Refining Network Lifetime of Wireless Sensor Network Using Energy-Efficient Clustering and DRL-Based Sleep...

Area of Science:

• Computer Science
• Electrical Engineering
• Network Engineering

Background:

• Wireless Sensor Networks (WSNs) are crucial for applications like environmental monitoring and military operations.
• Key challenges in WSNs include reducing network delay and extending network lifetime.
• Existing methods often struggle to balance energy efficiency and performance.

Purpose of the Study:

• To propose a novel algorithm, Energy-Efficient Scheduling using Deep Reinforcement Learning (E²S-DRL), to address WSN limitations.
• To enhance network lifetime and reduce data transmission delays in WSNs.
• To optimize energy consumption across sensor nodes.

Main Methods:

• The E²S-DRL algorithm employs a three-phase approach: clustering, duty-cycling, and routing.
• Zone-based Clustering (ZbC) using hybrid Particle Swarm Optimization (PSO) and Affinity Propagation (AP) for data aggregation.
• Deep Reinforcement Learning (DRL) for node duty-cycling and Ant Colony Optimization (ACO) with Firefly Algorithm (FFA) for routing.
• Network simulation using Network Simulator 3.26 (NS3).

Main Results:

• E²S-DRL effectively reduces energy consumption in WSNs.
• Network delay is decreased by up to 40%.
• Network throughput and lifetime are enhanced by up to 35% compared to existing methods (cTDMA, DRA, LDC, iABC).

Conclusions:

• The E²S-DRL algorithm presents a significant improvement in WSN performance.
• It offers a viable solution for extending network lifetime and minimizing delays.
• The proposed method demonstrates superior efficiency over conventional WSN protocols.