in the region of interest
||All nodes are energy constrained.
||The nodes are either a cluster head or ordinary head
||All nodes are static in nature
Radio energy model: The energy model for the sensors is based on the
radio model (Jiang et al., 2010).The radio model
equation is given by:
||No. of bits
|| Energy dissipated per bit to run the transmitter or the receiver circuit
||Energy dissipated during receiving data
||Energy dissipated per bit to run the transmit amplifier based
on the distance between the transmitter and receiver.
The simulation parameters for the evaluation of the LoPA algorithm shown in Table 1. Terrain dimensions for the evaluation are fixed as 500x500 m2 and network sized is 200 nodes. Base station for the clustering environment located at 250x750 m2. The experiment is conducted with energy considerations as shown in Table 1.
|| Simulation results and network parameters
RESULTS AND ANALYSIS
The scenario of the network model is simulated by considering the conditional utilization of the energy as shown in the Fig. 3. The graph in Fig. 4 takes time along X-axis as a measure of simulation specification and energy in Joules for performance evaluation along Y-axis.
Figure 3 explains that decreasing amount of nodes energy by the variation of time period of the nodes performance. The operation of the nodes differs from the battery level of the sensor. This condition depends completely on the transmission and reception of the sensors activity i.e., by checking the condition with threshold values. The sensor drops its energy after a long life time as a result the node reaches the state of death.
Energy decaying process as the variation of nodes alive and dead condition analysis starts is depicted in the Fig. 4. It also represents the characteristics of battery which V0→V1 for setting the threshold T1, V1→V2 for setting threshold T2 and V2→V3 for setting threshold T3.
The lifetime of the battery is represented as the equation:
Deployment and evaluation of performance of LoPA: The random deployment of sensor nodes in the terrain considered influences the sensor network performance in terms of Quality of Service (QoS) metrics. Figure 5 shows the random deployment scenario of sensor nodes set for this experiment.
Voronoi diagram represents the points that are connected in a region having
common boundary. Voronoi diagram helps in solving numerous geometric problems.
The common boundary of two Voronoi regions is called as Voronoi edge, The Voronoi
edges are called as Voronoi vertices. Voronoi diagram represents the clustering
environment of the proposed algorithm. The Voronoi diagram representing the
networking scenario as shown in Fig. 6.
Performance evaluation of the proposed algorithm is done using MATLAB protocol,
LEACH, PSO-C, LoPA and their performance are being compared.
|| LoPA conditional utilization of energy
|| Energy variations based on threshold
|| Random deployment scenario of sensor nodes
Initially 200 nodes we randomly deployed in 500x500 m2 terrain.
The data message size was 2000 bytes with 50 packet header. The energy efficiency
of the three algorithms with respect to network lifetime is examined.
|| Voronoi diagram for network model
||Alive nodes (vs) simulation rounds
|| Distribution of dead nodes and alive nodes
Figure 7 illustrates the improved network lifetime with the fault tolerant ability in case of LOPA-LEACH over the other two algorithms. The hotspots problem can be avoided causing a reliable communication to the network.
The dead nodes are equally distributed in terrain dimension so that the network
life time is increased as shown in Fig. 8. The load on the
network is equally distributed which avoids overloading of nodes near the base
station. The dead nodes in other two algorithms are concentrated in particular
region of terrain due to unequal load balancing strategy.
In this study, we have presented the results obtained for Lack of Power Avoidance (LoPA) using the heuristic algorithmic concept of fault classification based fault tolerant framework. The heuristic algorithm introduced in this paper could perform better compared to earlier works in reliable energy aware communication in WSN such as LEACH and PSO-C. It is found from our simulation experiments that the number of alive nodes is good over dead nodes. This situation can better prolong the lifetime of sensor nodes leading to reliable inter cluster communication. Our future work (in progress) aims at scalability and trustworthy issues for reliable communication in WSN and also to provide detail of the real time implementation.