Modeling Kinematic Cellular Automata:

Modeling Kinematic Cellular Automata:

An Approach to Self
An Approach to Self
-
-
Replication
Replication
Principal Investigator: Principal Investigator: Consultants: Consultants:
Tihamer Toth Tihamer Toth- -Fejel Fejel Robert Robert Freitas Freitas
Tihamer.Toth Tihamer.Toth- -Fejel@gd Fejel@gd- -ais.com ais.com Matt Moses Matt Moses
March 22, 2004 NIAC Fellows Presentation
NASA Institute for Advanced
NASA Institute for Advanced
Concepts
Concepts
Phase I: CP
Phase I: CP
-
-
02
02
-
-
02
02
March 22, 2004 NIAC Fellows Presentation
Modeling Kinematic Cellular
Modeling Kinematic Cellular
Automata
Automata

Rationale
Rationale

Benefits
Benefits

Applications
Applications

Project Goals
Project Goals

Strategy
Strategy

Accomplishments
Accomplishments

Conclusion and Future

Conclusion and Future
Directions
Directions

Additional Material
Additional Material
Rationale
Rationale

Why Self
Why Self
-
-
Replication?
Replication?

Why not Self

Why not Self
-
-
Assembly?
Assembly?

Why Kinematic Cellular

Why Kinematic Cellular
Automata?
Automata?

Why both macro and

Why both macro and
nano
nano
scale?
scale?
Rationale: Why Self
Rationale: Why Self
-
-
Replication?
Replication?

Revolutionary manufacturing
Revolutionary manufacturing
process
process

Nanotechnology
Nanotechnology

Massive reduction in costs per

Massive reduction in costs per
pound
pound

Controlled exponential growth

Controlled exponential growth
Rationale: Why not Self
Rationale: Why not Self
-
-
Assembly?
Assembly?
Examples have been demonstrated
Examples have been demonstrated
But
But

Not Genotype +
Not Genotype +
Ribotype
Ribotype
= Phenotype (GRP)
= Phenotype (GRP)

No theory
No theory

Against the principles of sound design
Against the principles of sound design
However
However
Use it for simple input parts
Use it for simple input parts
Rationale:
Rationale:
Why Kinematic Cellular Automata
Why Kinematic Cellular Automata
(KCA)?
(KCA)?

Combines Von Neumanns two

Combines Von Neumanns two
designs
designs

Increased flexibility
Increased flexibility

Decreased complexity
Decreased complexity

Large system work envelope

Large system work envelope

Sometimes better than smart dust

Sometimes better than smart dust
Rationale:
Rationale:
Why Both Macro and
Why Both Macro and
Nano
Nano
Scale?
Scale?

Abstract design
Abstract design

Macro:
Macro:

Possible with current technology

Possible with current technology

Useful products
Useful products

Proof of concept in short term

Proof of concept in short term

Nano
Nano
:
:

Quality of atoms (and molecules)

Quality of atoms (and molecules)

Self
Self
-
-
assembled input parts possible
assembled input parts possible

Significant financial payoff

Significant financial payoff
Benefit: Cost Reduction/Lb
Benefit: Cost Reduction/Lb
1
10
4
1 10
4
10
8
10
12
10
16
10
20
10
24
10
2
10
6
10
8
Wrench
Pentium
Potato
Lobster
Eiglers IBM Ad
Digital
Watch
Car
Yogurt
Salt
KCA SRS
Complexity (Parts and Interactions/lb)
C
o
s
t

(
$
/
l
b
)
Seven Seven
Magnitudes! Magnitudes!
G
A
I
N
Traditional Top Down Manufacturing vs Bottom-up Molecular Replication
Benefit: Programmable
Benefit: Programmable
Materials
Materials
Simple identical modules
Simple identical modules

Flow Mode
Flow Mode

Pixelated
Pixelated
Mode
Mode

Logic Processing Mode
Logic Processing Mode
Flow Mode Flow Mode
Pixelated Pixelated Mode Mode
Application: Space
Application: Space

