METHODOLOGY

A research team led by the Pacific Southwest Region University Transportation Center (UTC) developed a structured database that fused various datasets including real-world traffic state measurements, traffic accident records, roadway geometry, and weather information.

Using input data from the integrated database, two sets of traffic data were obtained: 1) a representation of normal traffic conditions without accident occurrence; and 2) the actual traffic conditions with the occurrence of accident, across a spatiotemporal region large enough to cover all of the potential impacts. After identifying the accidents’ impact region and predicting the traffic conditions under a “what-if” scenario (i.e., no accident) using machine learning techniques, the U.S. Environmental Protection Agency’s Motor Vehicular Emissions Simulator (MOVES) model was applied to conduct energy/emissions analysis. The benefit of the AEB system was estimated by calculating the difference in fuel and energy consumption and the amount of pollutants emitted in both the with and without traffic accident scenarios.

FINDINGS

Case I: Rear-end Collision that Occurred 5/16/2017 on SR-91 in Riverside, CA

The energy consumption within the spatiotemporal impact region would be reduced by up to 14.99 percent and the resultant pollutant emissions would decrease by as much as 17.59 percent, as shown in Table 1.

Table 1: Estimated Environmental Impact of Entire Spatiotemporal Region due to the Accident (Riverside, CA)

Case II: Rear-end Collision that Occurred 1/2/2020 on US-95 in Las Vegas, NV

The results showed that the AEB technology may improve energy economy by up to 34.6 percent and reduce pollutant emissions by as much as 22.5 percent if the selected accidents could be avoided. These results are shown in Table 2.

Table 2: Estimated Environmental Impact of Entire Spatiotemporal Region due to the Accident (Las Vegas, NV)