Hybrid K-Medoids with Energy-Efficient Sunflower Optimization Algorithm for Wireless Sensor Networks

Abstract

Wireless sensor network (WSN) sensor nodes should have adequate energy. Reduced energy usage is essential to maximize the endurance of WSNs. Combining WSN with a more significant energy source, a cluster head (CH), is another effective strategy for extending WSN durability. A CH is dependent on the communication inside and between clusters. A CH’s energy level extends the cluster’s life for the complete WSN. Determining the energy required in WSNs while developing clustering algorithms is challenging. For maintaining energy efficiency in WSNs, this research offers K-medoids with sunflower-based clustering and a cross-layer-based optimal routing approach. An efficient fitness function generated from diverse objectives is used to choose the CH. After CH selection, sunflower optimization (SFO) indicates the best data transmission line to the sink node. The proposed protocol, SFO-CORP, increased the network lifetime by 19.6%, 13.63%, 11.13%, and 4.163% compared to the LEACH, EECRP, FEEC-IIR, and CL-IoT protocols, respectively. The experimental results showed that it performed better for packet delivery ratio, energy consumption, end-to-end delay, network lifetime, and computation efficiency.

1. Introduction

The progress and increased availability of low-power wireless sensor networks have resulted from wireless communication technology. The sensor nodes in each application are responsible for sensing the target region and transmitting the gathered data to the sink node through wireless networks for further processing [1]. Typically, a node has four subsystems: detection, processing, communication, and a power unit [2]. These sensors communicate either internally or directly with the sink. As a result, the resource allocation and communication must be optimized to maximize the network’s lifetime [3].

WSN technology has evolved and is now extensively used in various fields, including manufacturing, medical care, network communication systems in smart homes, intelligent transportation, and public security. Today’s practical applications necessitate improving positioning accuracy, cost efficiency, and power consumption. However, developing a low-energy, accurate, stable, and functional algorithm is difficult [4]. A typical WSN comprises hundreds of small, self-contained sensor nodes with limited battery capacity. Since these sensor nodes are typically put in remote places, batteries cannot be readily replenished [5]. The energy of a sensor node determines its lifespan (battery). The main issue with WSNs is that the sensor nodes are quickly depleted and die. As a result, it is critical to develop an energy saving algorithm to extend the life of sensor nodes [6].

Protocols for cluster-based hierarchical routing are critical for minimizing the energy consumption of WSNs. Clustering enables this to be accomplished effectively and straightforwardly. Clustering groups many WSN nodes and creates communication between the group and the BS via the group leader, referred to as the CH. Selecting clusters and related CHs is a complex and intimidating undertaking in and of itself. Numerous tactics have been employed over the years to optimize the choice of CHs. Multiple optimization strategies have been tried to determine the optimal group of CHs in a network [7]. Fuzzy logic [8], ant colony optimization [9], genetic algorithms [10], particle swarm intelligence [11], and the harmony search algorithm [12] were among these strategies. The CH selection problem for WSNs is solved in this article using a k-medoid with sunflower optimization (SFO) approach.

Routing is critical in low-energy WSNs and is carried out using energy-efficient routing protocols [13]. A cross-layer optimized opportunistic routing approach is proposed to address the issues of effective clustering and data transmission in smart manufacturing. The issue of optimum cluster head (CH) selection is articulated in this study using the characteristics of various layers. Following the first cluster construction, the ideal CH for each cluster is determined by performing a cross-layer probability evaluation of each sensor node. Cross-layer parameters increase energy efficiency, while minimizing clustering and routing overheads. This research proposes a cross-layer optimized opportunistic routing protocol for WSNs that generates numerous pathways, measuring the node lifespan, communication reliability, and traffic intensity.

The remaining sections of this article are organized as follows: In Section 2, the related works with respect to energy efficiency in WSNs are discussed. Section 3 presents the proposed hybrid K-medoids sunflower optimization algorithm. Section 4 gives the simulation results and a discussion about the proposed algorithm. Finally, the proposed work is concluded in Section 5.

