Applying Biological Principles to the Design and Operation

Download Report

Transcript Applying Biological Principles to the Design and Operation

Functional biomimesis
*
Compliant
Sagittal
Rotary Joint
Active
Thrusting
Force
*[Cham et al. 2000]
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
1
Example: mapping from passive mechanical properties
of insects to biomimetic robot structures
Study biological materials,
components, and their roles in
locomotion.
Study Shape Deposition
Manufacturing (SDM) materials and
components.
viscoelastic
material
Servo
Motor
Roach
leg
stiff material
Displacement Input
Force Output
Models of material behavior and design rules for creating
SDM structures with desired properties
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
2
Example: mapping from passive mechanical properties
of insects to biomimetic robot structures
Study biological materials,
components, and their roles in
locomotion.
6
Force (mN)
4
Study Shape Deposition
Manufacturing (SDM) materials and
components.
Hysteresis loop
@10Hz
Data
Model
2
0
-2
-4
-6
-0.5-0.4-0.3-0.2-0.10 0.10.20.30.40.5
position (mm)
Models of material behavior and design rules for creating
SDM structures with desired properties
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
3
Self-tuning is needed
to adapt to changes
Velocity versus slope for
different stride frequencies
24 deg.
60
Sprawlita running on sloped track
Velocity (cm/s)
Frequency = 11 Hz
40
20
Frequency = 5 Hz
0
-10
0
downhill
slope
10
20
uphill
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
4
Biological approach
• Passive mechanical system and
predominantly feed-forward control
allow animal to run over rough
terrain.
• “Preflexes,” augmented by reflexes
and adaptation, account for changes
in system and environmental
conditions.
• The approach overcomes
limitations of slow neural pathways,
imperfect sensing, etc.
Task
Adaptation
model
FF model
Feed-Forward control
Learning
Reflex control
Command
signal
Mechanical
system
preflex
Sensory feedback
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
Environment
5
Adaptation in small biomimetic robots
• Use preflexes and open-loop
motor control for robust,
stable locomotion.
• Use simple, inexpensive
sensors to detect changes in
operating conditions.
• Use adaptation to tune the
parameters of the open-loop
system.
Adaptation
model
Feed-forward
activation pattern
and timing
Command
input
tripods
time
Mechanical
system
(actuators, limbs)
Contact
Sensor
preflexes
Environment
Locomotion
•No encoders, gyros,
tachometers...
•No tedious calibration
•No fancy filtering
•No sophisticated
closed-loop control.
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
Passive
stabilization
6
Thrust timing for max. height
y
T
Contact Time
Ground
Reaction
Force
Ton
Height
Thrust
Time
y
ttd
tc
tf
tl
maximize:
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
7
Effect of period
for “long thrust”
hopping
x 10
-3
2.5
Natural period:
tn = 0.21 sec
Hop
Height
2
1.5
1
0.5
Thrust magnitude:
F/mg = 1.50
0.15
0.2
0.25
0.3
0.35
0.4
Multiple Solutions
1
Damping: z = 0.20
0.8
Eigenvalues
0.6
0.4
1
2
3
0.2
“normal”
0.15
unstable period-1
hop-settle-fire
0.2
0.25
0.2
0.25
0.3
0.35
0.3
0.35
0.4
0
Velocity at
actuation
-0.5
-1
0.15
1
2
3
0.4
Period (ms)
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
8
Conclusions from 1 DOF model:
• Maximum hop height occurs if thrust is initiated
near maximum compression
• Stability requires thrust initiation before max.
compression.
• For long thrust (vs natural period) thrust should
begin before max. compression and
end essentially at liftoff.
• Therefore, measuring the interval between thrust
deactivation and liftoff is a good indicator of
whether the stride period is tuned correctly.
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
9
Adaptation algorithm
Dtn+1 = Ki - Kp(Td - Tl + Tv)
Lag
“Drift”
Trying to reduce
activation frequency
Td
Deactivation
Time
Tl
Loss of
Contact
Tv Const. offset
between deactivation
and lift-off times
ON
(Tdpiston
- Tl) activation
foot contact
50
0
100
140
180
220
260
300
Period (ms)
OFF
0
20
40
60
80
Gait Period
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
100
120
140
160
180
200
time (ms)
10
Slope adaptation demonstration
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
11
Hopping with variable stiffness
(1)
(2)
fn = sqrt(k/m)/(2*pi)
(3)
Discussion: Blue curve shows typical results when maximum stroke length is constrained.
Maximum period-1 hop height (1) is followed by range of non-period-1 hops (2) and then by
low amplitude, stable period-1 behavior (3).
At frequencies below (1) hopping reverts to “hop-settle-fire.”
J. Karpick 08MAR06
Stryker Interaction Design Workshop
January 2006
September 7-8, 2005
12