NASA Space Program Presents: The Role of Connective Tissue and Tensegrity in Enabling a Hybrid Distributed/Centralized Control Scheme for Complex Coordinated Human Motion February 20, 2008

The NASA @ the Brickhouse speaker series continues this month with Vytas SunSpiral, expert in robotics development currently working at NASA Ames. Come to relax, network, grab a bite, a beverage, and maybe just maybe learn something.

See you there! Space is limited so please rsvp to [[email protected]]

While we are seeing amazing strides in robotic mobility, the great potential benefit of robots will not be unlocked until they can reliably manipulate and physically interact with the natural environment. In order for robots to be safe and responsive in dynamic unstructured environments we must move beyond the current model of highly geared rigid mechanical structures. Biological systems, such as humans, offer a great source of inspiration for new ideas on the design and control of manipulation systems. The classical view of the human body, which focuses on the musculoskeletal system, is that the bones are structural load bearing members that articulate via the muscles; and this concept is reflected in the current design of most robots today. Drawing on concepts from diverse fields such as Robotic Manipulation, Human Anatomy, Machine Learning, Control Theory, Martial Arts, and Therapeutic Body Work, I will outline a new approach to understanding the human body that focus on the system of connective tissue which creates a tensegrity structure within which our bones and muscles operate. The nature of this system allows for efficient force distribution and communication throughout the whole body without involving central processing and can be mostly handled in a distributed manner by the spine. Finally I will describe preliminary ideas on how the brain and spine interact with each other to create a hybrid distributed/centralized control scheme that is computationally efficient, adaptive, and integrates input from a variety of sources such as the inner ear and eyes with the control of motion.
Bio:
Vytas SunSpiral is currently the Principal Investigator for the ArmLab, within the Intelligent Robotics Group, Intelligent Systems Division (Code TI), at NASA Ames Research Center; the Project Lead for the ESAS-12 Athlete Foot Fall Planning Project; and a Project Scientist at Carnegie Mellon University.

Prior to starting work at NASA, Vytas was the Robot Personality Engineer and Technical Manager at Mobot Inc. (1998-2000), a Pittsburgh, PA based robotics company which built unsupervised fully autonomous socially interactive robotic tour guides for museums. The robots built by Mobot achieved unprecedented records of long term autonomous deployment as they operated autonomously on a daily basis in public spaces for multiple years with no direct human supervision.