2. Literature Review

A novel energy-efficient WSN routing protocol is accomplished in the proposed protocol by optimizing the path’s minimal residual energy, while adhering to stated reliability limitations. The second protocol is a simplified variant of the first, in terms of duration and complexity. As illustrated by the simulation findings, the suggested procedures maximize the lifetime of WSNs, while adhering to reliability limits. In addition, they optimize how energy is distributed across the nodes in the network [14].

A CH selection technique for WSNs was created based on sampling and spider monkey optimization. This article offers a spider monkey optimization (SMO) approach based on sampling. If the sample population comprises nodes, CHs are chosen from the nodes. The simulation results were compared to those obtained using related protocols, including the threshold-sensitive energy-efficient delay-aware routing protocol (SMOTECP), particle swarm optimization clustering protocol (PSO-C), and low-energy adaptive clustering hierarchy centralized (LEACH-C), in both homogeneous and heterogeneous configurations [15].

An energy-efficient clustering technique was shown in action for magnetic induction-based underwater WSNs. The paper offered a jellyfish-breathing method for choosing the CH and an algorithm for automatically adjusting sensor nodes. Then, it used the current system model to provide a clustering protocol comprising two main components. The results of the simulation showed that the proposed clustering strategy offers a high network capacity rate, while preserving a high degree of energy normalization [16].

A hybrid strategy for optimizing the quality-of-service measures in WSNs was proposed. The study suggested a novel mixed-model approach for maximizing network longevity. The major objective of this work was to enhance the lifespan of WSNs by combining hybrid optimization techniques with active routing algorithms. Five hundred twelve bytes were processed every second. Each data transport packet was 100 bytes in size. According to the simulation results, DSDV alone was less effective in WSN scenarios than DSDV combined with the two optimization approaches and the routing protocol [17].

The authors suggested a novel dynamic energy efficiency routing (DEER) protocol that ensured message delivery. To send the message from the source to the destination, the DEER dynamically chose the node with the maximum amount of residual energy at predefined energy levels. Comparing this suggested method to probabilistic energy profiles, effective hop routing, digraphs, and random algorithms, it was shown to increase session life time and data flow efficiency [18].

Based on DS evidence theory, a routing method that conserves energy was described. The DS evidence theory uses three attribute indexes as evidence, considering among other things the traffic, the distance of the route by the shortest path, and the residual energy of the nearby nodes. Next, using the entropy weight approach, the weights of the three indices were objectively determined. The simulation findings demonstrated the strong transmission reliability of the DS evidence theory, which can significantly reduce network energy consumption compared to the MCRP and FLEOR algorithms, increase network lifetime, and minimize packet loss rates [19].

An integration of the fuzzy clustering technique with the EHO–Greedy algorithm for efficient routing in WSNs was proposed. It analyzed the distinct sink nodes of a fixed and moving sink, to optimize energy usage. Clustering was accomplished using an improved EM approach, which collected SNs with similar features. The choice of CH was essential to this clustering scheme’s success. The experimental outcomes indicated that the suggested system performed better than the existing techniques [20].

Harmony search provided energy-efficient clustering and routing for WSNs. The recommended method was broken down into two stages: clustering and routing. A new model for the clustering phase was developed, considering the energy consumption of both gateways and conventional nodes and routing. The suggested IHSCR outperformed the prior clustering algorithms in balancing the sensor nodes’ energy consumption (especially regular nodes), greatly enhancing the WSN energy efficiency and network lifespan [21].

The authors proposed a jellyfish dynamic routing protocol (JDRP) to protect location privacy and avoid minimum latency. A novel routing channel was planned to regularly combine the information from tentacle nodes, demonstrating how data were sent in a possible angle orientation, while using less energy and extending the lifetime. JDRP’s simulation results were compared to the QDVGDD, QWRP, and starfish routing protocols. The findings indicated that JDRP outperformed the other network characteristics and efficiently routed data packets to the dynamic sink [22].

