It is suggested that the source of congestion in the vicinity of freeway on-ramps might be an extended length of the freeway down-stream of the ramp, the cause of the congestion being a gradual increase in car-following headways as vehicles leave the tensions of the merge behind them. This causes a gradual diminution in capacity. Hence the term "capacity funnel". If flows are at or near capacity, the existence of the capacity funnel leads directly to gradual speed reductions for each vehicle. It is assumed that the capacity is a function only of position on the road regardless of speed and concentration and although this is doubtful in view of car-following work and experimental studies, it might be adequate for the present purpose. Using a deterministic model, speeds and concentrations are derived as functions of distance and time for a general capacity-position relationship, and for three such specific relationships. It emerges that both concentration and the reciprocals of the speed are linear or roughly linear in time and distance. A sampling procedure is outlined for detecting speeds and concentrations down-stream of an on-ramp. A computational procedure is suggested for using this information to predict freeway speeds and concentrations near the ramp, in sufficient time to reduce the flow by preventing the entry of ramp vehicles, and hence perhaps preventing congestion from developing on the expressway in the critical ramp area. Although this latter work rests upon the assumption that when congestion is imminent speed reciprocals and concentrations are approximately linear in time and distance, it does not depend upon deterministic assumptions about traffic behaviour. The computational procedure is based on regression theory, is fairly simple and rapid, and could be feasible for congestion-prone situations on expressways. (a) for the covering abstract of the symposium please see irrd abstract no. 224453.
Samenvatting