Abstract

Abstract

This document accompanies deliverable D3.4, which contains the software modules implementing the methods that form the core of the Mobility Pattern Mining module within the PETRA architecture, as presented in D2.2, devoted to deal with GPS and mobile phone (GSM) individual data. The rationale, motivations and some possible applications of such methods have been described in D3.3. The algorithms learn to identify the role or purpose of each trip or location within the history of a user, in terms of activity to be performed, whether it is a systematic trip or location, etc., and exploit such derived information for prediction purposes. The document briefly summarizes the interfaces and the functionalities provided.

Abstract

This document presents a set of methods for exploiting individual patterns and measures developed within WP3, and described in D3.3, to produce collective indicators. Such indicators will be used for various applications purposes, some of which are described as representative examples. This document provides preliminaries and the rationale of the methods, highlighting how they are used (or can be used) for the showcases of PETRA.

Abstract

This document presents the tools and framework developed within the PETRA project for simulating the mobility behaviour of crowds. The tools are mainly based on the modeling of individual users - possibly derived from real data - and allow to realize various kinds of simulations, from simple predictions over current traffic/crowd status to more involved what-if analyses. This document provides some preliminaries and the rationale of the methods, highlighting their usability over the PETRA showcases.

Abstract

This document presents the algorithms developed within PETRA to monitor the performances of the main predictive models built in the project, in order to provide feedbacks and possibly trigger updates or re-computation of the models. The algorithms are based on the comparison of the models against GPS mobility data. The document briefly summarizes the interfaces and the functionalities provided.

Abstract

This document presents the key algorithms that form the core of the Mobility Pattern Mining module within the PETRA architecture, as presented in D2.2, devoted to deal with GPS and mobile phone (GSM) individual data. The algorithms learn to identify the role or purpose of each trip or location within the history of a user, in terms of activity to be performed, whether it is a systematic trip or location, etc., and exploit such derived information for prediction purposes. This document provides some preliminaries, the rationale of the methods, highlighting the improvement over the state-of-art, and a brief summary of performances.