Upload
clifford-stone
View
224
Download
0
Embed Size (px)
Citation preview
8/11/2019 Kinema Tic Cellular Automata
1/48
Modeling Kinematic Cellular Automata:Modeling Kinematic Cellular Automata:
An Approach to SelfAn Approach to Self--
ReplicationReplication
Principal Investigator:Principal Investigator: Consultants:Consultants:
Tihamer TothTihamer Toth--FejelFejel RobertRobert FreitasFreitasTihamer.TothTihamer.Toth--Fejel@[email protected] Matt MosesMatt Moses
March 22, 2004 NIAC Fellows Presentation
NASA Institute for AdvancedNASA Institute for Advanced
ConceptsConcepts
Phase I: CPPhase I: CP--0202--0202
March 22, 2004 NIAC Fellows Presentation
8/11/2019 Kinema Tic Cellular Automata
2/48
Modeling Kinematic CellularModeling Kinematic Cellular
AutomataAutomata RationaleRationale
BenefitsBenefits
ApplicationsApplications
Project GoalsProject Goals
StrategyStrategy
AccomplishmentsAccomplishments
Conclusion and FutureConclusion and Future
DirectionsDirectionsAdditional MaterialAdditional Material
8/11/2019 Kinema Tic Cellular Automata
3/48
RationaleRationale
Why SelfWhy Self--Replication?Replication?
Why not SelfWhy not Self--Assembly?Assembly?
Why Kinematic CellularWhy Kinematic Cellular
Automata?Automata?
Why both macro andWhy both macro and nanonano scale?scale?
8/11/2019 Kinema Tic Cellular Automata
4/48
8/11/2019 Kinema Tic Cellular Automata
5/48
Rationale: Why not SelfRationale: Why not Self--
Assembly?Assembly?Examples have been demonstratedExamples have been demonstrated
ButBut
Not Genotype +Not Genotype + RibotypeRibotype = Phenotype (GRP)= Phenotype (GRP)
No theoryNo theory
Against the principles of sound designAgainst the principles of sound design
HoweverHowever
Use it for simple input partsUse it for simple input parts
8/11/2019 Kinema Tic Cellular Automata
6/48
8/11/2019 Kinema Tic Cellular Automata
7/48
Rationale:Rationale:
Why Both Macro andWhy Both Macro and NanoNano Scale?Scale?Abstract designAbstract design
Macro:Macro:Possible with current technologyPossible with current technology
Useful productsUseful productsProof of concept in short termProof of concept in short term
NanoNano::Quality of atoms (and molecules)Quality of atoms (and molecules)
SelfSelf--assembled input parts possibleassembled input parts possible
Significant financial payoffSignificant financial payoff
Ei l IBM Ad
8/11/2019 Kinema Tic Cellular Automata
8/48
Benefit: Cost Reduction/LbBenefit: Cost Reduction/Lb
1
104
1 104 108 1012 1016 1020 1024
102
106
108
Wrench
Pentium
Potato
Lobster
Eiglers IBM Ad
Digital
Watch
Car
YogurtSalt
KCA SRS
Complexity (Parts and Interactions/lb)
Co
st($/lb)
SevenSeven
Magnitudes!Magnitudes!
