vCare

To reduce the risk of relapse from a severe acute illness in old age, an active change of lifestyle motivated by the affected person is often necessary. To facilitate this change, the EU-funded Horizon2020 “vCare – Virtual Coaching Activities for Rehabilitation in Elderly” project aims to develop a virtual coaching system that suggests personalized care paths for patients transitioning from rehabilitation to the home environment. For doing so, the international research group links health information records with data from the home environment and regular interactions of the virtual coach with the affected person.

Within this project consortium, the FZI is among other things responsible for patient and rehabilitation modeling, pattern recognition for successful treatment paths, and learning as well as finding adequate treatment suggestions. Since the human being is at the center of the decision-making process, the explainability of the virtual coach’s suggestions is also an important factor. For enabling the person to adapt the virtual coach to the person’s personal preferences and experiences, an interactive machine learning approach is employed that combines reinforcement learning, imitation learning, and human-centric AI techniques.