Self-Transforming RoboticPlanetary Explorers

S. Dubowsky, MIT Field and Space Robotics LabI. Hunter, MIT Bioinstrumentation LabG. Chirikjian, Johns Hopkins University

In the future, 10 to 40 Years, society will wantrobots to explore the planets, moons, asteroidsand comets of the solar system, build roboticoutposts and ultimately the infrastructure forhuman colonies.

Current Rovers-(2000 c.e.)

Peak HeightsMars: Olympus Mons: 24,000 mEarth: Mount Everest: 8,000+ m

Todays Rovers Cannot:

Traverse Varying Terrain Obstacles

Climb Steep Cliff Faces

Cross Wide Ravines and Canyons

Assemble Structures, For example:Fuel extraction as a precursor to HEDSmissions

The Future

The current basic building blocks of robotic systemsare not up to the challenges of the future.They are: Unreliable Expensive Non-robust

Complex Heavy WeakInefficient Etc...

New Paradigms for the Design of SpaceRobotic Explorers and Workers areNeeded.

Our Vision: A Progression of Self-Transforming Planetary Explorers and Workers

Self-Transforming Explorer/WorkerRobot Concept (2010)

• Network of NodeElements

• Connected by ActiveBinary Elements(ABE’s)

The STX c.2010

A Rover vs. The STX Self TransformingExplorer

The STX Concept

Binary or Digital Robotics -- A KeyElement of STX

• Network of Flexible Memberswith Binary Embedded Actuators- (102 and 10?)

• Lightweight, simple and robust• Fault-Tolerant

G. Chirikjian

Digital Computer Analogy

The Dilemma:NIAC research focuses on problems that may not have solutionsfor 10 to 40 years (2010 to 2040). It would not be expected for aNIAC project to demonstrate practical solutions now.

Our approach:• In simulation, study the projected capabilities and

limitations the system level concepts.• By analysis and experiments bound the expected

capabilities of the component technologies.

The NIAC Research Dilemma

Phase II - Research ResultsOctober 1999 to May 2000

Results Efforts and Results - Outline• Simulation Studies• Component Technologies

Binary Bi-Stable Devices for STXActuator Technologies-Conducting PolymersNear Term Implementations of the TechnologyHyper-DOF Binary Devices - Step Toward CTX

The “Dilemma” Makes Simulations a Key Elementof Our Research.

Some Key Questions:• Can binary systems perform useful task in complex

planetary terrain?

• How many binary degrees-of-freedom are are required toachieve acceptable performance?

• How do you plan the behavior of these systems?

Simulation Studies

A Representative STX Robot Mission

Rough Terrain ExplorerCriteria: Accessibility

SpeedPowerSafety/Recovery, etc.

Material Transportation WorkerCriteria: Payload

SpeedEnergy, etc.

Construction WorkerCriteria: Dexterity

PayloadStrength

Etc…..

Task: An Explorer re-configures into a Cooperative RobotWorker Crew to Construct a Resource Extraction Facility

Initial Studies: Binary Gait (JHU)

JHU is studying• Different configurations of simple legged binary robotic formobility in benign terrain• Methods to plan simple motions (walking, turning, etc) inbenign terrain.

• Gait• Kinematics• Statics• Etc.

Find accessible area for a binaryrobot considering:– Statically stablity– Configurations are within the

the binary actuator ranges ofmotion and effort capabilties.

Initial Studies Rough Terrain AccessibilityStudies-(MIT)

2-D Results

A Key Observation

Binary Systems with Component technologieshaving Large Numbers of Degrees-of Freedom

Approach the Performance of ContinuousSystems.

Compliant Mechanisms,Embedded Actuators andSensors-ABEs

• Large Motions Without Motors,Bearings, Gears, etc.

• Greatly Reduced Number of MovingParts

• Lightweight• Binary Action• Greatly Reduced Number of Sensors

Component Technologies-Mechanisms

Conventional Technologies are Unable to Meet Demands forFuture Planetary Systems: heavy

complexfailure-prone, etc.

STX- Concept need large numbers of actuated bistable degrees-of-freedom. Our concept is to use:

Compliant elementsWith embedded on/off actuatorsInternal detents:

Discrete Binary Actuated Bistable Elements

The result would be a lightweight, simple,robust and fault-tolerant basic building blockfor STX.

