An algorithm for the recognition of levels of congestion in road traffic problems

Detection and recognition of the level of congestion at an intersection is a very important problem and a valuable source of information in traffic management. Although it is just one of all the aspects that make up a traffic management system, it seems to be a crucial point for gathering information. In this paper, we present a technique based on a k-means clustering algorithm for classification, which has been already successfully used in a number of pattern recognition problems, namely: as an algorithm for face recognition problems and in a number of medical diagnosis problems and it compares very well with the state of the art techniques.