Adaptation and Learning in Robots: Principles of Biological Control\n

Adaptation and Learning in Robots: Principles of Biological Control
Professor Slotine is the Director of the Nonlinear Systems Laboratory which studies general mathematical principles of nonlinear system stability, adaptation, and learning, and how they apply to robots and to models of biological control. The lab is particularly interested in how stability and performance constraints shape system architecture, representation, and algorithms in robots, and in whether similar constraints may in some cases lead to similar mechanisms in biological systems. Tools from nonlinear control, such as sliding variables, wave variables, and contraction theory also suggest a number of simple models of physiological motor control, which may help understand the specific roles of hierarchies, motor primitives, and nerve transmission delays.

Current projects include:
(1) Fast motion-vision coordination in robots; robotic catching of free-flying objects
(2) Models of the cerebellum and stability of biological feedback loops under nerve transmission delays
(3) Adaptive multiresolution approximation networks for real-time control and prediction
(4) Stable control using motion primitives; performance of combinations of local and centralized control mechanisms
(5) Entrainment models in mechanical and biological systems
(6) Nonlinear observer design techniques for real-time brain imaging