Kriging based on approximate road network distance for urban traffic state estimation

This paper improves Isomap algorithm, which is a nonlinear dimensionality reduction algorithm, and proposes a spatial alternation method based on metric multidimensional scaling. This method transforms road network distance in the original network space into approximate road network distance in a new Euclidean space, and then achieves Kriging based on this distance metric. The experiment of Nanchang's real data shows: this method has higher estimation accurate than Kriging based on Euclidean distance metric. Therefore, it is an effective solution to the problem of large-scale-road-network-level traffic state monitor.