4.1. Experimental Settings
In this section, a brief description of the experimental settings is provided, which were used to set up Aqua-Sim-enabled network simulator (ns2) environments for evaluating the performance of the proposed underwater relay node optimization framework W-GUN. The underwater simulation environment had utilized acoustic channels at Medium Access Control (MAC) and physical layers. Towards benchmarking a centric comparative experimental analysis, recent techniques in underwater relay node optimization were considered, including VBF [
27], HH-VBF [
28], VBVA [
29] and ES-VBF [
35]. Both VBF and HH-VBF possess the qualities of service-oriented modeling without considering underwater network characteristics in the modeling. VBVA and VBF both focused on void avoidance without topology knowledge using the geographic locations of relay nodes. Here, the significant impact of underwater characteristics on relay node locations is not considered. The adapted optimization framework W-GUN for underwater relay node selection with a dynamic duty cycle is implemented under realistic 3D underwater network environments of size
. Here, the underwater sensor nodes vary from 100 to 500 in the network region, which is quite a realistic assumption for scalability. Each sensor node in the region is charged with an initial energy of 100 j. For sending a single packet of size 512 bytes per node, the transmission energy per node is considered
, and receiving energy per node is considered to be
given that the ideal node needs 10 mw of energy. The transmission range of each underwater sensor node is 150 m, and the movement of underwater nodes in the horizontal dimension is considered to be 0–3 m/s. There is the consideration of an 100 s beaconing gape in underwater networking communication. Due to the longer propagation delay in underwater networking environments, the handling packet collision is significant. The packets colliding in one node may not collide with other nodes or arrive in different sequential orders in underwater environments. This is handballed effectively in our Aqua-Sim-based simulations where every node maintains a local copy of incoming packets and collision of packets are identified using the difference in received power levels locally. Therefore, the effect of collision only remains on local copies of a node and does not impact the copies of other nodes. This is the way in which collisions are handled locally at each node in our simulation experiments. For controlling the constrains of underwater physical layer implementation, we are essentially setting the exposed interface values of the simulator. For example, as attenuation model setting, spreading factor was considered to be two, along with the absorption coefficient calculated following Thorp’s equation with 1500 m/s propagation speed. An average of 50 simulation runs are averaged to get the results, and overall simulation time considered to be 1500 s. A confidence interval of 98% was considered for generating the results. The framework is implemented in the C++ environment in the simulator.
4.2. Analysis of Results
This section covers the comparative performance analysis part of the proposed approach with the existing methods. The analysis of the end-to-end delay versus the number of underwater sensor nodes is given in
Figure 3. Here, the number of underwater sensor nodes is varied as 100, 200, 300, 400, and 500. It shows that the delay is considerably less for the proposed underwater relay optimization method W-GUN than the state-of-the-art techniques. It is evident that the optimal selection of underwater relay nodes enhances the performance of the overall network, resulting in reduced delay. For a lower delay, the underwater routing path should be as optimal as possible. The aim of the proposed optimization method is to select the maximal number of underwater routes with a minimal number of underwater relay nodes. Therefore, the route consists of the minimum number of underwater relay nodes in our proposed approach. Thus, the delay is considerably minimal compared with existing schemes in the underwater literature.
A more detailed description of the end-to-end delay performance gain of the proposed framework is given in
Table 1 with the comparative investigation of the frameworks HH-VBF, VBVA and ES-VBF, as described in the literature. It can be observed that the average performance gain of W-GUN in terms of percentage is 27%, 42%, and 50% for ES-VBF, VBVA and HH-VBF, respectively. This can be attributed to the fact that the natural underwater characteristics are not considered in the literature, which rather majorly relies on service quality and location-centric underwater relay nodes. However, W-GUN utilized underwater characteristics by the consideration of whales movement, resulting in a considerable performance gain. This performance gain is further represented in a more readable way in
Figure 4, where percentage gain and end-to-end delay observations are shown to be in close relation. This verifies the results presented in
Table 1 and
Figure 3. Therefore, the proposed underwater framework outperforms the state-of-the-art techniques with considerably lower end-to-end delay.
