Urban Mobility and the Multiscale Structure of Road Networks

The development of future smart cities has driven a new research field where different disciplines cope with the complexity of modern cities: Science of Cities. In particular, urban mobility has become one of the most intensely studied topics, thanks to data sets provided by new Information and Communication Technologies (ICT), which make it possible to relate microscopic individual behavior to the statistical laws of urban mobility through the ubiquitous use of mobile phones. In the paper we cope with the problems of pointing out the multilayer structure of an urban road network using a mobile phone GPS data set (the MDT Tim data set) and of building data driven stochastic models for urban mobility using a Maximum Entropy Principle approach. We show as the individual path reconstruction using the GPS data set allows to apply a clustering procedure that divides the urban road network of Bologna city into four connected classes that are related to a hierarchical use of the road network by individuals to perform the urban mobility, and explain the statistical laws on path length and path duration. These results gives possibility of studying the universal properties of congestion formation in an urban road network by using by data driven Markov model that are inferred applying the Maximum Entropy Principle approach of Statistical Mechanics. In this way one reduces the complexity of models and control parameter number and it is possible to study control strategies to limit the congestion spreading.