grand challenges for autonomous mobile microrobots sarah bergbreiter dr. kris pister berkeley sensor...
Post on 21-Dec-2015
213 views
TRANSCRIPT
Grand Challenges for Autonomous Mobile Microrobots
Sarah Bergbreiter
Dr. Kris PisterBerkeley Sensor and Actuator Center, UC Berkeley
What is an Autonomous Mobile Microrobot?
• Size– Total size on order of millimeters
• Mobility– Should be able to move around a given
environment– Speeds of mm/sec
• Autonomous– Power and control on-board– Communication between robots (?)
Applications for Autonomous Mobile Microrobots
• Mobile Sensor Networks– Monitoring/surveillance– Search and rescue
• Cooperative Construction– Assisted assembly– Sacrificial assembly
Previous Microrobots
Seiko, 1992
Yeh, 1995-2001 Hollar, et al, 2002
Ebefors, et al, 1999
Sandia, 2001
Donald, et al, 2006
COTS Dust (Hill, et al. ACM OS Review 2000)
Making Silicon MoveRemove
Legs
Add Robot Body
Solar Cell Array
CCRs
XLCMOS
IC
Smart Dust (Warneke, et al. Sensors 2002)
1mm
Microrobots (Hollar, Flynn, Pister. MEMS 2002)
1mm
CotsBots (Bergbreiter, Pister. IROS 2003)
How Close Are We?
1995 2005
Sensing Small, but power hungry uW interface electronics
Low power sensors
Computation Clear trends
No COTS uPower
Talk of 1MIP/mW
10 MIP/mW COTS
100 MIP/mW demoed
Mechanisms Lab demos of toys Shipping products
Robust, reliable, …
DRIE, composites, …
Comm Cell phones taking off
WiFi?
Radios were >100mW
<100uA low rate spread-spectrum mesh networking
Power Material properties COTS thin film batteries
Efficient solar cell arrays
Lots of power conversion ICs
Motors Barely able to move themselves Inchworms, polymers
Solar Powered 10mg Silicon Robot
Why Is This So Hard?
1mm
Locomotion
Actuators
Power
Integration
Mechanisms
Challenge 1: Locomotion
1mm
Locomotion
Challenge 1: Locomotion
• Interaction with Environment– Obstacles are large
• Reduce Complexity– Difficult to actuate out of plane– Difficult to fabricate bearings
• Efficiency– Internal v. external work
Locomotion: Jumping
0 10 20 30 40 50-10
-5
0
5
10
15
20
25
30
35
Hopping Trajectory, Mass = 15 mg, Angle = 60 deg
distance (cm)
heig
ht (
cm)
5 uJ10 uJ25 uJ50 uJ
heig
ht (
cm)
distance (cm)
Hopping Trajectory, Mass = 15mg, Angle = 60deg
Locomotion: Comparison
• What time and energy is required to move a microrobot 1 m and what size obstacles can these robots overcome?
Proposed
(Jumping)
Hollar
(Walking)
Ebefors
(Walking)
Alice
(Rolling)
Time 1 min 417 min 2 min, 50 sec 25 sec
Energy 5 mJ 130 mJ 180 J 300 mJ
Obstacle Size 5 cm 50 m 100 m 5 mm
S. Hollar, "A Solar-Powered, Milligram Prototype Robot from a Three-Chip Process," in Mechanical Engineering: University of California, Berkeley, 2003. T. Ebefors, J. U. Mattsson, E. Kalvesten, and G. Stemme, "A walking silicon microrobot," presented at International Conference on Sensors and Actuators (Transducers '99), Sendai, Japan, 1999. http://asl.epfl.ch/index.html?content=research/systems/Alice/alice.php
Locomotion: Comparison
• What time and energy is required to move a microrobot 1 m and what size obstacles can these robots overcome?
A. Lipp, H. Wolf, and F.O. Lehmann., “Walking on inclines: energetics of locomotion in the ant Camponotus," Journal of Experimental Biology 208(4) Feb 2005, 707-19.S. Hollar, "A Solar-Powered, Milligram Prototype Robot from a Three-Chip Process," in Mechanical Engineering: University of California, Berkeley, 2003. T. Ebefors, J. U. Mattsson, E. Kalvesten, and G. Stemme, "A walking silicon microrobot," presented at International Conference on Sensors and Actuators (Transducers '99), Sendai, Japan, 1999. http://asl.epfl.ch/index.html?content=research/systems/Alice/alice.php
Ant (Walking)
Proposed (Jumping)
Hollar (Walking)
Ebefors (Walking)
Alice (Rolling)
Mass 11.9 mg 15 mg 10 mg 80 mg 10 g
Time 15 sec 1 min 417 min 2.8 min 25 sec
Energy 1.5 mJ 5 mJ 130 mJ 180 J 300 mJ
Obstacle Size climbing 1 cm 50 m 100 m 5 mm
Challenge 2: Actuators
1mm
Actuators
Challenge 2: Actuators
• Low Power• Small Size• Force/Displacement• Efficient• Simple Fabrication
and Integration• Power Supply
Compatibility• Robust
Pelrine, 2002
Yeh, 2001 Lindsay, 2001
Kladitis, 2000
Lu, 2003Wood, 2005
Actuators: Electrostatic Inchworm Motors
• High force at low power and moderate voltage
• Accumulate short displacements for long throw
• Fabricated in single mask process
• Hollar inchworm designed for 500 N of force and 256 m of travel in ~ 2.8 mm2
l
+-V d
t
k
F
Electrostatic Inchworm Motor
Challenge 3: Mechanisms
1mm
Mechanisms
Challenge 3: Mechanisms
• Simple Fabrication– Process Complexity– Batch v. Serial
• Efficient– Friction
• Robust
• Matching to ActuatorsWood, et al, 2003
Hollar, et al, 2002
Mechanisms: Silicon
• 2 months in the microlab, but very pretty!
Mechanisms: Assembly• Orthogrippers fabricated in
same process• Parts rotated 90o and
assembled out of plane• Thermal actuators and
rotation stages have been assembledClamp w/o Assembled Part
Clamp w/ Assembled Part
Challenge 4: Power
1mm
Power
Challenge 4: Power
• Small Mass and Volume
• Compatible with Actuators– Any converter
circuitry should be included
• Simple Integration
Nielsen, 2003
Cymbet
Roundy, 2003
Bellew, 2003
Power: Solar Cells
• Use isolation trenches to stack solar cells for higher voltages
• 0.5 – 100V demonstrated
• 10-14% efficiency• Small Size
– Chip area: 3.6 x 1.8 mm2
– Chip mass: 2.3 mg
• Complex Process
Challenge 5: Integration
1mm
Integration
Challenge 5: Integration
• Need to connect all of the pieces– Actuators, control,
power supply, sensors, radio…
• Robust• Compatibility• Serial v. Batch
Last, 2006
Srinivasan, 2001
What Next?
2005 2015
Mobility Tethered walking and autonomous pushups demonstrated
Autonomous
Walking, jumping, hopping, crawling
Actuators Inchworms Higher force, Larger displacements
COTS?
Mechanisms Complex fabrication More interesting materials
Microassembly
Power Solar Cells COTS solar cells
Batteries + Packaging
Integration Wirebonding Automated assembly
Self assembly
Sensing, Comm, Control
All there, but in pieces Integrated with microrobots to create bug networks
Thanks!