Effects of In-Vehicle and Infrastructure-Based Collision Warnings to Nonviolating Drivers at Signalized Intersections

Abstract

The potential effectiveness of warnings to drivers in imminent danger of collision with a red light violator was evaluated in an experiment that used a driving simulator. Three warnings were tested: (a) an infrastructure-based warning that immediately turned the traffic signal red and activated red wig-wag lights; (b) an in-vehicle warning that consisted of a brake pulse, voice annunciation, and activation of a red dashboard light; and (c) simultaneous activation of both warnings. Three warnings were tested with different groups of drivers. Drivers in each group were also exposed to one of three traffic conditions: (a) no other vehicles ahead or behind, (b) following closely behind another vehicle, and (c) being closely followed by another vehicle. Individual drivers received only one warning on only one trial. The warnings were given on approach to the intersection at a point where the probability that the driver would stop for a yellow change interval was less than 10%. All three warnings proved effective in delaying drivers' arrival at the intersection by at least 1 s. The simultaneous in-vehicle and infrastructure warning was significantly more effective than the other warnings and delayed 95% of drivers. The infrastructure-only and in-vehicle-only warnings delayed 67% and 80%, respectively. The presence of leading or following vehicles had no detectable effect on the probability response to the warning.