This paper describes the calibration and validation of TRARR (TRAffic on Rural Roads), a microscopic simulation model, for high-class two-lane rural roads and an experimental three-lane road in Finland. The experimental road has two-lane segments, alternating from one lane to the other, a design that is now being studied in many countries. The experimental road section is about 22km long, and has five three-lane sections with two lanes for one direction, and four such sections for the other direction. The calibration aimed to determine appropriate numerical values for the simulation model's parameters; it was done by comparing the simulation results with the extensive field measurement data of before-and-after studies. The validation aimed to ensure that the model was a valid representation of the real-world traffic process. Details are given of the analysis of: (1) a single pair of passing lanes; and (2) the three-lane road. According to both simulations and measurements, the experimental three-lane road slightly increased travel speeds, and substantially increased the number of overtakings, compared with the high-class two-lane design. Further simulations will analyse different designs of the three-lane road.
Abstract