Impact of Data Loss for Prediction of Traffic Flow on an Urban Road Using Neural Networks

Abstract

The deployment of intelligent transport systems requires efficient means of assessing the traffic situation. This involves gathering real traffic data from the road network and predicting the evolution of traffic parameters, in many cases based on incomplete or false data from vehicle detectors. Traffic flows in the network follow spatiotemporal patterns and this characteristic is used to suppress the impact of missing or erroneous data. The application of multilayer perceptrons and deep learning networks using autoencoders for the prediction task is evaluated. Prediction sensitivity to false data is estimated using traffic data from an urban traffic network.