Simulating human mobility patterns in urban areas.

With the rise of smart cities people are moving within urban spaces and still be able to pervasively interact with information, services, city’s resources and other people. In such a highly connected scenario, smartphones and other wireless portable devices are carried by humans, exhibit the same mobility behaviour of their human carriers and their movements strongly impact on the underlying network operation and performance. The understanding of human mobility in an urban space has become crucial to optimize the network management, to plan the adaptive allocation of critical resources and ensure constant quality of the user experience. This paper takes a first step in the direction of the design of a mobility model meeting behavioural and scale requirements of modern smart cities. We envision a smart city as a collection of places, each representing a Point of Interest (PoI) with specific value for single individuals and for a set of them. As a consequence, each individual has his/her own mobility footprint, while few of them share similar mobility patterns. By simulating the mobility of each individual across city’s places, we will be able to properly describe human mobility and social behaviour in urban spaces, and to extract all needed information about how city’s resources and services are accessed. The extensive use of CDR, GPS and WiFi traces, enables us to analyse the characteristics of city's Points of Interest (PoI), classify them for each individual according to their importance and study how the individuals move across them. The common features observed are the key points to build a metropolitan mobility simulator able to reproduce the regularity in spatio-temporal behaviour of mobile users and also how city sociality is built around PoIs of the city. The simulator exhibits high flexibility and can be applied in wide geographical and population scales.