EECH HEED an adaptive hybrid clustering protocol for energy efficient soil monitoring in heterogeneous wireless sensor networks

Abstract

Wireless Sensor Networks (WSNs) play a pivotal role in precision agriculture, especially for real-time soil health monitoring. However, the limited energy capacity of sensor nodes poses a significant challenge to achieving prolonged network lifetime and consistent data transmission. Existing clustering protocols often struggle with inefficient cluster head (CH) selection, limited adaptability to heterogeneous environments, and excessive control overhead. To address these issues, this paper proposes EECH-HEED, a hybrid and energy-efficient clustering model designed for soil monitoring in heterogeneous WSNs. The proposed protocol utilizes a dual-zone architecture: Zone 1, located near the base station, implements a HEED-based CH selection mechanism that considers both residual energy and communication cost for energy-aware clustering. Zone 2, positioned farther from the base station, employs an enhanced EECH-based hierarchical multi-hop clustering strategy, using residual energy and node degree to select primary and secondary CHs. Unlike the threshold-based hybrid clustering approach which, despite reducing redundant transmissions, relies on static thresholds—EECH-HEED introduces a dynamic threshold-based sensing mechanism. Here, hard and soft thresholds are continuously adjusted in real-time based on environmental change rates and node energy levels. This adaptive strategy significantly reduces energy consumption and extends the operational lifespan of sensor nodes. The model was rigorously tested using MATLAB simulations over 5000 rounds on a heterogeneous WSN setup inspired by real agricultural scenarios. Performance was benchmarked against leading protocols including LEACH, HEED, EECH, TCO, FPA, MIMO-HC, HTCCR, EEHP, and the HCRT model. EECH-HEED demonstrated marked improvements across multiple performance metrics—achieving a 33% reduction in Total Energy Consumption (TEC), a 15% increase in Packet Delivery Ratio (PDR), and a nearly 50% reduction in control overhead. It also recorded the lowest End-to-End Delay (190 ms) and the highest number of alive nodes by the end of the simulation. With its zone-based design, adaptive thresholding, and hierarchical energy-aware clustering, EECH-HEED presents a scalable and low-overhead solution for sustainable soil health monitoring. Future work includes its integration with UAV-based data collection systems and IoT platforms to support intelligent farming infrastructures.

Introduction

Wireless Sensor Networks (WSNs) have become essential in the era of smart agriculture, enabling real-time, autonomous environmental data collection across large geographical areas. These networks consist of spatially scattered, energy-constrained sensor nodes that monitor physical phenomena such as soil moisture, pH, temperature, and nutrient levels. By transmitting this information to a central base station (BS), WSNs facilitate the timely decision-making in precision farming, which leads to optimized irrigation forecast, fertilization control, and enhanced crop productivity. Despite their promise, WSNs face elementary challenges primarily being energy limitations. In field setups, the sensor nodes often run on non-rechargeable batteries and are usually deployed in distant or difficult-to-access regions. Therefore, minimizing energy consumption is vital to prolonging the operational lifespan of the network. Clustering-based protocols have emerged as a well-known energy-saving approach. These protocols categorized the sensor nodes into clusters, where a designated Cluster Head (CH) implements local data aggregation and forwards the aggregated information to the BS, thereby reducing redundant transmissions and balancing energy loads. The most renowned protocols like LEACH and HEED introduced the foundational ideas in probabilistic and energy-aware CH selection. However, these models exhibit considerable limitations when applied to large-scale and heterogeneous deployments. Issues such as irregular energy distribution, arbitrary CH election, static thresholds, and lack of multi-hop communication often cause in premature node failures, high control overhead, and inefficient network scalability. A recent improvement was proposed HCRT², who introduced the hybrid clustering and routing model with threshold-based data collection and zone-oriented CH election. Even as this model marked a step forward in reducing redundant transmissions and extending network longevity, it suffers from static threshold configurations, rigid CH election policies, and limited adaptability in dynamic environments like farming, where soil conditions change often due to irrigation, rainfall, and cultivation activities. To address these issues, this paper proposes EECH-HEED, a novel Hybrid Energy-Efficient Clustering Protocol designed for WSNs in soil monitoring applications. The protocol introduces a dual-zone clustering architecture:

• Zone 1: Located near the base station, which applies a HEED-based CH selection method which leverages residual energy and communication cost.

• Zone 2: Consisting of distant nodes with superior energy reserves, employs an EECH-based hierarchical multi-hop clustering mechanism, selecting primary and secondary CHs based on energy and node degree.

Additionally, EECH-HEED integrates an adaptive threshold-based data sensing strategy, which dynamically updates soft and hard thresholds using environmental trends and node energy levels. This ensures that only relevant data is transmitted, reducing unnecessary communication and conserving energy. The combination of zone-specific clustering, energy-aware CH selection, multi-hop routing, and adaptive sensing enables EECH-HEED to outperform existing protocols in terms of energy intake, packet supply ratio, end-to-end delay, and control overhead.

The above Fig. 1, illustrates a general architecture of a Wireless Sensor Network (WSN)-based Soil Monitoring System designed for precision agriculture applications. In this system, multiple sensor nodes are distributed across an agricultural field to observe soil parameters such as moisture, pH, and temperature. These nodes periodically sense data and pass on it to a nearby Cluster Head (CH), which acts as a local aggregator. The CH processes and forwards the collected data to a centralized Base Station for analysis and decision-making. The processed soil data is then made accessible to farmers and stakeholders through a mobile application or user dashboard, enabling real-time monitoring and informed agricultural interventions. This architecture ensures energy-efficient communication, scalability, and adaptability for large-scale soil health management. To address energy-related constraints, clustering-based protocols have been widely adopted in WSNs. These protocols divide the network into clusters, each governed by a Cluster Head (CH). It is the CH who is responsible for collecting data from member nodes and forwarding it to the BS, either directly or through hierarchical routing. This mechanism reduces redundant transmissions, balances the communication load, and enhances scalability.

Fig. 1

General Architecture of a Soil Monitoring using WSN.

Various clustering protocols have been proposed, each with unique features:

• Classical Protocols like LEACH offer simple and distributed architectures but suffer from short network lifetimes due to random CH selection or high communication overhead.

• Energy-Aware Protocols like HEED and EEHP improve CH selection by incorporating node residual energy, enhancing stability.

• Hierarchical Protocols like EECH, HTCCR and MIMO-HC enable multi-level communication structures that reduce long-range transmissions and improve routing efficiency.

• Hybrid Protocols: HCRT model and EECH-HEED.

• Metaheuristic-Based Protocols such as those using the Flower Pollination Algorithm (FPA) or Termite Colony Optimization (TCO) apply AI-inspired optimization techniques to achieve near-optimal CH selection, though often at the cost of higher computational complexity.

Figure 2 shows all these different types of Models of Wireless Sensor Network.

Fig. 2

Classification of Clustering –Based WSN Protocols.

A comparative analysis of existing clustering approaches is defined in Table 1, these Clustering algorithms have been proposed to address energy constraints in WSNs, each with distinct methodologies, advantages, and limitations.

Table 1 Classification of Clustering-Based WSN Protocols.

Despite significant advancements in clustering techniques, existing protocols exhibit inherent limitations such as uneven CH selection, lack of adaptive cluster formation, inefficient inter-cluster communication, and computational overhead. These challenges underscore the need for an optimized clustering model that integrates energy-aware CH selection with hierarchical clustering and adaptive load balancing.

Research objectives

Despite the advancements in clustering mechanisms, several critical gaps remain unaddressed, especially in the context of long-term, scalable, and energy-aware soil monitoring in agricultural environments:

• Existing models either focus on single-zone clustering or use static thresholds, limiting adaptability to environmental dynamics.

• Protocols such as that of HCRT improve data reduction but still lack adaptive cluster formation and continuous energy evaluation.

• There is a need for hierarchical, zone-based models that can balance energy consumption and extend network life without compromising communication quality.

The proposed EECH-HEED protocol aims to fill these gaps through a comprehensive hybrid design. This research is guided by the following core questions:

• Can a dual-zone adaptive clustering strategy, integrating HEED and EECH principles, significantly enhance the network lifetime compared to traditional homogeneous models?

• How does adaptive, threshold-tuned sensing influence control overhead, energy savings, and data delivery in environments characterized by fluctuating soil conditions?