GAIN
Traditional Top Down Manufacturing vs Bottom-up Molecular Replication
8/11/2019 Kinema Tic Cellular Automata
9/48
Benefit: ProgrammableBenefit: Programmable
MaterialsMaterialsSimple identical modulesSimple identical modules Flow ModeFlow Mode
PixelatedPixelated ModeMode Logic Processing ModeLogic Processing Mode
Flow ModeFlow Mode PixelatedPixelated ModeMode
8/11/2019 Kinema Tic Cellular Automata
10/48
Application: SpaceApplication: Space
ExplorationExploration
RobustRobust HyperflexibleHyperflexible
Resource UtilizationResource Utilization
Lower launch weightLower launch weight
ExpandableExpandable
TerraformingTerraforming
Politically feasiblePolitically feasible
Opens new frontierOpens new frontier
8/11/2019 Kinema Tic Cellular Automata
11/48
Project GoalsProject Goals
Characterize selfCharacterize self--replicationreplication
Quantify the complexity of SelfQuantify the complexity of Self--Replicating System (SRS) madeReplicating System (SRS) made
of Kinematic Cellular Automataof Kinematic Cellular Automata(KCA)(KCA)
Confirm approachConfirm approach Design a KCA SRSDesign a KCA SRS
Simulate designsSimulate designs
8/11/2019 Kinema Tic Cellular Automata
12/48
Project StrategyProject Strategy
Hybridize two selfHybridize two self--replicationreplication
modelsmodels Keep it simpleKeep it simple
Make it complicatedMake it complicated Refine approachRefine approach
Attempt designAttempt design
Imitate computersImitate computers
Imitate biologyImitate biology
8/11/2019 Kinema Tic Cellular Automata
13/48
AccomplishmentsAccomplishmentsGoalsGoals AccomplishmentsAccomplishments
CharacterizeCharacterizeunexplored areaunexplored area
Explored MultiExplored Multi --Dimensional SpaceDimensional Space
Quantify theQuantify the
difficultydifficultyNot trivial, but less than a PentiumNot trivial, but less than a Pentium
Confirm or refuteConfirm or refuteapproachapproach
Refined ApproachRefined Approach Useful SRSUseful SRS
Hierarchy of Subsystems, Cells, Facets, & PartsHierarchy of Subsystems, Cells, Facets, & Parts
Transporter, Assembler, & ControllerTransporter, Assembler, & Control ler
LowLow--level simpler than highlevel simpler than high--levellevel TopTop--DownDown vsvs BottomBottom--UpUp
SelfSelf--Assembly for input PartsAssembly for input Parts
Standard conceptsStandard concepts
Universal Constructor is approach, not goalUniversal Constructor is approach, not goal
Design a KCA SRSDesign a KCA SRS Developed RequirementsDeveloped RequirementsPreliminary DesignPreliminary Design
Simulate designsSimulate designs Modeled SimulationsModeled Simulations Sensor PositionSensor Position
NAND gate and opNAND gate and op--amp selfamp self--assemblyassembly FacetFacet
Transporter and AssemblerTransporter and Assembler
Characteri ing SelfCharacterizing Self Replication Adj stingReplication: Adjusting
8/11/2019 Kinema Tic Cellular Automata
14/48
Characterizing SelfCharacterizing Self--Replication: AdjustingReplication: Adjusting
thethe Freitas/MerkleFreitas/Merkle 116116--Dimension DesignDimension Design
SpaceSpace
Evolvability
Replicator
Information
Replicator
Energetics
Replication
Process
Product
Structure
Replicator
Performance
Replicator
Substrate
Replicator
Control
Manipulation
Autonomy
Manipulation
Redundancy
Manipulation
Centralization
Manipulation
Degrees of Freedom
Positional
AccuracyQuantity of Onboard
Energy TypesQuantity of
Manipulation
Types
Assembly
Style
Parts
Count
Parts
Scale
Parts
TypesQuantity
of Vitamin
PartsNutritional
Complexity
Parts
Preparation
Parts
ComplexityParts
Precision
MulticellularitySubunit
Scale
Subunit
Types
Subunit
Complexity
Replicator
Kinematics
Active
Subunits
Replicator
Parts
All
Replicators
Replicator
Structure
Subunit
Hierarchy
ReplicatorDesignability
Subunit
Complexity
Monotonicity
8/11/2019 Kinema Tic Cellular Automata
15/48
Quantifying Difficulty of SRS DesignQuantifying Difficulty of SRS Design
1.00E+00
1.00E+011.00E+02
1.00E+03
1.00E+04
1.00E+051.00E+06
1.00E+07
1.00E+081.00E+09
Units of
difficulty
# parts
+
# interactions
Wrench Automobile Pentium KCA SRS
8/11/2019 Kinema Tic Cellular Automata
16/48
8/11/2019 Kinema Tic Cellular Automata
17/48
Original Approach: FeynmanOriginal Approach: Feynman
methodmethod
1.1. Start with trivial selfStart with trivial self--replicationreplication
2.2. Move the complexity outMove the complexity out
