Transcript Biomimetic Robots for Robust Operation in Unstructured

BioMimetic Robotics
MURI
Berkeley-Harvard
Hopkins-Stanford
Biomimetic Robots for
Robust Operation in
Unstructured Environments
R. Howe
Harvard University
M. Cutkosky and T. Kenny
Stanford University
R. Full and H. Kazerooni
U.C. Berkeley
R. Shadmehr
Johns Hopkins University
http://cdr.stanford.edu/touch/biomimetics
ONR/DARPA MEETING ON LEGGED ROBOTS, COOPERATIVE BEHAVIOR, AND NAVIGATION
COSTAL SYSTMS STATION, PANAMA CITY, MAY 4-5, 1999
BioMimetic Robotics
MURI
Berkeley-Harvard
Hopkins-Stanford
Main ideas:
• Study insects to understand role
of passive impedance (structure and control),
study humans to understand adaptation and learning
(Full, Howe,Shadmehr)
• Use novel layered prototyping methods to create
compliant biomimetic structures with embedded
sensors and actuators
(Cutkosky, Full, Kenny)
• Develop biomimetic actuation and control schemes
that exploit “preflexes” and reflexes for robust
locomotion and manipulation
(Full, Cutkosky, Howe, Kazerooni, Shadmehr)
BioMimetic Robotics
MURI
Berkeley-Harvard
Hopkins-Stanford
Status - Locomotion
• First year of project
• Preliminary experiments to determine
insect leg properties
• Fabricated first prototypes of embedded
sensors and actuators
• Locomotion focus: rough terrain traversal,
inspired by cockroach running over
blocks surface ~3x “shoulder” height
MURI Interactions: Areas and Leadership
Motor Control
& Learning
Johns Hopkins
Rapid Prototyping
Stanford
Muscles and
Locomotion
UC Berkeley
Bob Full
MURI
Manipulation
Harvard
Sensors / MEMS
Stanford
Robots & Legs
UC Berkeley
Neuro-Mechanical Model
Higher
Centers
Sensors
Aero-, hydro-, terra-dynamic
Open-loop
Feedforward
Controller
(CPG)
Mechanical
System
(Actuators, limbs)
Feedback
Controller
Adaptive
Controller
Environment
Closed-loop
Sensors
Behavior
Neuro-Mechanical Model
Mechanical Feedback
Feedforward
Controller
(CPG)
Mechanical
Behavior
System
(Muscles, limbs)
Closed-loop
Sensors
Reflexive Neural Feedback
Contribution to Control
Mechanical System
Feedforward
Preflex
Motor program Intrinsic musculoacting through skeletal properties
moment arms
Predictive
Rapid acting
Passive Dynamic
Self-stabilization
Neural System
Reflex
Neural
feedback
loops
Slow acting
Active
Stabilization
Perturbation Response
Force Perturbation
Animal Strikes Obstacle
Working Hypotheses
Smaller Reaction Force
Joint Angles Altered
Less Stable
Decreased Speed
Larger Reaction Force
Joint Angles Similar
More Stable
Maintain Speed
Discoveries
•Preflex Present
•No Active Reflex Required
•Stiffness Varies During Cycle
Perturbation Experiments
Relative muscle length (%)
Muscle is Stiffest at Midstance
105
Active
100
95
0
50
Locomotion cycle (%)
100
Leg Stiffness
1st
Measures
of Leg
Stiffness,
Damping
Servo Motor
Roach leg
Length and
Force recording
Impact on Deliverables
1. Flexible Robot Leg
Could Reject Perturbations
2. Simplify Control (feedforward)
3. Suggest Design of Artificial Muscles
MEMS Instrumentation for biomechanics studies (Kenny/Full)
Micromachined Force Sensor for Adhesion Force
Measurement of Single Gecko Setae
Yiching Liang and Tom Kenny
Stanford University
~106 setae per animal, average 4.7 m diameter
Wall climbing mechanisms: Suction, Capillary (wet) adhesion,
Micro-interlocking, Electrostatic attraction - NOT;
van der Waals forces?
