Statistical analyses of traffic advisory systems on driving behavior

The effects of driving behavior using in-vehicle and out-of-vehicle traffic advisory information while driving through a computer animated representation of a mountain pass is explored in this dissertation. An initial analysis, using logit estimations on data already collected in the test area, was done to predict whether drivers were willing to use an in-vehicle system, and whether or not they would obey the system immediately or only if conditions warranted for two traffic messages (1: to slow down, or 2: to put on chains). This information was gathered prior to drivers seeing and using an in-vehicle system. These estimated models are compared to models estimated from data collected after drivers observe and use two types of message systems (in-vehicle and out-of-vehicle). To analyze how effective an in-vehicle system can be, a laboratory study was designed using a full-size driving simulator, where information on speed, lane changes, and braking was recorded. There were four driving conditions tested, driving with in-vehicle information, driving with on-road variable message information, driving with both types of messages present, and driving with no information present. The findings of this study showed that for long distances, no significant differences in speed and standard deviation existed regardless of the traveler system used. But for shorter distances, changes in speeds were identified. This research provides insight into the validity of using in-vehicle systems for traffic advisory as well as provide valuable information on the type of drivers willing to use and comply with information provided by traveler systems.