Delay Modeling of Un-signalized Roundabouts Using Neural Network and Regression

Delay is one of the primary variables of efficiency of intersections and roundabouts. There are not extensive experiences about delay in roundabouts in comparison with delay at intersections, and there haven’t been many studies in Iran particularly about un-signalized roundabouts. The main goal in traffic designs is to have a safe and efficient flow. In order to achieve this goal, a video has been prepared about traffic conditions of three un-signalized roundabouts in Rasht. Data related to geometric designs such as nearside legs width, turning width, far-side legs width and central roundabout diameter were measured by field observation. Data related to traffic volume such as nearside legs volume, turning volume and far-side legs volume were observed by already prepared videos. Eventually, analyses about right turning movements at un-signalized roundabouts were carried out by linear and nonlinear regression models. The best delay model of direct movements has linear relation, and left turning and direct movements has logarithmic linear relation. Because of many variables and the high accuracy of neural network, it was decided to model the data using this method, and the authors obtained the best model for the prediction of un-signalized roundabouts’ delay.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01598794
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Apr 21 2016 4:43PM