• How does EECH-HEED compare to existing benchmark protocols—including LEACH, HEED, EECH, and the HCRT model—across key performance indicators such as residual energy, alive nodes per round, end-to-end delay and packet delivery ratio (PDR)?

Motivation for the HEED-EECH hybrid model

In agriculture, soil health plays a critical role in determining crop productivity, resource utilization, and environmental sustainability. Soil parameters such as moisture, temperature, salinity, and pH require continuous monitoring to guide irrigation, fertilization, and crop rotation strategies. Conventional soil testing approaches are time-consuming and labour-intensive. WSNs offer a transformative solution by enabling non-intrusive, real-time monitoring over large agricultural areas. With strategic deployment of sensors across the field, WSNs support precision agriculture by delivering timely, localized soil data, leading to reduced water wastage, optimized fertilizer use, and improved crop yield. To address the limitations of existing clustering protocols, this research introduces a novel HEED-EECH hybrid clustering model, which integrates the distributed CH selection strategy of HEED with the multi-hop hierarchical clustering mechanism of EECH. The primary objectives of the proposed model are as follows:

• Energy-Efficient CH Selection: HEED-EECH dynamically selects CHs based on residual energy and communication cost, ensuring balanced energy distribution across sensor nodes.

• Hierarchical Multi-Hop Communication: EECH’s hierarchical clustering is incorporated to reduce direct long-range transmissions, optimizing inter-cluster communication.

• Adaptive Clustering Mechanism: Unlike static clustering models, EECH-HEED employs dynamic cluster formation techniques based on node density and energy availability, preventing energy imbalances.

• Extended Network Lifetime: By integrating efficient CH rotation and multi-hop data aggregation, the proposed model significantly enhances network longevity, making it well-suited for long-term soil monitoring applications.

Real-world relevance of EECH-HEED in agricultural soil monitoring

EECH-HEED was specifically designed for resource-constrained environments such as small-to-medium scale agricultural farms in rural regions, where:

• Battery-powered sensor nodes lack edge-computing or AI co-processors,

• Communication infrastructure is often unreliable or absent,

• Robust, interpretable, and lightweight routing models are required.

While AI-based protocols provide powerful adaptive capabilities, they often require central training, incur higher energy costs, and are difficult to explain or debug in critical agricultural deployments. In contrast, EECH-HEED:

• Integrates zone-aware clustering that reflects practical proximity and signal reliability,

• Uses adaptive thresholds tied to environmental and energy dynamics.

A novelty approach

While previous protocols have addressed energy-aware clustering or threshold-based sensing in isolation, EECH-HEED distinguishes itself through its dual-zone adaptive hybrid model. Zone 1 employs residual energy and cost-aware CH selection, while Zone 2 leverages hierarchical EECH-based routing with primary-secondary CH tiers. Coupled with a context-sensitive threshold adaptation mechanism, this integration provides a tailored and efficient solution for real-time, large-scale soil monitoring — a scenario underexplored in existing WSN designs.

Related work

Clustering-based routing methods play a pivotal role in enhancing the energy efficiency and scalability of Wireless Sensor Networks (WSNs). Over the past decade, numerous studies have proposed novel clustering mechanisms aimed at optimizing network lifespan, energy consumption, and data reliability, particularly in applications such as precision agriculture and soil monitoring. This section provides a critical review of existing clustering protocols, with a focus on their methodologies, strengths, and limitations, while identifying the research gaps that the proposed EECH-HEED model aims to address.

Recent developments in clustering-based WSN protocols

Energy-aware clustering continues to be a critical aspect of enhancing the longevity and efficiency of Wireless Sensor Networks (WSNs), particularly in resource-constrained environments. Aslam et al.¹ proposed an Adaptive Weighted Grid Clustering (AWGC) algorithm for renewable WSNs, which dynamically adjusts cluster formation based on residual energy and node distribution. Their approach demonstrated improved energy balancing across sensor fields. Building upon such foundational work, our proposed EECH-HEED protocol integrates adaptive thresholding with zone-based clustering to address spatial heterogeneity and energy disparity more effectively in static agricultural WSN deployments.

Bilal et al.² introduced a hybrid clustering and routing algorithm designed for heterogeneous WSNs. The proposed approach strategically deploys both homogeneous and heterogeneous sensor nodes while incorporating a threshold-based data collection mechanism to minimize unnecessary transmissions, thereby improving energy efficiency and network lifetime. However, the reliance on predefined thresholds may restrict the adaptability of the protocol to dynamically changing environmental conditions.

Recent advancements in mobile ad hoc networks (MANETs) and mobility-aware routing models also offer valuable insights for improving dynamic WSN clustering. For instance, Kour et al.³ reviewed various mobility models and routing strategies in MANETs, emphasizing the impact of node movement patterns on communication reliability. Building on this, Satveer et al.⁴ proposed a QoS-improved routing mechanism using an enhanced Manhattan mobility model integrated with Ant Colony Optimization, demonstrating significant improvements in packet delivery and latency under mobile scenarios. While our EECH-HEED protocol is designed for static deployments such as soil monitoring, the integration of mobility-based insights remains a promising future direction, especially for scalable or UAV-assisted WSN applications.

Wireless sensor network optimization have explored the integration of intelligent learning-based strategies to enhance energy efficiency and routing stability. For instance, Aslam et al.⁵ proposed an optimal wireless charging framework combined with SARSA-based intellectual routing in renewable WSNs, demonstrating significant improvements in network lifetime and data delivery reliability. Inspired by such learning-oriented approaches, our work focuses on adaptive thresholding and hierarchical clustering to address energy constraints in precision agriculture environments.

The integration of intelligent algorithms with energy-aware communication has shown promising potential for enhancing network longevity in IoT-integrated Wireless Sensor Networks (WSNs). Aslam et al.⁶ proposed an intelligent wireless charging path optimization framework tailored for critical nodes in renewable sensor networks, demonstrating how AI-driven approaches can significantly improve energy management and sustainability. Motivated by such innovations, our proposed EECH-HEED protocol introduces adaptive thresholding and dual-zone clustering to further improve routing efficiency and energy balancing in static WSNs, particularly for precision agriculture applications.

Table 2 given below shows a literature review/existing researches on various WSNs approaches.

Table 2 Existing Researches on various Wireless Sensor Network based approaches.

The above table systematically summarizes recent research contributions in clustering-based WSN protocols, highlighting their core objectives and corresponding limitations. While various approaches such as hybrid clustering, energy-efficient routing, mobility-aware protocols, and metaheuristic optimizations have shown notable improvements in isolated metrics like energy conservation or network lifetime, they often fall short in offering a holistic solution. Specifically, many existing models either rely on static configurations, exhibit limited adaptability to dynamic environmental conditions, or introduce computational overhead unsuitable for resource-constrained sensor nodes. These persistent research gaps reinforce the need for a dynamic, energy-aware, and lightweight clustering protocol that can efficiently operate in heterogeneous agricultural environments—directly motivating the design of the proposed EECH-HEED model³⁰.

To clearly establish the novelty of EECH-HEED, a comparative analysis with recent protocols published after 2022 is presented in Table 3.

Table 3 Comparative of EECH-HEED vs. Recent Protocols.

Protocols such as HCRT(2022), HTCCR (2023), and TCO-FPA (2024) have proposed energy-aware or hybrid clustering schemes. However, these methods often rely on either static thresholds, single-zone clustering, or lack real-world parameter validation. In contrast, EECH-HEED integrates a dual-zone adaptive clustering model, dynamic threshold tuning, and realistically simulated environmental parameters derived from agricultural deployments. This combination ensures higher network longevity, balanced CH distribution, and better applicability to field conditions.

Energy-efficient clustering techniques

Goyal et al. (2022) proposed a multi-hop hierarchical clustering protocol incorporating HEED and energy efficient data aggregation, which yielded improvements to the packet delivery ratio (PDR) and energy dissipation reduction rate, but unfortunately the use of static CH rotations resulted in an unbalanced energy consumption pattern among the sensor nodes³¹.

A novel adaptive clustering mechanism, proposed by Ghosh et al. (2022), dynamically optimizes CH election probability (residual energy conservation and network topology). As expected, such an approach can increase the stability of networks, but fails to solve communication bottlenecks among clusters (which leads to higher latency) in large-scale WSNs³².

Zhang et al. (2021) proposed an overall fuzzy logic clustering protocol for precision agriculture that incorporates hierarchical routing and AI-driven CH selection. Although there were quite significant improvements in network longevity and scalability, the computational burden associated with the fuzzy logic system poses challenges for real-time deployments in energy-constrained environments³³.

