Heterogeneous Energy Efficient Wireless Sensor Networks

Heterogeneous Energy Efficient Wireless Sensor Networks

Sapna Jindal

Electronics and Communication Engineering Department Vijay Arora

Applied Science and Research Department

ABSTRACT

Wireless sensor networks consist of nodes that transmit aggregated data to the base station by sensing the environment. In this paper we have used LEACH protocol on heterogeneity basis in which some nodes have higher energy as compare to other nodes. The higher energy nodes works as cluster heads which passes the information to the base station. Thus the lifetime of the system increases as the nodes having higher energy will die later. This system also performs data compression that helps in compressing the useless information and finally desired information is achieved on the base station.

Optimal no. of cluster heads are approximately five which is observed after simulation and also the nodes would die after a long time that would help in increasing system lifetime and would make a system more energy efficient. This simulation shows that energy can be saved approximately double as compare to previous protocol.

1. INTRODUCTION

   Wireless sensor networks (WSNs) uses nodes which contain memory and processors these nodes process the data and pass the collected information to the base station. Wireless sensor networks are used in various applications for example in vehicle tracking or in medical. The only problem faced by wireless sensor networks is limited energy of the nodes. This limited energy of the nodes limits the lifetime of the whole network. So in order to remove this problem we need efficient protocol which would help in increasing the lifetime of the network.

   All the protocols that are designed and implemented in WSNs should provide some real-time support as they are applied in areas where data is

   sensed, processed and transmitted based on an event that leads to an immediate action. A protocol is said to have real-time support if and only if it is fast and reliable in its reactions to the changes occuring in the network. It should provide useful information to the base station or sink using the data that is collected among all the sensing nodes in the network. The delay in transmission of data to the sink from the sensing nodes should be short, which leads to a fast response. The basic communication types considered send periodic data or event-driven data to the base station or to the sink. The other major type extracts data from a particular location or specific nodes or set of nodes. Routing protocols must fulfill these requirements along with energy conservation and focus on Quality of Service (QoS) factors. Such requirements can be fulfilled using clustering protocols. In clustering protocol, nodes are organized into clusters that communicate with their cluster head and the cluster head transmit the data to the base station.

   LEACH stands for low energy adaptive clustering hierarchy uses random distribution of energy to the nodes in which nodes forms a cluster randomly, nodes which have higher energy becomes cluster head and nodes which have lower energy gives the data to the cluster head after sensing the environment. These cluster head passes the collected data to the base station, it also performs data fusion which compresses the useless information and sends only desired amount of data which reduces amount of energy and thus increases the system lifetime. LEACH works according to homogeneous distribution of energy to the various nodes in the network. Due to which nodes those are cluster heads for a particular round die soon because of which the lifetime of the

   system reduces. This drawback is overcome by Heterogeneous LEACH protocol.

   Heterogeneous LEACH works on the principle of heterogeneity. In homogeneous each node utilizes same amount of energy while in heterogeneous each node uses different amount of energy. In wireless sensor network when we add more number of nodes in the network then these nodes are introduced with higher energy as compare to the other nodes in the network. We add more number of nodes in the network because lifetime of sensor networks are generally short due to which we add more number of nodes in the network which is basically performed by the heterogeneous LEACH. This simulation is based on Heterogeneous protocol.

2. Homogeneous LEACH Protocol

   LEACH uses random distribution of energy to the nodes in which nodes forms a cluster randomly, nodes which have higher energy becomes cluster head and nodes which have lower energy gives the data to the cluster head after sensing the environment. These cluster head passes the collected data to the base station,

   LEACH protocol works on randomization. Thus each node decides itself that whether it has to be cluster head for this round or not on probability basis. Now each cluster head passes a message to the other nodes in the network. Each node decides itself to which cluster head it wants to belong by searching the minimum transmission distance. Which reduces the amount of energy as the nodes passes the data to the nearest possible node.

   Energy saving also depends on the number of clusters in the network. If the number of clusters would be less then optimal number of clusters than non cluster head nodes will have to transmit the data on the longer distance which would increase the energy utilization also if the number of clusters would be more then the optimal number of clusters than

   would help in saving overall energy of the network. LEACH works in three phases: advertisement phase, set up phase and steady state phase. In advertisement phase, advertisement is passed to all the nodes. In set up phase where each node chooses its nearest cluster head to become the part of the cluster. In order to become the part of the cluster, node transmits its request to the cluster head in form of a message which consist of id of node and id of cluster head. After this cluster head passes a TDMA (time divison multiple access) schedule to each node of the particular cluster and accordingly collect the data from the other nodes. In steady state phase data is transmitted to the base station. Data is send into frames and each node sends data once per frame as it works according to the TDMA Technique.

