The need for developing fall detecting systems has become increasingly important, since older adults require a great care, especially in this matter. Falls are one of the most common problems among the elderly and can lead to serious health problems if not detected in time. There are several solutions for fall detection, among them the use of smart sensors, as part of the Internet of Things (IoT) architecture. Fall detection systems are thus mandatory in smart home environments, especially ambient assisted living (AAL) solutions, to ensure safe and secure living environments for the older adults. Also, machine learning (ML) has been included in these systems to considerably improve their ability to detect fall events. Currently, many fall detection systems lack the incorporation and data fusion between different sensing technologies, which can lead to more accurate results. In this way, the work presents the development of a smart sensing system that can be used to detect falls based on ultra-wide band (UWB) technology and inertial measurement unit (IMU) data. Long short-term memory (LSTM) models were implemented to detect such event based on the combination of both methods, having achieved high levels of accuracy of 95.8%.

--

Fall detecting system,Smart sensors,Internet of Things (IoT),Smart home,Ambient Assisted Living,Machine Learning (ML),Ultra-Wide Band (UWB) indoor localization,Inertial Measurement Unit (IMU)