Problem statement

Soil monitoring in smart agriculture relies on Wireless Sensor Networks (WSNs) to collect real-time data on soil moisture, temperature, and pH. However, energy-efficient data transmission remains a challenge, especially in heterogeneous and dynamic field environments. Traditional clustering protocols often use static thresholds and centralized control, which can lead to energy imbalance and early node depletion.

HCRT Model proposed a hybrid threshold-based clustering protocol dividing the WSN into two zones, with CHs statically defined and static thresholds for data transmission. While useful, their model lacks dynamic CH selection and adaptive threshold tuning based on real-time environmental or energy changes. These limitations reduce its applicability in evolving agricultural scenarios. One of the most recent experiments in this direction is the hybrid threshold-based clustering model suggested by Bilal et al. (2022).Their model divides the WSN into two physical zones: a centralized clustering zone with homogeneous nodes and a distributed clustering zone with heterogeneous nodes. The model also incorporates hard and soft thresholds to reduce unnecessary transmissions from nodes experiencing minor environmental changes. The difference between Bilal’s model and the proposed EECH-HEED model can be viewed as in Fig. 3 given below.

Fig. 3

HCRT Model VS Proposed EECH-HEED Model.

This contrast highlights how EECH-HEED not only divides the network physically but also tailors the CH selection strategy to the zone’s energy and communication profile as shown in Fig. 3.While effective in extending network lifetime, this approach reveals several limitations when applied to soil monitoring, which demands adaptive communication and cluster management strategies due to the dynamic nature of agricultural environments. The limitations of HCRT as compared our proposed model is depicted in Table 4.

Table 4 Limitations of the Base Model.

The proposed EECH-HEED: architectural overview

EECH-HEED introduces a dual-zone adaptive clustering protocol that combines the strengths of HEED and EECH. The dual-zone architecture in EECH-HEED reflects the spatial heterogeneity typical of real agricultural sensor deployments. Studies have shown that precision agriculture systems often use denser sensor deployments near fixed infrastructure (e.g., solar-powered weather stations, control units) and sparser, wider deployments across crop fields for monitoring variables like soil moisture, temperature, and salinity⁴⁰. HCRT deployed a WSN in a working agricultural field where 30% of nodes were near the BS (control region) and the rest were scattered across a 100 m × 100 m field. Similarly, Zhang et al. (2023) report heterogeneous energy profiles and communication needs across irrigation and sensing zones in smart farms. These deployments validate the use of region-based clustering strategies to optimize energy consumption and extend operational lifespan. To closely mimic such setups, our simulation adopts a zone-based model: 30 nodes are placed in a centralized circular area (30 m radius), and 70 nodes are uniformly distributed in the remaining field area. The centralized zone benefits from BS-controlled CH selection, while the outer zone uses probabilistic CH election, reflecting real-world constraints in communication accessibility and energy variability. This dual-zone strategy aligns with real deployments and provides a scalable, energy-aware mechanism suitable for large-scale agricultural WSNs. Each optimized for different network regions:

• Zone 1 (Inner region): Applies HEED-based CH selection. Nodes are homogeneous and closer to the base station (BS), with CHs elected based on residual energy and communication cost.

• Zone 2 (Outer region): Uses EECH hierarchical clustering with primary and secondary CHs for multi-hop transmission. Nodes here are heterogeneous and located farther from the BS.

This separation enables better load balancing, tailored energy handling, and efficient communication paths — as shown in Fig. 4.

Fig. 4

Proposed EECH-HEED Model for Soil Monitoring in WSNs.

The dual-zone clustering mechanism in EECH-HEED strategically partitions the monitoring field into two concentric zones to optimize energy usage and routing efficiency. Zone 1, positioned near the base station, employs HEED-based clustering with homogeneous nodes using residual energy and intra-zone communication cost for CH selection. In contrast, Zone 2 consists of heterogeneous nodes with varied energy levels and applies an EECH-based hierarchical clustering approach. Within this zone, nodes are grouped under Primary Cluster Heads (PCHs), which forward data to Super Cluster Heads (SCHs) before relaying to Zone 1 or directly to the base station. This hybrid model reduces transmission distance, balances energy consumption, and extends network longevity through adaptive, energy-aware CH election and tiered communication.

Mathematical model

This section describes the mathematical formulation used in the proposed EECH-HEED model for modelling energy consumption, CH selection, hierarchical communication, and adaptive thresholding in Wireless Sensor Networks (WSNs), which is for a heterogeneous WSN with randomly distributed sensor nodes deployed for soil monitoring over a fixed agricultural area.

Network Assumptions

• Total number of nodes: N

• Network field: A × A

• Nodes are either homogeneous (inner zone) or heterogeneous (outer zone).

• Nodes are energy-constrained; the BS is assumed to have unlimited energy.

• Multi-hop communication is employed in Zone 2 (EECH).

• Sensors monitor soil parameters and transmit data based on dynamic thresholds.

• Radio Energy Model

We assume the first-order radio energy model, as used in Eq. 1:

• Energy required to transmit k-bit data over distance d:

  $${\text{Etx}}\left( {{\text{k}},{\text{d}}} \right){ } = { }\left\{ {_{{k.E_{elec} + k. \in mp . d^{4} { },{\text{ d}} \ge {\text{d}}0}}^{{k.E_{elec} + k. \in fs . d^{2} { },{\text{ d}} < d0}} } \right\}$$

  (1)
  $${\text{Energy required to receivek}} - {\text{bit data}}:{\text{E}}_{{{\text{rx}}}} ({\text{k}}) = {\text{k}} \cdot {\text{E}}_{{{\text{elec}}}}$$

  (2)

• Data Aggregation Energy at Cluster Heads:

  $${\text{E}}_{{{\text{DA}}}} ({\text{k}}) \, = {\text{k }}.{\text{ E}}_{{{\text{agg}}}}$$

  (3)

where:

• k:Packet size.

• d: Distance between sender and receiver.

• d0: $\sqrt {\frac{{\varepsilon {\text{fs}}}}{{\varepsilon {\text{mp}}}}}$ is the threshold distance.

• E_{agg :} 5 nJ/bit.

• Cluster Head (CH) Selection

• Zone 1 (HEED):

The probability P_i​ of node i becoming a cluster head is proportional to its residual energy and inversely proportional to the average energy as shown Eq. 4:

$${{\varvec{P}}}_{{\varvec{C}}{\varvec{H}}}^{\mathbf{i}}={{\varvec{C}}}_{{\varvec{p}}{\varvec{r}}{\varvec{o}}{\varvec{b}}}\boldsymbol{ }\boldsymbol{ }\boldsymbol{ }.\boldsymbol{ }\boldsymbol{ }\boldsymbol{ }[\boldsymbol{ }\frac{{{\varvec{E}}}_{{\varvec{i}}}^{{\varvec{r}}{\varvec{e}}{\varvec{s}}}}{{{\varvec{E}}}_{{\varvec{a}}{\varvec{v}}{\varvec{g}}}}]$$

(4)

where:

• ${C}_{prob}$​ is the desired percentage of CHs.

• ${{\varvec{E}}}_{{\varvec{i}}}^{{\varvec{r}}{\varvec{e}}{\varvec{s}}}$​ is the current residual energy of node i.

• ${E}_{{\varvec{a}}{\varvec{v}}{\varvec{g}}}$ is the average residual energy of nodes in the neighbourhood.

• Zone 2 (EECH):

  $${P}_{CH}^{\text{i}}= \left[\frac{{E}_{\text{i}}}{{E}_{\text{max}}}\right] .\boldsymbol{ } [\frac{{D}_{\text{i}}}{{D}_{\text{max}}}]$$

  (5)

Dynamic threshold adjustment model

To ensure adaptive and energy-efficient CH selection, the EECH-HEED protocol incorporates a dynamic threshold function that varies based on the node’s residual energy, distance from the base station (BS), and current round. The dynamic threshold T_i​ for node i is defined as in Eq. 6:

$${\text{T}}_{{\text{i}}} ({\text{r}}) = \left\{ {\begin{array}{*{20}l} {\frac{P}{{1 - P X \left( {r mod \frac{1}{p}} \right)}}} \hfill & {{\text{X}} \alpha_{i} {\text{X}}\beta_{i } {\text{if}} i\hat{I}{\text{G}}} \hfill \\ 0 \hfill & {{\text{Otherwise}}} \hfill \\ \end{array} } \right.$$

(6)

where:

• P = desired percentage of CHs.