3. Heterogeneous LEACH Protocol

Heterogeneous LEACH works on the principle of heterogeneity. In homogeneous each node utilizes almost same amount of energy and we can not add additional nodes any time while in heterogeneous each node uses different amount of energy and we can add nodes any time. In wireless sensor network when we add more number of nodes in the network then these nodes are introduced with higher energy as compare to the other nodes in the network. We add more number of nodes in the network because lifetime of sensor networks are generally short due to which we add more number of nodes in the network which is basically performed by the heterogeneous LEACH. Heterogeneous LEACH also works in three phase advertisement phase, set up phase and steady state phase.

Advanced nodes which are denoted as m and the additional energy factor between the advanced nodes and normal nodes is denoted by a. So these advanced nodes get more chances to become cluster head thus overall energy of the network increases.

more number of cluster heads will utilize more

Suppose that E0

is the initial energy of each

amount of energy as cluster heads work on higher

energy as compare to non cluster nodes. Thus we need optimal number of clusters in the network which

normal sensor. Th energy of each advanced node is

then

E0 (1+a). The total (initial) energy of the new

1. N sensors are uniformly dispersed within a square

heterogeneous setting is equal to:

n . (1- m) E0 + n . m . E0 . (1 + a) = n . E0 . (1 + a. m) 1

As a solution, assume heterogeneous LEACH Protocol, which is based on the initial energy of the nodes. This solution is more applicable as compared to any other solution which assumes that each node knows the total energy of the network and then adapts itself to become a cluster head according to its remaining energy.

4. Wireless Sensor Network Model:

There are three main components in a wireless sensor network; sensor nodes, sink and monitored events as shown in figure 1.

Base station

Base station

field S. The BS is situated far away from the square

field S.

1. All sensors and BS (Base station) are stationary after deployment. Each node knows the location of BS. Each sensor with enough energy can communicate with BS directly.

2. Sensors can use power control to vary the amount of transmit power depending on the distance to the receiver.

3. Communication is symmetric and a sensor can compute the approximate distance based on the received signal strength if the transmission power is known.

4. All sensors are location-unaware and are heterogeneous in terms of initial energy.

1. Channel Propagation Model in Wireless Sensor Networks

   In wireless sensor networks channels are basically communication path between the transmitter and receiver. In a particular channel, the electromagnetic wave propagation can be modeled as a power law function of the distance between the transmitter and receiver.

   Network

   Various sensor nodes

   We assume certain cross over distance (d0) and now we will compare the distance between the transmitter and receiver with this cross over distance. If this distance is less than cross over distance then friss free space model is used which is based on d2

   attenuation and if the distance is more than the cross over distance then the two-ray ground propagation model is used which is based on d4 attenuation. The cross over distance is defined as:

   Figure 1: Wireless Sensor Network Model

   With the help of this model following assumptions that are adopted as below:

   4

   do

   Lhrht 2

   Where L 1 is the loss in the system which

   is not related to propagation, hr is the height of the

   loss), 914 MHz radios, and =3*108 / 914*106 =

   receiving antenna above ground, ht is the height of

   0:328m. Using the above parameters, we have:

   do =

   the transmitting antenna above ground, and is the wave length of the carrier signal.

   If the distance is less than do, the transmitted power is attenuated according to the Friss free space

   86.2m and on putting do = 86.2m, equations becomes:

   then the received power equation is given by:

   P

   6.82 10-4 Pt

   d 2

   : d 86.2 m

   5

   PGG 2

   r

   r

   P (d ) t t r

   (4 d )2 L

   r

   3

   2.25 Pt

   d 4

   : d 86.2 m

   Where Pr (d ) is the received power where d is the distance between the transmitter and receiver, Pt is the transmitted power, Gt is the gain of the transmitting antenna, Gr is the gain of the receiving

   antenna, is the wave length of the carrier signal

   and L 1 is the losses in the network which is not related to propagation.

   If the distance is greater than do then the transmitted power is attenuated according to the two- ray ground propagation equation given by:

2. Radio Energy Model in Wireless Sensor Networks

   Energy is utilized in transmitting and receiving of data from node to the base station. Radio energy is used for both transmission and reception. So radio energy model is used in which transmitter consumes energy via radio electronics and the power amplifier and receiver consumes energy via radio electronics.

