Assessment of Commercial Vehicle Emissions and Vehicle Routing of Fleets using Simulated Driving Cycles

Growing concern over greenhouse gas emissions has prompted research on identifying means to control and reduce these emissions. In transportation, vehicle emissions are strongly correlated with driver behaviour. Research relating driving behaviour with emissions uses driving cycles that are assumed to be constant regardless of vehicle type for a specific area. Using a single driving cycle, however, is simplistic for emission estimation purposes as there is significant variance in the driving behaviour of different vehicle types on various road types. In this research, microsimulation models are used to generate road and vehicle specific driving cycles to improve emission estimation. Typically, the calibration of microsimulation models are carried out using vehicle counts and average speed of vehicles. A genetic algorithm is used to show that calibrating these models against acceleration data in addition to average speed and vehicle count would provide a more accurate representation of driver behaviour. This claim is validated by comparing simulation results with observed values of different parameters related to driving behaviour. The model is then used to create improved driving cycles for various types of vehicles on different types of roads. Use of simulation allows data to be collected under consistent traffic conditions for all vehicle and road types. It is also shown that using simulated driving cycles produces emission factors that are closer to the observed compared to the average speed model. Finally, as a demonstration of the application of these driving cycles, a green routing problem has been used as a test case. The test case uses the developed driving cycles to optimize the routes of a hypothetical delivery company to minimize their emissions and costs under various circumstances incorporating the effect of vehicle load in estimating emissions. Results show statistically significant differences in total distance travelled, driving time, and CO2-eq emitted as the result of different minimization criteria of distance-, time-, and emissions optimal. It is also shown that by using the proposed approach, insightful analysis of emissions under various policies of cap-and-trade vs. carbon taxing can be conducted with good accuracy without requiring technology changes or other investments, which may be expensive.Ph.D