Analyse de signaux vibrotactiles et modèles flous de la perception : application aux interfaces tactiles pour l'automobile et l'aéronautique

摘要

In the field of automotive and aeronautical industries, human interaction with touch interfaces can be improved by using vibratory feedback or haptic effects, directly delivered to the user finger upon screen interaction. This new approach, which pro-actively stimulates the tactile sense, provides a safer and reliable way to interact with touch interfaces. Moreover, complex vibrational patterns can be designed in order to offer unique tactile sensations and thus increase user’s quality of experience. In this context, our first contribution focuses on a time-frequency analysis of vibrational patterns using the continuous wavelet transform of the signal delivered by the interface, in order to extract its most salient features, chosen based on a psychophysical study of the tactile sense, and which account for human perception. Using these features, our second contribution is a psychophysical model of vibrotactile perception developed using fuzzy logic and an original rule-base extraction method extending the classical Wang-Mendel approach. This model predicts the perceived comfort induced by a vibratory pattern according to its psychophysical properties, and shows good performances. The knowledge retrieved allowed the detection of several behavioral paradigms of vibrotactile perception, such as the effect of energy and texture. Our third contribution is the development of an ergonomic model of vibrotactile perception based on the evaluations of an automobile expert. Thanks to its good performances, the model was successfully adapted to the aeronautics area.The results of this thesis provide assistance to equipment suppliers by simplifying the conception of haptic effects intended to improve human interaction with touch interfaces.