DOES GAP ACCEPTANCE THEORY ADEQUATELY PREDICT THE CAPACITY OF A ROUNDABOUT

The traffic capacity and delay at roundabouts are important measures of an intersection's performance. The capacity of a roundabout is affected by drivers' behaviour and the driving code. The present Australian design codes have used "gap acceptance" techniques to quantify the behaviour at these intersections. However, there has been little questioning of the suitability of these techniques to estimate capacity or delays. Researchers at the TRRL in England have used a linear approximation to Tanner's model whereas other researchers have chosen to determine the gap acceptance parameters directly. This paper reviews the more important methods of estimating the capacity of a roundabout and discusses the development of a capacity model assuming the headways have a dichotomised distribution proposed by Cowan. The ability of the various models to predict the capacity of a particular single lane roundabout is also discussed. This required that the driver behaviour at the intersection be analysed in detail to determine appropriate gap acceptance parameters (critical gap and follow-on time), the appropriate headway distribution and the effect of exiting drivers on drivers entering at the same leg. Although the values for the various parameters are not suitable for use in design guides, this paper concludes that the gap acceptance techniques are appropriate for estimating capacity. The capacity fromula based on Cowan's Headway distribution is found to give results which are very similar to those given by Tanner's equation. As Cowan's model is more general, it is favoured by the author. Future ARRB research is outlined and the geometric and traffic parameters which are likely to affect the capacity are discussed. The number of the covering abstract for the conference is TRIS No. 393385. (Author/TRRL)