Exploration
Exploration

Robust
Robust

Hyperflexible
Hyperflexible

Resource Utilization
Resource Utilization

Lower launch weight
Lower launch weight

Expandable
Expandable

Terraforming
Terraforming

Politically feasible
Politically feasible

Opens new frontier
Opens new frontier
Project Goals
Project Goals

Characterize self
Characterize self
-
-
replication
replication

Quantify the complexity of Self

Quantify the complexity of Self
-
-
Replicating System (SRS) made
Replicating System (SRS) made
of Kinematic Cellular Automata
of Kinematic Cellular Automata
(KCA)
(KCA)

Confirm approach
Confirm approach

Design a KCA SRS

Design a KCA SRS

Simulate designs
Simulate designs
Project Strategy
Project Strategy

Hybridize two self

Hybridize two self
-
-
replication
replication
models
models

Keep it simple
Keep it simple

Make it complicated
Make it complicated

Refine approach
Refine approach

Attempt design
Attempt design

Imitate computers
Imitate computers

Imitate biology
Imitate biology
Accomplishments
Accomplishments
Goals Goals Accomplishments Accomplishments
Characterize Characterize
unexplored area unexplored area
Explored Multi Explored Multi - -Dimensional Space Dimensional Space
Quantify the Quantify the
difficulty difficulty
Not trivial, but less than a Pentium Not trivial, but less than a Pentium
Confirm or refute Confirm or refute
approach approach
Refined Approach Refined Approach
Useful SRS Useful SRS
Hierarchy of Subsystems, Cells, Facets, & Parts Hierarchy of Subsystems, Cells, Facets, & Parts
Transporter, Assembler, & Controller Transporter, Assembler, & Controller
Low Low- -level simpler than high level simpler than high- -level level
Top Top- -Down Down vs vs Bottom Bottom- -Up Up
Self Self - -Assembly for input Parts Assembly for input Parts
Standard concepts Standard concepts
Universal Constructor is approach, not goal Universal Constructor is approach, not goal
Design a KCA SRS Design a KCA SRS
Developed Requirements Developed Requirements
Preliminary Design Preliminary Design
Simulate designs Simulate designs
Modeled Simulations Modeled Simulations
Sensor Position Sensor Position
NAND gate and op NAND gate and op- -amp self amp self - -assembly assembly
Facet Facet
Transporter and Assembler Transporter and Assembler
Characterizing Self
Characterizing Self
-
-
Replication: Adjusting
Replication: Adjusting
the
the
Freitas/Merkle
Freitas/Merkle
116
116
-
-
Dimension Design
Dimension Design
Space
Space
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
Accuracy
Quantity of Onboard
Energy Types
Quantity of
Manipulation
Types
Assembly
Style
Parts
Count
Parts
Scale
Parts
Types
Quantity
of Vitamin
Parts
Nutritional
Complexity
Parts
Preparation
Parts
Complexity
Parts
Precision
Multicellularity
Subunit
Scale
Subunit
Types
Subunit
Complexity
Replicator
Kinematics
Active
Subunits
Replicator
Parts
All
Replicators
Replicator
Structure
Subunit
Hierarchy
Replicator
Designability
Subunit
Complexity
Monotonicity
Quantifying Difficulty of SRS Design
Quantifying Difficulty of SRS Design
1.00E+00
1.00E+01
1.00E+02
1.00E+03
1.00E+04
1.00E+05
1.00E+06
1.00E+07
1.00E+08
1.00E+09
Units of
difficulty
# parts
+
# interactions
Wrench Automobile Pentium KCA SRS
Hierarchy
Hierarchy
Computer
Biology
KCA SRS
Horse
Self-replicating System:
Useful
Processor
Brain and
Muscles
Subsystems:
Transporter, Assembler, and Controller
Bus/Memory, ALU,
and Controller
Cells
Cells: Cubic devices with only three
limited degrees of freedom
Organelles
Facets: Symmetrical implementation
Finite State
Machines,
Shift Registers,
Adders, and
Multiplexers
Proteins
Parts: Inert, Simpler than higher levels
NAND gates
Genes
Self-assembling Subparts:
Wires, Transistors, actuator components
Transistors, Wires
Molecules
Molecules
Molecules
Original Approach: Feynman
Original Approach: Feynman
method
method
1.
1.
Start with trivial self
Start with trivial self
-
-
replication
replication
2.
2.
Move the complexity out
Move the complexity out
of the environment and
of the environment and
into the SRS by
into the SRS by
doubling parts count of
doubling parts count of
the component (Trivial
the component (Trivial
+1 +1
case)
case)
3.
3.
Reiterate
Reiterate
Original Approach: Feynman
Original Approach: Feynman
method
method

