Background:

It is nowadays widely recognized how biology can represent an inspiration source in robotics, both for the development of biomimetic components and for new control principles for robotic systems (Brooks, 1991; Dario et al.,2005) to the final aim of developing robots with better sensory motor performance, especially in real world scenarios (Laschi et al., 2008; Guglielmelli et al., 2007; Pfeifer et al., 2007). According to neurophysiological findings, human motor control is based on sensory predictions more than on sensory feedback (Berthoz, 2002; Johansson, 1998). Due to the delays in the transmission of the nervous signals, fast and coordinated movements cannot be explained by pure feedback (Kawato, 1999; Miall et al, 1993; Wolpert et al, 1998). Current neuroscientific literature provides a rich landscape of anticipation-based possibilities of explanation, for various aspects of sensory-motor coordination. Generally speaking, anticipation is regarded as a way for overcoming a significant difficulty related to feedback-based models, when they are used for explaining sensory-motor coordination capabilities in humans and animals in general. The rapidity of sensory-motor coordination in humans and animals is unlikely to be explained by feedback based models, as sensory signal conduction and processing in the CNS requires too much time. If motor commands were generated only on the basis of the last sensory feedbacks perceived by the subject, reaction to external stimuli would be always late with respect to the actual event that caused the sensory stimulation. This issue, diffusely discussed in (Wolpert et al., 1998), can be retrieved in several models of locomotion and gaze control. In fact, prediction is a mandatory factor during smooth pursuit eye movements. During maintained smooth pursuit, the lag in eye movements can be reduced or even cancelled if the target trajectory can be predicted (Fukushima et al. 2002).

Aim and Scope of the Workshop:

This workshop aims at discussing the scientific issues and the technological challenges related to the implementation of predictive models in the control of humanoid robots, with reference to (but not limited to) locomotion and the control of gaze, which is strictly related to locomotion and which is strongly based on anticipation mechanisms.

Organizers

Paolo Dario, Scuola Superiore Sant’Anna, Pisa, Italy

Alain Berthoz, College de France, Paris, France

José Santos-Victor, Istituto Superior Tecnico, Lisbon, Portugal

Atsuo Takanishi, Waseda University, Tokyo, Japan

Schedule

Sunday, June 27, 2010