• r = current round number.

• G = set of nodes that have not been CHs in the last ​ rounds $\frac{1}{p}$ rounds.

• α_i = $\frac{{E}_{\text{resdual},\text{ i}}}{{E}_{\text{initial}}}$  = normalized residual energy factor of node i.

• β_i = 1−$[\frac{{d}_{\text{ i}}}{{d }_{max}}]$  = distance-based proximity factor to the BS.

• d_i​ = distance from node i to the BS.

• d_max​ = maximum possible distance to the BS in the network.

• Update thresholds:

  $$HT_{t} = HT_{0} + \lambda \frac{dS}{{dt}}and ST_{t} = ST_{0 } + \mu \cdot \left( {1 - \frac{{E_{res} }}{{E_{init} }}} \right)$$

  (7)

This threshold (see flowchart of dynamic threshold in Fig. 5) ensures that nodes with higher remaining energy and closer proximity to the BS have a higher probability of becoming CHs, leading to balanced energy consumption across the network.

Fig. 5

Flowchart Dynamic Threshold.

Assuming P = 0.1, E_residual,i = 0.4 J, E_initial = 0.5 J, d_i = 50 m and d_max = 100 m, the dynamic threshold for a node in round 5 becomes:

α_i = $\frac{0.4}{0.5}$  = 0.8 and β_i = 1−$\frac{50}{100}$ = 0.5

T_i(5) = $\frac{0.1}{1-0.1 X (5 mod 10 )}$  × 0.8 × 0.5 = $\frac{0.1}{1-0.5}$  × 0.4 = 0.2 × 0.4 = 0.08.

Such adjustment dynamically controls CH selection probability, reducing overuse of low-energy or distant nodes. Below in Fig. 6 shows Dynamic threshold over rounds.

Fig. 6

Dynamic threshold T_i over Rounds.

Here is the first visualization: a graph showing how the dynamic threshold T_i ​evolves over simulation rounds as the residual energy decreases. This graph reflects:

• A decreasing trend in T_i as node energy depletes.

• Periodic spikes due to the modulo-based rotation of cluster head candidacy.

• The impact of energy-awareness and proximity in threshold calculation.

The presented mathematical model provides a comprehensive framework for analyzing the energy dynamics, cluster head selection probability, multi-hop communication, and adaptive threshold-based sensing in the EECH-HEED protocol. By modelling these factors explicitly, the proposed approach ensures a balanced energy distribution, minimizes redundant data transmissions, and optimizes communication paths, ultimately contributing to enhanced network longevity and efficiency in soil monitoring applications. Below Fig. 7 shows EECH-HEED based network simulation for monitoring soil.

Fig. 7

Flowchart of proposed EECH-HEED Protocol.

Step by step procedure of proposed EECH-HEED model is represented in Algorithm 1.

Algorithm 1

EECH-HEED Hybrid Clustering Protocol

Simulation setup

To authenticate the performance and efficiency of the anticipated EECH-HEED protocol, extensive simulations were conducted using MATLAB R2023a in a synthetic Wireless Sensor Network (WSN) environment representing a typical agricultural soil monitoring scenario. The simulations aim to compare EECH-HEED against established protocols—LEACH, HEED, EECH, TDEEC, HADCC, and the HCRT model in terms of energy consumption, longevity, data delivery, and communication efficiency.

Research methodology and experimental setup

The experimental parameters used in this study were carefully chosen to ensure a realistic evaluation environment for the proposed EECH-HEED protocol. Although this work is entirely simulation-based, the parameter configuration—such as network area, node distribution, energy consumption models, and threshold-based sensing—has been adapted from the real-world deployment described in HCRT Model. Their setup, implemented in an actual agricultural field, provides empirically validated values that align with practical soil monitoring needs. By incorporating these values, including a 100 m × 100 m network area, a central base station, 100 heterogeneous sensor nodes, and first-order radio energy parameters, the simulation aims to reflect real environmental and network constraints.

To evaluate the robustness of EECH-HEED under real-world uncertainty, the simulation setup was extended to include node heterogeneity and environmental dynamics. Specifically, the initial energy of sensor nodes was randomized between 0.3 J and 0.6 J, simulating practical differences in node power supply, harvesting, or aging. Furthermore, to mimic environmental variations such as soil moisture fluctuation, the hard and soft sensing thresholds were made adaptive, varying periodically to reflect real-time field conditions. These enhancements enable a realistic testbed to assess the adaptability of the proposed model.

Adaptive threshold-based sensing mechanism

To minimize redundant data transmission and reduce network congestion, EECH-HEED employs an adaptive threshold-based sensing mechanism. This approach utilizes two types of thresholds:

• Hard Threshold (HT): The absolute value beyond which a sensor triggers data transmission (e.g., drastic change in soil moisture).

• Soft Threshold (ST): The minimum change since the last transmission that justifies sending an update.

Sensor nodes continuously monitor their environment but transmit data only when:

• The current sensed value exceeds the Hard Threshold, and

• The change since the last reported value exceeds the Soft Threshold.

This dual-check mechanism ensures that only significant environmental changes are reported, reducing unnecessary traffic.

Irrelevant transmissions refer to packets generated due to minor or negligible changes in sensor readings—such as a 0.1% shift in soil temperature or moisture—that do not influence control actions or decision-making. By filtering out such low-priority data, the protocol conserves energy, reduces contention at the MAC layer, and enhances network longevity.

The thresholds themselves are dynamically adjusted based on contextual parameters such as residual energy and rate of environmental change, enabling the system to adapt to both energy availability and monitoring urgency.

The mobility model assumes static deployment, as is standard in agricultural soil monitoring where sensors are fixed in the ground. Further, the adoption of hard and soft sensing thresholds for field behaviour as well as the implementation of a zone-based clustering structure directly approximates the behaviour in the real-world environment. This approach provides a verification basis for the simulation performance. It provides an opportunity to validate the model under widely-used conditions in order to ensure its usefulness in real-time soil monitoring systems. More specific work is required to deploy the model in a physical testbed to further validate the results. The detailed experimental factors used in the simulation of EECH-HEED are given in Table 5.

Table 5 Simulation Parameters.

As shown in Table 5, the sensor field is modelled as a 100 × 100 m area with the base station positioned centrally to facilitate balanced communication. The network comprises 100 nodes, partitioned into two zones: Zone 1 (30 nodes) using HEED-based CH selection and Zone 2 (70 nodes) employing EECH-based multi-hop clustering. A heterogeneous energy model is adopted, where advanced nodes possess 50% more energy than normal nodes to reflect real deployment scenarios. Adaptive threshold-based sensing is regulated using soft and hard thresholds to suppress redundant transmissions. The choice of transmission, reception, and aggregation energy values aligns with the first-order radio model, widely adopted in WSN research. This setup ensures that the protocol is evaluated under rigorous and practical constraints, providing a solid foundation for assessing its applicability in real-world soil monitoring systems.

Performance metrics

The following Table 6 defines the metrics used to evaluate each protocol. These together measure energy efficiency, communication quality and network lifetime.

Table 6 Performance Metrics.

Performance evaluation and results discussion

To validate the performance, scalability, and energy-efficiency of the proposed EECH-HEED clustering protocol, extensive MATLAB R2023a simulations were conducted in a synthetic WSN environment that mirrors a practical agricultural soil monitoring scenario. The simulation parameters were carefully adapted from the empirical setup presented in HCRT to ensure realistic comparisons and reproducibility. The experimental network comprises 100 sensor nodes (N = 100) uniformly distributed in a 100 m × 100 m agricultural field. A zone-based clustering architecture is adopted, where 30 nodes (m = 30) are located in the first-level circular region near the Base Station (BS), with a radius of 30 m, and the remaining 70 nodes (n = 70) are located in the second-level distributed area. The BS is located at the center of the field and remains static throughout the simulation.

Nodes in the first-level zone are homogeneous and have an initial energy of 0.5 Joules, while nodes in the second-level region are heterogeneous—20% of them (14 nodes) are assigned higher initial energy levels to simulate real-world deployment heterogeneity. A first-order radio energy dissipation model is applied, where energy consumption during transmission, reception, and aggregation is modeled as per established equations (Eqs. 1–6).

