SIDCA: Sistema multiagente de detección de caídas vía móvil en ancianos y personas de movilidad reducida

Abstract

Nowadays, fall detection for elderly people and people with mobility disabilities is a major concern for public and private institutions. This paper presents a multiagent system aimed at detecting falls through mobile devices, and provide a response in execution time. The system incorporates a novel fall detection algorithm based on machine learning and decision trees techniques. The core of the proposed system are three agent types which coordinate themselves to obtain the posture of the user taken into consideration from the data that the mobile device provides, as well as the GPS position, which can be sent via SMS or automatic phone call to the relatives or care givers. The proposed system is self-adaptive and personalizes the decision frontiers established for the user.