Calibrated time-dependent two-regime traffic flow models

Two traffic flow regimes are clearly observed in the flow vs density data points: the free flow regime (cars can easily manoeuvre) and the congested flow regime (cars are limited in their speed and manoeuvering by delays and congestion). The objective of this study is to calibrate models that would represent changes in traffic flow and density with respect to the time factor on freeways. Relating these characteristics to the time factor results in a clear and reliable distinction between data points in the flow-density plane of the two different flow regimes. Incorporating the time factor to traffic flow characteristics may assist in traffic control in order to mitigate congestion and improve traffic safety. The data analysis has resulted in a new model of the flow relative to time, separated into a zone influenced by stable increasing demand and a zone of oscillatory traffic flow. A point of discontinuity apparently exists between these two zones. This is the transition point between the linear and the oscillatory models. Results of model calibrations in the flow-time plane have shown a good linear fit in the free flow zone. The congested flow regime is better investigated by an oscillatory model. A stable high traffic flow is obtained for a continuous period of time (one-half to one hour), which is the stabilising oscillatory zone, and ranges from 1,710 to 1,850 vehicles/hour per lane. This range may represent a steady-state freeway lane capacity as opposed to higher traffic flows observed for a short time interval, which are considered as capacity in the Highway Capacity Manual ('Highway' 2000).