Driver En Route Guidance Compliance and Driver Learning with Advanced Traveler Information Systems: Analysis with Travel Simulation Experiment

Abstract

If real-time driver en route guidance advice does not meet driver preferences (e.g., preference for taking the freeway) or the advice is not correct, drivers are very likely to ignore the information, and the guidance system becomes ineffective in their route choice no matter how advanced the system. There is a need to investigate the factors affecting driver compliance with en route guidance advice. A travel simulation experiment was used to investigate significant factors affecting driver route choice behavior. A linear mixed model was developed for describing the factors affecting driver compliance with guidance advice using the compliance rate over several simulated trips as a dependent variable. The issue of repeated observations is addressed. The system accuracy and subjects’ learning experience in their spatial experience at the same intersection and temporal experience in the same day are also taken into account. The model results show that significant factors are involved: freeway advice, turning advice, congestion occurrence, incident occurrence, subjects’ spatial experience, subjects’ temporal experience, and subjects’ education level; there are several important interactions as well.

Leave the expressway or not? Impact of dynamic information

“…The alternate arterial road route is an imaginary route. The arterial road route can be deemed as an alternate route that a real-world VMS-based ATIS recommends to drivers [5,25]. Thus, our SP settings are reasonable, though the employed network at first sight seems simple.…”

“…Some other studies used SP data from travel simulator experiments (e.g., [17][18][19][20][21][22][23][24][25]). Others used revealed preference (RP) data to model drivers' response behavior (e.g., [26][27][28][29]).…”

“…With the increasing popularity of simulation-based estimation, panel data models that address repeated observations are gaining more attention. Methodologies that have been applied to address repeated observations include mixed logit models (e.g., [19,23]), random effect models (e.g., [14,29]), multinomial probit (e.g., [13,18]), normal mixing distributions (e.g., [6]), generalized estimating equations (e.g., [5,15]), and mixed linear models (e.g., [25]). …”

From user equilibrium to system optimum: a literature review on the role of travel information, bounded rationality and non-selfish behaviour at the network and individual levels

“…Many other studies contain findings on compliance rates towards advice. Chen and Jovanis (2003) found a compliance rate of 89% in their simulator study. Ben-Elia, Di Pace et al (2013) found average compliance rates between 62% and 82%, depending on the accuracy level.…”

Neural Agent (Neugent) Models of Driver Behavior for Supporting ITS Simulations

“…For example, a number of studies [7] indicated that a driver's route and the shortest route differ in significant ways and that most drivers use several criteria in selecting their routes. Furthermore, drivers' spatial and temporal experiences produce variations in route selection [8].…”