Multiscale Pattern Recognition for Multimodal Network Predictions
Subject
Short-term predictions is aimed at creating a transportation system that can provide services proactively. These predictive systems only require a rough prediction but at a higher scale such as the scale of an entire urban area in order to support strategic decisions by traffic operators. Traditional network-wide traffic predictions use time series data, where the spatial correlations of the network is not considered. Since spatial correlations are extremely relevant for transport applications, this research aims at developing network-wide traffic predictors using 3D maps where the spatial relations are inherently included. This research is funded by SETA, a Horizon 2020 EU project that is aimed at providing better metropolitan mobility.
Scientific challenges
As there are no previous studies on the use of 3D maps for prediction, the main scientific challenge of this work is creating these 3D maps and classifying the 3D data for prediction. There are various gaps in the literature that needs to be addressed to achieve this. We need to reduce the complexity of using such a large network by generating multiscale networks. Another challenge of working with an urban scale or national scale network is the data reliability. The missing and spurious data needs to be addressed. Finally, the traditional clustering and classification techniques needs to be extended to work with 3D data to be used for prediction.
Societal relevance
Network-wide traffic predictions have two-fold applications. It can assist individual travellers in making better route choice and departure time decisions. For professionals, such traffic state information will provide criteria with which to better manage and control traffic to reduce congestion. Reducing congestion have a positive effect on the economy as it implies increased mobility or traffic volume which leads to decrease in travel time. However, increased traffic volume leads to road safety issues and also can adversely affect the environment.
Panchamy KrishnakumariStart/end date: 15th February, 2016 – 15th February, 2020Daily supervisor: Oded Cats Promotor: Hans van Lint |