The authors suggested the use of a cluster-based dynamic energy-aware routing system for agricultural WSNs. A routing protocol based on the gateway cluster energy efficient centroid (GCEEC) was provided in this work. In this protocol, gateway nodes were chosen from each cluster, and the cluster chiefs were chosen from the centroid locations. Data from cluster head nodes was routed to base stations through gateway nodes. The suggested methodology located and span the CH at the cluster’s energy centroid. The simulation results showed that the protocol outperformed the EECRP regarding network longevity, throughput, and energy usage [23].

A protocol for energy-efficient clustering and immune-inspired routing for a WSN-assisted IoT system was proposed based on adaptive fuzzy rules. An FEEC-IIR protocol was proposed in this research as an adaptive fuzzy rule-based, energy-efficient clustering, and immune-inspired routing mechanism for WSN-assisted IoT devices. Good performance results were shown for the quality-of-service measures, such as packet delivery rate (99%), packet loss rate, throughput (0.95 Mbps), net-work lifespan (5500 rounds), and end-to-end latency compared to the current routing protocols (45 mJ) [24].

A WSN-based approach for efficient fuzzy-based cross-layer routing protocol was explained. The study focused on locating sensor nodes and promoting cross-layer effective route finding for data transmission reliability in the network. The objective was to follow the sensor nodes’ position and activity. A new technique named the fuzziness-based contiguous node refining algorithm was proposed to analyze its behavior. NS2 simulations demonstrated that it outperformed current approaches of ELOER, EFLOR, SSPRA, DRPC, and TSTP in connectivity ratio, end-to-end latency, throughput, overhead, and energy usage [25,26].

An effective energy-saving system offered a reduction of energy use and extension of network life. One of the most typical methods for lowering WSN energy usage is clustering technology. The fundamental concept behind this approach is to employ clustering methods to shorten the communication distance between sensor nodes; using the K-medoids approach to identify the ideal intersections of the sensor nodes and to obtain flawless cluster results, and then to decide on the proper cluster heading. By distributing the network load among clusters, it is possible to increase the energy efficiency, while increasing the network’s lifespan. According to simulation data, the approach surpassed the other widely-used algorithms regarding power usage and network longevity [27].

The authors suggested a multimodal population-based iterative heuristic global optimization method called the sunflower optimization algorithm (SFO). SFO uses concepts such as root speed and pollination to achieve improved robustness compared to conventional algorithms. The issue of identifying structural degradation in composite laminated sheets in reverse was then addressed using this new methodology. Due to optimization, the newly proposed optimization approach was able to locate the ideal point in the standard test, proving its superior performance [28].

3. Proposed Methodology

The SFO approach and the cross-layer based optimized opportunistic routing protocols (CORP) are used in this study. A new cross-layer-based optimized opportunistic routing strategy for WSNs decreases power consumption and balances it across nodes. K-medoid and sunflower optimization techniques reduce the detrimental effects of extreme values and determine the best medoid among the sensor nodes for the optimal clustering results. This technology can enhance energy efficiency and effectively increase the network’s life by balancing the network burden.

3.1. System Infrastructure

3.1.1. Network Model

The system infrastructure comprises a single base station and many sensor nodes. There are two categories for each sensor node. The primary cluster node and the common node are the two different types. The public node’s duties include monitoring the surroundings and providing perception data to the CH nodes. Common nodes are frequently selected as CH nodes, which gather, integrate, and send their data to the BS. Figure 1 depicts the design of the WSN architecture.

3.1.2. Energy Model

The system’s architecture comprises many sensor nodes connected to a single base station (BS). The “common node” is one type of node, whereas the “cluster head node” is the another. The CH node receives sensor data from the common nodes, monitoring the surroundings. The cluster head node carefully selects a common node from among them and gathers, combines, and transmits data to the BS.

$$E_{Tx}\left(L,d\right)=(E_{elec}\times L)+({\epsilon }_{amp}\times L)$$

(1)

$$E_{Rx}\left(L,d\right)=E_{elec}\times L$$

(2)

where d represents the distance between the transmission and receiving nodes. In Equation (1) $E_{Tx}\left(L,d\right)$ denotes the energy consumed through the transmission of an L-bit packet; and in Equation (2), $E_{Rx}\left(L,d\right)$ is the energy consumed through the reception. $E_{elec}$ denotes the energy consumed per bit in the transmitter and receiver sensor nodes. ${\epsilon }_{amp}$ represents the energy the amplifier consumes during the transmission phase, which may be computed using Equation (3).

