Summary
Existing models used by researchers to estimate greenhouse gas and pollutant emissions from vehicle fleets are fed by traffic models and emissions factors from fixed site measurements. However, evidence suggests that this approach does not accurately predict emissions from vehicles operating on real roads by actual drivers. This is both because vehicles are calibrated by manufacturers to unrealistically mild driving conditions to meet regulations and because drivers routinely drive much more aggressively than in standard cycles. Real-world driver behavior, represented primarily by vehicle speed and acceleration, may increase emissions factors compared to what is predicted by models. This project will study how the Minnesota vehicle fleet?s greenhouse gas (CO2) and pollutant (NOx and particulate) emissions compare to modeled values to quantify potential increases in emissions attributable to driver behavior. In one aspect of the research, a very large time-series dataset owned by the University of Minnesota of over 1 million vehicles will be used to study real driving cycles on Minnesota roads over a range of ambient conditions and vehicle types. Laboratory measurements and literature values will be used to estimate the increase in emissions over those predicted by models like the U.S. EPA Motor Vehicle Emission Simulator, a framework used by MnDOT and most other departments of transportation to estimate emissions. In another part of the project, on-road emissions will be calculated using drive cycles from vehicles on roadways using an embedded instrumented vehicle equipped with radar sensing to correlate emissions to local traffic speed and acceleration, a novel approach compared to conventional fixed site measurements.