IPV6 routing protocol for low power lossy network (RPL) is a standardized routing protocol developed by the IETF ROLL working group. RPL has attained a lot of attraction to the research community as it demands modification adding to its improvement. RPL constructs a Destination Oriented Directed Acyclic Graph (DODAG) considering the metrics and constraints through explicit Objective Functions (OFs). The OFs chooses the best parent for reliable data communication. The existing objective function due to a single metric lacks in satisfying the requirement of real-time IoT applications. Still, the mobility aspect of the node is not considered in the objective function for optimal path selection. In the proposed methodology a Fuzzy Supervised Learning algorithm for quality node prediction using the composite metric based OF from the candidate parent set. The testing phase shows 93% accuracy of prediction with Artificial Neural Network (ANN) classifier which outperforms Naïve Bayes and Decision Tree (DT).
Paper
Full text
FSS: Fuzzy Supervised Learning for Optimal Path Selection in RPL
Semantic Scholar · Computer Science · 2021
Abstract
IPV6 routing protocol for low power lossy network (RPL) is a standardized routing protocol developed by the IETF ROLL working group. RPL has attained a lot of attraction to the research community as it demands modification adding to its improvement. RPL constructs a Destination Oriented Directed Acyclic Graph (DODAG) considering the metrics and constraints through explicit Objective Functions (OFs). The OFs chooses the best parent for reliable data communication. The existing objective function due to a single metric lacks in satisfying the requirement of real-time IoT applications. Still, the mobility aspect of the node is not considered in the objective function for optimal path selection. In the proposed methodology a Fuzzy Supervised Learning algorithm for quality node prediction using the composite metric based OF from the candidate parent set. The testing phase shows 93% accuracy of prediction with Artificial Neural Network (ANN) classifier which outperforms Naïve Bayes and Decision Tree (DT).