Automatic generation of personal maps

Abstract

Current collection of data on urban space usage has to rely on paper based surveys that require a huge investment to put in place. New technologies, like positioning systems and handheld devices, provide us with means to build a system that is able to that same data, with less costs and with extended possibilities. For such a system to be successful, it must be aware of the user’s activities in the less intrusive way possible. Activities are all associated with the places where they occur. This information associated with the modes of transport used to travel between such places forms the basic components of a person’s personal map. In order to build this map, we base ourselves on data collected through the multiple sensors present on a smartphone and use it in conjunction with several artificial intelligence and statistic techniques to build a model capable of inferring activities in a non intrusive way, after a short period of