There is a shortage of effective and feasible procedures for validating microscopic, stochastic traffic simulation models. Many times when the micro-simulators are used on real networks, it has led to traffic gridlock, or to simulation failures. The failures could provide clues about where there may be deficiencies in the simulation model.
Outcomes & Results
NISS analyzed the data from the CORSIM simulation model to discover where failures occurred. A failure was identified as one or more traffic links on the network where vehicles were unable to discharge for an unusually long period. These malfunctions can be detected through the use of link-based time traces of vehicle trips. Identifying locations where malfunctions arose required further spatial analyses. The result was that there would be less failures in the micro-simulation model, thus improving traffic management.
NISS examined emission models for integrating on road emissions data with traffic parameters and produced a usable method for traffic management operations. The team verified results by integrating results from remote sensor data on emission measurements and traffic data using area wide detectors. They also designed ITS system architecture for performing field trial experimentation. NISS assembled a small forum of technical experts and representative users in the region to review the methodology and recommend revisions, if any, before performing field experimentation.
Technical Report 154: Failure Detection and Diagnosis in Micro-Simulation Traffic Models
Examine emission models for integrating on road emissions data with traffic parameters and produce a usable method for traffic management operations.
Principal Investigator(s): Nagui Rouphail, NCSU
Senior Investigator(s): Jerome Sacks, NISS; Alan Karr, NISS
Post Doctoral Fellows: Todd Graves, Piyushimita Thakuriah