Automating synthetic trip data generation for an agent-based simulation of urban mobility
Abstract: This paper explores the use of an automated pipeline to construct synthetic (artificially derived) trip data from aggregate socio-demographic sources to build a simulation of individual vehicles interacting with one another. The study shows that quality data sources are required in order to do this effectively and accurately. It is shown that aspects of typical patterns and behaviours may still not be represented within the final simulation. It is often a complex, expensive and impractical exercise to obtain in situ measurements across an entire city to build simulation scenarios to help with effective planning and understanding of emissions at a fine resolution. Since road traffic is a major source of harmful pollutant emissions, we explore the use of the simulation to generate emission outputs at a per vehicle level through simulation time. The paper concludes that although a valid simulation scenario can be constructed from the derived synthetic dataset, new techniques need to be developed in order to obtain an equilibrium in the simulation to allow it to not only behaves as a valid urban mobility scenario but can also be calibrated to align to represent reality.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)