$${\epsilon }_{amp}=\left\{\begin{array}{c}{\epsilon }_{fs}\times d^2, when d\le d_0\\ {\epsilon }_{amp}\times d^4, when d>d_0\end{array}\right.$$

(3)

where $d_0$ is a threshold value associated with the sensor node’s transmission model. The sensor node will use the free-space propagation model if its distance, d, is less than or equal to $d_0$. The multipath fading channel model is employed. ${\epsilon }_{fs}$ and ${\epsilon }_{amp}$ are parameters for communication energy. Equation (4) is used to determine $d_0$.

$$d_0=\sqrt{\frac{{\epsilon }_{fs}}{{\epsilon }_{mp}}}$$

(4)

3.2. K-Medoids-Based Clustering Algorithm (KCA)

A KCA is provided to obtain a universal clustering strategy that consumes less energy and increases the network longevity. In our suggested approach, a base station collects data from all nodes sensing through the cluster head nodes. The proposed approach is divided into two phases: setup and communication phases.

The setup phase involves grouping all nodes into appropriate clusters and identifying the CH nodes. This stage determines the number of clusters, selects the first CH nodes, and obtains the outcome. The number of clusters, K, is computed using Equation (5).

$$K=\sqrt{\frac{N}{2}}$$

(5)

where N is the number of nodes. Let O be the central location for all nodes, and it can be calculated using the expression below.

$$O=\frac{\sum _{i=1}^N{X}_i}{N}$$

(6)

where $X_i$ is the coordinate of sensor i.

Let R be the average distance between the central location and all sensor nodes, and this can be calculated using Equation (7).

$$R=\frac{\sum _{i=1}^N{│X}_i-O│}{N}$$

(7)

The central circle is calculated using the values of O and R. The K points are selected uniformly as the initial mean point, $I_{ix}$ and $I_{iy}$ can be computed using Equation (8).

$$\left\{\begin{array}{c}{I}_{ix}=R\times \cos \left(\frac{360}{k}\times \left(i-1\right)\times \frac{\pi }{180}\right)+O_x\\ I_{iy}=R\times \sin \left(\frac{360}{k}\times \left(i-1\right)\times \frac{\pi }{180}\right)+O_x\end{array}\right.$$

(8)

As the starting mean points are chosen randomly in the K-medoids method, they always have worse results when two or more mean points are quite near, which results in a considerable increase in the iteration time. These critical procedures result in the uniform first mean points being calculated. Thus, the partitioning technique reduces the distance between the ordinary and represented nodes. Equation (9) precisely denotes the absolute-error criteria E.

$$E={\sum }_{i=1}^k{\sum }_{x\in C_i}dist(x,r_i)$$

(9)

The replacement method is given by Equation (10).

$$I^{(t+1)}=\left\{\begin{array}{c}{I}^{*}, E^{*}-E^{\left(t\right)}<0\\ I^{\left(t\right)}, otherwise\end{array}\right.$$

(10)

Figure 2 illustrates the flow chart for the proposed KCA. The excess energy of the cluster head nodes must be addressed during the iteration time. The cluster controller node needs to be replaced if its remaining energy at the beginning of the iteration is lower than the average residual energy of all sensor nodes. The final clustering result is obtained by periodically repeating Equation (10).

3.3. Sunflower Optimization Method

Every day, similarly to the hands of a clock, the sunflower cycle repeats: they awaken and follow the light. They turn around at night and wait. The following morning, they start again. In the context of biology, the pollination process of flowering plants is the act of reproduction. The authors considered the unique characteristics of sunflowers in this work. The best sun-facing angle is determined. In pollination, the goal is to obtain the lowest possible value at random. A small distance separates flower i and flower i + 1. Every flower on the planet releases millions of pollen gametes. On the other hand, the sunflowers only create and develop one pollen gamete per flower for simplicity.

