This course will introduce the basics of sensorimotor learning (Piaget Theory etc.) by focusing on the contribution of interactive systems to know-how transmission. In industrial or cultural field, in order to learn «how to perform » an expert gesture the transmission ‘in person’ is used. It means that the learner observes his master, tries to imitate his gestures and assimulation of the teached information is based on the interaction installed between them. However for various reasons the master can’t assure his presence for the entire learning process and the learner needs to train himself alone. In industrial context this need is even more important, since productivity issues appear when the master spends time with the trainee. New technologies can bring a partial solution to this issue, by providing interactive systems able to assist gestural learning during the self trainings.
For the creation of such systems it is essential to identify and understand all the gestural know-how components, to model and analyze them. Students will thus discover different movement modelling and analysis methods as well as different types of feedback that can be provided in order to guide the learner and permit him to adjust his gestural errors.
In a second part of the course, a methodology will be presented to highlight gesture recognition technologies for vocational training. This methodology consists of several stages going from gestural vocabulary definition to motion capture, data processing and feature extraction, comparison of gestural data and use of sensorimotor feedback for the guidance of gestural performance. Several applications in different fields will be presented (learning of artistic-musical gesture, technical gesture in the factory etc.).
Alina Glushkova|MINES ParisTech