Description

Research Goals

Over the past three decades, aging of the European population and effects of unhealthy lifestyles have greatly increased the number of patients with high metabolic risk (e.g. type–2 diabetics and/or obese) who may benefit from physical activity to improve their motor performance and physiological/stress–related profile. Their motor rehabilitation and clinical supervision require specialised centres, such as sanitary gyms, and healthcare professionals (physiotherapists/rehabilitators) whose role has become even more critical during the Covid–19 pandemic. To overcome these issues,  I-TROPHYTS proposes to develop a visionary IoT and robotic framework that enables real–time and semi–autonomic monitoring, training and supervision of motor rehabilitation activities in indoor environments. The project employs humanoid robotics to support the healthcare professional and patients during the rehabilitation activities through the interaction of the robot with a smart space, where it acquires a grid of physiological and stress–related conditions of the patients’ activity. Based on these contextual data, the robotic system adapts the current motor routines to other more suitable ones, ensures a safe interpersonal distance between the acting subjects and, in critical cases, alerts the physiotherapist. Due to its structure, I-TROPHYTS addresses significant knowledge advances in multiple fields (rehabilitation, IoT, humanoid robotics, AI for healthcare) and involves a highly interdisciplinary team of Physical Engineering (PE) and Life Science (LS) researchers. Within the PE area, the project proposes the design and proof–of–concept implementation of a cutting–edge hw/sw platform, composed of:

  • a sensing layer, aimed at creating an integrated smart space fed by IoT wireless sensor networks for the characterization of patients’ vital and stress–dependent conditions;  
  • a knowledge layer, representing sensing information through new ontologies for the rehabilitation domain;
  • a planning and reasoning layer, which designs the robot adaptive decision–making module on a novel multi–agent architecture.

As for the LS field, the project introduces proper methodologies for motor routines definition and assessment of physiological and stress–related parameters, in a robotic–aided rehabilitation scenario. Finally, the project involves a close collaboration with an SME (CMG srl) dedicated to clinical activity and providing resources for the platform installation/calibration/evaluation and for the user acceptance rating.