   PG G h h

   PG G h h

   2 2

   Pr (d ) t t r t r 4

   d 4

   Where Pr (d ) is the received power where d is the difference between transmitter and receiver,

   Figure 2: Radio energy model

   Pt is the transmitted power, Gt is the gain of the

   transmitting antenna, Gr is the gain of the receiving

   Thus in order to transmit a B-bit message over a

   distance d, radio energy at the transmitter is given by:

   antenna, hr

   is the height of the receiving antenna

   above ground, and ht is the height of the transmitting

   E (B,d ) E (B) E

   (B,d) 6

   antenna above ground.

   In order to calculate cross over distance we use an omni directional antenna with the following parameters: Gt = Gr =1, ht = hr = 1.5m, L =1 (without

   Tx Tx _ elec Tx _ amp

   BE

   • B

   d 2 : d d

   large. Therefore, the energy dissipated in the cluster-

   E (B, d )

   Tx

   Tx

   elec friss _ amp o 7

   4

   4

   head node during a single frame is:

   o

   o

   BEelec Btwo _ ray _ ampd : d d

   E BE N BE N

   Now radio energy at the receiver is given by:

   CH elec n

   B d 4

   BF n

   10

   two _ ray _ amp to _ BS

   ERx (B) ERx _ elec (B)

   E (B) BE

   where B is the number of bits in each data message and dto_BS is the distance from the cluster-head node to the base station. Each non-cluster-head node transmits its data to the cluster-head once during a

   8 frame. Now the distance between the non-cluster-

   Rx elec

   The energy utilized by a cluster during sending of the frame is:

   head node and its cluster-head is less than the cross- over distance, so the energy dissipation follows the Friss free-space model (e.g., d2 power loss).Thus, the energy used in each non-cluster-head node is:

   E E

   • N E 9

     E BE

   • B

   d 2 11

   cluster CH

   n non _ CH

   non _ CH elec friss _ amp to _ CH

   Where Ecluster is the energy utilized by a cluster during

   The energy dissipated in a cluster during the frame is:

   sending of the frame, ECH is the energy utilized by

   E E

   N E 12

   the cluster-head node during a single frame, Enon_ CH is the energy utilized by a non cluster-head node during a single frame, N is the number of nodes distributed randomly in an M*M region, n is the

   expected number of clusters per round.

   cluster CH n non _ CH

   Thus the total energy for the frame is:

   Etotal nEcluster

3. Optimum Number of Clusters in Heterogeneous LEACH Protocol

   B(E N E N n

   d 4 NE

   elec BF two _ ray _ amp to _ BS elec

   In heterogeneous LEACH, the cluster formation algorithm is used which makes sure that the expected

   N 1

   friss _ amp 2

   M )13

   2

   2

   n

   number of clusters per round is n. We need to define

   optimum number of clusters so as to minimize utilization of energy in the network. We can assume there are N nodes distributed uniformly in an M*M region. If there are n clusters, we get average N/n nodes per cluster. Each and every cluster head in the network consumes energy in order to receive data from various other nodes and then finally to transmit the aggregate data to the base station. The distance between the cluster head node and base station is

   The optimum number of clusters can be determined by setting the derivative of Etotal with respect to n to

   zero:

   dEtotal 0

   dn

   d

   d

   4

   two _ ray _ amp to _ BS

   N

   1 M

   2

   2

   )

   )

   friss _ amp ( 2

   Putting the values

   Eelec = 50 nJ/bit, EBF = 5

   2

   2 nopt

   nJ/bit,

   friss _ amp

   = 10 pJ/bit/m ,

   two _ ray _ amp =

   p>N

   N

   nopt

   2

   friss _ amp M 14

   d 2

   0.0013 pJ/bit/m4, N = 100, M = 100 m, n = 5, B = 4000

   bits and Rb = 1Mbps bits gives:

   E

   two _ ray _ amp to _ BS

   N frames/round

   start

   9mJ

   In this work, N=100nodes, M=100m, friss _ amp =10pJ,

   two _ ray _ amp =0:0013pJ, and 75< dto _ BS <185, so the expected optimum number of clusters are:

   tmsg 4ms

   t frame 80ms

   t 0.08 seconds * Estart

   1 n

   1 n

   opt

   6 15

   round

   9mJ

   When there is only 1 cluster, the non-cluster-head nodes will have to transmit data very far to reach the cluster-head node, and when there are more than 5 clusters, there is not as much local data aggregation

   Therefore, given the initial energy of the nodes, E start

   the cluster-heads and associated clusters should be

   being performed.

