A lightweight semantic-location system for indoor and outdoor behavior modelling

In intelligent environments, one of the most common available input data is location. It can be easily captured with little additional infrastructure thanks to the ever-present smartphones or smartwatches that enable new opportunities and services in the field of pervasive computing and sensing. However, in some cases, such as in an elderly care context, it is useful to infer additional information, such as identifying unusual activities or abnormal behaviors of monitored users. In this paper, a system that uses location data to infer additional semantic information about a user's behavior is presented. The semantic location data can then be transformed into behavioral indicators that can be used to analyze the user's activities. In order to infer user activities, the proposed system requires a minimal infrastructure.