SMART VR for Commercial motor vehicles Safety: A Scalable Virtual Reality Framework with AI-Driven Hazard Simulation and Physiological Monitoring

Abstract

Commercial motor vehicles (CMVs) are vital to national logistics but remain disproportionately involved in high-severity crashes, with human factors such as fatigue, distraction, delayed hazard recognition, and cognitive overload contributing significantly to crash risk. Despite advancements in regulation and vehicle technologies, conventional training methods still fall short in preparing CMV drivers for unpredictable, high-risk environments. These approaches often rely on passive instruction or low-fidelity simulation, offering limited realism, adaptability, and behavioral insight. As a result, they struggle to address evolving hazards, monitor physiological states such as fatigue or attentional lapses, support effective skill transfer, or replicate critical scenarios for evaluation and intervention. To address these gaps, we present SMART VR, a scalable and modular virtual reality framework for CMV safety training and human factors research. Built on the CARLA simulator and Unreal Engine, SMART VR provides a unified, high-fidelity platform that integrates immersive simulation, AI-driven hazard generation, and physiological monitoring, supporting deployment through VR headsets and full-scale cockpit hardware with force-feedback steering and operational controls. A configurable scenario engine dynamically injects hazards, from lane incursions, visibility loss, erratic traffic behavior, and auditory distractions, based on predefined or adaptive logic, with each event precisely time-aligned with vehicle telemetry (speed, braking, steering, lane position) and real-time physiological monitoring via wearable sensors capturing eye gaze, heart rate variability, and electrodermal activity. These synchronized data streams enable multidimensional assessments of driver state, including attentional focus, cognitive workload, and stress response, addressing a critical gap in conventional training systems. The framework’s modular design enables the import of custom road environments, integration with external tools such as decision-support systems, and development of targeted training protocols. This flexibility supports the replication of high-risk operational scenarios under controlled conditions and enables repeatable, simulations for validating safety interventions, driver-assist technologies, and human–machine interface designs, advancing CMV training, behavioral evaluation, and intelligent transportation systems.

Keywords: Virtual Reality Simulation, Commercial Motor Vehicle Safety, Driver Behavior Modeling, Training and Safety Interventions

DOI: 10.54941/ahfe1006905