Training data of city Tunnel traffic situation awareness

Abstract

The network of city Tunnel monitoring systems has accumulated a great deal of multi-source heterogeneous monitoring data, which mainly consists of video data, traffic data, environmental data, sensor data and so on. This paper discusses how to use the monitoring data to realize city tunnel traffic situation awareness. First, we need to mine a set of strong associated characteristic groups according to the association rules, and then determine the corresponding traffic awareness for each characteristic group to build our training data set for machine learning methods. Finally we can aware the situation of city tunnel traffic by using machine learning methods. We solve the problem of how to build the training data in the background of complex multi-source data, which is a prerequisite for many machine learning methods. © Springer-Verlag Berlin Heidelberg 2013.