Plenty of room at the bottom , top

Plenty of room at the bottom , top
-
-
down, fractal
down, fractal
approach
approach
Mystery
Refine Approach (by 180
Refine Approach (by 180

)
)

We should start at the

We should start at the
bottom level and work
bottom level and work
up
up

Imitate Mother Nature

Imitate Mother Nature

The Trivial+2 case has

The Trivial+2 case has
already been done
already been done
Molecules
The Bottom
The Bottom
-
-
up Approach
up Approach
Well
Well
-
-
ordered environment,
ordered environment,
Simple inert
Simple inert
parts
parts
Symmetric
Symmetric
facets
facets
Modular
Modular
cells
cells
Assembler, Transporter, and Controller
Assembler, Transporter, and Controller
subsystems
subsystems
Self
Self
-
-
Replicating
Replicating
System
System
Subsystem Requirements
Subsystem Requirements
If atoms are analogous to bits,
If atoms are analogous to bits,
then:
then:

Memory/Bus
Memory/Bus
--
--
> Transporter
> Transporter

Moves Parts
Moves Parts

ALU
ALU
--
--
> Assembler
> Assembler

Connects Parts
Connects Parts

Control
Control
--
--
> Controller
> Controller

Decides which Parts go where
Decides which Parts go where

Standardized
Standardized
Transporter Subsystem
Transporter Subsystem
(pink corner structural part)
Assembler Subsystem
Assembler Subsystem
(light blue preparation tool)
(yellow edge structural part)
(pink corner structural part)
Controller Subsystem
Controller Subsystem
Cell Design Requirements
Cell Design Requirements

Structure:
Structure:

Lock, 1
Lock, 1
-
-
D slide, disconnect
D slide, disconnect

Actuators:
Actuators:

Transform
Transform

Move
Move

Sensors:
Sensors:

Detect Position
Detect Position

Transmit messages
Transmit messages

Logic:
Logic:

Decode messages
Decode messages

Accept, store, forward messages
Accept, store, forward messages

Activate commands
Activate commands
Unit Cell
Unit Cell
(center structure, motors, sensors, and tabs omitted)
Facet Design Requirements
Facet Design Requirements

Structure:
Structure:

Insert or retract
Insert or retract

Actuators:
Actuators:

Transform
Transform

Move tabs
Move tabs

Sensors:
Sensors:

Transform
Transform

Logic:
Logic:

Decode
Decode
Unit Facets
Unit Facets
Parts Design Requirements
Parts Design Requirements

Structure:
Structure:

Solid
Solid

Motors:
Motors:

Rotary
Rotary

Linear
Linear

IMPC
IMPC

Sensors:
Sensors:

Translate signals
Translate signals

Detect parts position
Detect parts position

Logic
Logic

Activate messages to motors
Activate messages to motors

Aggregate digital logic
Aggregate digital logic
Parts: Structure, Sensors &
Parts: Structure, Sensors &
Solenoids
Solenoids
Software Simulation
Software Simulation

