Experimental Findings with VISSIM and TransModeler for Evaluating Environmental and Safety Impacts using Micro-Simulations

Abstract

Micro-simulation packages provide an efficient and systematic approach to depicting traffic dynamics. Nonetheless, many of these models used by the micro-simulation packages are only calibrated with respect to observed traffic indicators such as average speed, traffic count, and so forth, while omitting non-traffic indicators. This paper aims to investigate the performance of VISSIM and TransModeler when depicting non-traffic indicators such as fuel consumption, emissions, and safety. A model was first calibrated for traffic indicators based on Next Generation SIMulation (NGSIM) trajectories. Results indicated that after calibration, simulation accuracy was still unsatisfactory with regard to energy consumption and emission measurements, with errors of up to 38.23% in VISSIM. In assessing safety, the relative error of VISSIM increased from 12.36% to 59.92% after calibration. The error in TransModeler increased to almost 100%. Furthermore, this study explored the simulation accuracy of VISSIM and TransModeler under different traffic conditions and discovered that the models’ accuracies were relatively high when simulating stop-and-go traffic. We also explored the causes of these observed differences through a regression model. This study presents practical insight into the deficiencies of micro-simulation related research, and based on error analysis, provides a theoretical reference for optimizing simulation accuracy from a novel perspective.