Can results of car-following model calibration based on trajectory data be trusted?

Calibration of car-following models against trajectory data has been widely applied as the basis for several type of studies ranging from the investigation and benchmarking of models, to the study of parameters correlation, or other theoretical issues like the inter/intra driver heterogeneity or the multi-anticipative driving behaviour. However, very few of these studies attempted also to analyze and quantify the uncertainty entailed in the calibration process and its impacts on the accuracy and reliability of results. A thorough understanding of the whole calibration problem (against trajectory data), as well as of the mutual effect of the specific problems raised in the field literature, indeed, does not yet exist. In this view, a general methodology to assess a calibration procedure was proposed and applied to the calibration of the Gipps’ car-following model. Compact indicators were proposed to evaluate the capability of a calibration setting to find the “known” global solution, in terms of both the accuracy and the robustness as to the variation of the starting conditions of the optimisation algorithm. Then, a graphical inspection method, based on the so called cobweb plots, was proposed to explore the existence and the nature of the local minima found by the algorithms, as well as to give insights into the measure of performances and the goodness of fit functions used in the calibration experiments. The methodology has been applied to all the calibration settings (i.e. combinations of algorithms, measure of performances and goodness of fit functions) utilized in the field literature so far. The study allowed us to highlight and motivate, for the model under investigation, the limits of some of these calibration settings. Research lines towards the definition of robust settings for the problem of car-following models calibration based on real trajectory data, are finally outlined.