For video footage from past you can visit the individual event pages, or go to our YouTube Channel.
To filter by event category, click on the event category link in the table below or use the menu on the right.
List of Past Events
The Dynamics of Perception & Action: Of Legged Locomotion & Bouncing Babies
William H. Warren
Tuesday, March 06, 2007, 01:00pm - 02:00pm
Brown University, Department of Cognitive & Linguistic Sciences
How do perception and action yield organized behavior? I will develop the view that stable, adaptive behavior emerges from the interaction between a structured environment and an embodied agent with a few simple control laws. The behavioral dynamics of this interaction can be formalized as a nonlinear dynamical system whose behavior is organized around attractors, repellers, and bifurcations. Two case studies exemplify this approach.First, consider a six-month old learning to bounce in a �jolly jumper.� A longitudinal study reveals that following random exploration, infants suddenly achieve stable bouncing by matching their leg stiffness to the spring stiffness and driving the system at its resonant frequency. The behavioral dynamics can be modeled as an intrinsically forced mass-spring system, in which leg stiffness and kicking frequency are specified by sensory information.Second, consider an adult guiding their locomotion through a complex, changing environment. Locomotor behavior can be decomposed into four basic components: (a) steering toward a stationary goal, (b) avoiding a stationary obstacle, (c) intercepting a moving target, and (d) avoiding a moving obstacle. We study each behavior during walking in a virtual environment and model it as a dynamical system. By combining model components, we can predict locomotor paths in more complex environments. Organized behavior can thus be understood as emerging on-line from the agent-environment interaction, making explicit planning unnecessary. Rather than acting like a centralized controller, the nervous system discovers simple control laws that exploit physical and informational regularities to stabilize behavior.