Sensor Position Simulation Tool

Sensor Position Simulation Tool

NAND gate & op

NAND gate & op
-
-
amp Self
amp Self
-
-
Assembly
Assembly
Tool
Tool

Facet Animation
Facet Animation

Transporter and Assembler Simulation

Transporter and Assembler Simulation
Position Sensor Simulation
Position Sensor Simulation
Self
Self
-
-
assembly of
assembly of
NAND gate and op
NAND gate and op
-
-
amp
amp
Facet Animation
Facet Animation
Simulation of
Simulation of
Transporter and Assembler
Transporter and Assembler
Conclusion and Future
Conclusion and Future
Directions
Directions
No roadblocks!
No roadblocks!

Final Design for macro physical prototypes
Final Design for macro physical prototypes

Build physical prototypes
Build physical prototypes

Build and run small cell collections
Build and run small cell collections

Build and run subsystems
Build and run subsystems

Build macro scale SRS
Build macro scale SRS

Write Place and Route software
Write Place and Route software

Refine concept at
Refine concept at
nano
nano
scale
scale
Acknowledgements
Acknowledgements

NASA Institute for Advanced Concepts
NASA Institute for Advanced Concepts

J ohn
J ohn
Sauter
Sauter

Altarum
Altarum

Rick Berthiaume, Ed Waltz, Ken Augustyn, and
Rick Berthiaume, Ed Waltz, Ken Augustyn, and
Sherwood Spring
Sherwood Spring

General Dynamics AIS

General Dynamics AIS

J ohn McMillan and Teresa Macaulay
J ohn McMillan and Teresa Macaulay

Wise Solutions
Wise Solutions

Forrest Bishop
Forrest Bishop

Institute of Atomic
Institute of Atomic
-
-
Scale Engineering
Scale Engineering

J oseph Michael
J oseph Michael

Fractal Robots, Ltd.

Fractal Robots, Ltd.
Additional Material
Additional Material

Assumptions
Assumptions

Previous and Related Work

Previous and Related Work
KCA SRS Assumptions
KCA SRS Assumptions

Parts supplied as automated cartridges

Parts supplied as automated cartridges

Low rate of errors detected in code

Low rate of errors detected in code
Previous and Related Work
Previous and Related Work

Freitas
Freitas
and Long
and Long
-
-
NASA Summer Study:
NASA Summer Study:
Advanced Automation for Space Missions (1980)
Advanced Automation for Space Missions (1980)

Michael
Michael
-
-
Fractal Robots
Fractal Robots

Chirikjian
Chirikjian
and
and
Suthakorn
Suthakorn
-
-
Autonomous Robots
Autonomous Robots

Moses
Moses
-
-
Universal Constructor Prototype
Universal Constructor Prototype

Zyvex
Zyvex
-
-
Exponential Assemblers
Exponential Assemblers

Freitas
Freitas
and
and
Merkle
Merkle
-
-
Kinematic Self
Kinematic Self
-
-
Replicating
Replicating
Machines (2004)
Machines (2004)
Previous Work:NASA Summer Study
Previous Work:NASA Summer Study
Advanced Automation for Space
Advanced Automation for Space
Missions
Missions
-
-
Freitas and Long
Freitas and Long
-
-
(1980)
(1980)

Strengths
Strengths

First major work since 1950s
First major work since 1950s

Cooperation of many
Cooperation of many
visionaries
visionaries

Weaknesses
Weaknesses

short, no follow
short, no follow
-
-
up
up

paper study only
paper study only

pre
pre
-
-
PC technology
http://www.islandone.org/MMSG/aasm/
FOR MORE INFO...
PC technology
Previous Work: Joseph
Previous Work: Joseph
Michael
Michael

