Analysis and modelling the effects of information accuracy on travellers' behaviour

The investigation of the effects of information on travellers' behaviour at individual level is a pre-requisite for any analysis on the impacts of ATIS on traffic networks. It is widely expected that travellers' behaviour can be strongly influenced by the ability of the information system in making accurate estimations of the actual travel times they will experience on the network; however, relatively few experimental studies have been developed to this regard. In this paper such an analysis is carried out; in particular, the paper investigates the effect of the information accuracy (and of its dispersion over time) on travellers' behaviour. The final objective was to specify, calibrate and validate different modelling approach for simulating in which way the respondents react to ATIS and how the travellers' concordance depends on the level of information accuracy. At this aim a stated preference experiment were built, discrete choice models based on random utility theory were developed, switching and holding approaches were investigated, homoscedastic and heteroscedastic models were estimated.