Título: Statistical model of traffic flow at an unsignalized intersection and its use for capacity estimation
Ponente: Tomáš Hobza (Czech Technical University in Prague, Czech Republic)
Fecha y hora: 23/05/2024, 12:00
Lugar: Thinking Lab (UMH Campus de Elche, Edificio Torretamarit) y online
Organizador: Domingo Morales González
Abstract:
We present a probabilistic model of traffic flow at an uncontrolled intersection of the T-type and its possible use for evaluation of intersection capacity. The statistical method used is based on the generally accepted practice formulated by Werner Siegloch (1973), in which an estimation of capacity is influenced by the so-called gap acceptance rule, when the driver of a minor road vehicle accepts/rejects a gap (time clearance) between two successive main-road vehicles for his/her inclusion maneuver. In this contribution, such a stochastic scheme is reformatted strictly into a language of random variables. This makes it possible to validate/reject older approaches by which various authors have estimated the probabilities for gap acceptance and other related statistical characteristics, including a capacity itself. The performed analysis convincingly shows that some traditional regression techniques used in practical applications of the Siegloch’s method can fatally fail. Instead, the authors propose a verified methodology based on the application of modern regression methods. The contribution of the new methodology lies in a fundamental refinement of capacity estimates, which is finally demonstrated by applying it to recent empirical data sets.