4. Total Energy consumed per round

E start

rotated approximately every 0.08*

0.009

seconds.

The total energy drained from each node per round depends on the average number of frames per round, Nframes/ round , as follows:

9 Radio Parameters

In Wireless Sensor Networks simulations are made with MATLAB. According to these simulations, we have assumed network which consists of 100 nodes which are randomly distributed. The base station which is the

E N E

CH /round frames/round CH / frame

E N E 16

E N E 16

non _ CH /round frames /round non _CH / frame

destination and is far from the network and the closest

distance which is taken is 75 meters from the nearest node. The bandwidth of the channel which is used to transmit the data is considered to be 1Mbps. The

where

ECH / frame is the energy to receive all signals

length of the data message is considered to be 500 bytes. The cross over distance which decides the

from non-cluster-head nodes, the energy to aggregate

the signals, and the energy to send the aggregate data to the base station which means total energy from cluster head to the base station, while

attenuation of signal is considered to be 87m. The energy in order to aggregate the data is considered to be 5nJ/bit. Also the size of the network is 100*100

square of meters.

Enon _ CH / frame

cluster-head.

is the energy to send a data to the

1. Simulation Results

   Simulations are performed by considering the 100 node network. Number of clusters are compared with the average energy consumed per round.

   Figure 3: Average energy dissipation per round in LEACH versus the number of clusters.

   Various nodes decide according to the probability that whether the particular node would become cluster head for this round or not.

   Figure 4: Cluster head nodes at time t.

   Figure 5: Cluster head nodes and cluster formation at time t.

2. Comparison between Direct Transmission and LEACH Protocol

   Direct transmission consists of nodes which directly transmit the information to the clusters. While LEACH protocol works according to clusters. Figure 6 shows the average amount of energy consumption verses number of cluster heads in the complete network. This graph shows that LEACH achieves approximately 7 times reduction in energy compared to direct communication, when using optimal number of cluster-heads. The main energy saving of LEACH protocol is due to combining compression with the data routing. There is a trade-off between the quality of output and the amount of compression achieved. In this case, some data from the individual signals is lost, but this results in a reduction of the overall energy dissipation of the system.

   Figure 6: Average energy dissipation versus percentage of cluster head nodes.

3. Energy Dissipation Comparison of heterogeneous- LEACH and LEACH Protocols

   We have discussed the energy consumption of various protocols related to LEACH protocol. But now the comparison is related to the heterogeneous LEACH. Now we will compare the performance of both protocols via graphs. As in LEACH, nodes transmit their data to the cluster head node during each frame of data transfer, and the cluster head aggregates the data and sends the resultant data to the BS. When the cluster head nodes energy is depleted, the nodes in the cluster lose communication ability with the BS and are essentially dead. Whereas, in hetero-LEACH protocol some of the nodes are of higher energy nodes (advanced nodes) that takes more time to get depleted, hence they hetero-LEACH protocol system has larger lifetime in comparison to LEACH.

   The results of hetero-LEACH simulations are shown in Figure 7 and 8 for m = 0.1 and a=0.9 with initial energy 0.5J and 1J respectively.

   Figure 7: Number of alive nodes using LEACH in the presence of heterogeneity with advanced nodes m=0.1 and normal nodes a=0.9 with initial energy 0.5J.

   Figure 8: Number of alive nodes using LEACH in the presence of heterogeneity with advanced nodes m=0.1 and normal nodes a=0.9 with initial energy 1J.

   We observe that LEACH takes some advantage of the presence of heterogeneity (advanced nodes), as the first node dies after a significantly higher number of rounds (i.e. longer stability period) compared to the homogeneous case (m = a= 0). The lifetime of the network is increased.

   The simulations are performed to determine the number of rounds of communication when 1%, 20%, 50% and 100% of the nodes die using direct transmission, MTE, LEACH and heterogeneous LEACH with each node having the same initial energy level. Once a node dies it is considered dead for the rest of the simulation. The nodes begin to die at a more uniform rate after about 20% nodes die. This is because the distances between the nodes become greater, and nodes have to become leaders more often causing the energy to drain rapidly. As can be expected, the number of rounds doubles as the

   3000

   2500

   Number of Rounds

   Number of Rounds

   2000

   1500

   1000

   500

   0

   Direct MTE LEACH

   hetero-LEACH

   1 20 50 100

   %age of Nodes Die

   energy/node doubles for a given size of network.

