Application of Drone Video to Multi-intersection Facility Simulation Modeling

Traffic simulation is used to evaluate current and future alternatives and other complex traffic phenomena by decision-makers. However, the accuracy of the output is highly dependent on the quality of the input, calibration, and validation datasets. Calibration is an essential topic of intense debate in the simulation community. The challenge is multifold and originates from different venues, such as lack of simultaneous data for the stimulus and response of the system, lack of consensus on global calibration parameters and their definition, and discord on the meaning of the phenomena (calibration and validation). Recently, drone-captured videos have created a uniquely powerful medium to aid this process.This paper illustrates how modeling teams can employ drone data to build accurate and reliable traffic simulation models for complex signalized intersection facilities. Two real-world settings, an offset intersection and a Continuous Flow Intersection (CFI), with associated drone data, are used as case studies. For each facility, two macroscopic and a microscopic traffic analysis tools are used. Results of the initial calibration efforts revealed the tendency of overestimation in the analytical models stemming from faulty assumptions regarding continuous arrivals at the main intersection throughout the cycle. A solution to the issue is presented, and future research directions on the topic are identified.