The PLUG-IN project is aimed at the design and development of a platform for urban mobility able to integrate and manage information coming from heterogeneous sources. This can be exploited to compute the current state of the traffic, estimate its evolution in the short-mid term, define possible strategies to react to congestions and provide indications in real time (possibly personalized) to users.The estimation and prediction of the state of the traffic will be improved with respect to the state of the art through both macroscopic and statistical models of the traffic network, that will take advantage of the available integrated heterogeneous information.Then, an innovative decision support system will help the managers of the urban network in choosing strategies to improve traffic conditions.Both truck and rail transportation will be addressed by the project, also considering their reciprocal interactions. Furthermore, specific strategies will be developed for the communication issues arising from the context.Three case studies will be addressed:- efficient management of logistic fluxes, with a focus on the city of Genova. In the considered scenario, a logistics operator must carry by truck goods from the port through the city;- efficient management of the heterogeneous communication networks of professional operators;- rail traffic control, integrated with constraints from the public transport.
TYPE AND DATE
STARTING DATE – Mar 1, 2013
ENDING DATE – Feb 28, 2015
COST € 95.000,00
FUNDING € 62.000,00
ISSIA MAIN ACTIVITIES
The activities of ISSIA will focus mainly on these two objectives:
- definition of predictive models for the state of the traffic based on techniques of statistical learning from data coming from the different information sources considered in the PLUG-IN project. This will allow to expand the informative set typically used in this kind of application (mainly vehicular density and speed), which will result in more accurate prediction.
In particular, the effort will be devoted to:
- definition of the parametric and non-parametric statistical learning algorithms for traffic prediction;
- development of ad-hoc active learning techniques for the selection of input data;
- implementation of the algorithms for performance evaluation.
- definition of strategies for the real-time optimization of the traffic network based on predictive control algorithms, able to exploit the short and mid-term predictions on the state of the traffic developed in the above-described phase.
SELEX ES SPAvarious SMEsUniversity of Genoa