   Figure 9 and 10 shows the number or rounds the complete sensor network take until 1%, 20%, 50%, and 100% nodes die with initial energy per node of 0.5J and 1J respectively for a 100m x 100m network. It is clear that as the initial energy of the sensor node decreases then the nodes drain out quickly and follows the inverse square law.

   1500

   Direct MTE LEACH

   hetero-LEACH

   Number of Rounds

   Number of Rounds

   1000

   500

   0

   1 20 50 100

   %age of Nodes Die

   Figure 9: Performance results with initial energy 0.5J/node for a 100m x 100m network

   Figure 10: Performance results with initial energy

   1J/node for a 100m x 100m network

   The simulation shows that heterogeneous LEACH achieves:

   • Approximately 2x the number of rounds decreases compared to direct transmission, when 1%, 20%, 50%, and 100% nodes die for a 100m x 100m network.

   • Approximately 2x better than MTE for a 100m x 100m network.

   • As the energy level doubles the number of rounds approximately doubles for all cases.

   • Near optimal performance.

1. CONCLUSION

We have used clustering based algorithm in which load is equally distributed among the nodes. Heterogeneous protocol is used which uses additional nodes which are equipped in the network with higher energy these additional nodes becomes cluster heads and all other nodes becomes non cluster heads at any time. This protocol does not require any control information from the base station and distributes the energy in the network which is very effective in order to increase system lifetime. The Simulation results show that heterogeneous-LEACH is much more efficient in cmparison to direct, MTE and LEACH routing protocol. Heterogeneous-LEACH achieves approximately 2x reduction in energy as compared with LEACH protocol. Thus heterogeneous type of protocol can be applied to various applications which needs energy savings and increased lifetime of the system. Based on these simulations heterogeneous

LEACH provides energy dissipation, quality of the network, and lifetime of the system.

REFERENCES

1. A. Cerpa, J. Elson, D. Estrin, L. Girod, M. Hamilton and J. Zhao, (April 2001), Habitat monitoring: Application driver for wireless communications technology, ACM SIGCOMM Workshop on Data Communications in Latin America and the Caribbean

2. C. Schurgers and M.B. Srivastava,(2001) Energy efficient routing in wireless sensor networks, in the MILCOM Proceedings on Communications for Network-Centric Operations: Creating the Information Force, McLean, VA.

3. C. Li, M. Ye, G. Chen and J. Wu, (2005), An energy- efficient unequal clustering mechanism for wireless sensor networks, Proceedings of the 2nd IEEE International Conference on Mobile Ad-hoc and Sensor Systems.

4. C. Merlin and W. Heinzelman, (November 2005), A Study of Safety Applications in Vehicular Networks, Proceedings of the IEEE International Workshop on Heterogeneous Multi-Hop Wireless and Mobile Networks.

5. E. Woodrow and W. Heinzelman, (September 2002), SPIN-IT: A Data Centric Routing Protocol for Image Retrieval in Wireless Networks, Proc. International Conference on Image Processing (ICIP '02).

6. M. Yebari, T. Addali, A.Z.Sadouq and M. Essaaidi, (January-june 2008), Energy conservation challenges in Communication Protocols for Wireless Sensors Networks: State-of-The-Art Study, International journal on Information and Communication Technologies, Vol. 1, No. 1-2 , pp. 29-35.

7. M. Perillo and W. Heinzelman, (2004), DAPR: a protocol for wireless sensor networks utilizing an application-based routing cost, Proceedings of the IEEE Wireless Communications and Networking Conference (WCNC04).

8. M. Qin and R. Zimmermann, (2007), An energy- efficient voting-based clustering algorithm for sensor networks, Journal of Universal Computer Science.

9. M. J. Handy, M. Haase and D. Timmermann, (2002), Low energy clustering hierarchy with deterministic cluster head selection, Proceedings of IEEE MWCN.

10. Nitin Mittal, Davinder Pal Singh, R.S. Chauhan and Dr. Rajesh Goel, (September 11,2009), Optimal Energy Efficient Routing Protocol for WSN, Proceeding of International conference on Emerging Trends in Electronics and Communications, pp-8.5.