Motor control has a predominant role in the understanding of cognition [ref 1]. We employ various methods ranging from machine learning and non-linear optimization to investigation of human behavior (in collaboration with the human behavior lab). On the computational side, this is done using parametric and non-parametric techniques. The human behavior aspects are covered relying on the equipment of the human behavior lab for measuring kinematics of the arms and body, eye movements, and electromyography.
|Examples of the acquisition and reproduction of human kinematics using the VICON system (left) and the iCub simulator (right).|
|Unfolding in time of the approximating ANN of a non-linear optimal control task (e.g. generation of generic point to point trajectories).||Point to point reaching using minimum jerk cost functions and non-linear optimal control methods. Here optimal control implemented through a cascade of ANN is used to approximate a set of control variables from start to end of the trajectory.|
|Summary of methods (top row) and results (bottom row) of the “perceptual-only” (left panel) and “driving motor act” (right panel) experiments, for different force field conditions. In the perceptual task no significant difference was found among all different force fields conditions, while in the motor task the dynamics stability and a gravitational environment improved prediction.|
von Hofsten, C.An action perspective on motor development. Trends in cognitive sciences, 2004. 8(6): p. 266-272.
F. Nori, G. Metta, G. Sandini Exploiting Motor Modules in Modular Contexts . In Robust Intelligent Systems. Schuster, Alfons (Ed.) 2008, XII, 299 p. 81 illus., Hardcover ISBN: 978-1-84800-260-9.
A. Sciutti, F. Nori, G. Metta, T. Pozzo, G. Sandini Motor and perception-based prediction Perception 37 ECVP Abstract Supplement, 2008.
Bentaleb, T, Sakka, S, Metta, G & Sandini, G Using Whole Body Captured Human Motion to Control a Humanoid Robot in In Proceedings of the Journees Nationales de la Robotique Humanoide (JNRH'09). Nantes, France, May 5-6, 2009.
Ivaldi, S.; Baglietto, M.; Metta, G.; Zoppoli, R. An application of receding-horizon neural control in humanoid robotics. 2009 L. Magni et al. (Eds.): Nonlinear Model Predictive Control, LNCIS 384, Springer-Verlag Berlin Heidelberg pp. 541-550 .