Journal of information and
communication convergence engineering
(J. Inf. Commun. Converg. Eng.)

Wireless sensor networks (WSNs) are used to oversee harsh environments and collect information from both reachable and unreachable locations [1]. The nodes of WSNs sense variables in the target domain and deliver data to a base station (BS) [2]; they fulfill many needs in the environmental, health, military, and industrial fields. The growth potential of WSNs is of particular importance for distributed terrain monitoring, especially for terrains that are inaccessible, hostile, or physically isolated. Recent advances in WSNs have decreased implementation costs and increased functionality [3]. However, the sensor nodes (SNs) are battery-driven and possess limited available energy. It is generally impossible to recharge or replace sensor batteries, owing to the costs and resources required to access the harsh geographic areas. The low-range communication processing capability, scalability, restricted memory, and deployment in unfriendly environments are further challenges.

When a node drains its energy completely, it fails to function and is considered a dead node. Then, target areas that are serviced by those nodes are neglected, even if nearby nodes have sufficient energy to operate. Furthermore, in addition to coverage, maintaining network connectivity is important so that every sensor can send data to the next node, a cluster head (CH), or a BS, depending on the network topology [4]. Hence, while tracking targets, enhancing the network lifespan and sustaining connectivity are crucial tasks.

SNs spend more energy during communication (i.e., sending and receiving data) than during routine computation processes [5]. With indirect communication, SNs dissipate more energy, owing to transmission distances. By contrast, multihop communication saves energy by relaying data via the nearest nodes or CHs to the BS. With multi-hop communication, SNs near the sink/BS have extreme communication overhead, resulting in power issues in the target area [6],[7].

To overcome battery issues and enhance network lifespan, numerous clustering techniques have been proposed for WSNs [8-11] to prolong the life of the network [12-15]. The low-energy adaptive clustering hierarchy (LEACH) protocol improves network lifespan over direct or multi-hop transmissions, but it still has limitations. The CH is chosen randomly, which does not guarantee sufficient node dispersal and optimal performance. Sensors having low or high residual energy enjoy an equal chance of becoming the CH. Hence, those with low residual energy will quickly die as a CH, thereby shortening the life of the network [16].

The load-balancing energy-efficient sleep–awake aware (EESAA) protocol [17] considers multiple parameters, such as stability period, throughput, and network life, by enhancing and examining the clustering algorithm performance of WSNs. The CH collects data from the cluster nodes, fuses them, and sends them to the BS. However, the EESAA does not guarantee an even distribution of CHs across the network.

A stable energy-efficient network (SEEN) was also presented [18] in which the CH selection is based on the node residual energy and distance parameters. With this, a few SNs are facilitated by the best communication and processing competencies. Srivastava et al. [19] presented the optimized zone-based energy-efficient routing protocol (OZEEP) in which node density, residual energy, distance, and node mobility are used as input parameters for a fuzzy inference system to select an appropriate CH. However, it is the BS that stipulates all network controls and operations, which leads to poor scalability.

The problems with WSN routing schemes from the literature can be summarized as follows:

• • Reselecting the same node as CH quickly drains its energy.

• • The CHs are not evenly distributed across the network.

• • Multiple CHs in the same cluster result in additional collisions that lead to increased energy consumption.

To resolve these complications, we propose a novel improved energy-efficient cluster-based routing protocol (IECRP) that is designed for useful clustering, which includes CH selection based on residual energy, neighbor distance, and number of neighbors. The significant contributions of this paper are as follows:

• • We evaluate and increase clustering algorithm performance in terms of network lifetime, throughput, and stability period for WSNs.

• • To ensure uninterrupted and collision-free communication, we focus on time-division multiple-access (TDMA)-based node transmissions.

• • CH selection is based on current residual or maximum average energy.

• • Active nodes transmit packets to other nodes or the BS, whereas the rest remain in a sleep state.

Many clustering algorithms have been proposed to increase network lifespan, such as LEACH [20], kernel-based fuzzy C-means clustering [21], and dual-cluster heads technique based on Krill herd optimization [22]. WSN cluster routing emerged from these three protocols. The initial cluster routing technique for WSNs was LEACH. This algorithm presented the notion of rounds, which reflects the periodic implementation of cluster-based routing. A node is nominated based on its remaining energy during CH selection and whether it was nominated as a CH earlier.

