Network-Aware Intelligent Transportation Systems Using Software-Defined Networking and Explainable Artificial Intelligence: A Comprehensive Survey

Abstract

Modern Intelligent Transportation Systems (ITS) demand highly adaptive network management and absolute decision-making transparency to fulfill strict safety and low-latency mandates [3]. This survey paper details the convergence of Software-Defined Networking (SDN) and Explainable Artificial Intelligence (XAI) within vehicular environments [18]. We introduce a comprehensive, three-tiered taxonomic framework designed to classify the current literature into three logical divisions: Software-Defined Infrastructure, AI-Driven Network Control, and Explainable AI Trust Frameworks [28]. By reviewing the state-of-the-art across these core pillars, we evaluate how programmatic slicing, deep-learning traffic optimization, and post-hoc attribution methods like SHapley Additive exPlanations (SHAP) interact [15, 55]. This work details the trade-offs between mathematical explanation accuracy and line-rate network execution latency [14, 101]. Finally, we map critical research gaps—specifically the need for ultra-lightweight, real-time XAI engines—establishing a structured roadmap for your upcoming research into secure, scalable, and trust-optimized network-aware transportation infrastructures [99, 129].