Strengths
Strengths

The DOS of Utility Fog
The DOS of Utility Fog

Working macro modular
Working macro modular
robots
robots

Limited DOF = better
Limited DOF = better
structure
structure

Weaknesses
Weaknesses

Fractals just push problem to
Fractals just push problem to
lower, less
lower, less
-
-
accessible level
accessible level

no detailed methodology for
no detailed methodology for
self
self
-
-
replication
replication
http://www.fractal-robots.com/
FOR MORE INFO...
Previous Work: Forrest
Previous Work: Forrest
Bishop
Bishop
Strengths Strengths
Very Limited DOF Very Limited DOF
Clear macro design Clear macro design
Weaknesses Weaknesses
Nanoscale Nanoscale
implementation clearly implementation clearly
implied, but not clearly implied, but not clearly
designed designed
no detailed methodology no detailed methodology
for self for self - -replication replication
http://www.iase.cc/html/overtool.htm
FOR MORE INFO...
Related Work:
Related Work:
Chirikjian/Suthakorn
Chirikjian/Suthakorn

Strengths
Strengths
Autonomous implementation Autonomous implementation
of Trivial of Trivial
+2 +2
case (4 parts) case (4 parts)
Directed towards Directed towards
extraterrestrial applications extraterrestrial applications
Lego isomorphic with Lego isomorphic with
molecules molecules

Weaknesses
Weaknesses
Small UC envelope Small UC envelope
Depends on non Depends on non- -replicating replicating
jigs jigs
High entropy environment High entropy environment
limits extension to Trivial limits extension to Trivial
+3 +3
case case
http:// http://caesar.me.jhu.edu/research/self_replicating.html caesar.me.jhu.edu/research/self_replicating.html
FOR MORE INFO...
Related Work:
Related Work:
Zyvex
Zyvex

Projects
Projects
Applying MEMS and Applying MEMS and nanotubes nanotubes
Parallel Micro and Exponential Parallel Micro and Exponential
Assembly Assembly

Strengths
Strengths
First and only funded company trying First and only funded company trying
to build a to build a Drexlerian Drexlerian assembler assembler

Weaknesses
Weaknesses
MEMS is 1000X too big MEMS is 1000X too big
surfaces too rough surfaces too rough
Exponential Assembly is machine self Exponential Assembly is machine self - -
assembly (not Universal Constructor; assembly (not Universal Constructor;
not GRP paradigm; not Utility Fog) not GRP paradigm; not Utility Fog)
FOR MORE INFO...
http:// http://www.zyvex.com www.zyvex.com/ /
Related Work:
Related Work:
Freitas/Merkle
Freitas/Merkle
Kinematic Self Kinematic Self- -Replicating Machines Replicating Machines
( (Landes Landes Bioscience, 2004) Bioscience, 2004)
First comprehensive review of field First comprehensive review of field
1. 1. The Concept of Self The Concept of Self - -Replicating Replicating
Machines Machines
2. 2. Classical Theory of Machine Replication Classical Theory of Machine Replication
3. 3. Macroscale Macroscale Kinematic Machine Kinematic Machine
Replicators Replicators
4. 4. Microscale Microscale and Molecular Kinematic and Molecular Kinematic
Machine Replicators Machine Replicators
5. 5. Issues in Kinematic Machine Replication Issues in Kinematic Machine Replication
Engineering Engineering
6. 6. Motivations for Molecular Motivations for Molecular- -Scale Machine Scale Machine
Replicator Design Replicator Design
(c) 2004 Robert Freitas and Ralph Merkle
Freitas is a technical consultant for this project
Related Work: Matt Moses
Related 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
Why Universal Constructors?
Why Universal Constructors?
Envelope =
everything
Rock
UC
Envelope =
nothing
Assembly Line
Envelope =
one thing
Robot
Envelope =
many things
SRS
Envelope = every
constituent part

UC is the ability to build anything

UC is the ability to build anything

Uses
Uses
Genotype+Ribotype
Genotype+Ribotype
= Phenotype
= Phenotype

Construction envelope includes itself

Construction envelope includes itself

Atoms equivalent to bits

Atoms equivalent to bits

SRS only needs limited UC capability

SRS only needs limited UC capability