Thesis : Allan HENRY

We assume that fundamental principles of voice guidance have universal aspects that we will try to discover and exploit, in order to design devices for controlling a robot by voice. We will first explore how people spontaneously interact to guide to an invisible 3-D object a) a person directly, b) a flying robot pilot. Of particular interest are the non-verbal components of speech that contribute to the most effective (intuitive and rapid) instructions.
The phonic sequences used by the guides and the joystick manipulation sequences used by the pilots in this first phase will have to be matched by end-to-end network learning. The emerging intelligence will then be embedded in the teleoperation system.
Our primary objective is to use the most relevant voice features to facilitate drone control by the operator's voice, and possibly to facilitate sound guidance for a blind person.
The project combines the respective expertise of three supervisors, in speech processing, cognitive neuroscience and robotics, from three Grenoble laboratories. The PhD student will benefit from synergies with an ongoing thesis on visual-auditory substitution (SamGuide project).

Supervisors :
Christian GRAFF christian.graffuniv-grenoble-alpes.fr (christian[dot]graff[at]univ-grenoble-alpes[dot]fr) - LPNC
Sylvain HUET sylvain.huetgipsa-lab.grenoble-inp.fr (sylvain[dot]huet[at]gipsa-lab[dot]grenoble-inp[dot]fr) (Co-encadrant) - GIPSA-Lab
José-Ernesto GOMEZ BALDERAS jose-ernesto.gomez-balderasgipsa-lab.grenoble-inp.fr (jose-ernesto[dot]gomez-balderas[at]gipsa-lab[dot]grenoble-inp[dot]fr) (Co-encadrant) 
Solange ROSSATO solange.rossatouniv-grenoble-alpes.fr (solange[dot]rossato[at]univ-grenoble-alpes[dot]fr)(Co-encadrant)

Keywords : robot,Human Machine Interface,target reaching,acoustics,voice,drone,