We will present a conceptual model of a smart system that uses data collected by personal mobile devices for the early detection of diseases and the subsequent issuing of an alert. Using data obtained from mobile devices, preferably smartphones or smartwatches, such as heart rate, number of daily steps, cough or breathing sounds, blood oxygenation level, etc., it is possible to build a diary of biometric signals, usable in the most diverse situations. In particular, it is possible to build a status configuration that characterises the user of the device and to monitor this/her general status. This configuration, once obtained, can then be used by a model specially trained for anomaly detection that can issue an alert of a change in state with possible medical significance. This alert can be emitted only to the user or, with permission, to alert a health platform and the professional who usually follows the individual. This conceptual model is being used as the central component in the pilot of a research project called AIM Health, taking place at Istar_Iscte, in close collaboration with the Santa Maria Hospital through the Association for Research and Development of the Faculty of Medicine AIDFM/FM/ULisboa, CIS_Iscte, the Institute of Telecommunications, the Instituto Superior Técnico and the Cardio Vascular Centre of the University of Lisbon. The AIM Health project aims to create a reliable and safe smartphone application (App) capable of early detection of symptoms due to COVID-19 infection, i.e. as soon as there is a change in biorhythm indicators that can be associated with lung diseases. The chosen smart detection solution is based on a distributed computing platform and secure data storage and communication.

This work was partially supported by Fundac ̧a ̃o para a Cieˆncia e a Tecnolo- gia, I.P. (FCT) with ISTAR Projects: UIDB/04466/2020 and UIDP/04466/2020 and DSAIPA/AI/0122/2020 AIM Health. The authors would also like to thank Hospital de Santa Maria.

Artificial Intelligence,Digital Health,Anomaly detection

• Computer and Information Sciences - Natural Sciences