of the environment andof the environment andinto the SRS byinto the SRS bydoubling parts count ofdoubling parts count ofthe component (Trivialthe component (Trivial+1+1
case)case)
3.3. ReiterateReiterate
Original Approach: FeynmanOriginal Approach: Feynman
8/11/2019 Kinema Tic Cellular Automata
18/48
Original Approach: FeynmanOriginal Approach: Feynman
methodmethod Plenty of room at the bottom , topPlenty of room at the bottom , top--down, fractaldown, fractal
approachapproach
Mystery
8/11/2019 Kinema Tic Cellular Automata
19/48
Refine Approach (by 180Refine Approach (by 180))
We should start at theWe should start at thebottom level and workbottom level and work upup
Imitate Mother NatureImitate Mother Nature
The Trivial+2 case hasThe Trivial+2 case has
already been donealready been done
Molecules
8/11/2019 Kinema Tic Cellular Automata
20/48
The BottomThe Bottom--up Approachup Approach
WellWell --ordered environment,ordered environment,Simple inertSimple inert partsparts
SymmetricSymmetric facetsfacetsModularModular cellscells
Assembler, Transporter, and ControllerAssembler, Transporter, and Controllersubsystemssubsystems
SelfSelf--ReplicatingReplicating SystemSystem
8/11/2019 Kinema Tic Cellular Automata
21/48
Subsystem RequirementsSubsystem Requirements
If atoms are analogous to bits,If atoms are analogous to bits,
then:then:Memory/BusMemory/Bus ----> Transporter> Transporter
Moves PartsMoves PartsALUALU ----> Assembler> Assembler
Connects PartsConnects Parts
ControlControl ----> Controller> Controller
Decides which Parts go whereDecides which Parts go where
StandardizedStandardized
8/11/2019 Kinema Tic Cellular Automata
22/48
8/11/2019 Kinema Tic Cellular Automata
23/48
Assembler SubsystemAssembler Subsystem
(light blue preparation tool)
(yellow edge structural part)
(pink corner structural part)
8/11/2019 Kinema Tic Cellular Automata
24/48
Controller SubsystemController Subsystem
8/11/2019 Kinema Tic Cellular Automata
25/48
Cell Design RequirementsCell Design Requirements
Structure:Structure: Lock, 1Lock, 1--D slide, disconnectD slide, disconnect
Actuators:Actuators: TransformTransform
MoveMove
Sensors:Sensors: Detect PositionDetect Position
Transmit messagesTransmit messages
Logic:Logic: Decode messagesDecode messages
Accept, store, forward messagesAccept, store, forward messages
Activate commandsActivate commands
8/11/2019 Kinema Tic Cellular Automata
26/48
Unit CellUnit Cell
(center structure, motors, sensors, and tabs omitted)
8/11/2019 Kinema Tic Cellular Automata
27/48
Facet Design RequirementsFacet Design Requirements Structure:Structure: Insert or retractInsert or retract
Actuators:Actuators:
TransformTransformMove tabsMove tabs
Sensors:Sensors:TransformTransform
Logic:Logic:
DecodeDecode
8/11/2019 Kinema Tic Cellular Automata
28/48
Unit FacetsUnit Facets
8/11/2019 Kinema Tic Cellular Automata
29/48
Parts Design RequirementsParts Design Requirements Structure:Structure:
SolidSolid Motors:Motors:
RotaryRotary
LinearLinear IMPCIMPC
Sensors:Sensors: Translate signalsTranslate signals Detect parts positionDetect parts position
LogicLogic Activate messages to motorsActivate messages to motors
Aggregate digital logicAggregate digital logic
Parts: Structure Sensors &Parts: Structure Sensors &
8/11/2019 Kinema Tic Cellular Automata
30/48
Parts: Structure, Sensors &Parts: Structure, Sensors &
SolenoidsSolenoids
8/11/2019 Kinema Tic Cellular Automata
31/48
Software SimulationSoftware Simulation
Sensor Position Simulation ToolSensor Position Simulation Tool NAND gate & opNAND gate & op--amp Selfamp Self--AssemblyAssembly
ToolTool
Facet AnimationFacet Animation
Transporter and Assembler SimulationTransporter and Assembler Simulation
8/11/2019 Kinema Tic Cellular Automata
32/48
Position Sensor SimulationPosition Sensor Simulation
SelfSelf assembly ofassembly of
8/11/2019 Kinema Tic Cellular Automata
33/48
SelfSelf--assembly ofassembly of
NAND gate and opNAND gate and op--ampamp
8/11/2019 Kinema Tic Cellular Automata
34/48
Facet AnimationFacet Animation
Simulation ofSimulation of
8/11/2019 Kinema Tic Cellular Automata
35/48
Simulation ofSimulation of
Transporter and AssemblerTransporter and Assembler
Conclusion and FutureConclusion and Future
8/11/2019 Kinema Tic Cellular Automata
36/48
Conclusion and FutureConclusion and Future
DirectionsDirectionsNo roadblocks!No roadblocks!