The radiation intensity is inversely proportional to the square of the distance between the sensor nodes. The strength is lowered by one factor when the space is doubled. Factor 4 quadruple, and factor 9 decrease. The larger the distance between the sunflower plants, the less heat it absorbs, and the more comes from the sun. As a result, the same may be said for this research.

The amount of heat $Q_i$ received by each plant is then determined using

$$Q_i=\frac{P}{4\pi {r_i}^2}$$

(11)

where P is the power source, and $r_i$ is the distance between them.

The sunflowers face the sun in the following direction,

$$\overrightarrow{S_i}=\frac{X^{*}-X_i}{││X^{*}-X_i││}, \mathrm{w}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{e} i=1,2,3,\dots .,n$$

(12)

The direction of the sunflower is computed using Equation (13)

$$d_i=\alpha \times Pi(Xi+Xi-1)\times \left|\right|Xi+Xi-1\left|\right|$$

(13)

where $\alpha$ is the inertial displacement, and the chance of pollination is Pi($\left|\right|Xi+Xi-1\left|\right|$), which means that the sunflower “i” pollinates its nearest neighbor i−1 and produces a new individual in a random position. Those nearest to the sun aim for local development in small stages, whereas those further away from the sun move frequently. Limiting the maximum number of steps per individual is also critical, so that possible global minimum candidates are not overlooked. The maximum step is expressed as follows:

$$d_{max}=\frac{││X_{max}-X_{min}││}{2\times N_{pop}}$$

(14)

where X_max and X_min are the values at the maximum and minimum.

N_pop is the total number of plants in the population. The new plantation is determined using Equation (15):

$$\overrightarrow{X_{i+1}}=\overrightarrow{x_i}+d_i\times \overrightarrow{s_i}$$

(15)

3.4. Cross-Layer Based Optimized Opportunistic Routing Protocol

The steps of the newly developed algorithm are shown in Figure 3. This population may be a random sample or a representative sample. Each individual’s evaluation enables us to choose which will be converted into the sun with the highest evaluation.

The model’s unique optimization technique is as follows:

• The network layer can create gradients within the monitored region, calculate the distance between each slope, and create score nodes within the monitored area using topology information provided by the physical layer.
• The network layer uses the physical layer’s information on node density and the remaining energy to select the CH node.
• The network layer clusters the nodes relatively well based on the signal intensity between the nodes in the link layer.
• The link-layer protocol can be correctly implemented using knowledge of the node cluster structure from the network layer. The link-layer protocol nodes are based on enhancing the channel efficiency. Nodes communicate with one another using a scheduling-based link layer protocol, to lessen conflict inside the cluster.
• The network layer chooses the routing and optimal path based on the physical layer’s information about the remaining energy and the link layer’s data on the link quality during the routing selection process.
• If the member node is used as a relay node during routing, the primary cluster node must offer the member node more time slots to reduce the routing delay.
• The connection layer employs error control strategies that maximize the channel state and link quality, to assure data transfer accuracy.

4. Experimental Results

The performance of the proposed technique was estimated with existing routing and clustering techniques, such as LEACH [29], EECRP [30], FEEC-IIR [24], and CL-IoT [31]. Compared to the previous approaches, 500 nodes were used to calculate the performance characteristics for energy consumption, network life, end-to-end latency (E2ED), PDR, communication cost, and communication overhead. In these simulations, 100 homogenous sensor nodes and nine primary cluster nodes with infinite battery energy were positioned in 1000 m². The parameters used for the experimental setup are given in

Table 1. Experimental Setup.

Parameter	Value
Area	1000 m2
Nodes count	100
Initial energy	0.1 J
Bandwidth	20 kbps
Size of packets	500 bytes
Distribution of nodes	Random
Total nodes count	500

.

