A model of route perception in urban road networks

ATIS and new technologies are attracting increasing attention towards understanding and modeling the behavior underlying drivers' route choice. The new approach to route choice modeling is also having a significant impact on network assignment models, traditionally based on simple hypotheses of route choice. Models based on explicit route enumeration, allowing more realistic behavioral assumptions, were recently proposed. Random utility has been the standard theoretical framework for explicit models of route choice. Random utility models are based on two conceptual steps: identification of available alternatives (choice set) and choice from a given choice set (specification of systematic utility and functional form). The first step is particularly relevant in route choice where several paths are, in principle, available in the network, and many empirical studies seem to suggest that only a subset of these are actually perceived by trip makers, i.e., belong to their choice set. While in the literature there are papers dealing with the analysis of route perception in networks from the cognitive point of view, most operational models of path availability/perception in connection with network assignment are implicit and/or indirect. In other words, models of route perception (enumeration) are seldom explicitly specified and, explicit or implicit, are calibrated on indirect information, which is both disaggregated (routes chosen by a sample of drivers) and aggregated (measured flows). This study proposes an operational model explicitly simulating route perception by drivers in an urban network and presents some calibration results based on a sample of routes stated as available/perceived by students and university workers in the city of Reggio Calabria. The results seem to confirm that few routes are actually perceived as feasible alternatives and that topological, level of service and socio-economic attributes influence users' perception. The model could be integrated within a route choice simulation procedure in network assignment models.