Quantifying effects of ramp metering on freeway safety

This study presents a real-time crash prediction model and uses this model to investigate the effect of the local traffic-responsive ramp metering strategy on freeway safety. Safety benefits of ramp metering are quantified in terms of the reduced crash potential estimated by the real-time crash prediction model. Driver responses to ramp metering and the consequent traffic flow changes were observed using a microscopic traffic simulation model and crash potential was estimated for a 14.8 km section of I-880 in Hayward, California and a hypothetical isolated on-ramp network. The results showed that ramp metering reduced crash potential by 5-37% compared to the no-control case. It was found that safety benefits of local ramp metering strategy were only restricted to the freeway sections in the vicinity of the ramp, and were highly dependent on the existing traffic conditions and the spatial extent over which the evaluation was conducted. The results provide some insight into how a local ramp metering strategy can be modified to improve safety (by reducing total crash potential) on longer stretch of freeways over a wide range of traffic conditions.