Final Design for macro physical prototypesFinal Design for macro physical prototypes
Build physical prototypesBuild physical prototypes
Build and run small cell collectionsBuild and run small cell collections
Build and run subsystemsBuild and run subsystems
Build macro scale SRSBuild macro scale SRS Write Place and Route softwareWrite Place and Route software
Refine concept atRefine concept at nanonano scalescale
8/11/2019 Kinema Tic Cellular Automata
37/48
AcknowledgementsAcknowledgements NASA Institute for Advanced ConceptsNASA Institute for Advanced Concepts
JohnJohn SauterSauterAltarumAltarum
Rick Berthiaume, Ed Waltz, Ken Augustyn, andRick Berthiaume, Ed Waltz, Ken Augustyn, and
Sherwood SpringSherwood SpringGeneral Dynamics AISGeneral Dynamics AIS John McMillan and Teresa MacaulayJohn McMillan and Teresa Macaulay
Wise SolutionsWise Solutions
Forrest BishopForrest Bishop
Institute of AtomicInstitute of Atomic--Scale EngineeringScale Engineering
Joseph MichaelJoseph MichaelFractal Robots, Ltd.Fractal Robots, Ltd.
8/11/2019 Kinema Tic Cellular Automata
38/48
Additional MaterialAdditional Material
AssumptionsAssumptions
Previous and Related WorkPrevious and Related Work
C S S
8/11/2019 Kinema Tic Cellular Automata
39/48
KCA SRS AssumptionsKCA SRS Assumptions
Parts supplied as automated cartridgesParts supplied as automated cartridges Low rate of errors detected in codeLow rate of errors detected in code
8/11/2019 Kinema Tic Cellular Automata
40/48
Previous and Related WorkPrevious and Related Work FreitasFreitas and Longand Long -- NASA Summer Study:NASA Summer Study:
Advanced Automation for Space Missions (1980)Advanced Automation for Space Missions (1980) MichaelMichael -- Fractal RobotsFractal Robots
ChirikjianChirikjian andand SuthakornSuthakorn --Autonomous RobotsAutonomous Robots
MosesMoses -- Universal Constructor PrototypeUniversal Constructor Prototype
ZyvexZyvex -- Exponential AssemblersExponential Assemblers
FreitasFreitas andand MerkleMerkle -- Kinematic SelfKinematic Self--ReplicatingReplicatingMachines (2004)Machines (2004)
8/11/2019 Kinema Tic Cellular Automata
41/48
Previous Work:NASA Summer StudyPrevious Work:NASA Summer StudyAdvanced Automation for SpaceAdvanced Automation for Space
MissionsMissions -- Freitas and LongFreitas and Long --(1980)(1980)
StrengthsStrengths
First major work since 1950sFirst major work since 1950s
Cooperation of manyCooperation of manyvisionariesvisionaries
WeaknessesWeaknesses
short, no followshort, no follow--upup paper study onlypaper study only
prepre--PC technology
http://www.islandone.org/MMSG/aasm/FOR MORE INFO...
PC technology
P i W k J hP i W k J h
8/11/2019 Kinema Tic Cellular Automata
42/48
Previous Work: JosephPrevious Work: Joseph
MichaelMichael StrengthsStrengths
The DOS of Util ity Fog The DOS of Util ity Fog
Working macro modularWorking macro modularrobotsrobots
Limited DOF = betterLimited DOF = betterstructurestructure
WeaknessesWeaknesses
Fractals just push problem toFractals just push problem tolower, lesslower, less--accessible levelaccessible level
no detailed methodology forno detailed methodology for
selfself
--replicationreplication
http://www.fractal-robots.com/
FOR MORE INFO...
