The Swedish national road administration's methods for calculating capacity, delay, queue length, etc at unsignalised intersections are based on a queuing theory model and on early German work; these methods are used in the CAPCAL computer program, completed in 1977, which now needs to be updated and improved. This paper presents a detailed proposal for a set of calculation procedures, intended for the new version of CAPCAL. The objectives of these procedures are essentially the same as before, to provide (1) tools for evaluating different control strategies and geometrical designs; (2) methods for estimating the more important performance measures for intersections, compatible with similar methods for uninterrupted flow conditions. The paper first considers the performance measures, including: (1) capacity and degree of saturation; (2) average delay; (3) proportions of stopped and delayed vehicles; (4) queue lengths; (5) noise and air pollution; (6) effects of saturation. It then describes the proposed new calculation procedures, which are based on a new queuing theory model using two gap acceptance parameters. Formulae are given for the mean and variance of the service time for one movement in one lane, with first order conflicts only, in terms of the headway distribution and other characteristics of the intersection. The effects of higher order conflicts and short lanes are then discussed. Finally, some mathematical results are given about capacity and degree of saturation, delay and stops, and queue lengths.
Abstract