Energy-Aware Routing Optimization for 5G Vehicular Fog Networks

Abstract

The rapid expansion of 5G-enabled Intelligent Transportation Systems (ITS) is accelerating global vehicular connectivity, with projections exceeding 1.9 billion connected vehicles by 2030. Despite this growth, energy-efficient routing in dynamic fog-assisted vehicular networks remains a critical challenge. Traditional routing protocols such as Ad hoc On-Demand Distance Vector (AODV), Dynamic Source Routing (DSR), and Greedy Perimeter Stateless Routing (GPSR) rely on static heuristics and lack adaptability to fluctuating vehicle density, high mobility, and real-time fog node workloads, often leading to energy inefficiencies, load imbalance, and increased latency. To address these challenges, this study proposes an intelligent Classification and Regression Tree (CART)-based framework for predicting optimal load-balanced routes and energy efficiency scores using real-world vehicular network datasets. The proposed system integrates Decision Tree (DT), Support Vector Machine (SVM), and a novel Brain Branch model, which combines a Multi-Layer Perceptron (MLP) with Extra Trees (ET), within a user-friendly Tkinter-based graphical user interface (GUI) for real-time analysis. The pipeline incorporates data preprocessing, exploratory data analysis (EDA), load distribution modeling, and multi-model training. Among the evaluated approaches, the Brain Branch model demonstrates superior performance by effectively capturing complex routing patterns and energy dynamics. The proposed framework enhances routing precision, reduces unnecessary energy consumption, and enables efficient fog node selection with lower latency.