NCRDAP: An AI-driven clustering routing and data aggregation protocol for energy-efficient wireless sensor networks

Wireless Sensor Networks (WSNs) play a pivotal role in numerous Internet of Things (IoT) applications; however, their performance remains constrained by limited energy resources, inefficient clustering, suboptimal routing, and redundant data transmissions. To address these persistent challenges, this study hypothesizes that integrating intelligent optimization techniques can simultaneously improve energy efficiency, network longevity, and data reliability in WSNs. Accordingly, we propose a novel AI-driven framework titled Neural-optimized Clustering, Routing, and Data Aggregation Protocol (NCRDAP). The framework combines an Artificial Neural Network (ANN)-based Cluster Head (CH) selection mechanism, an enhanced Quantum Particle Swarm Optimization (QPSO) for multi-hop routing, and a dual-step data aggregation strategy using edge computing to reduce redundancy and minimize communication overhead. The methodology was implemented and evaluated through extensive simulations using MATLAB R2022a, incorporating widely accepted radio energy models and comparative benchmarks including Low Energy Adaptive Clustering Hierarchy (LEACH), LEACH-Centralized (LEACH-C), LEACH with Genetic Algorithm (LEACH-GA), Cluster-Based Data Aggregation (CBDA), Power-efficient and Scalable Adaptive Network (PSAN), and Energy-aware Network Selection (ENS) protocols. The experimental results demonstrate that NCRDAP extends network lifetime by 19–23 %, enhances throughput by 20–30 %, and reduces overall energy consumption and packet loss ratio compared to existing techniques. Furthermore, QPSO exhibited faster convergence behavior and superior routing efficiency, while the dual-step edge processing strategy significantly reduced redundant transmissions without imposing substantial computational overhead. These findings confirm that the proposed NCRDAP framework offers a scalable, energy-efficient, and reliable solution for real-time, resource-constrained WSN applications.