Intelligent Model-Driven Internet-of-Things Architecture for Connected Vehicles: Improving Quality of Experience in Smart Cities

The rapid growth of Internet of Things (IoT) technologies is transforming connected systems, particularly in domains such as Connected and Autonomous Vehicles (CAVs). However, maintaining a high Quality of Experience (QoE) for end users remains a critical challenge as system complexity and real-time demand increase. Existing IoT development approaches based on Model-Driven Architecture (MDA) provide structured abstraction and model transformation, but largely prioritize interoperability and functionality over user-centric quality considerations. This limitation hinders the design of adaptively responsive IoT systems capable of sustaining QoE under dynamic operating conditions. This paper proposes an intelligent MDA framework that embeds QoE modeling across the Computation Independent Model (CIM), Platform Independent Model (PIM), and Platform Specific Model (PSM). The framework incorporates a multi-metric QoE evaluation model that combines statistical and machine-learning techniques to monitor and predict quality variations in real time. Simulation-based evaluation of representative CAV scenarios using NS-3 and XGBoost classifiers shows that the framework achieves 90%–93% overall accuracy across datasets of 1000 to 3000 records. Additionally, the findings confirm that the integration of QoE within the MDA life cycle enhances the adaptability, responsiveness, and overall quality of the service. The outcomes highlight how extending MDA with QoE-aware mechanisms can enable more reliable and user-focused IoT-enabled connected vehicle systems in smart city environments.