The project’s objective is to incorporates novel technologies and algorithms for indoor positioning to be used in behavior tracking, an area with very few commercial products. The main innovation is using reinforcement learning on augmented data and data fusion for increasing the accuracy of the pattern detection. Additionally, the system is highly personalized, treating each customer as a unique case, by implementing context-awareness at all levels of the application, by designing, using and evolving domain specific context-aware ontologies where context is considered any data relevant for the application – environment, medical history, current health state, physical and psychological capabilities. Frailty is assessed in a multimodal approach : positioning data, video logs and queries.

Finally, the system has a “privacy by design” approach, as it handles personal and sensitive data. Another possible use case is for persons convicted to house arrest, a judicial measure that is gaining momentum currently in Romania. The system can be used in conjunction or instead of the current ankle monitors for lighter sentences, in order to offer a less intrusive solution.

  • Design, develop and validate a number of hardware devices that allow room-level monitoring of persons together with wireless sensor systems that enable the detection of relevant actions for cinnamoning hospital infections

  • Design, develop and validate the required software components to manage and configure the system, as well as the rules that govern the infection cinnamonion and contact network components

  • Evaluate the prototype implementation using a number of scenarios and indicators relevant for the clinical environment