COD-STAR: The Energy Efficient Vehicular Communication Using Chinchilla-Based Adaptive Path Selection and Trust Evaluation

Abstract

Vehicular Ad Hoc Networks require routing protocols which are able to handle high mobility, uncertain connectivity, and changing environmental conditions. Proposed COD-STAR presents a hybrid method involving Chinchilla Optimization and Dempster-Shafer Theory that can support intelligent routing that is trust-aware and context sensitive. The protocol starts with building anchor paths based on the structured urban topology, and then real-time contextual data of the mobile nodes and infrastructure is collected. Various routing metrics are assigned belief values and fused with the help of DST to determine how reliable the paths are when there is uncertainty. Adaptive exploration and exploitation of candidate routes is done by Chinchilla agents, which are informed by dynamic fitness models that combine fused beliefs and environmental volatility. The continuous updating of belief scores is facilitated by feedback mechanisms, which allow the recalibration of the path in real-time and refinement of decisions. The combination of optimization intelligence and belief-based reasoning allows COD-STAR to deliver valid and effective communication routes in demanding conditions of VANETs, which guarantees the routing stability and scalability in the ever-changing traffic situations.