Searcharxiv⌕ Search

arXiv subjects

Protiva Sadhukhan Prova

Publications and source records attributed to Protiva Sadhukhan Prova.

2 recordsLinked to original sources

A Bias-Variance Framework for Estimating Stochastic Speed-Density Relationship Considering Constrained Sample Size

Accurate calibration of traffic speed-density fundamental diagram models is essential for freeway traffic flow analysis, yet existing approaches estimate parameters from fixed datasets without examining how estimates evolve as data availability increases. This study develops a comprehensive Bayesian stochastic calibration framework integrating sequential Markov Chain Monte Carlo (MCMC) estimation via a Metropolis-Hastings within Gibbs sampling scheme with a novel Bias-Variance Cone analysis for asymptotic parameter identification. The framework is applied to 17 macroscopic speed-density models using one full year of 5-minute PeMS loop detector data from three California freeway stations. Sequential MCMC progressively increases the number of randomly sampled days from 1 to 365, enabling systematic evaluation of parameter stability and convergence as a function of data availability. Among the 17 models, nine achieved stable posterior convergence with physically interpretable estimates, while four exhibited catastrophic structural failures due to logarithmic singularities, unit incompatibilities, and initialization sensitivity. The Bias-Variance Cone framework, representing the first application of asymptotic bias-variance decomposition to fundamental diagram calibration, establishes that free-flow speed parameters are reliably identified within approximately 30 days, whereas density and shape parameters require 100 to 182 days, providing the first quantitative minimum data requirements in the literature. Collectively, the findings provide an empirical identifiability ranking of 17 fundamental diagram models, a validated minimum data sufficiency framework, and a replicable, automated platform for large-scale stochastic fundamental diagram calibration.

stat.AP↗

How does the Performance of the Data-driven Traffic Flow Forecasting Models deteriorate with Increasing Forecasting Horizon? An Extensive Approach Considering Statistical, Machine Learning and Deep Learning Models

With rapid urbanization in recent decades, traffic congestion has intensified due to increased movement of people and goods. As planning shifts from demand-based to supply-oriented strategies, Intelligent Transportation Systems (ITS) have become essential for managing traffic within existing infrastructure. A core ITS function is traffic forecasting, enabling proactive measures like ramp metering, signal control, and dynamic routing through platforms such as Google Maps. This study assesses the performance of statistical, machine learning (ML), and deep learning (DL) models in forecasting traffic speed and flow using real-world data from California's Harbor Freeway, sourced from the Caltrans Performance Measurement System (PeMS). Each model was evaluated over 20 forecasting windows (up to 1 hour 40 minutes) using RMSE, MAE, and R-Square metrics. Results show ANFIS-GP performs best at early windows with RMSE of 0.038, MAE of 0.0276, and R-Square of 0.9983, while Bi-LSTM is more robust for medium-term prediction due to its capacity to model long-range temporal dependencies, achieving RMSE of 0.1863, MAE of 0.0833, and R-Square of 0.987 at a forecasting of 20. The degradation in model performance was quantified using logarithmic transformation, with slope values used to measure robustness. Among DL models, Bi-LSTM had the flattest slope (0.0454 RMSE, 0.0545 MAE for flow), whereas ANFIS-GP had 0.1058 for RMSE and 0.1037 for flow MAE. The study concludes by identifying hybrid models as a promising future direction.

cs.LG↗