Previous Work: ForrestPrevious Work: Forrest
8/11/2019 Kinema Tic Cellular Automata
43/48
Previous Work: Forreste ous o o est
BishopBishop StrengthsStrengths
Very Limited DOFVery Limited DOF
Clear macro designClear macro design WeaknessesWeaknesses
NanoscaleNanoscaleimplementation clearlyimplementation clearly
implied, but not clearlyimplied, but not clearlydesigneddesigned
no detailed methodologyno detailed methodologyfor selffor self--replicationreplication
http://www.iase.cc/html/overtool.htm
FOR MORE INFO...
R l t d W kR l t d W k
8/11/2019 Kinema Tic Cellular Automata
44/48
Related Work:Related Work:
Chirikjian/SuthakornChirikjian/Suthakorn
StrengthsStrengths Autonomous implementationAutonomous implementation
of Trivialof Trivial+2+2 case (4 parts)case (4 parts)
Directed towardsDirected towardsextraterrestrial applicationsextraterrestrial applications
Lego isomorphic withLego isomorphic withmoleculesmolecules
WeaknessesWeaknesses
Small UC envelopeSmall UC envelope Depends on nonDepends on non--replicatingreplicating
jigsjigs
High entropy environmentHigh entropy environment
limits extension to Triviallimits extension to Trivial+3+3
casecasehttp://http://caesar.me.jhu.edu/research/self_replicating.htmlcaesar.me.jhu.edu/research/self_replicating.html
FOR MORE INFO...
R l t d W kR l t d W k ZZ
8/11/2019 Kinema Tic Cellular Automata
45/48
Related Work:Related Work: ZyvexZyvex
ProjectsProjects Applying MEMS andApplying MEMS and nanotubesnanotubes Parallel Micro and ExponentialParallel Micro and Exponential
AssemblyAssembly
StrengthsStrengths First and only funded company tryingFirst and only funded company trying
to build ato build a DrexlerianDrexlerian assemblerassembler
WeaknessesWeaknesses MEMS is 1000X too bigMEMS is 1000X too big
surfaces too roughsurfaces too rough
Exponential Assembly is machine selfExponential Assembly is machine self--assembly (not Universal Constructor;assembly (not Universal Constructor;not GRP paradigm; not Util ity Fog)not GRP paradigm; not Util ity Fog)
FOR MORE INFO...
http://http://www.zyvex.comwww.zyvex.com//
8/11/2019 Kinema Tic Cellular Automata
46/48
Related Work:Related Work: Freitas/MerkleFreitas/MerkleKinematic SelfKinematic Self--Replicating MachinesReplicating Machines
((LandesLandes Bioscience, 2004)Bioscience, 2004)First comprehensive review of fieldFirst comprehensive review of field
1.1. The Concept of SelfThe Concept of Self--ReplicatingReplicating
MachinesMachines2.2. Classical Theory of Machine ReplicationClassical Theory of Machine Replication
3.3. MacroscaleMacroscale Kinematic MachineKinematic MachineReplicatorsReplicators
4.4. MicroscaleMicroscale and Molecular Kinematicand Molecular KinematicMachine ReplicatorsMachine Replicators
5.5. Issues in Kinematic Machine ReplicationIssues in Kinematic Machine ReplicationEngineeringEngineering
6.6. Motivations for MolecularMotivations for Molecular--Scale MachineScale MachineReplicator DesignReplicator Design(c) 2004 Robert Freitas and Ralph Merkle
Freitas is a technical consultant for this project
Related Work: Matt MosesRelated Work: Matt Moses
8/11/2019 Kinema Tic Cellular Automata
47/48
Related Work: Matt MosesRelated Work: Matt Moses
Strengths:
CAD to physical
implementation
Large envelope UC
Weaknesses:
Strain bending under load
Manual control
Moses is a technical consultant for this project
8/11/2019 Kinema Tic Cellular Automata
48/48