Reconstruction Quality of Congested Freeway Traffic Patterns Based on Kerner’s Three-Phase Traffic Theory

Abstract—This paper discusses the reconstruction quality of spatio-temporal congested freeway traffic patterns depending on the information provided by different equipment rates of probe vehicles. In this research Kerner’s three-phase traffic theory is applied, which distinguishes two different phases in congested traffic: synchronized flow and wide moving jam. In the presented approach spatio-temporal congested traffic patterns are reconstructed from intelligent probe vehicle in-formation generated by an on-board traffic state detection, identifying traffic states along a vehicle’s trajectory at any time. With a data fusion algorithm combining the data of several probe vehicles, a detailed picture of spatio-temporal congested traffic patterns is revealed. Comparing Ground-Truth with the reconstructed traffic pattern shows that a reconstruction quality comparable to that of established traffic flow models is achievable with probe vehicle equipment rates of about 0.5 %. At higher equipment rates of about 1-1.5 % the achievable quality already exceeds the quality established traffic flow models are able to offer based on a dense detector network with average detector distances of 1-2 km. Keywords-Traffic monitoring; Traffic state detection; Traffic data fusion; Probe vehicle data; Three-phase traffic theory; Models ASDA and FOTO; Traffic data quality; Traffic quality indices. I