4.1. Packet Delivery Ratio (PDR)

The PDR measures the proportion of packets transmitted and received by a network. An evaluation of the current and new choices using PDR is shown in Figure 4. The proposed system was superior to the alternative methods, as evidenced by this graph. The suggested approach accomplished a high PDR (99%) compared to the previous designs. The number of sensor nodes increased the PDR. The PDRs of the conventional methods such as LEACH, EECRP, FEEC-IIR, and CL-IoT were 86%, 89%, 93%, and 95%, respectively. A percentage comparison of PDR is given in

Table 2. Comparison of PDR (%).

Node Counts	LEACH	EECRP	FEEC-IIR	CL-IoT	SFO-CORP
100	86	89	93	95	99
200	83	87	90	92	94
300	80	84	88	89	93
400	77	80	85	87	90
500	73	77	81	85	89

.

4.2. Estimation of the Energy Consumption

The network energy was determined using the sum of the energy for the various nodes. Figure 5 compares the recommended system’s overall energy use to the existing practices. The novel method used 0.30 nJ less energy per 500 nodes than the previous methods. The graph shows how the new strategy performed better than the alternatives in terms of expressiveness. As demonstrated in

Table 3. Comparison of the Energy Consumption (nJ).

Node Counts	LEACH	EECRP	FEEC-IIR	CL-IoT	SFO-CORP
100	0.23	0.2	0.17	0.15	0.1
200	0.25	0.24	0.2	0.19	0.14
300	0.32	0.27	0.25	0.22	0.20
400	0.4	0.38	0.35	0.3	0.28
500	0.5	0.43	0.37	0.33	0.30

, the frequency of the sensor node upsurges influenced the amount of energy used.

4.3. Evaluation of E2ED

This is the ratio of the number of packets a receiver receives to the time needed to convey each packet. Figure 6 shows an end-to-end delay (E2ED) analysis for the proposed and existing methods. The suggested method outperformed the previous systems, with a lower E2ED (8.2 ms). When the number of nodes increases, the E2ED will rise. According to

Table 4. Comparison of E2ED (ms).

Node Counts	LEACH	EECRP	FEEC-IIR	CL-IoT	SFO-CORP
100	4	3.8	3.5	3	2
200	5	4.8	4.6	4.2	3
300	7.8	7	6	5	4
400	10	9	7.8	7	6
500	13	12	10	9	8.2

, the E2ED of the existing approaches, LEACH, EECRP, FEEC-IIR, and CL-IoT was 13, 12, 10, and 9 ms, respectively.

4.4. Network Lifetime

A system’s life cycle is how long it can operate and fulfil a specific mission. Figure 7 compares the performance of the newly discovered approaches to the current technologies throughout a network’s life cycle. The graph below shows that the suggested technique resulted in a system lifespan of 3300 rounds, significantly longer than the previous methods. As the number of nodes in a network grew, the network’s lifetime decreased. The lifespan of the system using the current approach, LEACH, EECRP, FEEC-IIR, and CL-IoT, was 1500, 2800, 2500, and 3000 cycles, respectively, as mentioned in

Table 5. Comparison of Network Lifetimes (rounds).

Number of Nodes	LEACH	EECRP	FEEC-IIR	CL-IoT	SFO-CORP
100	4200	4400	4500	4800	5000
200	3800	4200	4300	4600	4900
300	3200	3900	4000	4400	4500
400	2500	2800	3500	3700	4000
500	1500	2000	2500	3000	3300

.

4.5. Computational Efficiency

The computing efficiency of the protocols was measured in terms of their communication cost (CC) and communication overhead (CO).

The packet loss ratio (PLR) determines the communication cost (CC) of a network as a result of congestion and frequent route disconnection, and it is expressed as:

$$Communication cost=1-\frac{P_r}{P_g}$$

(16)

where P_r refers to the energy consumed by a node when transmitting data to another node or a base station. P_g refers to the energy consumed by a node when it receives data from another node or the base station. The average communication cost between the SFO-CORP protocol and the present protocols is compared in Figure 8.

Table 6. Communication Cost (rate).

