Commuters' enroute diversion and return decisions: Analysis and implications for advanced traveler information systems

Incident-induced congestion is a major source of delay and frustration for drivers in large urban areas. Advanced Traveler Information Systems (ATIS) have been proposed within the framework of Intelligent Vehicle Highway Systems (IVHS) to address one component of the incident-induced congestion problem: diversion of drivers to alternate routes. To fully utilize the potential of ATIS, transportation managers need to understand driver response to such congestion. This study examines short-term commuter response to unexpected (incident-induced) congestion. It investigates factors which influence diversion from the regular route and return to the regular route after diversion. Discrete choice models of diversion and return behavior show that the following information and trip factors increase the probability of diversion: delay information received from radio traffic reports as opposed to observation of congestion, longer delays and longer travel times, and number of alternate routes used in the past. Further, drivers were more likely to divert if they lived in the city as opposed to the suburbs, were risk seekers, had a higher stated propensity to divert and were male. However, anticipated congestion on the alternate route inhibited drivers from diverting. Finally, drivers who had longer commute trips were more likely to return to their regular route. The most important implication for designing ATIS is that traffic information must be "customized" to account for individual differences. Specific design implications are discussed in this paper.