Fuzzy logic-based methodology for quantification of traffic congestion

Abstract

This paper presents a fuzzy-based methodology for quantification of congestion level for traffic control on expressways using traffic flow speed and density. Inductive loop detector data on the Interstate 880 obtained through the Freeway Performance Measurement System were used to estimate congestion levels following the fuzzy logic approach. In comparison with the Highway Capacity Manual, the results generally show a good correspondence. However, unlike the Highway Capacity Manual that defines step-wise measurement of levels of service based entirely on density, the proposed fuzzy inference system allows a flexible combination between speed and density to provide a more detailed indication of congestion intensity to describe the gradual transition of traffic state. For comparison, the congestion indices evaluated with both density and speed were compared to those evaluated with either speed or density using the same data set. Results from this comparative study reinforce the statements from previous studies that expressway speed is conservative under free-flow and light traffic conditions, but decreases significantly just before the flow rate approaches the road capacity. The results also show significant differences between the congestion indices evaluated using a single quantity, while the congestion indices using both density and speed tend to neutralize in between and scale up in a stable manner with the levels of service. Considering the abstract nature of congestion terminology, it is necessary to quantity traffic congestion on the expressways using both variables to minimize the potential bias in representing the operation of expressway traffic properly, which is particularly important under heavy congested conditions.