A physiological characterization of motion sickness disease and a methodology to detect its arousal

According to some epidemiological studies, one-third of the global population is highly susceptible to motion sickness. The main reason for this disease is assigned to the mismatch between the real and the perceived motion. People suffer this kind of conflict that can be characterized by different symptoms, such as nausea, pallor, cold sweating, increased salivation and stomach awareness. The sensitiveness to this disease can be debilitating. This thesis work addresses the identification of motion sickness disease. An in-depth analysis of the available literature on motion sickness disease has been carried out, with attention to the physiological signals that are usually analysed in this context and that show a physiological reaction to the felt disease. A data analysis framework has been developed to analyse the Electrocardiogram signal. The target signal has been preprocessed by applying filtering techniques with tuned frequencies according to the expected physiological range. A set of features has been then extracted both in time and in frequency domain, whose correlation has been analysed to prune the starting features set. The most promising classification algorithms have been adopted to create and validate a model able to recognize the well-being or the state of illness of the subjects. The best configuration of these algorithms has been set through a grid search analysis. Subjects have been also analysed from a demographic point of view, to assess the impact of the subject context on the model performance.