Russ Tedrick: Robotics, Underactuation, and Control

The Essence of Robotic Movement

Russ Tedrick explores the intersection of mechanics and control theory, emphasizing that movement in robotics is not just about active control, but understanding how to leverage the natural physical world.

Passive Dynamics and Nature’s Efficiency

• Passive dynamic walkers represent the pinnacle of efficient motion, operating solely on gravity rather than complex control systems.
• Biological systems exhibit extreme efficiency, such as the rainbow trout which uses the vortex wake behind rocks to conserve energy, a phenomenon that even a dead fish can mimic.

The Philosophy of Control

Tedrick highlights that we should avoid overly rigid boundaries between mechanics, control, and intelligence.

"We shouldn't draw a line as clearly as we tend to."

• Underactuated systems are those where we have fewer actuators than degrees of freedom; Tedrick argues this is the natural state of the world.
• Rigid body assumptions in contact mechanics often lead to complex simulation paradoxes and discontinuities, requiring robust optimization-based control strategies.

Challenges in Robotics

• Testing and Validation: The difficulty of identifying corner cases in complex systems—notably seen during the DARPA Robotics Challenge—necessitates a shift toward better simulation environments like Drake.
• Manipulation: Moving beyond simple "pick and place" to complex tasks like tying shoelaces or buttoning shirts remains a primary goal, requiring a richer understanding of tactile sensing and soft contact interactions.

Future of Human-Robot Collaboration

• Fleet Learning: Rather than relying on a single robot to master all scenarios, distributed learning across a fleet is the scalable path forward.
• Societal Integration: Rather than fearing a "robot apocalypse," the future likely involves deep integration (human-machine hybrids) and robots serving as collaborative tools in the home, particularly for the elderly.