Sentry: development of an IMU-based holter for the ecological rehabilitation of the upper limb.

Movement analysis has become increasingly crucial in rehabilitation as a critical reference to plan therapeutic intervention in recent years. Despite being highly accurate, most of the available solutions on the market are limited to the clinical environment due to their cost and portability. During the rehabilitation period, the physiotherapist cannot follow the patient in the various delicate phases of the day. In periods of forced social distancing like the one we have recently experienced, the already limited medical visits may be further reduced, with the risk of decreasing treatment effectiveness. A possible solution may come from modern technologies, but at the moment, the available motion analysis devices are not affordable for private professionals to allow for ecological measurements, i.e. to collect data directly from the daily living scenario of the patient. Sentry is a wearable, non-invasive device that aims to answer most of these problems and is based on Inertial Measurement Units (IMUs). With an optimal trade-off between costs and portability, recent evidence promotes the use of IMUs in ecological measures to improve the effectiveness of the treatment. IMUs allow studying joints with more degrees of freedom at once for prolonged periods and without hindering the person's natural movement. This thesis presents the development and the performance estimation of an IMU-based device, with the development of both hardware and software to obtain a working prototype. The Sentry device includes two IMU sensors that can estimate the absolute orientation between the arm and the shoulder. Multiple tests have been performed using a robotic arm that executed repeated movements to evaluate repeatability, accuracy, and drift. These parameters were assessed under different speed values, positions and accelerations, reproducing the everyday conditions that a wearable device must withstand on a patient in the clinical phase. The sensors were also tested on a person in an actual use scenario to evaluate the effects of magnetic disturbance due to the environment, i.e. fields generated by electronic devices. Tests were carried out to evaluate the feasibility of this solution for ecological use, analysing aspects such as ease of use and repeatability of positioning (essential for use without the vigilance of the clinician) and battery life. The device was compared with the optoelectronic motion analysis system to analyse the planar shoulder movements of healthy people. The measurement uncertainties seem more than acceptable for the clinical applications targeted by the study, i.e., post-surgical limb rehabilitation. The limitations of the sensors are mainly related to magnetic disturbances such as hard and soft iron effects, which incidence can vary by application. It has been observed that this solution can respond to the needs of a rehabilitation environment that has to deal with an increasingly older population in developed countries and therefore requires all possible solutions to make the rehabilitation process more efficient.