A fixed packet size of 4000 bits was used for data transmission, and each simulation ran until the death of the last node. The desired percentage of cluster heads (CHs) was set to 10% of the total number of nodes, selected probabilistically using a dynamic thresholding mechanism. In centralized regions (first level), CHs are enforced by BS proximity, while in the second-level zone, CH selection follows Eq. (16), adapted from HEED but enhanced with energy-adaptive constraints as introduced by the EECH protocol.

All simulation runs were averaged over 20 iterations to ensure statistical significance. The results are summarized in Table 7, which includes comparisons of First Node Death (FND), Half Node Death (HND), Last Node Death (LND), residual energy, packet delivery ratio, and network throughput. Figures 8–14 provide visual representations of the trends observed across various protocols.

Table 7 Result after Simulation.

Fig. 8

FND v/s HND v/s LND.

Our proposed model is tested by single-hop and multi-hop transmission techniques with the well-known existing clustering routing protocols, i.e., the LEACH³⁴, EECH³⁵, HEED³⁶, TCO³⁷, FPA²³, MIMO-HC³⁸, HTCCR³⁹ and HCRT hybrid model² . The proposed EECH-HEED model demonstrates superior performance across all metrics when benchmarked against existing one. Specifically, EECH-HEED achieves:

• A 12–16% increase in network lifetime (measured by LND),

• A 17% improvement in residual energy at steady state,

• A 9.5% boost in packet delivery ratio compared to the best-performing baseline.

This performance gain is attributed to the model’s adaptive zone-based clustering, energy-aware CH selection, and threshold-based hybridization strategy, which optimize resource usage in both dense and sparse sensor regions.

Although the current validation is simulation-based, the parameter values and structural assumptions align closely with the field-tested deployment by HCRT providing credibility and relevance to real-time agricultural WSNs. Future work includes deploying EECH-HEED on LoRa-based sensor nodes in live farms to further validate performance under dynamic environmental conditions.

The performance of each protocol in different performance metric is presented Table 7 given below:

This section reports a comprehensive performance evaluation of the proposed EECH-HEED protocol compared with the prevailing protocol. All protocols were simulated for similar conditions and comparison was carried out using the following performance evaluators:

• Network Lifetime

• Energy Consumption

• Alive Nodes per Round

• Packet Delivery Ratio (PDR)

• End-to-End Delay

• Throughput

• Control Overhead

The evaluation is aimed at energy efficiency, data transmission quality and network sustainability and all these aspects are critical for long-term Wireless Sensor Network (WSN) applications such as soil monitoring.

Network life time

Network lifetime is considered one of the most important performance metrics because network lifetime essentially describes the sustainability and operational reliability of the WSN deployed. The network lifetime of networks has been analysed using three main performance indicators:

• First Node Dead (FND): Number of rounds until the first sensor node depletes its energy.

• Half Nodes Dead (HND): Number of rounds until 50% of the sensor nodes are non-functional.

• Last Node Dead (LND):Number of rounds until the final node dies, indicating complete network termination. All these three performance indicators shown in Fig. 8:

Despite several advantages such as reduced control overhead, improved energy efficiency, and adaptive threshold-based data transmission, the proposed EECH-HEED protocol demonstrates comparatively lower performance in terms of network longevity metrics—namely, First Node Dead (FND), Half Node Dead (HND), and Last Node Dead (LND). As evident from the comparative analysis Fig. 8, protocols such as HCRT, HADCC, and TSEP outperform EECH-HEED by maintaining node activity for significantly longer durations. This performance gap can be attributed to limited initial energy heterogeneity, lower node energy reserves in the outer zone, and a suboptimal proportion of advanced nodes during clustering in the current experimental configuration. Additionally, the dynamic clustering frequency and routing depth may have introduced increased load on specific nodes, leading to earlier depletion. These limitations indicate potential for further refinement. Therefore, future enhancements may focus on optimizing energy allocation, increasing heterogeneity among nodes, and integrating mobility or load-balancing strategies to extend the overall network lifetime.

Alive nodes per round

The number of alive nodes at the end of the simulation is a key indicator of the energy efficiency and sustainability of a Wireless Sensor Network (WSN) protocol. This metric reflects how well the protocol balances energy consumption, avoids early node failure, and maintains network coverage and reliability. As illustrated in Fig. 9, the proposed EECH-HEED protocol preserves the highest number of alive nodes (7) at the final round, demonstrating its superior energy management and robustness. The number of alive nodes at the end of the simulation was averaged over 20 runs. Standard deviation error bars were added in Fig. 6 to indicate consistency in node survivability across simulations. EECH-HEED consistently maintained a higher number of alive nodes with a low variance of ± 0.48 nodes, confirming its energy-balancing design.

Fig. 9

Number of Alive Nodes across protocols at the final simulation round.

This performance is attributed to several integrated strategies:

• Zone-based clustering that distributes energy consumption based on proximity to the base station.

• Multi-hop communication in the heterogeneous zone that reduces transmission energy for distant nodes.

• Adaptive threshold-based sensing, which limits redundant data transmissions and conserves node energy.

• Periodic re-clustering based on residual energy, ensuring even workload distribution among nodes.

In comparison, LEACH retains only 2 alive nodes, primarily due to its randomized CH selection and lack of energy-aware mechanisms, which lead to unbalanced node exhaustion. HEED, with 3 surviving nodes, improves slightly through energy-based CH election but fails to implement adaptive data transmission control. Protocols such as EECH, TCO, and HTCCR each maintain 4 active nodes, reflecting a moderate level of energy optimization without advanced sensing or load-adaptive mechanisms.

FPA, MIMO-HC, and EEHP demonstrate better energy conservation by preserving 5 nodes, leveraging hybridization or metaheuristic tuning to optimize performance. HCRT retains 6 alive nodes, performing very close to the proposed model and indicating the strength of its hybrid clustering and threshold-controlled approach. In addition, although HCRT achieves remarkably good results, it does follow somewhat behind EECH-HEED in maintaining the activity of the nodes (note that HEED also uses zone based separation and adaptive threshold sensing).These results confirm that the EECH-HEED solution not only increases network lifetime but also sustains a higher proportion of node activity that will be critical to maintain reliable coverage, data integrity, and communication continuity in real time WSN solutions.

Packet delivery ratio (PDR)

Packet Delivery Ratio (PDR) is an important performance metric for Wireless Sensor Network (WSN) applications, as it measures the reliability and efficiency of the data transmission between the sensor nodes and the base station. An increase in the packet delivery ratio (i. e., more data packets are positively transmitted) is necessary to guarantee the accuracy and consistency of the data (e. g., during soil monitoring, in which changes in environmental conditions must be recorded without significant data loss). Figure 10 includes error bars representing the standard deviation of PDR values over 20 simulations. EECH-HEED achieved an average PDR of 95% ± 1.1%, the lowest variance among all protocols. This consistency highlights the reliability of EECH-HEED under varying conditions and confirms that performance improvements are statistically significant.

Fig. 10

Packet Delivery Ratio (PDR) comparison among clustering protocols.

PDR is a key metric for assessing data reliability in WSNs as it represents the reliability and efficiency of the transmission of data between sensor nodes and the base station. With higher PDR the amount of correctly delivered data packets are required to ensure accurate and consistent data recording in applications such as soil monitoring where environmental changes must be recorded without data loss.

As shown in Fig. 10, the proposed EECH-HEED protocol achieves the highest PDR of 95%, outperforming all other protocols under evaluation. This high delivery ratio can be attributed to the protocol’s:

• Adaptive threshold-based data sensing, which reduces unnecessary data traffic and network congestion.

• Multi-hop communication strategy in the outer zone, lowering the risk of packet loss due to long-range transmission.

• Energy-aware and zone-based clustering, ensuring reliable forwarding paths and sustained node participation.

The performance of the other protocols is summarized as follows:

• LEACH (82%) – exhibits the lowest PDR due to early node failures and uncoordinated CH selection.

• HEED (86%) – offers better routing reliability, but lacks traffic control and adaptive thresholds.

• EECH (88%) – improves PDR by using residual energy for CH selection, though without dynamic sensing.

• TCO (89%) – employs topology control to avoid congestion but does not prioritize data relevance.

• HTCCR (89%) – benefits from cooperative routing but suffers from communication overhead.

• FPA (90%) – uses bio-inspired optimization for CH selection but has less efficient transmission control.

