1. By Naeem Amjad 1

2.  Challenges  Introduction  Motivation  First Order Radio Model  Proposed Scheme  Simulations And Results  Conclusion 2

3.  Energy efficiency  Maximum Network Lifetime  Coverage Ability  Increased Throughput  Low delay  Efficient routing  Cluster Head selection technique 3

4.  Wireless Sensor Network is a collection of small randomly dispersed devices that provide the ability to monitor physical and environmental conditions in real time.  WSNs are independent when deployed into the field because they have the ability of self- configuration and survival. 4

5. There are two types of Clustering:  Clusters once constructed and never be changed throughout network lifetime, are called Static Clusters.  Clusters based on some sort of network characteristics and are changing during network operation are known as Dynamic Clusters.  DREEM-ME is based upon static clustering. 5

6.  All the nodes are homogeneous  All nodes are proactive (continuously monitoring data).  Base station is in the center of the network.  Maximum energy based cluster head selection is used. 6

7.  Clustering technique of LEACH does not assure a fix number of CHs in each round.  In LEACH protocol, the number of CHs selected are not optimum.  We select the nodes as CHs which carry maximum energy in a particular region. So, this technique assures the optimum number of cluster heads in every round. 7

8.  For efficient use of energy and improvement of coverage, DREEM-ME divides the total area into small sub-regions  These sub-regions are treated separately for the nodes distribution. 8

9. Here, k= length of message in bits d= distance of transmission Eelec= Energy to run the transmitter or receiver’s circuitry Eamp= Energy to amplify the signal 9

10.  The target is to maximize the network coverage.  There exists a trade off between coverage and the energy.  So we have localized the whole network and divided the network into sub-regions that helps in avoiding the coverage hole. 10

11.  In DREEMME, the CH selection is entirely based upon the maximum energy.  In a particular region, the node with maximum energy is selected as the CH for that region in the current round. 11

12. 12 Total Nodes = 90 Nodes in each Region = 10

13.  All non-CH nodes of outer regions (6, 7, 8, 9) check their distances from CHs of six regions which are close to them.  For example, each node of Region 6 checks its distance from CHs of its nearby regions (2, 3, 5, 6, 7, 9) and then finds the minimum of these six distances.  So every non-CH node of outer regions sends its data to the CH which is at minimum distance. 13

14.  In DREEMME, all nodes of Region 1 are using Direct communication because they are at smaller distance to the BS as compared to the nodes of other regions.  All the other regions are considered as static clusters. 14

15.  Multi-hop technique is used for CH-CH communication.  CH of region 9  CH of region 5  BS  CH of region 8  CH of region 4  BS  CH of region 7  CH of region 3  BS  CH of region 6  CH of region 2  BS  Means sending the data 15

16.  We take a 100m x 100m area for our network  Total nodes are 90.  We divided the area into three concentric circles with 20m, 35m and 50m radii.  Simulations are done in MATLAB.  Packet Drop Model is also implemented.  Average of 5 simulations is plotted in the results. 16

17. 17

18.  The results fluctuate to-and-fro around a mean value in every simulation.  Confidence interval is the interval in which we are pretty confident about our results.  We have observed the range of variance of our desired results and then defined their upper and lower values and the mean. 18

19.  Nodes of regions of outermost circle die first because area of each outer region is greater than middle or inner circle..  Then nodes of regions of middle circle start to die.  And the Direct Communication nodes die in the last because they are much closer to the BS. 19

20. 20

21.  According to our network strategy packets sent to the BS per round should ideally follow the explanation below: Packets sent to BS by 1st Region DT nodes  10 Packets Sent to BS by 2nd Region CH Node  1 Packets Sent to BS by 3rd Region CH Node  1 Packets Sent to BS by 4th Region CH Node  1 Packets Sent to BS by 5th Region CH Node  1 Total Packets Sent to BS per round  14  As long as all the nodes are alive the packets sent remain 14. 21

22. 22

23.  In DREEM-ME, Packet Drop concept is used which makes it more close to the reality situation.  In reality the wireless links are not perfect or ideal.  There is always a probability that some of packets may be dropped on their way. 23

24. 24

25.  We calculated the dropped packets with the probability of dropping as 0.3 out of 1.  But it is also practically possible that the probability of packet loss is less than 0.3.  Packets received are not the same as the packets sent in each round because of the packet drop technique. 25

26. 26

27.  In this paper, we have proposed a new clustering technique for WSNs.  DREEM-ME uses static clustering and maximum energy based CH selection.  Multi-hop route is used for the CHs at long distance to sink.  The network field is divided evenly into circles and sectors to reduce the distance between CHs and BS.  In MATLAB simulations we compared our results with LEACH and LEACH-C.  CH selection technique of DREEM-ME provides better results than its counterparts, in terms of network lifetime, stability period, area coverage and throughput. 27

28. 28