Abstract

The paradigm of Smart Cities is based on the idea of enhancing citizens’ life by means of digital technologies. The widespread generation of data is seen as the main enabling factor of this paradigm, as it makes possible to create data-driven services to be offered to the citizens. While objective measurable data are generally considered (eg, measurement of air pollution), we argue that subjective data may also play a relevant role in the context of Smart Cities. Indeed, services could be better tailored to citizens, provided that their subjective experience can be effectively assessed. In this paper, we consider citizens’ emotions as the subjective data that can be exploited to improve services, and we propose a system for inferring this data from the movements of the citizens themselves. We note that movements can be acquired in various ways, such as by using cameras or non-invasive motion capturing technologies (eg, the Kinect) that can be easily deployed in a Smart City. Specifically, we propose a system that, from the analysis of movements acquired using the Kinect 2.0, can effectively i) disambiguate between portions of movements characterized by non-negative or negative emotions and ii) identify in real time the instants of transitions between such states. Results, obtained after extensive simulations, make us confident that the proposed system can find application in a real Smart City context (eg, to automatically assess if a person in a public place is too much nervous). Finally, we also observe that citizens’ movements are not as privacy sensitive as other types of data that are generally considered in the emotion recognition task (eg, facial …

Date

March 19, 2026

Authors

Davide Andreoletti, Felipe Cardoso, Andrea Arzillo, Luca Luceri, Achille Peternier, Silvia Giordano