The results in
Figure 5 show the comparison of packet delivery ratio performance between the proposed framework and state-of-the-art techniques with varying underwater network density in the range of 100–500 sensor nodes. It can be observed that the packet delivery ratio is significantly higher for the proposed underwater relay optimization framework compared to state-of-the-art techniques. The better packet delivery ratio for the proposed framework can be attributed to the utilization of underwater environmental characteristics in relay node optimization, ultimately paying as a higher packet delivery rate. Furthermore, an underwater environmental scenario-centric delivery path is the target of the proposal’s discovery mechanism, in order to have more stable nodes in the path for a higher packet delivery ratio. This means that the proposed underwater relay node optimization framework selects the more stable underwater routes with the optimal number of underwater relay nodes. Therefore, the delivery path consists of stable underwater relay nodes in our proposed approach. Thus, the packet delivery ratio is considerably higher compared with existing schemes in the underwater literature.
A more detailed analysis of performance gain in terms of packet delivery ratio of the proposed framework is presented in
Table 2. This comparative investigation considered the underwater state-of the-art literature for highlighting the respective performance gains against the framework in this study. It can be observed that the average performance gain of the proposal in terms of percentage is 15%, 23%, and 32% for ES-VBF, VBVA and HH-VBF, respectively. The reason behind this the utilization of underwater characteristics for identifying network dynamics which is not considered in the existing literature. The existing techniques have utilized service quality and location of underwater relay nodes for making data delivery decisions. However, the proposed framework considers underwater network dynamics resulting in significant performance gain in the packet delivery ratio. This performance gain is further shown in a more scientifically understandable way in
Figure 6. Here, the percentage gain as well as packet delivery ratio observations are presented close together to make it easy to highlight the performance benefits of the proposal in comparison with the literature. This is also helpful in validating the results shown in
Table 2 and
Figure 5. Thus, the proposed underwater framework provides a better packet delivery ratio in underwater environments compared to state-of-the-art techniques.
A comparative analysis between the proposed framework and state-of-the-art techniques is presented in
Figure 7 for energy consumption performance as a function of the number of underwater sensor nodes in the underwater network. It can be observed that the energy consumption is considerably lower for the proposed underwater relay optimization method W-GUN than the state-of-the-art techniques. It is evident that the optimal selection of underwater relay nodes reduces the energy consumption of the overall network, resulting in better utilization of the energy of underwater nodes. For lower energy consumption, the underwater routing path should be as optimal as possible. The proposed optimization framework selects the optimal number of underwater relay nodes for optimizing the energy usage at the node level. Therefore, the route consists of stable underwater nodes as well as the minimum number of underwater relay nodes in our proposed framework. In particular, it can be observed that energy consumption reaches more than 7000 j for VBVA with 500 sensor nodes in the underwater network. In the case of our proposal, it reaches up to approximately 5000 j with similar 500 sensor nodes underwater network density. Here, we want to clarify that the energy consumption is of the overall underwater network, considering all the sensor nodes’ energy consumption. As we have specified in our experimental setting description, total simulation time considered for each simulation experiment was 1500 s and each point considered in the result is an average 50 simulation runs. Essentially, after our simulation time, overall 14% of underwater network energy has been consumed in the case of VBVA and approximately 10% of underwater network energy has been consumed in the case of the proposed framework. Therefore, energy consumption is lower considerably compared with existing schemes in the underwater literature.
