A flexible architecture for cognitive sensing of activities in ambient assisted living

Abstract : We discuss our experience with the design, implementation, deployment and evaluation of a Human Activity Detection application. This application relies on a software architecture for IoT-based monitoring in Ambient Assisted Living (AAL), based on a Semantic Message Oriented Middleware (SeMoM). Going beyond traditional sensor management systems, our architecture addresses the heterogeneity challenge in IoT healthcare systems and ensures semantic and technical interoperability. SeMoM is able to handle the data acquisition process for a variety of heterogeneous devices, and to provide a reasoning mechanism based on expert knowledge and logic through cognitive sensors. Sensor data and observations are annotated using an extended version of the Cognitive Semantic Sensor Network (CoSSN) ontology built on top of the Semantic Sensor Network (SSN) ontology. CoSSN provides a formal representation that supports a semantic detection of activities.