Application of Probe Vehicle Data for Traffic Delay Estimation at Signalize Intersections under Varying Degrees of Saturation

Abstract

Currently, traffic monitoring relies heavily on road-installed detection devices, which involve high installation and maintenance costs and often fail to comprehensively cover all areas. An alternative approach is to use data from vehicles equipped with GPS devices (Probe Cars), which offers a cost-effective and flexible solution for data collection. However, this method faces limitations in accuracy due to the inability to capture data from all vehicles. This study investigates the error in using Probe Car data to estimate traffic delays under varying Degrees of Saturation (DoS) and different percentages of GPS-equipped vehicles, using a microscopic traffic simulation model in Aimsun to simulate isolated intersections. The analysis shows that increasing the proportion of Probe Car data significantly reduces delay estimation errors. It also indicates that the optimal proportion of Probe Cars depends on the road's Degree of Saturation (DoS). For roads with low DoS or light traffic, a higher proportion of Probe Cars is needed for accurate data. Specifically, for very light traffic (DoS = 0.10), up to 50% of Probe Cars may be required. In contrast, roads with high DoS or heavy traffic need fewer Probe Cars to achieve accurate results. For high-density traffic (DoS = 0.90), just 10% of Probe Cars is sufficient to keep delay estimation errors below 10%. This study demonstrates the potential of Probe Car data to reduce reliance on additional traffic monitoring devices and provides a framework for improving traffic condition assessments