An enhanced optimal velocity model for car-following with connected and autonomous vehicles

Accurate modeling of car-following behavior is crucial for improving traffic flow, safety, and energy efficiency in connected and autonomous vehicle (CAV) systems. This study investigates vehicle dynamics using the Optimal Velocity Model (OVM) for both single vehicles and multi-vehicle platoons, with a focus on stability, accuracy, communication delay, and energy consumption. Simulation results indicate that the classical OVM exhibits persistent stop-and-go oscillations across all intervals, reflecting inherent instability under high-sensitivity parameters. Single-vehicle tracking initially achieves high accuracy, with an RMSE of 0.74 m/s, but degrades over time, rising to 1.21 m/s, highlighting cumulative error under dynamic conditions. Extending the model to a vehicle string reveals disturbance amplification and sustained limit-cycle oscillations, demonstrating string instability and the critical influence of inter-vehicle interactions. Communication delay analysis shows that minimal latency maintains stable tracking, whereas delays exceeding 500 ms induce high-amplitude oscillations in relative velocity, providing quantitative bounds for V2V and V2I systems. Energy consumption is analyzed over time, showing consistent cumulative trends and robust handling of transient high-power events. Calibration against empirical spacing data improves model fidelity, maintaining RMSE below 16 m across all scenarios. Overall, the framework enhances velocity tracking and reproduces realistic traffic patterns, offering a robust platform for evaluating car-following behavior. The results provide a foundation for future work on adaptive, learning-based controllers, delay-aware coordination, and real-time validation in complex traffic environments.