From Sensors to AI: How Rehabilitation Becomes Personalized

From Sensors to AI: How Rehabilitation Becomes Personalized

• December 10th, 2025

Upper-limb rehabilitation is a therapeutic pathway focused on restoring the functionality of the shoulder, elbow, wrist, and hand following trauma, surgery, or pathological conditions. In this context, technological innovation can serve as a valuable ally for improving treatment quality and tailoring recovery to each patient's needs.

As part of a PhD program in bioelectronics carried out in collaboration between the Medical Devices team of the Institute of Digital Technologies for Personalized Healthcare (MeDiTech) at SUPSI and the University of Luxembourg (Uni.Lu), the project led to the development of wearable sensors that integrate two key technologies: surface electromyography (sEMG), which measures the electrical activity produced by muscles, and inertial measurement units (IMU), used to track kinematic movements.

The platform also includes a web and mobile application that allows real-time recording and visualization of data collected via Bluetooth-connected sensor nodes. Thanks to this intuitive interface, therapists can adapt the rehabilitation pathway to the specific needs of each patient.

To validate the approach, the project also carried out an experimental measurement campaign involving 20 volunteers. This resulted in a pseudonymized dataset-where personal data are replaced with identification codes-based on physiotherapeutic movements inspired by the standardized Fugl-Meyer clinical assessment scale. The dataset was then released as open data for the benefit of the scientific community.

Data analysis enabled the training of algorithmic models capable of accurately recognizing different movements (e.g., elbow flexion and extension, shoulder abduction at 90° and 180°, vertical shoulder extension at 90°, external and internal shoulder rotation, wrist flexion and extension), confirming the potential of the platform for future applications in personalized rehabilitation.

Looking ahead, the project will move toward even more advanced technologies, including high-density electromyographic sensors (HDsEMG) and algorithms embedded directly into wearable devices, with the goal of bringing these solutions into everyday clinical practice. Furthermore, once the rehabilitation clinic for the continuation of the project has been identified and the statistical sample selected, the team will determine the specific pathology on which to focus the next phase. This will ensure a targeted development pathway and maximize the project's potential to deliver meaningful therapeutic impact.