An Adaptive Multi-Metric Fuzzy Logic-Based Parent Selection Strategy for Scalable RPL Routing in Heterogeneous WSNs

In wireless sensor networks (WSNs), the Routing Protocol for Low-Power and Lossy Networks (RPL) is the de facto standard for routing. However, its performance degrades in large-scale, dynamic, or resource-constrained deployments owing to unbalanced load distribution, premature node failures, and limited scalability. This study proposes an adaptive parent selection framework that integrates three critical parameters - residual energy, queue length, and parent capacity into a unified decision model. A Mamdani-type fuzzy inference system processes these inputs to compute an optimal parent score, while an autoregressive moving average (ARMA) model monitors packet arrival intervals to regulate parent reselection. This ensures that route changes are triggered only by significant traffic deviations, avoiding unnecessary switching and improving stability. Additionally, by incorporating the MAC-layer feedback from IEEE 802.15.4, the method enhances route reliability, balances energy consumption, and reduces retransmissions. Simulation studies over a $500\times 500$ m2 region with 100–1000 nodes suggested significant improvements compared to benchmark protocols, achieving 83.91% throughput efficiency, a 46.7% reduction in energy consumption, and extended network lifetime.