Shape Deposition Manufacture of Mesoscale Robotic Devices

1
Shape Deposition Manufacture of Mesoscale Robotic
Devices

• Sean Bailey
• Stanford University
• Pontedera, Italy
• June 11, 2002

2
Roadmap

• Who am I?
• Initial biomimetic robotics work
• Recent animal work
• What I want to doin Pontedera
• Summary

Intro
3
Thanks

• Professor Dario and everyone else here
• Pietro and Lisa
• Mark Cutkosky

Intro
4
Who am I?
Intro
5
Common Themes

• Robotics
• Intelligence
• Animal/artificial
• Structures
• What I am not
• Manufacturing expert
• Controls/AI expert
• Biologist/Neuroscientist

Intro
6
Common ground with other projects

• Next generation of neuroscience and clinical
  tools
• MiTech (aka CRIM)
• Cesares grippers (soft materials, sensors)
• Intestinobot, spinal cord probe
• ARTS
• Neuro-prosthetics (neural interface)
• Sensorized hands
• Percro
• Robotic-human interaction with embedded sensing

Intro
7
Shape Deposition Manufacturing

• Layered manufacturing technology
• Deposit and shape materials
• Better surface finish
• Arbitrary geometries
• Embedded components
• Low requirements for deposition
• Polymers, metals
• Well-suited for mesoscale
• Gap between machining and lithography
• 3D assemblies with features on 1-100 m scale

Early work
8
Biomimetic Inspiration

• Bob Full (UC Berkeley)
• Passive mechanical properties dominate
• Compliance
• Damping
• Sprawled posture
• Feedforward motor signals open loop control

Early work
9
The Sprawl Family

• Biomimetic mapping
• Kinematics
• Compliance
• Built a series of robots
• Sprawl (video)
• Mini-Sprawl (video)
• Sprawlita (video)

Early work
10
Sprawlita
Early work
11
Biomimetic Locomotion?

• Locomotion dynamics
• Energy profiles, ground reaction forces
• Differences, but what does it mean?
• What is the animal trying to regulate while it
  is running?
• What, how fast, and in what way
• Sensor and adaptation schemes for robotics

Early work
12
Animal Experiments

• Paradigm
• Environment-modifying computer in the loop during
  natural behaviors
• Running naturally, perturb, look for changes (in
  actively controlled parameters)
• Why cockroach?
• Excellent runner
• Relatively simple nervous system easy to
  interpret EMG

Watson and Ritzmann, 1998
Cockroach EMG
Recent work
13
Animal Experiments
Kugle setup
Force-displacement actuator
Universal joint (aka candle)
EMG electrodes (50mm silver wire)
Cockroach (Blaberus discoidalis)
Kugle (hollow foam on air bearing)
14
Animal Experiments

• Videos

Recent work
15
Animal Experiments
EMG recording electrodes in extensor tibia

• Videos

Recent work
16
Preliminary Results

• Animal data is hard to interpret
• Running is inherently irregular what is
  steady-state?

EMG
VerticalDisplacement
One Stride
Recent work
17
Preliminary Results

• Changes are apparent
• Quantify? Validate?

Stride Period
Before
Perturbation
After
Recent work
18
Biomimetic Structures

• Amazing actuators
• Passive mechanical properties
• Heavily sensorized

Recent work
19
Biomimetic Structures

• In fact, begs the question of why not just use
  the animal instead of trying to rebuild it
• Makes sense in some cases, but
• Animals have more constant needs
• Engineered devices are more convenient, flexible
  in other ways precision, non-biological tasks,
  etc

Recent work
20
What I want to do while Im here

• Make SDM more 3D currently 2 ½ D
• Shape soft materials
• 3D wiring
• Inter-layer embedded components
• Anything else interesting
• Cool machines focused ion beam
• Neuroscience medical device projects

In Italia
21
Shape Soft Materials

• Right now,
• Shape clear stuff (70D shore)
• Plane white stuff (90A shore)
• Squeegy blue stuff (10A shore)
• Shaping attempts
• Failure during machine
• Poor surface finish
• Cooling shrinks part, and releases
• Ideas
• High speed (material damping)
• Localized freezing (liquid nitrogen jets) and
  global warming

In Italia
22
3D Wiring sheets vs. bundles

• Flexible circuits (sheets, 2 ½ D)
• Limited aspect ratio
• Constant aspect ratio
• Stiff traces
• 3D placement of sensors
• Vias!
• Want more flexibility (bundles, 3D)
• Flexures
• Aspect ratio changes
• Higher density
• Compliant traces

exit point at top of fingertip
In Italia
23
3D Wiring sheets vs. bundles

• Component connections
• Die-packaged ICs
• Ideas
• Electroplating (Localized? Electroless?
  Sputtering?)
• Conductive inks/polymers (in situ SDM circuits)

In Italia
24
Inter-layer Embedded Components

• Cross layer boundaries
• Actuators, sensors
• Wiring
• Fibers (for strength)
• Need protection from
• Encapsulation
• Shaping
• Current methods are hacks
• Thin channels, dams, spacers, hiding, protective
  sacrificial material (wax)
• Ideas
• Selectively removable materials (from
  lithography)
• Selectively permeable materials (sponges)

In Italia
25
Summary

• Interests
• Previous work
• Activities here
• Interest in finding out more about other
  projects

Summary
26
(No Transcript)