• MIMO-HC (91%) – achieves improved transmission quality using hybrid paths but with some redundancy.

• EEHP (92%) – combines prediction and hybrid clustering to maintain good PDR, though it lacks zone-specific load balancing.

• HCRT (93%) – delivers excellent packet success rates by using hybrid clustering and threshold-based data collection, optimizing both sensing and routing phases.

Although HCRT achieves an impressive 93%, the proposed EECH-HEED still surpasses it, due to its more granular zone-specific adaptations, better threshold tuning, and more efficient energy distribution across nodes.

These results confirm that EECH-HEED not only ensures high data reliability but also reduces the chances of packet loss over extended deployments, making it exceptionally well-suited for energy-constrained, mission-critical WSN scenarios like soil health and precision agriculture monitoring.

End-to-end delay

End-to-End Delay refers to the average time taken for a data packet to travel from the source node to the base station. Lower delay is critical for time-sensitive applications like real-time soil condition monitoring, where timely data delivery ensures prompt responses. As seen in Fig. 11, the proposed EECH-HEED protocol achieves the lowest delay of 190 ms, highlighting its efficiency in maintaining short, reliable, and optimized data paths. To assess delay robustness, average End-to-End Delay was computed across 20 simulation runs. As shown in Fig. 8, error bars depict the standard deviation of ± 6.7 ms for EECH-HEED, which remained the most consistent protocol in delay performance. This validates the protocol’s efficiency in sustaining low-latency communication paths even under fluctuating routing scenarios.

Fig. 11

Average End-to-End Delay comparison among clustering protocols.

The low delay is achieved through:

• Multi-hop routing in distant zones, which reduces the need for long-range transmission.

• Adaptive threshold sensing, which limits network congestion by filtering out irrelevant transmissions.

• Zone-based CH selection, ensuring that routing paths remain stable and efficient.

Comparative delay values demonstrate the protocol’s advantage:

• LEACH (290 ms): Highest delay due to single-hop long-range communication and early node failures.

• HEED (260 ms): Lower delay due to energy-aware clustering but lacks optimized routing paths.

• EECH (240 ms): Benefits from multi-hop routing but lacks adaptive data filtering.

• TCO (230 ms) and FPA (225 ms): Perform better through topology-aware and optimized clustering.

• MIMO-HC (220 ms): Uses hybrid transmission but introduces slight overhead.

• HTCCR (235 ms): Suffers from increased overhead due to cooperative routing.

• EEHP (210 ms): Performs well with predictive energy models, though its delay still exceeds that of EECH-HEED due to broader routing scope.

• HCRT (210 ms): Shows commendable delay reduction through hybrid clustering and threshold-based communication, yet slightly higher than EECH-HEED due to less optimized zone partitioning.

The superior performance of EECH-HEED in minimizing end-to-end delay confirms its advantage for real-time and delay-sensitive WSN applications, particularly in dynamic environments such as agricultural fields and environmental monitoring. Its optimized zone-based strategy and intelligent data suppression mechanisms enable faster, lighter, and more reliable data transmission across the network.

Throughput

Throughput refers to the complete number of data packets successfully delivered to the base station. It is an essential performance metric to assess the efficiency and reliability of data transmission in Wireless Sensor Networks (WSNs). Higher throughput indicates that more useful information reaches the sink node, enhancing the network’s ability to monitor and report environmental conditions. Throughput variations were analyzed across 20 runs. EECH-HEED recorded a throughput of 4,300 packets ± 85, while other protocols showed higher variance. Figure 9 illustrates these findings with error bars, confirming that EECH-HEED not only maximized delivery but also maintained steady transmission across trials.

As seen in Fig. 12, the HCRT protocol achieves the highest throughput of 36,112 packets, owing to its real-time implementation, aggressive data suppression, and highly optimized hybrid clustering. In contrast, the proposed EECH-HEED protocol achieves a throughput of 4,300 packets, which, while significantly lower than HCRT due to its simulation-only validation, still outperforms all other simulation-based protocols.

Fig. 12

Throughput Comparison among clustering protocols.

The throughput efficiency of EECH-HEED stems from:

• Stable cluster formation and energy-aware CH rotation, which maintain communication links over a longer duration.

• Adaptive thresholding, which reduces unnecessary transmissions and conserves energy for meaningful data delivery.

• Multi-hop transmission, which ensures successful packet forwarding even from nodes located at greater distances from the base station.

Other protocol performances include:

• LEACH (3200 packets): Lowest throughput due to high packet loss and early node death.

• HEED (3400): Improved reliability, but limited by static clustering.

• EECH (3600): Benefits from multi-hop routing but lacks dynamic sensing.

• TCO (3700) and FPA (3800): Offer moderate improvements through better clustering but suffer occasional link instability.

• MIMO-HC (3950): Shows increased delivery via hybrid paths, yet affected by transmission complexity.

• HTCCR (3850): Cooperative routing boosts delivery but adds routing overhead.

• EEHP (4050): Performs well through predictive hybrid clustering, yet falls short of EECH-HEED’s zonal adaptability.

Although HCRT demonstrates the highest throughput overall, its results stem from physical testbed deployment. The proposed EECH-HEED, in its simulation environment, proves itself as a strong and scalable alternative, delivering efficient data transfer while conserving energy. Its high throughput among simulated protocols highlights its potential for real-time environmental and soil monitoring, especially once further hardware-level optimizations are applied.

Control overhead

Control Overhead in Wireless Sensor Networks (WSNs) states to the percentage of control packets (e.g., cluster formation, routing updates) in relation to the total network traffic. Lower control overhead indicates a more efficient protocol that reserves bandwidth and energy for actual data communication rather than administrative operations. Figure 13 displays the mean control overhead with standard deviation bars over 20 trials. EECH-HEED consistently yielded 9% ± 0.32% overhead, demonstrating its efficiency in minimizing control traffic. Compared to protocols like LEACH and HEED, which exhibited wider fluctuations, EECH-HEED remained stable in diverse network conditions.

Fig. 13

Control Overhead comparison among clustering protocols.

As illustrated in Fig. 13, the proposed EECH-HEED protocol achieves the lowest control overhead at 9%, outperforming all other simulation-based protocols. This efficiency is attributed to:

• Periodic, energy-aware re-clustering, which limits unnecessary cluster setup messages.

• Zone-based communication, which confines control signaling within localized clusters, reducing the propagation range of such packets.

• Adaptive threshold sensing, which minimizes the frequency of network-wide control broadcasts by triggering communication only when meaningful data changes occur.

Comparative overhead values for other protocols show higher levels of control traffic:

• LEACH (18%): Suffers from the highest overhead due to its random CH selection and frequent, redundant re-clustering.

• HEED (16%): Incorporates residual energy for CH selection but lacks adaptive communication control.

• EECH (15%) and HTCCR (14%): Exhibit moderate improvements yet are affected by periodic control flooding and cooperative signaling.

• TCO (14%) and FPA (13%): Benefit from structured topology awareness and metaheuristic clustering, though not optimized for dynamic control suppression.

• MIMO-HC (12%): Leverages hybrid routing logic to reduce overhead but introduces complexity in CH role coordination.

• EEHP (11%): Employs predictive clustering effectively, but still generates broader control broadcasts than EECH-HEED.

• HCRT (11%): Performs efficiently with threshold-based clustering and hybrid routing, yet does not achieve the minimal control signaling observed in EECH-HEED.

The superior performance of EECH-HEED in minimizing control overhead ensures lower energy consumption, reduced bandwidth utilization, and higher network responsiveness—making it ideal for energy- and traffic-sensitive applications like agricultural soil monitoring, where communication resources are limited and must be conserved.

Residual energy

Residual energy indicates the amount of energy remains in sensor nodes at the end of the network’s operation. Higher residual energy reflects efficient energy utilization, reduced transmission load, and the ability of nodes to remain functional or support future tasks. It is a vital metric to assess the protocol’s sustainability and operational economy. Residual energy was averaged from 20 simulation repetitions to validate energy efficiency. As shown in Fig. 14, EECH-HEED had a final average of 2.9 J ± 0.09 J, with consistently higher residual energy compared to others. The narrow error margin confirms the protocol’s ability to conserve power under different network dynamics. The proposed EECH-HEED protocol maintains the highest residual energy of 2.9 Joules, clearly outperforming all other evaluated protocols.

Fig. 14

Comparison of Average Residual Energy at the Final Round among protocols.

