Cornell Machines and
Organisms Group:
Focus Area 3 of the Cornell IGERT
program on Nonlinear Systems
Faculty Participants
John
Guckenheimer: Mathematics
John
Hermanson: Veterinary College
Jane Wang: Theoretical and Applied
Mechanics
Andy Ruina: Theoretical and Applied
Mechanics
Hod Lipson: Mechanical and
Aerospace Engineering
Ephrahim
Garcia: Mechanical and Aerospace Engineering
Francisco Valero-Cuevas: Mechanical
and Aerospace Engineering
Our
interdisciplinary group at Cornell combines experimental, theoretical and
computational approaches to understand complex mechanical function in moving
machines and organisms. Arthropods and vertebrates are complex and nonlinear
biological systems whose versatility and robustness are the inspiration, if not
the envy, of engineers developing machine locomotion, flight and manipulation.
How and why can a fly with a handful of neurons outperform autonomous aircraft?
Why did four legged animals evolve to transition from trot to gallop? How can
the human hand be the epitome of dexterous manipulation when it has nonlinear
viscoelastic actuators and sensory delays?
A goal of
the group is to establish an interdisciplinary, synergistic research
environment where machines and organisms are seen as part of the continuum of
solutions to the mechanical challenges of locomotion, flight and manipulation.
Comparing and contrasting moving machines and organisms enables us to
understand both better.
In
particular, we will focus on three aspects of the mechanical function of
machines and organisms:
1. Structure and
function:
The group investigates
the materials, mechanics and structural topology that give rise to complex
mechanical function. The mechanical capabilities of muscles and tendons are
remarkable and still poorly understood compared to engineering materials and
actuators. Conversely, the best use of engineering materials and actuators in
machine design is a challenging problem.
2. Control and
regulation: The
group studies stability and control of mechanical function. Arthropods and
vertebrates have evolved to move quickly, efficiently, and stably. We will seek
to understand organism function and optimize machine design by identifying the
mechanical characteristics that make animals fast, efficient, stable and
dexterous.
3. Optimization
and evolution: The
group considers the evolution of machine design and the design consequences of
biological evolution. That is, how step-by-step changes in a design subject to
multidimensional constraints can add functionality and improve performance.
We are
opening open research avenues and educational opportunities in these areas via
a synergistic combination of mathematics, neuroscience, robotics, mechanics,
anatomy, physiology, and engineering.
Some
examples of our interdisciplinary group at Cornell are:
Structure and Function: Andy Ruina studies locomotion in
people and animals through robotic-style models, both computational and
physical. In recent work, he has shown that intermittent pendulum swinging
explains basic aspects of brachiation in arboreal primates and noted that the
three-beat foot-fall pattern of a galloping horse reduces collisional losses.
Francisco Valero-Cuevas studies the functional contributions of the structure
of complex anatomy of the human hand to manipulation. He also investigates the
functional consequences of variability, uncertainty and damage of structural
elements in the context of able, impaired and surgically modified hand
function. He interacts with clinical centers such as the Alberto Vilar Center
for Research on the Hand and Upper Extremity at the Hospital for Special
Surgery in New York City.
John Hermanson is investigating the elastic storage
mechanisms in the leg muscles of horses during locomotion.
Jane Wang investigates the intricacies of unsteady
aerodynamics of insect flight. She seeks to understand the fundamental physical
principles of design and control in flapping flight through studies of insect
flight, and to create virtual insects on computer.
John Guckenheimer seeks to develop computational
tools that will make it easier to create computational models for mechanical
devices and biomechanical systems to improve accuracy in their simulation. He
further strives to create algorithms that will automate the analysis of these
systems, especially in studying periodic orbits that correspond to steady
gaits.
Control and regulation: Francisco Valero-Cuevas' research
uses a rigorous analytical and experimental approach to understand dexterous
manipulation in humans and robots. This understanding will be instrumental to
revolutionize the design and performance of robotic hands and the clinical
rehabilitation of injured or diseased human hands.
Ephraim Garcia investigates the use of "smart
materials," such as super-elastic elements, in devices for adaptive morphological
changes, distributed sensing of environmental forces on machines, and
piezoelectric antennae for direct tactile sensing. Distributed sensing enables
organisms to gather information from their environment in ways that have been
difficult to reproduce in machines. For instance, adding tactile sensing to an
extender with end-effector (a robotic arm and hand), would allow force
reflective feedback to be generated and fed to a remote driver of the machine,
thus permitting more robust teleinspection or manipulation of objects.
Optimization and evolution: Hod Lipson studies robot mechanics
and its control through evolutionary dynamics. He seeks to understand
biological complexity and how it can be recreated. To reproduce the
functionality of a hand, for example, we clearly need to understand not only
how mechanisms work, but the processes by which mechanisms can add new
functionality relative to their predecessors, and increase their complexity
when beneficial without losing robustness.
Along these lines, Lipson, Valero-Cuevas and Garcia
are developing bio-inspired legged morphologies for clinical and robotic
autonomous robots to travel over uneven terrain. We are also recreating
fundamental steps in the co-evolution of hand and brain that endow the human hand
with its manipulation dexterity.
Program
requirements consist of two courses in nonlinear dynamics and computational
methods, a year-long interdisciplinary project, participation in an IGERT
seminar, a summer internship and completion of a Ph.D. minor.
Applications are coordinated through
participating graduate fields at Cornell. Applicants should describe their
interest in the IGERT program as part of the statement of purpose in their
Cornell graduate school application. They should also complete the contact form
on the web site
(click on IGERT Fellowship and then
Application).
Direct Inquiries to John
Guckenheimer at gucken@cam.cornell.edu