💻 computer science

WSN-Derived Node Deployment and Data Transmission Model Using HEED Routing Protocol for Large-Scale IoT Network Applications

This paper proposes an adaptive Wireless Sensor Network-based IoT node deployment and data transmission model that integrates the HEED clustering protocol with IEEE 802.11x standards and deep learning analytics to optimize energy efficiency, extend network lifetime, and enhance Quality of Service in large-scale IoT applications.

Original authors: Abirama Sundari Purshothaman, C. Sankar Ram

Published 2026-06-26

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Abirama Sundari Purshothaman, C. Sankar Ram

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a massive city where thousands of tiny, battery-powered messengers (these are your IoT sensors) are scattered everywhere to watch over things like traffic, weather, or factory machines. The problem? These messengers have very small batteries. If they all try to shout their messages at once to a central headquarters, they run out of power quickly, and the city goes dark.

This paper proposes a smarter way to organize these messengers so the city stays alive and working for much longer. Here is how they did it, broken down into simple concepts:

1. The "Neighborhood Captain" System (Clustering)

Instead of every single messenger shouting directly to the main headquarters (which wastes energy), the paper suggests organizing them into neighborhoods.

The Analogy: Think of a large office building. Instead of every employee walking all the way to the CEO's office to drop off a memo, they give their memo to a Team Captain in their department. The Captain gathers all the memos, summarizes them into one big report, and walks it to the CEO.
The Paper's Twist: In this system, the "Team Captains" (called Cluster Heads) are chosen based on who has the most battery left and is in the best spot. If a Captain gets tired (low battery), a new one is picked from the neighborhood so no one burns out too fast.

2. The "Smart Selector" (HEED Protocol)

The paper uses a specific rulebook called HEED to pick these captains.

The Analogy: Imagine a game where you need to pick a team leader. Old methods might just pick the person standing closest to the door. But HEED is smarter: it looks at two things:
1. Energy: "Do you have enough juice in your battery to run this errand?"
2. Distance: "Are you close enough to the other team members to talk easily?"
By balancing these two, the system ensures the leaders are strong and the team doesn't have to shout too loudly to be heard.

3. The "Recharging Station" (Energy Management)

The paper mentions a strategy where the system doesn't just wait for batteries to die.

The Analogy: It's like a relay race where runners don't just run until they collapse. The system has a plan to "recharge" or swap out nodes before they completely die, keeping the network alive longer. It also uses a "sleep mode" (like a cat napping) where messengers turn off their radios when they aren't talking, saving precious battery.

4. The "Super-Brain" at the End (Deep Learning)

Once the Team Captains gather all the reports and send them to the main headquarters (the Base Station), the paper introduces a Deep Learning model (specifically an LSTM).

The Analogy: Imagine the headquarters receives a mountain of paperwork. Instead of a human trying to read every single line, a Super-Brain computer (the AI) instantly reads the summary, spots patterns (like "it's going to rain" or "a machine is breaking"), and makes a decision.
The paper tested this using a dataset of network traffic (including fake cyber-attacks) and found that this AI brain was very good at spotting the difference between normal activity and trouble.

5. The Results: Who Won the Race?

The authors ran a computer simulation (a virtual test drive) to see how their new system compared to older methods.

The Outcome: Their new system (which they call DL-HEED) kept more messengers alive for longer than the old systems.
- More Survivors: In a test with 500 messengers, their system kept about 250 of them alive and working, while the old system only kept about 100 alive.
- Faster Delivery: Messages arrived with less delay (less traffic jams).
- Better Accuracy: The AI brain correctly identified "attacks" or problems 96.4% of the time, which is better than the previous best models.

Summary

In short, this paper says: "Don't let your tiny sensors shout randomly until they die. Group them into neighborhoods, pick smart leaders based on who has the most energy, let them sleep when they aren't talking, and use a super-smart computer at the end to make sense of the data." This keeps the whole network running longer, faster, and more reliably.

Technical Summary: WSN-Derived Node Deployment and Data Transmission Model Using HEED Routing Protocol for Large-Scale IoT Network Applications

Problem Statement
The rapid expansion of Internet of Things (IoT) applications in domains such as healthcare, smart cities, and industrial automation has led to the deployment of large-scale Wireless Sensor Networks (WSNs). However, these networks face significant challenges regarding energy efficiency, reliability, and scalability. Traditional centralized cloud processing models suffer from high communication overhead, latency, and bandwidth constraints, particularly in dynamic environments where node availability and traffic patterns fluctuate. Furthermore, existing routing and clustering protocols often fail to adapt to resource-constrained conditions, leading to uneven energy consumption, premature node death, and data loss. The core problem addressed is the need for an adaptive framework that optimizes node deployment, clustering, and data transmission to extend network lifetime while ensuring reliable, low-latency communication in large-scale IoT environments.

