Delay models for Roundabouts under mixed Traffic flow Conditions in Developing Countries

Abstract

The prime focus of this study is to develop empirical roundabout entry delay models for cars, two wheelers, bicycles & heavy vehicles as a function of traffic flow and geometric variables. Multivariate regression (MVR) analysis is employed to develop the above four delay models under mixed traffic flow conditions. Required data were collected from 8 roundabouts located in five cities of India. The M/M/1 delay model is validated under mixed traffic flow conditions and found that it is showing high variability for high delay ranges. The Co-efficient of determination (R2) and Nash-Sutcliffe co-efficient (E) are found to be (0.97, 0.89), (0.91, 0.91), (0.85, 0.84) & (0.82, 0.81) for cars, bicycles, two wheelers & heavy vehicles delay models respectively. The independent variables such as weaving length (), circulating flow ( ), Weaving width (Ww), Entry width ( ), diameter of central island (D) and entry Flow ( ) are found to be statistically significant at 95% confidence level. In sensitivity analysis, entry flow variable (Ef) is found to be highest contributing variable such as the percentage (%) of sharing are 32.46 and 31.71 in cars and bicycle models respectively. Also Weaving width (Ww) and diameter of central island (D) are contributing 41.52 %, 33.37% in two wheelers and heavy vehicle delay models respectively.