Performance Evaluation of Fuzzy Logic and DQN for Energy Optimization in WSNs with variable BS positions
TL;DR
This study proposes a novel Energy-Aware Cluster Optimization Routing (EACO-R) protocol, which creates adaptive energy-aware routing paths de-pending on residual energy, connection stability, transmission distance, and network congestion.
Abstract
Wireless Sensor Networks (WSNs) play an im-portant role in Internet of Things (IoT) appli-cations because they enable continuous envi-ronmental monitoring and data collection. Nu-merous previous studies introduced clustering and routing algorithms for efficient data trans-fer in WSN. However, they still face issues such as high latency, minimum network life-time, high energy consumption, and commu-nication delay. To address these issues, this study proposes a novel Energy-Aware Cluster Optimization Routing (EACO-R) protocol. Ini-tially, Fuzzy Enhanced Hierarchical Clustering (FEHC) is utilized to generate stable and bal-anced clusters by accounting for uncertainty in sensor node distribution. The best cluster heads are then selected by an Adaptive Levy Mutation-based Dynamic Enzyme Action Op-timizer (ALM-DEAO) employing residual en-ergy, local node density, and Euclidean dis-tance to the base station. Finally, the Multi-Head Attention-assisted Dynamic Priority Ad-justment Deep Q-Network (MHA-DPA-DQN) creates adaptive energy-aware routing paths de-pending on residual energy, connection stability, transmission distance, and network congestion. The simulation results demonstrate that the pro-posed model obtained 42.4J in energy consumption analysis, which is lower than the existing approach