engAGE results in a nutshell

Karde technological results


Karde has now finalised its work on Memas in engAGE. Memas is a life mastering assistant intended for elderly people and people with MCI. In its basic version Memas contains a calendar for structuring daily activities, instruction videos for operation of household appliances, pleasure functions like private photo albums and videos and also other type of videos in addition to contact information. For engAGE Memas has been augmented with self-reporting, cognitive games and assessment of cognitive level. Self-reporting and scoring in cognitive games have been input to the machine learning algorithms provided by TUC in addition to information from the activity tracker developed by Tellu. The assessments in Memas are based on output from the machine learning algorithms.
In the figure, the home page of Adminweb, on which you configure the app for the user.

Memas Adminweb

Tellu technological results


Tellu is wrapping up their technical development in the engAGE project, and has completed a final version of the Dialogg mobile gateway app. Dialogg is a cross-platform mobile application connected to the TelluCare Remote Patient Monitoring service. In engAGE, an engAGE-specific version of the app has been developed and tested, as part of the Monitoring, Self-Reporting and Big Data Processing (MSRBD) service. The app functions as a gateway, automatically transferring the data collected by fitness trackers into the ML-based Cognitive Decline Assessment (MLCDA) service for processing.

Dialogg screenshot

TUC technological results


TUC is finalizing the development of the engAGE ML-based Cognitive Decline Assessment (MLCDA) service that is integrated in the engAGE final prototype. This service analyses and correlates data acquired by different engAGE services, through machine learning algorithms, to determine insights into the cognitive state and potential decline of the older adults. The employed machine learning technique is based on a Graph Learning and Convolutional Neural Network (CNN) ensemble that is trained on historical trials data and that provide results directly in the Memas communication platform where they can be analysed by the healthcare professionals. An additional management dashboard has been developed to help visualize MLCDA outputs in an interactive and user-friendly way as charts and graphs.

Machine learning architecture