Towards the design of augmented feedforward and feedback for sensorimotor learning of motor skills

Paraskevi Kritopoulou (1) , Sotiris Manitsaris (1) , Fabien Moutarde (1)

1 CAOR - Centre de Robotique

Creating a digital metaphor of the “in person transmission” of manual-crafting motor skills, is an extremely complicated and challenging task. We are aiming to achieve the above by creating a mixed reality environment, supported by an interactive system for sensorimotor learning that relies on pathing techniques. The gestural instruction of a person, the Learner, arises from the reference gesture of an Expert. The concept of the system is based on the simple idea of guiding with the projection of a gesture depicting path in 2D space and in real time. The path is projected either as a feedforward that describes the gesture that has to be executed next, either as a feedback that amends the gesture while taking into account the time needed to correct the mistake. This projection takes place in the exact area where the object lies and the Learner is being trained, to avoid any distraction from the crafting task.

Mots clés: Pathfinding; gesture recognition; sensorimotor learning; scene analysis

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]; Environnements Informatiques pour l'Apprentissage Humain

Paraskevi Kritopoulou, Sotiris Manitsaris, Fabien Moutarde. Towards the design of augmented feedforward and feedback for sensorimotor learning of motor skills. 3rd International Symposium on Movement and Computing (MOCO'2016), Jul 2016, Thessalonique, Greece. pp.38, ⟨10.1145/2948910.2948959⟩. ⟨hal-01437136⟩