Application of the KNN algorithm based on KD tree in intelligent transportation system

Abstract

The intelligent transportation system has demonstrated its strong advantages in solving the urban transport problem. One of its important roles is able to reflect the traffic conditions timely through the floating car. The key problem is to find out the candidate road sections from the vast road network quickly. Then we make the floating car match to the corresponding road by the map-matching algorithm. So we can get the real location of the floating car on the map. Every floating car needs to select candidate road sections from the whole road network, so the computing time is an important factor in affecting the real-time performance of the whole system. The commonly used method is to build an ellipse according to the probability criterion. It needs to determine the size of the ellipse, which is based on the statistic theory. It also needs to find these road sections which are in the ellipse from the whole road network. The whole process is complicated and time-consuming. Therefore, this paper proposes the k-nearest neighbors algorithm based on KD tree to get the candidate road sections.