2-Axis Micromachined Force Sensor
Special 45 ion implantation to
embed piezoresistors on surfaces
and side walls.
la te ra l
sensor
L ig h tly d o p e d
(p ie zo re s is tive )
ve rtic a l
sensor
H e a vily d o p e d
(h ig h ly c o n d u c tive )
Attachment point
Gecko measurements now underway...
MURI Interactions: Areas and Leadership
Motor Control
& Learning
Johns Hopkins
Reza Shadmehr
Manipulation
Harvard
Rapid Prototyping
Stanford
Muscles and
Locomotion
UC Berkeley
MURI
Sensors / MEMS
Stanford
Robots & Legs
UC Berkeley
Neuro-Mechanical Model
Higher
Centers
Sensors
aero- , hydro, terra-dynamic
Open-loop
Feedforward
Controller
(CPG)
Mechanical
System
(Actuators, limbs)
Feedback
Controller
Adaptive
Controller
Environment
Closed-loop
Sensors
Behavior
Relating Limb Impedance and Learning
General Goal:
Understand human arm impedance strategies when learning
tasks in unstructured environments
Challenges:
The biomechanics of the human arm are dominated by
multiple time delays in feedback.
How do time delays affect measures of arm impedance?
Humans learn internal models to learn control.
How does a change in the internal model affect measures
of arm impedance?
Results
In general, time delays in feedback reduce apparent viscosity
and add apparent mass to a system.
Example: m x  b x  kx ( t   )  0
Using Taylor series expansion
x (t   )  x (t ) 
dx
dt
The resulting
on the delay :
2
(  )  1 / 2
d x
dt
2
(  )  
2
system has apparent impedance
( m  1 / 2 k  ) x  ( b  k  ) x  kx  0
of :
Human Arm Motor Control Model
A model of the human arm’s time-delayed control processes
were used to derive bounds on the impedance changes that
should occur as a function of learning.
Implications for Robot Control
• Relates delays to variation in limb
impedance - convenient means of
analyzing mechanical interactions
• Method for trading off “costs” of
higher-level processing delay vs. passive
impedance
MURI Interactions: Areas and Leadership
Motor Control
& Learning
Johns Hopkins
Rapid Prototyping
Stanford
Muscles and
Locomotion
UC Berkeley
MURI
Manipulation
Harvard
Robert Howe
Sensors / MEMS
Stanford
Robots & Legs
UC Berkeley
Impedance in Manipulation
STIFFNESS
Muscle Impedance
FORCE
Example: Grasping in an
unstructured environment
• Before contact:
No interaction force =>
Low arm stiffness k
• Collision with object
produces small
disturbance force
f=kx
Variable Impedance Manipulation Testbed
Whole-Arm Manipulator
(Barrett Technologies)
• Low moving mass
• Negligible friction
• Back driveable
=> Low impedance robot
Goal: Minimum Impedance Grasping
and Maniplation
Combine biologically-inspired
elements:
• low-impedance manipulator
• feedforward dynamic models
(limit feedback gains to
reduce impedance)
“Intrinsic” tactile sensing
(contact location from forcetorque measurements)`
• simple contact sensing
=> Ability to probe and grasp
objects with minimum forces in
unstructured environments
MURI Interactions: Areas and Leadership
Motor Control
& Learning
Johns Hopkins
Rapid Prototyping
Stanford
Muscles and
Locomotion
UC Berkeley
MURI
Manipulation
Harvard
Sensors / MEMS
Stanford
Robots & Legs
UC Berkeley
Hami Kazerooni
Objectives
• Create a robust, simple, and fast legged
platform, able to traverse rough block surface
• Use off-the-shelf fabrication technology
• Explore role of open-loop impedance and
mechanical design
• Serve as early testbed for control concepts
Initial Focus: Leg Mechanism
Full has shown that a substantial portion of
locomotor control is simple and resides in the
mechanical design of the system
Biological Observations
• Control results from the
properties of the parts and their
morphological arrangement.
