Cyclical Dynamics and Control of a Neuromechanical System
Abstract
In this project, we used computational models to analyze how the intrinsic dynamical properties of neural and mechanical systems interact to produce stable, but adaptable locomotion. Animal locomotion is a rhythmic behavior that requires the effective coupling of multiple feedback loops, including mechanical coupling between the animal's body and the environment, coupling between muscular force production and body movement, and sensory feedback. Floquet theory, a branch of nonlinear dynamics, includes ways to analyze how such rhythmic systems respond to perturbations. We developed several robust ways of estimating the Floquet modes of a rhythmic system, which are canonical patterns of activity after a perturbation. We found that when a block of muscle is forced to change length sinusoidally and is cyclically activated, it is strongly self-stabilizing, even with no sensory feedback. When two muscles act antagonistically, as they do around most vertebrate joints, then the system is less stable naturally. However, with sensory feedback, the joint can be stabilized very easily. This research may be extended to analyze Floquet modes based on empirical data, to examine the stability properties of real muscle, and to study the stability of fish swimming and control potential of fish fins.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2012
- Accession Number
- ADA584529
Entities
People
- Eric D Tytell
Organizations
- Johns Hopkins University