Constructing Two-Layered Freight Traffic Network Model From Truck Probe Data

The road network in the Keihanshin area (the second most populated urban region in Japan encompassing the metropolitan areas of the cities of Osaka in Osaka prefecture, Kobe in Hy_go Prefecture, and Kyoto in Kyoto prefecture) is modeled as two layers from the viewpoint of the frequency of use of roads by freight vehicles. This model was derived ,tested and validated from the probe car data of 300 trucks operated by 21 freight companies in the Keihanshin area. The two layers are called high-frequency-network and low-frequency-network layers. Characteristics, such as the density of probe data, average driving time, and speed of both network layers, were estimated. This analysis revealed that trucks spend a duration about 6 times longer on the high-frequency network even though it is 1/4 the size of the low-frequency network. The instantaneous velocity histograms indicated a significant difference in speeds between the two network layers of 25.03 km/h in the high-frequency layer and 16.87 km/h in the low-frequency layer. Service-level evaluations, reliability analysis, and route-selection modeling should become easier to deal with in the future by introducing these layered networks.