Methods and Tools for Supporting the Use, Calibration and Validation of Traffic Simulation Models
These models are extensively used in applications that have great potential impact on the safety, capacity and environmental efficiency of the road system. However, the fidelity of results and conclusions drawn from a simulation study, as well as the range of possible applications the tools can reliably be used for, are questionable: the same simulation study carried out by different people, even when using the same tool, is likely to give different results.
The main objective of the action is to develop, implement and promote the use of methods and procedures for supporting the use of traffic simulation models, especially on the topics of model calibration and validation.
To this date, the bulk of resources and effort in the field of traffic simulation have focused on "model development", leading to many simulation models being available on the market.
For these reasons the main objective of this Action is to develop, implement and promote methodologies and procedures to support the use of traffic simulation, especially on the topics of calibration and validation. To this aim the sharing and exchanging of available traffic datasets will also be a key task of the Action.
The project examined issues such as data availability and quality, as well as the relationship between data accuracy and calibration, as well as developing and testing methodologies suitable for calibration and validation of deterministic as well as stochastic models. Major accomplishments were the definition of a general framework for the management of uncertainty in traffic simulation, the development and application of Global Sensitivity Analysis techniques to traffic simulation, the building of a new simulation platform to benchmark dynamic origin-destination (OD) estimation and prediction algorithms, and the development of a new methodology to reconstruct trajectory data from noisy measurements, applied to NGSIM I80-1 dataset.