A Grid-Based Spatial Index for Matching between Moving Vehicles and Road-Network in a Real-time Environment

Moving vehicles in a street network react to and reflect the actual traffic situation. Therefore, they can be considered as an important source of the dynamic traffic information, especially from large cities where it would be too expensive to seamlessly install stationary sensors for data acquisition. Without extra instrument to be set up along the roadway, traffic information can be derived from the behavior of moving vehicles at a low cost. Nevertheless, moving vehicles as such do not provide valuable information straightforwardly. Their scattered locations in a rather large space must be aligned with the street network, which is in essence a point-to-line matching task. So far, a large number of researchers worldwide have addressed various matching issues. Many of the reported results reveal a high matching rate with a high accuracy on certain data types or test areas. In case of matching individual vehicles with the streets along which they are dynamically moving, however, the matching quality is influenced not only by the accuracy, but also by the computing efficiency. It does not make any sense if the matching algorithm is unable to yield results before the vehicles move to substantially different locations. To realize the point-to-line matching in real time, the authors proposed an innovative grid-based spatial index for linear data on the basis of decomposition principles.