AN INVESTIGATION OF HEADWAY DISTRIBUTION MODELS ON TWO-LANE ROADS PASSING THROUGH PERI-URBAN AREA (Paper 647)

This paper demonstrates a simplistic approach of modellingtime headways especially on roads that pass through peri-urban areas. Since, public transport facilities in those areas are not adequate, the gaps in such facilities are usually met by privately own vehicles and different types of para-transit modes of transportation including paddle tri-cycle. As a result, the prevailing traffic exhibits extreme heterogeneity in its mix which has an impact on vehicle-arrival characteristics and inter-arrival time or headways as a consequence.Frequent formation of platoons makes distribution of headways somewhat skewed and conventional negative exponential distribution cannot describe them. The current study, accordingly, aimed at developing a method of modelling headways and thereby, alleviating the implications on capacity and level-of-service on these roads. Since, it is difficult to calibrate ‘mixed models’ under mixed traffic, the paper proposed combined distribution as an alternative wherein shorter and longer headways are modelled separately. On the basis of field data collected on a two-lane highway passing through moderately developed peri-urban area, the study investigated the compatibility of combined distributions. Three distribution functions, namely Exponential, Lognormal and Gamma models were chosen and fitted to the observed headway data using maximum-likelihood estimation and K-S test technique with 5% level-of-significance.