Parts | Days | Selection | Search | Downloads | Help
SYPH: Physik im Hirn - Physical Approaches to Brain Function
SYPH 2: Physik im Hirn - Physical Approaches to Brain Function
SYPH 2.10: Poster
Thursday, March 14, 2002, 15:30–18:00, Poster D
Self-organization in sensor-motor loops by the homeokinetic principle — •Ralf Der1, J. Michael Herrmann1, and Michael Holicki2 — 1Institut für Informatik, Universität Leipzig, Augustusplatz 10/11, D-04109 Leipzig — 2Mathematisches Institut, Universität Leipzig, Augustusplatz 10/11, D-04109 Leipzig
Brain-style computing in animate agents profits from the exploitation of the effect of motoric activity onto sensory input via the environment. In order to predict these effects in face of the vast complexity of the external world, the brain must both invoke dynamical models of the sensor-motor loop as well as control the behavior such that prediction of sensory input becomes feasible and the complexity gap between model and true behavior is minimized. We consider model systems where it can be shown explicitly that advanced predictors naturally stabilize behaviors while retarded predictors can lead to the emergence of new behaviors by amplification of internal or external noise. The homeokinetic principle suggest to combine different predictors such that behaviors emerge which are well-suited to the robot. The analysis of the model systems is accompanied by experiments with real robotic systems.