Simulation Results Analysis

In Table 4.3, it shows that our proposed ESCH is 23.1 % better than LEACH for the time of FND. And the time of HND in ESCH is 12.1 % better than that in LEACH.

The time of FND and HND in ESCH are longer than LEACH, because the ESCH adopts the manner that balances the energy load of each node. At the beginning of ESCH, we examine the utilization level which means the amount of energy consumption in term of how much percentage of original energy in the node. The indicator of energy balancing is the utilization level, which points out how much energy of the node is utilized from the beginning until now.

FND HND LND

LEACH (seconds) 405 630 1070

ESCH (seconds) 527 990 1200

Performance 23.1 % 57.1 % 12.1 %

Table 4.3 The time comparison for FND, HND and LND.

From Figure 4.1, at the same time, the number of the nodes surviving in ESCH is more than that in LEACH. Since the cluster head selection of LEACH doesn’t take the residual energy into account, the node with little energy may become cluster head which results in heavy energy consumption, and the node would die quickly. But the

32

node with higher energy can’t represent any meaning, thus we pay attention on utilization while considering the original energy of each node. So, we can select the less utilized node to be cluster head to distribute energy load to avoid overusing on the node with high energy.

Figure 4.1 also shows that the lifetime of ESCH is longer than LEACH. The factor of prolonging the system lifetime is adapting the number cluster head according to the number of nodes surviving in a sub-region. After some nodes were dead, the number of cluster head in most of rounds is more than ESCH because the less number of nodes will increase the probability to be cluster head in the current round. So, if the number of cluster heads is larger than the best number of cluster heads, the energy consumption will increase. In ESCH, we use the best number of cluster heads to be a reference for the number of cluster-heads in the current round. Also, the less number of cluster heads will increase energy waste since the communication distance between the cluster head and its members may become larger so as to increase transmission power. However, using the adaptive control of the number of cluster heads can reduce the consumption of energy.

33

Figure 4.1 Number of alive nodes over time.

4.3.2 Energy Consumption

Figure 4.2 shows the energy consumption over simulation time. In Figure 4.1, we can see the gap of LEACH and ESCH is obvious for number of alive nodes at 500 second. The number of nodes which survives will be the factor that affects the number of cluster head in LEACH so that the energy consumption is also influenced. The details about energy consumption that we will discussed in later paragraphs.

Before 500 second, the diversities of energy consumption in LEACH and ESCH are almost the same. There are some factors for this result. One is that the ESCH have more probabilities to elect the cluster heads with higher bandwidth so the energy consumption will be higher. At the same time, we also control the number of cluster heads to the best one, and it can save power for communications with the nodes and

34

base station. Another factor is that the LEACH randomly selects the nodes to be cluster heads that doesn’t consider the bandwidth, so the energy consumption is usually lower than ESCH. But the number of cluster heads is not managed by LEACH, the energy consumption can be higher when the number of cluster heads is not equal to the best one. Since the different factors for increasing or decreasing energy consumption affect each other, the energy consumption for both strategies are almost the same before 500 second.

After 500 seconds, we can compare Figure 4.1 and Figure 4.2. In Figure 4.1, after 500 seconds, the nodes surviving are decreasing, and the energy consumption is raised. When the number of nodes decreases in LEACH, the total energy consumption will increase. One reason is that the number of cluster head may be raised due to the increasing probabilities of selecting cluster head. Another factor is that the number of members in a cluster is reduced, and thus the length of TDMA schedule will be shortened. Because a shorter TDMA schedule can place more transmissions than long TDMA schedule for transmission phase, the data transmission of sensed data will increase so that the energy consumption is raised. In ESCH, because of the energy balancing and adapting the number of cluster heads, the energy consumption is lower and much stable than LEACH.

35

Figure 4.2 Total energy consumption over time

4.3.3 Number of Packet Received

From the observation of Figure 4.3, the number of packets that the base station received in ESCH is always better than that in LEACH. The factor is that the LEACH election strategy never takes the bandwidth of the node into account, so for the same period of time the LEACH can’t achieve the same performance of ESCH. In contrast to the election strategy of LEACH, the ESCH selects the top CH of nodes with higher bandwidth after filtering the node by utilization level checking. Once the cluster is created, the transmission time for transmitting data to the base station can be reduced and thus the TDMA schedule is shorten. This makes the TDMA schedule have more probability to fit more schedules into the data transmission phase, so the number of packets received at the base station can be raised.

36

Figure 4.3 Number of packets received at BS over time.

The Figure 4.4 shows the relationship of energy consumption and the number of received packets at the base station. For the same energy consumption, ESCH performs better than LEACH. The reason is that ESCH adopts the method of adapting cluster head to save energy for communications, and selects the higher bandwidth to make the cluster heads transmit more sensed data to the base station.

37

Figure 4.4 Energy versus Number of Received Packets at BS

4.3.4 Number of Cluster Head

The ESCH controls the number of cluster heads to avoid the energy depleting quickly. If too many cluster heads exist in the sub-region of WSN and these cluster heads communicate directly with the base station, since the base station is usually far away from the cluster heads, it causes the cluster heads to raise its transmission power for communicating with the base station. If too few cluster heads exist in the sub-region, we still can’t get the benefit for energy saving. This is because the communication distance between cluster members and the cluster head is raised since there are too few cluster heads to make a good choice. Because the transmission power is proportional to square of the distance, the energy depletion for communication within a cluster needs to be concerned.

38

From Figure 4.5, the ESCH has a stable number of cluster head. In contrast to ESCH, the number of cluster head in LEACH is varying. The reason is that LEACH doesn’t consider controlling the number of cluster head. By observing Figure 4.1, after 600 seconds, the number of surviving node in LEACH is decreasing but the number of cluster head may be bigger than the best number of cluster heads. In ESCH, it adapts the number of cluster head according to the number of node existing in a sub-region. So, the number of cluster heads decreases as the number of surviving node drops. Through this way we can save the energy on communications, no matter within the cluster or from the cluster head to the base station.

Figure 4.5 Number of cluster heads over time.

39 head to save energy for communications. The ESCH sets the number of cluster head to the best one in the current round. Finally, each node shows its own bandwidth to decide who can become a cluster head. By ranking their bandwidth in descendent order, and choose the best of CH nodes which is the number of cluster head calculated from the last step, the base station can always receive more packets than LEACH.

And because of distributing energy load, the number of nodes surviving in ESCH is more than that in LEACH. Furthermore, the ESCH saves more energy than LEACH via controlling the number of cluster heads in each round.

In this work, he best number of cluster head that ESCH adopted is from the experiment of LEACH. This optimal value is based on the homogeneous WSN, but it may not be optimal in the heterogeneous WSN. Because the diversities of bandwidth in heterogeneous WSN, the rate for energy depletion is much different from that in homogeneous WSN. Thus, we can take the rate of energy depletion into account to obtain the optimal number of cluster heads in the future work.

While the ESCH is focusing on a sub-region of whole WSN, the further work is how to select a local base station that is responsible for transmit sensed data to the global base station in this sub-region.