The SN generates a random number between zero and one. It is selected as the CH if the number is smaller than T(n), after which the node broadcasts the information over the network. Every cluster member receives information from the other CHs. Afterward, the cluster member joins the CH with a strong signal and sends a join request to the CH. Upon receiving the request (approval), the CH assigns time slots using TDMA and codes using code division multiple access.

The probability of a node being selected as a CH based on each node's residual energy was later modified by Thein et al. [23], enhancing network lifespan by 40–50 %. In [24], another CH selection scheme was proposed for data aggregation, also enhancing the network lifetime by eliminating redundant data. The protocol considers a hotness factor to modify the threshold value, which describes a specific node's relative hotness in the network.

The CH selection in [25] focused on particle swarm optimization (PSO). The parameter selection is based on the residual energy, node degree, quantity of optimal CHs, and distance from the intra-cluster. This technique achieves better efficiency in different network metrics when assessed with other routing algorithms. A PSO based on energy-efficient CH selection was implemented in [26], where the CH is chosen based on the PSO by considering the residual energy, distance from the BS, and node-to-node distance. Using an optimization scheme (i.e., clustered gray-wolf search optimization), a security-aware CH is selected to increase the lifespan of the network in [27].

CH selection based on residual energy was then proposed as a LEACH improvement [28]. A modest multi-hop technique to LEACH was also introduced, and it was found that both algorithms enhanced the network lifetime better than LEACH after a given period. A non-probabilistic multi-criteria approach was presented for CH selection in [29], where the entire network is divided into different zones. The decision tool analytical network process is used to select the CH, and the best parameters are selected from a set of collected parameters for CH selection.

To continuously monitor the environment, nodes are organized randomly in a network-targeted region/area. The total number of nodes is represented by N, where N = {n₁, n₂, n₃, ……, n_n}. The network model assumption is promulgated into the sensing area by deploying nodes to design the IECRP protocol.

• • A single BS is placed above the sensing area.

• • The SNs and the BS are static in the network region.

• • At the time of deployment, all nodes have the same energy; therefore, uniformity is sustained.

• • After acknowledgment, the BS interacts with the CH.

• • Depending on the distance of the receiver, the transmission power is adjusted.

The architecture of IECRP is shown in Fig. 1. In each round, a new CH is selected based on its residual energy. Communication takes place based on the near-optimal path technique.

Figure 1. Network structure.

The energy required for data transmission depends on the transmission range of the transmission circuit. Similarly, the energy needed to receive the data also relies on the same factors. The energy necessary to transmit a data packet is

(1)$$E_{Tr}(l,d)=E_{ele}(l)+E_{amp}(l,d),$$

(2)$$E_{Tr}(l,d)=\text{ {}l{E}_{ele}+l{\epsilon }_{fs}{d}^2;d<d_0\},$$

(3)$$E_{Tr}(l,d)=\text{ {}l{E}_{ele}+l{\epsilon }_{mp}{d}^4;d\ge d_0\}.$$

The energy needed to receive a data packet is

(4)$$E_{Rr}(l)=E_{ele}(l)=l{E}_{ele},$$

where E_ele(l) is the per-bit energy dissipated by transceiver circuitry. The free space (d²) or multipath (d⁴) propagation is used, depending on the transmission range. E_amp (l,d) is the amplification energy of the data at distance d.

In this section, we propose the IECRP load-balancing technique, which improves the CH selection procedure by considering the node residual energy, node locations, and the number of neighbors.

To simulate and analyze performance, MATLAB 2017a was used for the proposed IECRP protocol, comparing it with two other benchmarks: SEEN and OZEEP. Table 1 lists the proposed IECRP simulation scheme.

Table 1 .

Parameters	Values
Targeted area	100×100 m
Sensor nodes initial energy (E0)	1 J
Energy cost during data aggregation (EDA)	5 nJ / bit/ signal
Total number of sensor nodes	100
Size of packet	4,000 bits
Energy cost during transmission (EelectTx)	50 nj/bit
Energy cost during receiving (EelecRx)	50 nj/bit
Transmit amplifier (Eamp)	/bit /m2

We introduced a new load-balancing protocol, IECRP, which ensures efficient energy depletion and enhances the network lifetime of WSNs. The proposed method minimizes energy consumption, which leads to an increased network lifespan. Using this protocol, the live node concentration is increased and the dead node concentration is decreased. In terms of CH election, network stability, and network lifetime, the IECRP protocol was found to offer better efficiency than SEEN and OZEEP.