Node Counts	LEACH	EECRP	FEEC-IIR	CL-IoT	SFO-CORP
100	0.15	0.12	0.1	0.07	0.05
200	0.18	0.15	0.14	0.09	0.08
300	0.23	0.19	0.15	0.13	0.10
400	0.25	0.2	0.18	0.16	0.12
500	0.35	0.25	0.2	0.18	0.17

explains the communication costs for the different algorithms. As density grows, so does the communication cost performance, owing to frequent route disconnections, frequent retransmissions, and congestion.

The communication overhead (CO) is computed by dividing the total number of routing packets counted throughout the simulation time using the number of data packets. It is calculated as:

$$Communication overhead=\sum _t(\frac{{RT}^t}{{DT}^t})$$

(17)

where ${RT}^t \mathrm{i}\mathrm{s} \mathrm{a} \mathrm{c}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{t}\mathrm{h}\mathrm{e} \mathrm{r}\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{e} \mathrm{p}\mathrm{a}\mathrm{c}\mathrm{k}\mathrm{e}\mathrm{t}\mathrm{s}$ and ${DT}^t$ represents the number of data packets at time t. Figure 9 shows the efficiency of communication overheads with a variable number of nodes. Due to the above reasons, SFO-CORP had a lower communication overhead than any other protocol. Compared to the previous protocols, the parameters utilized in SFO-CORP limited the procedures necessary for effective CH selection and route construction, resulting in a lower communication cost and overhead for the SFO-CORP protocol. The network’s routing packets increased due to routine reclustering and route reconstruction tasks. When compared to the cutting-edge protocols shown in

Table 7. Communication Overhead (rate).

Node Counts	LEACH	EECRP	FEEC-IIR	CL-IoT	SFO-CORP
100	1.5	1.2	1	0.8	0.5
200	2.3	2	1.9	1.2	1.0
300	4	3.5	3	2.3	2.0
400	5.8	4.5	4.2	3.5	3
500	8	6.2	5.5	4.8	4.5

, the SFO-CORP protocol performed better overall, in terms of network communication overheads, by reducing the number of routing packets.

In highly mobile situations, the suggested strategy seeks to increase the packet transfer rate and decrease end-to-end latency. Temporary connections can result in transmission errors and retransmission in a WSN based on high portability. The sensor hub might use more energy in this situation and can also boost the throughput, while lowering packet delivery rates. The suggested approach could ensure a secure connection, while conserving the energy of the tuned system. Finally, there is reason to expect reduced data gathering costs. Sensors concentrate the remaining energy, extending the life and dependability of the system. It is designed for unique, adaptive circumstances, commensurate with the product’s quality.

4.6. Time Analysis

Figure 10 depicts a timing analysis of the CH selection approach. The proposed method took less time than the existing CH selection methods. The period would be extended as the number of CHs increases. The proposed schedule, as shown in

Table 8. Timing analysis for CH selection.

Node Counts	LEACH	EECRP	FEEC-IIR	CL-IoT	SFO-CORP
5	120	110	95	98	65
7	131	120	115	105	78
9	144	128	120	112	83
11	149	135	138	118	90
13	153	140	145	127	105

, yielded a reduced (65 s) execution time of 5 CH compared to the existing methods.

5. Conclusions

This study proposes a brand-new method called SFO-CORP. It tackles the issue of short-lived sensor nodes close to the base station. The high traffic volume between the many CHs and the sink causes the low-life problem. SFO-CORP uses a system for choosing energy-efficient cluster heads that considers various factors, including closeness, cost, residual energy, and coverage. SFO-CORP was compared to the existing optimization-based routing methods. The strategy was successfully developed and validated in a MATLAB Tool simulation. Compared to the LEACH, EECRP, FEEC-IIR, and CL-IoT protocols, the suggested protocol, SFO-CORP, enhanced the network lifespan by 19.6%, 13.63%, 11.13%, and 4.163%, respectively. Compared to the LEACH, EECRP, FEEC-IIR, and CL-IoT protocols, the suggested protocol SFO-CORP outperformed them regarding PDR, network lifetime, E2ED, energy consumption, and computational efficiency.