The real-time traffic control technique of urban expressways is proposed. This technique is composed of an estimation model of O-D proportions between on-ramps and off-ramps and an on-ramp control model. The O-D proportions are estimated dynamically and recursively. Therefore, even when traffic flows change with time, this model is suitable for a real time application. The on-ramp control model formulated as a fuzzy linear programming problem is appropriate for treating the vagueness of the traffic capacity or the queue length. Experiments with real world data observed on Hanshin Expressway suggest that the proposed traffic control technique is ensured to have feasible solutions even if no steady state conditions are attained. It can be seen that results of traffic control are different by setting fuzzy variables and the proposed technique is effective to reduce the congestion of the expressway.
Abstract