ETIS Lab.

neurocybernetic team



Home | Publications | Members | Internships | Medias | Videos | Access


Active perception and multimodality

Keywords: multimodal integration, facial expression, mirror neurons, artificial skin, tool-use, body image, imitation learning, cerebellum, sensorimotor coordination.

Current theses

Exploration and active visual scene recognition : Impact of the use of an event based camera.
The visual analysis of a scene seems inherent to a lot of autonomous inteligent systems. Nowadays: - It is often passive, computing the frame of an image. - It can be very ressource consuming, in case of complex scenes and high quantity of data to process. Thus, this thesis aims at introducing and evaluating the impact of recognition algorithms, using more computing-friendly event-based datas, where the system actively participate to the recognition with a sensori-motor loop.
Low-level synchronization to intentional interaction : intuitive interactions with a companion robot
Nils Beaussé (Dr. Ghilès Mostafaoui, Prof. Philippe Gaussier)

Succeeded theses

Incremental Activation of Multiple Learning Systems From Motor Babbling to Tool-Use
The main topic of the thesis is to model the neural mechanisms underlying the emergence of tool-use in infants development for robots.
Artificial Skin for a Humanoid Robot
The sense of touch is considered as an essential feature for robots in order to improve the quality of their physical and social interactions. For instance, tactile devices have to be fast enough to interact in real time, robust against noise to process rough sensory information as well as adaptive to represent the structure and topography of a tactile sensor itself – i.e. the shape of the sensor surface and its dynamic resolution. In this paper, we conducted experiments with a self-organizing map neural network that adapts to the structure of a tactile sheet and spatial resolution of the input tactile device; this adaptation is faster and more robust against noise than image reconstruction techniques based on electrical impedance tomography. Other advantages of this bio-inspired reconstruction algorithm are its simple mathematical formulation and the ability to self-calibrate its topographical organization without any a priori information about the input dynamics. Our results show that the spatial patterns of simple and multiple contact points can be acquired and localized with enough speed and precision for pattern recognition tasks during physical contact.
Interactive Learning in Autonomous Robotics: Toward New Kinds of Human Robot Interface
Dr. Antoine de Rengerve 2014 (Dr. Pierre Andry, Prof. Philippe Gaussier)
Robotic hand gripping screwdriver
Electronic skin
Tino tracking ball

© 2011–2018 — ETIS UMR 8051 — webmaster Arnaud Blanchard