Tesis (Magíster en Ingeniería Electrónica)

Physical exercise (PE) contributes to achieving a successful rehabilitation program and rehabilitation processes assisted through social robots. However, the amount and intensity of
exercise needed to obtain positive results are unknown. Several considerations must be kept
in mind for PE implementation in rehabilitation as monitoring of patients’ intensity, which
is essential to avoid extreme fatigue conditions, which may cause physical and physiological complications. Machine learning models have been implemented to fatigue management
but limited in practice due to the lack of understanding of how an individual’s performance
deteriorates with fatigue accumulation; that can vary based on the physical exercise, environment, and individual’s characteristics. As a first step toward realizing the human-centered
approach to artificial intelligence and expert systems, this master thesis lays the foundation
for a data analytic approach to managing fatigue in walking tasks. The proposed framework
capitalizes on continuously collected human performance data from wearable sensor technologies. It establishes criteria for a feature and machine learning algorithm selection for fatigue
management, classifying four fatigue diagnoses state. Based on the proposed framework and
a large number of test sets used during the evaluation of the classifiers, we have shown that
(i) the random forest model presented the best performance with an average accuracy of
≥ 98% and an F-score of ≥ 93%, this model was comprised of ≤ 16 features; and (ii) the
prediction performance was analyzed by limiting the sensors used from four IMUs to two
or even one IMU with an overall performance of ≥ 88%; hence, only one wearable sensor
is needed for fatigue detection. This research presents an initial approach to a promising
tool for physical rehabilitation, and regarding classification accuracy, it presents remarkable
results according to the literature. We provide links to our data and code as supplementary
materials to encourage future work in this crucial area.