<![CDATA[Real-time travel time estimation is essential for intelligent transportation systems (ITS), yet operational traffic data streams are often incomplete due to sensor failures, communication delays, and limited coverage. This paper investigates the effectiveness of interpolation techniques for reconstructing temporally continuous travel-time profiles from real-time speed and density observations. Two approaches—linear interpolation and spline interpolation—are implemented and evaluated across varying traffic regimes (normal flow, dense traffic, and extreme congestion). Model performance is assessed using Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) against reference travel-time measurements. The results show that interpolation-based methods consistently outperform a conventional baseline relying on average observed speeds, improving estimation accuracy by up to approximately 15%. Linear interpolation yields competitive performance under stable conditions, while spline interpolation achieves lower MAE and RMSE under congestion, indicating stronger robustness to nonlinear traffic dynamics. Additionally, interpolation improves service availability and estimated time of arrival (ETA) reliability with minimal computational overhead, supporting practical deployment in resource-constrained environments. These findings suggest that interpolation provides a lightweight and effective enhancement for real-time travel time estimation and can serve as a reliable preprocessing layer for advanced predictive models in future work.]]>
Copyrights © 2025
