Recognition of human activities through eSteps insoles

Recognition of everyday human activity through mobile personal sensing technology plays a central role in the field of pervasive healthcare. The Bologna-based American company eSteps Inc. addresses the growing motor disability of the lower limbs by offering pre-, during and post-hospitalisation monitoring solutions with biomechanics and telerehabilitation protocol. It has developed a smart, customised and sustainable device to monitor motor activity, fatigue and injury risk for patients and a special app to share data with caregivers and medical specialists. The objective of this study is the development of an Artificial Intelligence model to recognize the activity performed by a person with Multiple Sclerosis or a healthy person through eSteps devices.

Sensor-based evaluation of Circadian motor behavior in people with Dementia. Development and validation of analysis strategies

[ITA]La demenza consiste nel deterioramento, spesso progressivo, dello stato cognitivo di un individuo. Chi è affetto da demenza, presenta alterazioni a livello cognitivo, comportamentale e motorio, ad esempio compiendo gesti ossessivi, ripetitivi, senza uno scopo preciso. La condizione dei pazienti affetti da demenza è valutata clinicamente tramite apposite scale e le informazioni relative al comportamento vengono raccolte intervistando chi se ne occupa, come familiari, il personale infermieristico o il medico curante. Spesso queste valutazioni si rivelano inaccurate, possono essere fortemente influenzate da considerazioni soggettive, e sono dispendiose in termini di tempo. Si ha quindi l'esigenza di disporre di metodiche oggettive per valutare il comportamento motorio dei pazienti e le sue alterazioni patologiche; i sensori inerziali indossabili potrebbero costituire una valida soluzione, per questo scopo. L'obiettivo principale della presente attività di tesi è stato definire e implementare un software per una valutazione oggettiva, basata su sensori, del pattern motorio circadiano, in pazienti affetti da demenza ricoverati in un'unità di terapia a lungo termine, che potrebbe evidenziare differenze nei sintomi della malattia che interessano il comportamento motorio, come descritto in ambito clinico. Lo scopo secondario è stato quello di verificare i cambiamenti motori pre- e post-intervento in un sottogruppo di pazienti, a seguito della somministrazione di un programma sperimentale di intervento basato su esercizi fisici. --------------- [ENG]Dementia involves deterioration, often progressive, of a person's cognitive status. Those who suffer from dementia, present alterations in cognitive and motor behavior, for example performing obsessive and repetitive gestures, without a purpose. The condition of patients suffering from dementia is clinically assessed by means of specific scales and information relating to the behavior are collected by interviewing caregivers, such as the family, nurses, or the doctor. Often it turns out that these are inaccurate assessments that may be heavily influenced by subjective evaluations and are costly in terms of time. Therefore, there is the need for objective methods to assess the patients' motor behavior and the pathological changes; wearable inertial sensors may represent a viable option, so this aim. The main objective of this thesis project was to define and implement a software for a sensor-based assessment of the circadian motor pattern in patients suffering from dementia, hospitalized in a long-term care unit, which could highlight differences in the disease symptoms affecting the motor behavior, as described in the clinical setting. The secondary objective was to verify pre- and post-intervention changes in the motor patterns of a subgroup of patients, following the administration of an experimental program of intervention based on physical exercises.

Analysis and detection of partial discharges in an electric motor of a hybrid supercar.

The objective of this thesis was the development of a new detection method of partial discharge (PD) activity in the stator of an electrical hybrid supercar fed by a silicon carbide converter, for which detection with common methods make it very difficult to separate PD pulses from switching noise. This work focused on the analysis and detection of partial discharges making use of an antenna, a peak detector, and an oscilloscope capable of capturing the electromagnetic pulses emitted during PD activity. Validation of the proposed method was done by comparing the partial discharge inception voltage (PDIV) detected by this system with the one obtained from an optical method of proven accuracy, with different rise times and samples. Further development of this method, if proved successful on a full stator, can help increasing the overall reliability of the car, potentially allowing for real time detection of PD activity and predictive maintenance before failure of the insulation system in a hybrid vehicle.

Motor Unit Spike Trains Recontruction from sEMG Signals for Gesture Classification on a Low-Power Processing Platform

Hand gesture recognition based on surface electromyography (sEMG) signals is a promising approach for the development of intuitive human-machine interfaces (HMIs) in domains such as robotics and prosthetics. The sEMG signal arises from the muscles' electrical activity, and can thus be used to recognize hand gestures. The decoding from sEMG signals to actual control signals is non-trivial; typically, control systems map sEMG patterns into a set of gestures using machine learning, failing to incorporate any physiological insight. This master thesis aims at developing a bio-inspired hand gesture recognition system based on neuromuscular spike extraction rather than on simple pattern recognition. The system relies on a decomposition algorithm based on independent component analysis (ICA) that decomposes the sEMG signal into its constituent motor unit spike trains, which are then forwarded to a machine learning classifier. Since ICA does not guarantee a consistent motor unit ordering across different sessions, 3 approaches are proposed: 2 ordering criteria based on firing rate and negative entropy, and a re-calibration approach that allows the decomposition model to retain information about previous sessions. Using a multilayer perceptron (MLP), the latter approach results in an accuracy up to 99.4% in a 1-subject, 1-degree of freedom scenario. Afterwards, the decomposition and classification pipeline for inference is parallelized and profiled on the PULP platform, achieving a latency < 50 ms and an energy consumption < 1 mJ. Both the classification models tested (a support vector machine and a lightweight MLP) yielded an accuracy > 92% in a 1-subject, 5-classes (4 gestures and rest) scenario. These results prove that the proposed system is suitable for real-time execution on embedded platforms and also capable of matching the accuracy of state-of-the-art approaches, while also giving some physiological insight on the neuromuscular spikes underlying the sEMG.