Musculoskeletal units and legs
do much of the computations
on their own by using segment
mass, length, inertia, elasticity,
and damping as “primitives”.
Engineering Equivalence
• System performance is
function of the physical
system; no feedback
control has been used to
alter the dynamics of the
system.
Biological Observations
• Position control using reflexes
is improbable if not impossible
Engineering Equivalence
• No need for sensors for
position speed, or force
control
• During climbing, turning, and
maneuvering over irregular
terrain, animals use virtually
the same gait as in horizontal
locomotion - an alternating
tripod. The animals appear to
be playing the same
feedforward program for
running.
• A one degree of freedom
system only. No need to
design elaborate multivariable robotic legs.
1-DOF Linkage Design Example
f
g
b
c
a
d
MURI Interactions: Areas and Leadership
Motor Control
& Learning
Johns Hopkins
Rapid Prototyping
Stanford
Mark Cutkosky
Muscles and
Locomotion
UC Berkeley
MURI
Manipulation
Harvard
Sensors / MEMS
Stanford
Tom Kenny
Robots & Legs
UC Berkeley
Application: Small robots with
embedded sensors and actuators
Shaft coupling
Shaft
Motor
Leg links
Building small robot legs
with pre-fabricated
components is difficult…
Is there a better way?
Shape Deposition Manufacturing
(CMU/SU)
Embedded Component
Part
Deposit (part)
Support
Shape
Shape
Deposit (support)
Embed
Embedded Components
+ Soft materials =>
•Improved robustness
•Simplified construction
Robot leg example
(http://cdr.stanford.edu/biomimetics)
Steel leaf spring
Piston
Part Primitive
Outlet for valve
Valve Primitive
Circuit Primitive
Inlet port primitive
Designer
composes the
design from library
of primitives,
including
embedded
components
Robot Leg: compacts
The output of the
software is a
sequence of 3D
shapes and
toolpaths.
Embedded components
Part
Support
Robot Leg: embedded parts
Steel leaf-spring
Piston
Sensor and circuit
Valves
A snapshot just after valves and pistons were inserted.
Pressure Control in Small Pneumatic Systems
• SDM allows fabrication of small integrated mechanisms
• Control of small pneumatic systems with off-the-shelf
components (solenoid valves) is in a challenging regime
• Miniature analog servo-valves needed for smooth
performance are not available
t
A ir V o lu m e
P isto n
Solenoid Pressure
Valves Control Impossible
A tm o sp h e ric P re ssu re
E xh a u st va lve
P re ssu re S e n so r
In le t va lve
S u p p ly P re ssu re
PWM Control
Small Pneumatic
Systems
Usual regime of
Operation
t volume
SDM Considerations for Embedded Sensors/Actuators
Different Sensors and Actuators have different
considerations for embedding, generally these
include:
 Coupling and Adhesion
 Fixturing, Positioning, Placement
 Protection and Encapsulation
 Multiplexing, Connectivity, Interconnect Integrity
and Strain Relief
 Thermal energy generation and cooling
Sensor circuit boards interconnect pins protected
in wax before embedding
Circuit boards embedded
with pressure sensor--sensor
ports protected with wax
Embedded sensor and
circuitry with sacrificial wax
removed
Assembled into pneumatic
system
Robot Leg: completed
Finished parts ready for testing
MURI Interactions: Areas and Leadership
Motor Control
& Learning
Johns Hopkins
Rapid Prototyping
Stanford
Muscles and
Locomotion
UC Berkeley
MURI
Manipulation
Harvard
Sensors / MEMS
Stanford
Robots & Legs
UC Berkeley