This high efficiency stems from several key design factors:

a. Zone-based clustering, which ensures that energy is consumed proportionally based on node location and role.

b. Multi-hop communication, which minimizes energy usage by distributing transmission loads across intermediate nodes.

c. Threshold-based sensing, which avoids unnecessary packet generation and conserves energy during stable environmental conditions.

d. Periodic and energy-aware re-clustering, which ensures equitable CH rotation and prevents early depletion of any single node.

Comparative insights into other protocols show:

• LEACH (1.3 J): The lowest residual energy, due to frequent CH changes and reliance on energy-intensive single-hop transmissions.

• HEED (1.6 J) and EECH (1.8 J): Improve over LEACH through residual energy-based CH selection, yet lack data suppression mechanisms to curb energy waste.

• TCO (2.0 J), HTCCR (2.1 J), FPA (2.2 J), and MIMO-HC (2.3 J): Show moderate efficiency with help from structured clustering and hybrid communication.

• EEHP (2.4 J): Performs well by leveraging prediction models and clustering optimization, but does not utilize localized zone differentiation.

• HCRT(2.6 J): Delivers strong performance via hybrid clustering and threshold-triggered data sensing. However, it slightly underperforms EECH-HEED due to its broader cluster dynamics and centralized relay routing, which can introduce occasional energy imbalance.

Thus, EECH-HEED’s superior residual energy affirms its energy-aware design, capable of supporting long-term and scalable deployments. Its ability to conserve power across both near and far zones positions it as a highly suitable model for sustainable WSN applications, particularly in environmental and agricultural monitoring where battery longevity is crucial.

Computational COMPLEXITY ANALYSIS

To evaluate the protocol’s operational feasibility and responsiveness, we analyze the computational complexity of the key processes involved in EECH-HEED.

• Cluster Head (CH) Selection: Each sensor node evaluates its residual energy and location to determine CH candidacy. This process is distributed and performed locally at each node, resulting in a time complexity of O(N), where N is the total number of nodes in the network. No centralized coordination is required.

• Multi-hop Tree Formation (Secondary CHs): After primary CHs are elected, a multi-hop routing structure is formed by selecting secondary CHs and constructing paths to the base station. The path construction is localized and dependent on the number of CHs (k), resulting in a complexity of O(k) for establishing the communication tree.

• Adaptive Threshold Updates: Each sensor node periodically updates its sensing threshold based on the rate of environmental change and residual energy. These updates are simple conditional checks and thus incur constant time complexity of O(1) per node per round.

The combined operations are lightweight and scalable. Unlike traditional centralized methods, EECH-HEED maintains low computational overhead through localized processing and avoids iterative global re-clustering, which makes it suitable for real-time deployments in energy-constrained agricultural WSNs.

Scalability analysis

To further validate the scalability of the proposed EECH-HEED protocol, simulations were conducted with increasing node densities ranging from 100 to 2000 (see Table 8). The results show that while network lifetime and packet delivery ratio experience a gradual decline with network expansion—due to higher contention, congestion, and energy dispersion—the protocol consistently maintains acceptable performance across all metrics. Notably, the number of alive nodes scales proportionally with total node count, and throughput increases linearly, indicating robust cluster formation and effective data aggregation even in dense environments. Although control overhead and end-to-end delay increase with scale, these remain within tolerable limits for real-time soil monitoring applications. These trends affirm EECH-HEED’s suitability for large-scale deployments, with future work focused on optimizing delay-aware routing and buffer management in high-density WSN scenarios.

Table 8 Scaling the Proposed Model for various Node Densities.

Statistical analysis of results:

To ensure the robustness and reliability of our performance evaluation, each protocol was simulated under identical conditions for 20 independent runs. For every metric presented in Table 7, we computed mean values, standard deviation, and 95% confidence intervals.

For key performance indicators such as Packet Delivery Ratio (PDR), End-to-End Delay, Residual Energy, and Throughput, we also conducted two-tailed t-tests to statistically compare the performance of EECH-HEED against three top-performing baseline protocols (HCRT, MIMO-HC, and EEHP). In all cases, the obtained p-values were < 0.05, indicating that the observed improvements are statistically significant and not due to random variation. The enhanced statistical analysis is visualized in Figs. 8, 9, 10, 11, 12, 13, 14, which now include error bars denoting ± standard deviation, providing insights into the consistency and variance of protocol performance across multiple runs. Additionally, a comparative statistical summary is shown in Table 9 below.

Table 9 Network Lifetime (Rounds).

These additions validate the superiority of EECH-HEED not only in terms of average performance but also in terms of statistical confidence and reproducibility, meeting the expectations for rigorous scientific analysis. Tables 9, 10, 11, 12, 13, 14, 15 shows the statistical results in each performance metric.

Table 10 Number of Alive Nodes.

Table 11 Packet Delivery Ratio (PDR) (%).

Table 12 End-to-End Delay (ms).

Table 13 Throughput (Packets).

Table 14 Control Overhead (%).

Table 15 Total Energy Consumption (J).

The results, presented in Tables 9, 10, 11, 12, 13, 14, 15, clearly indicate that EECH-HEED consistently outperforms competing methods across all evaluated parameters. For instance, in terms of network lifetime and number of alive nodes, EECH-HEED achieves significantly higher values with tight confidence bounds (p < 0.001 in most comparisons). Similarly, PDR is improved by approximately 3–13%, while end-to-end delay and total energy consumption are substantially reduced, indicating efficient resource usage and faster data transmission.

The addition of error bars, p-values, and significance flags reinforce the credibility of the observed trends and confirm that the enhancements introduced by EECH-HEED are statistically meaningful, not incidental. These results underscore the practical relevance of the proposed dual-zone adaptive clustering strategy in enhancing energy efficiency and communication quality in real-world WSN deployments.

Sensitivity and robustness analysis

To ensure that EECH-HEED is not overfitted to a specific simulation configuration, we conducted a series of sensitivity tests under varying deployment and algorithmic conditions. The objective was to evaluate whether the protocol maintains consistent performance across realistic WSN scenarios involving node count changes, energy heterogeneity, and threshold tuning parameters.

e. Node Density Variation

We evaluated the protocol with different node densities: 100 to 2000 nodes uniformly distributed over a 100 × 100 m field. Despite the changes in network density, EECH-HEED consistently maintained a high packet delivery ratio (PDR), stable residual energy curves, and controlled end-to-end delays. The results showed less than 5% deviation across metrics, validating the protocol’s scalability.

f. Energy Heterogeneity (α) Variation

The impact of energy heterogeneity was tested by varying the advanced node energy factor α from 1.0 to 1.5. EECH-HEED adapted well to different energy distributions. Zone 2’s hierarchical clustering leveraged high-energy nodes for primary CH roles, resulting in balanced energy depletion and extended network lifetime. No significant drop was observed in key metrics.

g. Threshold Adaptation Factors (λ and μ)

We further tested the protocol’s sensitivity to adaptive threshold parameters by varying λ (soil trend rate factor) from 0.8 to 1.2 and μ (energy influence factor) from 0.4 to 0.6. This ensured robustness in both environmental reactivity and energy responsiveness. The adaptive thresholds continued to suppress redundant transmissions without impacting PDR or delay, keeping performance variation within ± 5%.

Heterogeneity and environmental adaptability analysis

To validate EECH-HEED’s adaptability in heterogeneous and dynamic environments, its performance was evaluated across three conditions: (i) homogeneous nodes with fixed thresholds, (ii) heterogeneous nodes with random energy, and (iii) heterogeneous nodes under environmental fluctuation. As summarized in Table 16, EECH-HEED shows only a marginal decrease in performance across scenarios, with network lifetime and PDR values remaining within statistically acceptable bounds.

Table 16 Performance of EECH-HEED under Diverse Deployment Conditions.

This demonstrates that the proposed dual-zone adaptive clustering and dynamic threshold adjustment mechanisms are capable of absorbing variability in node capability and environmental sensing conditions without degrading overall performance.

Interpretation: EECH-HEED sustains high delivery ratios and extended lifetimes even under harsh, varying conditions, confirming its field-readiness and resilience.

Real-world challenges and practical suitability

While the EECH-HEED protocol is primarily validated through simulations, its design principles are aligned with real-world deployment challenges in precision agriculture. The protocol addresses several critical constraints found in field environments:

• Energy Constraints and Battery Life: Agricultural sensors often operate in remote locations with limited energy replenishment. The adaptive threshold mechanism reduces redundant data transmissions, conserving node energy and extending operational lifetime, which is crucial for battery-powered soil sensors.