Bistable Mechanisms

Eliminates the need forbearings, lubrication

Inherent spring characteristicsprovide bias forces, compliance

Internal detent structure latchesinto discrete states:

– eliminates need to keepactuators powered

– Improve disturbance rejection

Bistable Mechanisms

Eliminates the need forbearings, lubrication

Inherent spring characteristicsprovide bias forces, compliance

Internal detent structure latchesinto discrete states:

– eliminates need to keepactuators powered

– Improve disturbance rejection

Miniature Rotary Joint

• Antagonistic pair of SMA wires• Bistable elements sandwiched by flexure beams• ± 25o deflection

flexure SMAbistableelement

SMA fixture

Pantograph Mechanism

flexures

SMA

• Flexures replacebearings

• SMA actuated• Considerable motion

amplification

SMA are a Surrogate forConducting Polymers

Component Technologies-Conducting Polymers

Conducting Polymers for:• Embedded Muscles

(Actuation )• Sensing• Signal Transmission• Computation

Polypyrrole

Anions (PF6-)

•Low Voltage (0.5-10V)

•High Force (30 Mpa)

•Inexpensive ($1.50/kg)

Conducting Polymers

Cantilever

MPa

Electro -magneticElectro-static

Muscle -cardiacPolymer - gel

Muscle -skeletalPneumatic

Piezo -PolymerHydraulic

Piezo -CeramicMagneto-strictive

Polymer -conductingSMA - NiTi

0.01

0.1 1 10

100

1000

Electro -magneticElectro-static

Muscle -cardiacPolymer - gel

Muscle -skeletalPneumatic

Piezo -PolymerHydraulic

Piezo CeramicMagneto-strictive

Polymer -conductingSMA - NiTi

Achieved

Anticipated

Active Stress

Achieved

Electro-static

Polymer - gel

SMA -

0.01

0.1 1 10

100

1000

kW/kg

Muscle -cardiac

Muscle -skeletal

Piezo Polymer

Piezo - Ceramic

Magneto-strictive

Polymer -conducting

NiTi

Anticipated

Power to M

ass

Active Strain

( % )

Electro -magnetic

Pneumatic

HydraulicMuscle -cardiac

Polymer -gel

Muscle -skeletalElectro-static

SMA - NiTiPolymer -

conductingMagneto-strictivePiezo -

PolymerPiezo -

Ceramic

0.01

0.1 1 10

100

Electro -magnetic

Pneumatic

HydraulicMuscle -cardiac

Polymer -gel

Muscle -skeletalElectro-static

SMA - NiTiPolymer -

conductingMagneto-strictivePiezo -

PolymerPiezo -

Ceramic

Conducting Polymers and TheCTX Vision

• Actuators• Wire (cf Cu)• Transistors• Sensors• Batteries• Super Capacitors• Memory• Light-emitting diodes• Photodetectors & cells• Electrochromic

displays

While “engineering”problems remain to besolved, the results todate suggest theapproach is feasible inthe 10 and 10+ year timeframe.

A Near Term Practical Implementation of BinaryMechanisms

A Reconfigurable Rover Rocker-Bogie Suspension (in cooperationwith JPL)

Objective: To adapt to difficultterrain and improve vehiclestability

Experimental systemimplemented with ShapeMemory Alloys

A Binary Gripper

A Near Term Practical Implementation ofBinary Mechanisms

Structure layout

Physical system

Minimum Channel Binary Controller:

A Near Term Practical Implementation of BinaryMechanisms

A Deployable Rover CameraMount

Hyper-Degree of Freedom Binary Elements

• Motivation: Provide thebuilding blocks for the CTXConcept

• Distributed Flexibility toAchieve Large Motions

• Hyper-Binary DOF throughEmbedded Actuation and Sensing

• N approaches ∞

Analytical Questions:

Finite Element and OptimizationStudies

0123456789

10

Max

imum

Def

lect

ion

(mm

)

Col

umn

Voi

d

Dia

mon

d

Dog

leg

Elbo

w

Deflection Comparison

• How to Shape Structures with Distributed Flexibility for largeDeformations?• How to Place Binary Actuators to Achieve Hyper DOF?• Large Deformations = 100 Time Actuator Deformation.

Prototyping

SMA Based ActuatorExperiments

Experimental Studies

Embedded Sensing

Sensor ExperimentsPolymer based Internal StateSensors for Control.

Adding Embedded Polymer Sensing toEmbedded Binary Polymer Actuators inElastic Polymer Members is a step toward toproof of concept of CTX Systems.

A New Paradigm for Planetary RoboticExplorers and Workers

MIT

The Contributors

MIT-The Field and Space Robotics• Steven Dubowsky• Kate Andrews• Olivier Chronon• Matt Litcher• Vivek SujanMIT - The Bioinstrumentation Lab• Ian Hunter• John MaddenJohns Hopkins University• Greg Chirijkan• Zhou YuJPL - Technical Assistance• Paul Schenker