Abstract
The SORTT project aimed to develop an innovative framework enabling healthcare professionals to supervise a larger number of patients while reducing dependence on costly healthcare infrastructures. The project focused on supporting motor rehabilitation directly in patients' homes through the integration of real-time monitoring, edge intelligence, and ontology-based reasoning. The underlying challenge was the integration of heterogeneous information sources—including physiological signals, inertial measurements, diagnostic information, patient conditions, and rehabilitation goals—into a unified system capable of supporting personalized rehabilitation and clinical decision making. To address this challenge, SORTT combined wearable Internet of Things (IoT) technologies, edge-computing architectures, Deep Learning (DL) techniques, and semantic knowledge representation within a unified platform. Wearable sensors continuously collected patient motion and physiological data, while edge-based algorithms processed the acquired information locally to assess exercise execution and generate high-level indicators. These outputs were integrated into an ontology-based knowledge model capable of representing patients, rehabilitation exercises, clinical conditions, and therapeutic objectives. The resulting framework supports context-aware monitoring, adaptive rehabilitation supervision, and the generation of meaningful information for healthcare professionals. The project also delivered a proof-of-concept implementation integrating sensing, intelligent data processing, and semantic reasoning, demonstrating the feasibility of a distributed and scalable approach to home-based physiotherapy.
Results achieved
The SORTT project successfully developed and validated an integrated technological framework combining IoT sensing, edge computing, AI, and ontology-based knowledge representation for home-based motor rehabilitation. A comprehensive analysis of the rehabilitation domain was first carried out, covering physiotherapy protocols, wearable sensing technologies, digital twins, semantic technologies, and healthcare ontologies. This activity enabled the identification of representative rehabilitation scenarios, suitable sensing technologies, and the physiological parameters required for continuous patient monitoring. The project designed and implemented an interoperable IoT infrastructure based on wearable sensors capable of acquiring patients' physiological and movement data during rehabilitation sessions. Particular attention was devoted to interoperability among heterogeneous sensing devices based on Web of Things (WoT) approaches and to the development of edge-computing architectures able to process sensor data locally. Machine and Deep Learning algorithms were integrated into the edge platform to analyse rehabilitation exercises, assess the quality of the motor routines, and generate high-level information suitable for subsequent semantic processing. A major outcome of the project was the development of an ontology-based framework capable of integrating heterogeneous patient information into a coherent semantic representation. The proposed Ontology-based Human Digital Twin (OHDT) models patients, rehabilitation exercises, clinical conditions, therapeutic goals, physiological observations, and contextual information while explicitly representing data provenance, contextual dependencies, and possible inconsistencies among heterogeneous data sources. This semantic layer enables a comprehensive representation of the patient's evolving condition and provides the knowledge foundation for context-aware reasoning and adaptive rehabilitation support. The project also developed a complete input/output infrastructure integrating wearable sensing, edge intelligence, semantic reasoning, and monitoring services into a unified rehabilitation platform. The resulting proof of concept was experimentally evaluated using voluntary healthy participants, allowing the consortium to validate the complete data flow, assess the responsiveness and reliability of the IoT infrastructure, evaluate different sensing configurations, and verify the overall integration of the software and hardware components. The outcomes of the project were disseminated through multiple channels, including: the publication of more than ten articles in PE-related and LS-related editorial venues, the organization of a workshop (IEEE W3R-HEALTH) on topics closely related to the SORTT project, the participation to public engagement events.Project details
Unibo Team Leader: Marco Di Felice
Unibo involved Department/s:
Dipartimento di Informatica - Scienza e Ingegneria
Coordinator:
CNR - Consiglio Nazionale delle Ricerche(Italy)
Total Unibo Contribution: Euro (EUR) 79.574,00
Project Duration in months: 27
Start Date:
30/11/2023
End Date:
28/02/2026