The overall energy consumption of the proposed framework is investigated in detail in
Table 3 with a comparison to the literature, focusing on respective performance gain percentage. It can be noted that the average energy consumption performance gain of the proposed framework in percentage is 27%, 53%, and 46% for ES-VBF, VBVA and HH-VBF, respectively. This can be attributed to the fact that the optimal and stable underwater nodes are considered in the proposed framework, in comparison with the literature majorly relying on the quality of service and location-centric underwater relay nodes. W-GUN utilized underwater characteristics resulting in a considerable energy performance gain. This energy performance gain is further represented in a more understandable way in
Figure 8, where percentage gain and energy consumption observations are shown in close relation for better clarity and understanding. This verifies the results presented in
Table 3 and
Figure 7. Therefore, the proposed underwater framework outperforms the state-of-the-art techniques with considerably lower energy consumption.
The overall network lifetime as a function of the number of sensor nodes in the underwater network is comparatively investigated in
Figure 9, considering the proposed framework and the state-of-the-art techniques. Similar network density in the range of 100–500 sensor nodes is considered for this experiment. It can be easily noticed that the network lifetime is effectively longer for the proposed underwater relay optimization framework than the compared state-of-the-art techniques. It is evident that the optimal selection of underwater relay nodes enhances the communication performance of the overall network in terms of a longer network lifetime. For a durable network lifetime, the underwater communication path should be optimal. The aim of the proposed optimization method is to select the maximal number of underwater communication routes with a minimal number of underwater relay nodes. Therefore, the route consists of the minimum number of underwater relay nodes in our proposed approach. Therefore, the network lifetime is longer considerably compared with existing schemes in the underwater literature.
The longer network lifetime-centric performance benefits of the proposed framework are explored in detail in
Table 4 in terms of percentage gain compared with the literature. The average performance gain of W-GUN in terms of network lifetime percentage can be notes as 12%, 32%, and 46% for literature including ES-VBF, HH-VBF and, VBVA respectively. The performance benefits can be reasoned to the fact that the natural underwater characteristics are not considered in the literature, which rather majorly relied on the quality of service and location information of underwater relay nodes. However, the proposed framework has utilized underwater characteristics, resulting in considerable performance gain and a longer network lifetime. This network lifetime performance gain is further represented in a more readable way in
Figure 10. Here, the percentage gain and network lifetime observations are shown in close relation so that they can be analyzed relatively. This is also verifying the results presented in
Table 4 and
Figure 9. Therefore, the proposed underwater framework shows a longer network lifetime compared to the state-of-the-art techniques in the underwater literature.
The throughput performance of the proposed framework is comparatively studied in the results presented in
Figure 11, with varying numbers of sensor nodes in the underwater network environment. The throughput of the proposed framework is significantly higher for the considered underwater relay optimization environment compared to state-of-the-art techniques. The better throughput of the proposal is due to the utilization of stable and lower delay-centric underwater relay nodes. In the framework, an underwater environmental scenario-centric routing path is discovered to have more stable nodes in the delivery path. In other words, the proposed underwater relay node optimization framework selects the more stable underwater routes compared to the considered existing literature with the optimal number of underwater relay nodes. The delivery path consists of stable underwater relay nodes in our proposed framework. Thus, the throughput performance of the proposal is considerably higher compared with existing schemes in the underwater literature.
The benefits in terms of the throughput of the proposed framework are analyzed as performance gain in
Table 5. It is a comparative investigation between the proposed framework and underwater state-of-the-art techniques. The average throughput performance gain of W-GUN in terms of percentage is 10%, 33%, and 45% for the underwater literature including ES-VBF, HH-VBF and, VBVA respectively. The reason behind the better throughput is the utilization of underwater characteristics for identifying network dynamics. However, W-GUN considers underwater network dynamics, resulting in significant performance gains in throughput. These performance benefits are further evident in
Figure 12 in a more scientifically understandable way. Here, the percentage gain and throughput observations are presented in close relation so that they can be analyzed relatively. This is helpful in validating the results shown in
Table 5 and
Figure 11. Thus, the proposed underwater framework provides higher throughput in underwater environments compared to the state-of-the-art techniques.