Methodology
The paper proposes an adaptive, multi-layer IoT-WSN architecture that integrates a Hybrid Energy-Efficient Distributed (HEED) routing protocol with deep learning-based data analytics. The methodology is structured as follows:

Network Architecture and Deployment:
- The network is modeled as a multi-layer system comprising device, cluster, edge gateway, and cloud layers.
- IoT nodes are deployed using a probabilistic distribution model within a defined region ( $N \times M$ ).
- Communication utilizes IEEE 802.11x standards for node deployment and routing, while low-power standards (IEEE 802.15.4/ZigBee) are used for sensing and local buffering.
- A closed-loop feedback mechanism dynamically updates node, route, and transmission behaviors.
Clustering and Routing (HEED Protocol):
- The study implements a modified HEED protocol for clustering. Unlike standard probabilistic methods, HEED selects Cluster Heads (CHs) based on two primary parameters: residual energy and node degree (proximity/communication cost).
- CH Selection Process: The protocol uses an iterative, distributed approach. Nodes calculate a probability of becoming a CH based on residual energy ( $E_{residual}$ ) and maximum energy ( $E_{max}$ ). The communication cost is defined by the distance between nodes ( $d_{ij}$ ).
- Data Aggregation: CHs act as data aggregators, collecting data from intra-cluster nodes, performing redundancy removal and fusion, and transmitting aggregated data to the Base Station (BS).
- Energy Management: The model incorporates a charging and recharging system to maintain node energy levels. It employs Time Division Multiple Access (TDMA) for intra-cluster scheduling to minimize collisions and idle listening. A dynamic duty-cycling mechanism switches nodes between active and sleep states.
Deep Learning Integration:
- Aggregated data is transmitted to a cloud-based server where a Deep Learning (DL) model processes the time-series data.
- Specifically, a Long Short-Term Memory (LSTM) model is deployed to analyze temporal dependencies in the IoT data for prediction and decision-making.
- The paper also explores the integration of Graph Neural Networks (GNN) to enhance CH selection by analyzing node features (energy, location, RSSI, degree) as a graph structure ( $G = (V, E)$ ).
Simulation and Evaluation:
- The network model is simulated using Network Simulator 2 (NS2).
- The BoT-IoT dataset (specifically a 5% filtered subset containing normal and attack traffic) is used to validate the system's performance in detecting anomalies and managing traffic.
- Performance metrics include network lifetime, throughput, packet delivery ratio (PDR), end-to-end delay, and energy consumption.

Key Contributions
The paper outlines the following specific contributions to improving Quality of Service (QoS) and energy efficiency in WSN-based IoT networks:

Adaptive Node Deployment: A random and hierarchical deployment model that organizes nodes for regional monitoring, coupled with a charging/recharging strategy to sustain energy levels.
Enhanced HEED Protocol: Implementation of HEED for energy-efficient clustering, cost-effective inter-cluster communication, and time-division management to avoid data collisions and redundancy.
Optimized Data Aggregation: Selection of energy-efficient data aggregators (CHs) to collect and forward data to a hub, improving bandwidth usage.
Deep Learning Integration: Deployment of an LSTM model on the server to process continuous time-series data, enhancing decision-making accuracy in dynamic networks.
Closed-Loop Feedback: A model to dynamically monitor and update node behavior, routes, and transmission parameters.
Simulation Framework: A comprehensive simulation in NS2 with Python-based DL processing, comparing results against existing models like LEACH, PEGASIS, and TEEN.

Results
The simulation results demonstrate that the proposed DL-HEED model outperforms the standard HEED protocol and other existing models (LEACH, PEGASIS, DEEC) across several metrics:

Network Lifetime: DL-HEED significantly extends the number of rounds before node death. For 500 nodes, DL-HEED maintains approximately 250 live nodes beyond 1500 rounds, whereas standard HEED drops below 100 nodes.
Packet Delivery Ratio (PDR): DL-HEED achieves higher PDR rates. At 500 nodes, DL-HEED reaches 91.3%, compared to 82.5% for standard HEED.
Energy Efficiency: The proposed model maintains higher average residual energy. At 500 nodes, DL-HEED retains 0.09 J compared to 0.07 J for HEED.
Latency: DL-HEED exhibits lower end-to-end delay compared to HEED across varying network sizes.
Classification Performance: Using the BoT-IoT dataset, the proposed model achieved an accuracy of 96.4%, a precision of 95.2%, and an F1-score of 94.9%, outperforming the Hybrid Clustering Model (93.6% accuracy) and DEEC-based models.

Significance and Claims
The authors claim that this research bridges the gap between theoretical energy-efficient models and scalable, real-world IoT solutions. By integrating the HEED protocol with deep learning (LSTM and GNN), the framework offers a robust solution for next-generation IoT applications that require:

Extended Network Lifetime: Through balanced energy consumption and recharging strategies.
Scalability: The ability to maintain performance as network density increases from 50 to 500 nodes.
Reliability: Improved data delivery and reduced latency in dynamic environments.
Intelligent Processing: The ability to perform real-time anomaly detection and predictive analytics directly on aggregated sensor data.

The paper concludes that while the current model relies on simulation and assumes a stable initial clustering structure, it provides a foundational adaptive architecture. Future work is identified as necessary to address mobility, hardware implementation, and enhanced security mechanisms.