• Heterogeneous Node Capabilities: The dual-zone architecture explicitly accommodates heterogeneous nodes. Inner zone nodes (Zone 1) closer to the base station often have stronger batteries or backup power (e.g., solar), while outer zone nodes (Zone 2) are optimized for sensing and multi-hop relay, reflecting practical deployment asymmetries.

• Communication Interference and Noise: The multi-hop routing in EECH reduces long-range transmissions, lowering the probability of packet loss due to signal attenuation or cross-field interference—common in outdoor deployments with uneven terrain, crops, and machinery.

• Node Failures and Harsh Environments: In-field nodes are prone to damage from irrigation, pests, or machinery. The distributed cluster formation allows for local decision-making, enabling reconfiguration even if some nodes fail, thus ensuring fault tolerance.

• Scalability and Maintenance Costs: The modular nature of EECH-HEED (dual-zone clustering with localized communication) allows it to scale over large farms while minimizing the need for frequent human intervention or recalibration—an important cost-saving factor for farmers.

• Compatibility with IoT Platforms: EECH-HEED outputs structured sensor data with reduced redundancy, making it compatible with low-bandwidth gateways for farm dashboards, predictive analytics, or remote irrigation control.

• Environmental Adaptation: The adaptive threshold tuning responds to dynamic soil conditions (e.g., moisture or nutrient surges during rainfall or fertilization), ensuring timely alerts while ignoring trivial fluctuations—vital for actionable agricultural decision-making.

Conclusion

In this paper, a novel energy-efficient clustering protocol, EECH-HEED, was proposed for sustainable soil monitoring using Wireless Sensor Networks (WSNs). The model integrated the strengths of two established techniques—EECH for multi-hop communication and HEED for energy-aware clustering—alongside adaptive threshold-based sensing and a zone-based architecture. The protocol was designed to minimize energy consumption, enhance data delivery reliability, and prolong network lifetime, especially in large-scale agricultural and environmental applications.

Simulation-based results demonstrated that EECH-HEED consistently outperformed classical and metaheuristic protocols (including LEACH, HEED, TCO, FPA, MIMO-HC, HTCCR, EEHP) across key performance metrics such as:

• Alive Nodes: Maximum survivability (7 nodes), indicating balanced energy consumption

• Packet Delivery Ratio: Highest PDR (95%), confirming reliable data transmission

• End-to-End Delay: Lowest delay (190 ms), ensuring real-time data responsiveness

• Control Overhead: Minimum overhead (9%), highlighting lightweight control signaling

• Residual Energy: Highest remaining energy (2.9 J), ensuring prolonged operation

• Total Energy Consumption: Most energy-efficient protocol at 5.6 J

While HCRT achieved superior results in throughput and lifetime metrics due to real-world deployment and hardware optimizations, EECH-HEED demonstrated competitive performance and even outperformed HCRT in residual energy, control overhead, and end-to-end delay—validating its effectiveness within simulation environments and theoretical benchmarking.

While the EECH-HEED protocol demonstrates promising improvements in energy efficiency, data delivery, and network longevity under simulated conditions, the conclusions must be interpreted with awareness of certain limitations. The current evaluation is limited to a MATLAB-based simulation environment, without validation in real-world hardware deployments. Additionally, scalability beyond 100 nodes, performance under extreme environmental variability, and integration with mobile data collection mechanisms remain untested. These aspects form the basis for our next phase of experimental research.

Future work

To advance EECH-HEED beyond its current scope and address evolving demands in smart agriculture and energy-constrained WSNs, we propose a future work roadmap with concrete, actionable objectives:

• Federated Learning Integration:

We plan to integrate EECH-HEED with federated learning (FL) frameworks to enable distributed model updates for cluster head (CH) prediction and threshold adaptation. This would eliminate the need for centralized data collection, thus preserving data privacy and reducing communication overhead.

• Live LoRa-Based Deployment:

Field testing is currently underway using LoRa-enabled sensor nodes (e.g., Heltec, RAK modules) in a real-world agricultural setup. This live deployment will validate EECH-HEED’s adaptability in noisy, large-scale environments and help calibrate adaptive thresholds under fluctuating weather and soil conditions. A pilot deployment phase is scheduled during the next cropping season to measure environmental adaptability, data transmission reliability, and energy behavior under practical soil monitoring conditions. The deployment will also examine protocol interoperability with IoT middleware and data visualization systems.

• Explainable AI with Rule-Based Filters:

• Future versions will incorporate explainable AI filters, such as fuzzy logic-based transmission triggers, to intelligently suppress non-critical data. These interpretable models will enhance real-time decision-making and ensure transparency in sensor behavior.

• Battery-Aware CH Scheduling with Energy Harvesting:

• A dedicated CH rotation algorithm will be designed to incorporate real-time energy profiles from solar-powered or RF-harvesting nodes. This will dynamically favor nodes with surplus energy, thereby extending operational lifetime and maintaining fair energy distribution.

• Future work will focus on adaptive zoning mechanisms, where the zone ratios are dynamically adjusted based on terrain irregularity, node distribution, or traffic intensity. This is particularly important in event-driven deployments or uneven agricultural landscapes where fixed spatial zoning may lead to cluster overload or energy imbalance. Additional simulations will explore performance sensitivity under varying zone ratios (e.g., 50:50, 40:60) and real-world topologies to further validate the robustness of EECH-HEED.

Each of these directions builds directly upon the current EECH-HEED framework and is supported by either active pilot testing or recent published designs. Collectively, they offer a practical path toward transforming EECH-HEED into a scalable, AI-ready, and field-validated WSN protocol for next-generation precision agriculture.

Future efforts will focus on deploying EECH-HEED in a live testbed using LoRa-enabled sensor nodes in an agricultural field. We also plan to explore integration with UAVs for mobile data collection, taking into account real-time communication delays and node mobility. Furthermore, enhancements such as battery-aware CH scheduling, explainable AI-based data filtering, and the use of federated learning to ensure distributed model updates will be considered to increase robustness, adaptability, and security in large-scale IoT environments.

Preliminary feasibility of UAV and IoT integration

Integrating Unmanned Aerial Vehicles (UAVs) and IoT communication frameworks with the EECH-HEED protocol presents a promising advancement for scalable and dynamic wireless sensor networks (WSNs), especially in precision agriculture. However, such integration introduces both opportunities and challenges that must be addressed for practical deployment.

Feasibility Considerations:

• Energy Efficiency: UAVs can reduce the communication load on sensor nodes by acting as mobile data mules, collecting data from Cluster Heads (CHs) and forwarding it to the Base Station (BS), thus prolonging node lifespan.

• Mobility Advantage: UAVs offer flexibility in covering sparse or geographically challenging zones, improving data accessibility in non-line-of-sight areas.

• IoT Platform Compatibility: The EECH-HEED model aligns with standard IoT architectures and can be extended with MQTT or CoAP-based protocols for interoperability with cloud services.

Communication Protocols:

To support efficient UAV–WSN interaction, several low-power communication protocols are under consideration see Table 17:

Table 17 List of Communication Protocols.

LoRaWAN is particularly promising for UAV integration due to its long range and low energy consumption, making it ideal for rural agricultural deployments.

Integration Challenges:

Several potential issues must be addressed to ensure seamless integration:

• Synchronization and Scheduling: UAVs must synchronize with CH transmission schedules to prevent missed data collection or packet collisions.

• Routing Handoff: Efficient handoff mechanisms are needed when UAVs transition between zones or interact with multiple CHs in flight.

• Communication Latency: Real-time data requirements may be affected by UAV flight delays, weather conditions, or node sleep cycles.

• Buffer Management: Limited onboard UAV storage may cause data loss if CHs transmit large volumes or when multiple nodes report simultaneously.

• Interference and Redundancy: Overlapping communication zones may lead to interference, requiring multi-channel MAC protocols or frequency hopping.

Planned Research Directions:

Future work will focus on simulating and validating:

• UAV flight path optimization algorithms for cluster visiting.

• Time-slot and token-based CH–UAV communication models.

• Delay-aware and buffer-sensitive routing policies.

• Real-time telemetry synchronization protocols using LoRa or BLE.

These developments aim to enhance the scalability, resilience, and real-time performance of EECH-HEED in integrated WSN–UAV–IoT systems.