Modifiering av trafiksimuleringsmodellen SUMO för studie av trafikföring och förarbeteende på mötesfria 2+1-vägar
Information
Författare: Ella Calais, Filip KristoferssonBeräknat färdigt: 2021-06
Handledare: Johan Olstam
Handledares företag/institution: Statens väg- och transportforskningsinstitut (VTI)
Ämnesgranskare: Hans Rosth
Övrigt: -
Presentationer
Presentation av Ella CalaisPresentationstid: 2021-06-08 14:15
Presentation av Filip Kristofersson
Presentationstid: 2021-06-08 15:15
Opponenter: Ida Adolfsson, Kristin Boman
Abstract
Traffic simulation is a growing research field which enables study and evaluation of the dynamics of a traffic system with the help of today’s various traffic simulation models. Rural roads with separated oncoming lanes and alternating lanes for overtaking, so-called 2+1 roads, are becoming increasingly common in Sweden due to its positive safety effects. In Sweden, there is today only one traffic simulation model developed for simulation of these road types, RuTSim (Rural Traffic Simulator). RuTSim is a research tool which makes it difficult to apply for others than the developers. This study therefore investigated whether the more accessible open-source code traffic simulation model SUMO (Simulation of Urban MObility) developed for urban and motorway traffic can be modified for modeling of traffic operations on 2+1 roads. Simulations of a virtual 2+1 road section were examined qualitatively and quantitatively using various comparison metrics aimed at capturing the characteristic traffic operations and human driver behavior on 2+1 roads. The results were further compared with an identical road section simulated in RuTSim. By adjusting parameters in the SUMO model and using its source code in its original form, the results show that it is to some extent possible to reflect the traffic operations and human driver behavior on 2+1 road, while some important mechanisms cannot be fully captured. By further modifying the source code in such a way that the speed of overtaking vehicles is temporarily increased, some of these missing mechanisms are successfully captured. The conclusion is that SUMO to some extent can model 2+1 roads, but the simulation model still lacks some basic mechanisms for capturing some of the traffic operations. For future studies, it is recommended to focus on further developing the sub-models for the merging behavior at the end of the two-lane sections, which differs compared to merging at on-ramps or lane closures since merging on 2+1 roads are part of an overtaking process.