Antenna Optimization
Technology and Applications
CMU
Jason D. Lohn, PhD
Assoc Research Professor
Carnegie Mellon University
Silicon Valley Campus
NASA Research Park
Mountain View, CA
2011 DMI Workshop
Mountain View, CA, May 2011
Late 19th
Century
Innovators &
Innovations
Maxwell, Hertz,
Marconi, …
Characterization
100% Manual,
Hunches, Trial-andError
Kraus, Yagi, King,
Altshuler, …
70s-80s
NEC, HFSS, …
90s
Present
Optimization
Synthesis
Advent of CAD
tools, Trial-andError
Improved tools
Next generation tools
Traditional
Design Process
Semi-Automated
Design Process
Fully Automated
Design Process
Requirements
Requirements
Requirements
Prior Design Search
Prior Design Search
Create Initial Design
Create Initial Design
Compare to
Requirements
Automated
Run Setup
GA Optimizer
GA Optimizer
CEM
Tool
eg, IE3D
Run Optimization
standard design
evolved design
•Slow, labor intensive
•Expensive
•Can be limited by design bias
•Requires significant expertise
•Faster, yet still requires much
labor to configure
•Requires substantial GA &
computer programming
•Current system is not
commercial-quality
Automatic
Simulate & Analyze
Performance
Refine & Iterate
Program Optimization
Software
Automated
Design
Selection
CEM
Tool
eg, IE3D
standard and/or evolved design
•Fast, automatic, powerful
•Less expensive
•Easy to use
•Optimize across multiple
objectives: RF specs, cost
•Minimal design bias
•Requires less expertise
•Finds untraditional antennas
•Augments engineer’s abilities
•Commercial-quality CAD tool
antenna generator
Traditional
Design Process
Semi-Automated
Design Process
Fully Automated
Design Process
Requirements
Requirements
Requirements
Prior Design Search
Prior Design Search
Create Initial Design
Create Initial Design
Compare to
Requirements
Automated
Run Setup
GA Optimizer
GA Optimizer
CEM
Tool
eg, IE3D
Run Optimization
standard design
evolved design
•Slow, labor intensive
•Expensive
•Can be limited by design bias
•Requires significant expertise
•Faster, yet still requires much
labor to configure
•Requires substantial GA &
computer programming
•Current system is not
commercial-quality
Automatic
Simulate & Analyze
Performance
Refine & Iterate
Program Optimization
Software
Automated
Design
Selection
CEM
Tool
eg, IE3D
standard and/or evolved design
•Fast, automatic, powerful
•Less expensive
•Easy to use
•Optimize across multiple
objectives: RF specs, cost
•Minimal design bias
•Requires less expertise
•Finds untraditional antennas
•Augments engineer’s abilities
•Commercial-quality CAD tool
Technology
CMU
Charles Darwin
James Clerk Maxwell
Outline
• Why is antenna design hard?
• What are evolutionary algorithms and can
they help?
• Space Technology 5 Mission
• Future Antenna Design Software you may
want to try
CMU
Introduction
CMU
What is an antenna?
Simply a transducer
Transforms waves from electronics domain to
air/space propagation and back
To space
To radio
Introduction
• Designing antennas by hand can:
– Be time and labor intensive
– Limit complexity and performance
– Require significant experience
• Electromagnetics theory:
– Exact, deceptively simple
– Solving equations becomes difficult
very quickly
– Cannot be solved in closed-form for
even simple problems
• Can’t underestimate bending metal in a chamber
CMU
Introduction
• Antenna design is technocery:
– The less you know, the more like sorcery
– The more you know, the more it is like technology
• But you never really k`now enough to completely
remove the sorcery – you just learn better what you
don’t know
• If you try to understand it all before designing and
building, you get left behind most practitioners
• Intuition can help – some discoveries made through a
hunch and trial and error
• But intuition can also be wrong
• Problems are highly multimodal and multivariate
CMU
Introduction
CMU
• Alternative to traditional design is computer-automated
design
• Evolutionary Algorithms (EA) have potential
• EAs have been applied to antenna design for over a
decade (first paper 1994):
– Wire antennas [Linden & Atshuler ‘96]
– Antenna arrays [Haupt ’96]
– Quadrifilar helical antennas [Lohn, Kraus & Linden ’00]
– In-situ [Linden ’00]
• Antennas a favorable domain…
Nice Properties of Searching
Antenna Design Spaces
CMU
• Search spaces contain continuous (analog) regions
• Rich, topological search spaces
• In general small changes genotype  small phenotypic
changes
• Complex, real-world, non-linear, search space
• Black Art for human designers; relatively open attitude
(breaking conservation of energy)
• High-fidelity simulators; evaluations generally take on the
order of seconds; less need to worry about process
corners; evolving with noise effective for manufacturability
• Many designs prototype-able  cross reality gap
• Exhaustive testing possible
• (Our) representations biased for parsimony
• On-line evolution experiments not onerous
How Difficult is Antenna Design and
Optimization?
feed
Driven
Element
0.5 λ
Separation
Distance
[0.04, 2.00] λ
Reflector Element
Length [0, 4] λ
Linden, 1997
CMU
Search Space
CMU
10
8
6
4
1.88
1.68
2
1.48
1.28
0
1.08
0.88
-2
0.48
3.82
3.62
0.28
3.22
3.42
Length
0.68
2.82
3.02
0.02
0.22
0.42
0.62
0.82
1.02
1.22
1.42
1.62
1.82
2.02
2.22
2.42
2.62
Gain
0.08
Separation
Computing Needs
Helpful to have a large
amount of CPU cycles
available – you are
solving large, complex
physics simulations,
millions of times!
Linux cluster
• 40 racks (higher
density)
• dual CPU
• no disks
• some are dual core
• commodity server
market hardware
CMU
Evolved Antennas… Which Ones?
CMU
NASA’s Space Technology 5
Mission (2006)
ST5 Mission
• Three nanosats (20in
diameter).
• Measure effect of solar
activity on the Earth's
magnetosphere.
CMU
Original ST5 Antenna Requirements
•
•
•
•
•
•
•
•
•
•
•
•
Transmit Frequency: 8470 MHz
Receive Frequency: 7209.125 MHz
Antenna RF Input: 1.5W = 1.76 dBW = 31.76 dBm
VSWR: < 1.2 : 1 at the antenna input port at Transmit Freq, <
1.5 : 1 at the antenna input port at Receive Freq
Antenna Gain Pattern: Shall be 0 dBic or greater for angles 40
<= theta <=80; 0 <= phi <= 360
Antenna pattern gain (this shall be obtained with the antenna
mounted on the ST5 mock-up) shall be 0.0 dBic (relative to
anisotropic circularly polarized reference) for angles 40 <= theta
<=80; 0 <= phi <= 360, where theta and phi are the standard
spherical coordinate angles as defined in the IEEE Standard
Test Procedures for Antennas, with theta=0 to direction
perpendicular to the spacecraft top deck. The antenna gain
shall be measured in reference to a right hand circular polarized
sense (TBR).
Desired: 0 dBic for theta = 40, 2 dBic for theta = 80, 4 dBic for
theta = 90, for 0 <= phi <= 360
Antenna Input Impedance: 50 ohms at the antenna input port
Magnetic dipole moment: < 60 mA-cm^2
Grounding: Cable shields of all coaxial inputs and outputs shall
be returned to RF ground at the transponder system chasis.
The cases of all comm units will be electrically isolated from the
mounting surface to prohibit current flow to the spacecraft
baseplate.
Antenna Size: diameter: < 15.24 cm (6 inches), height: <
15.24 cm (6 inches)
Antenna Mass: < 165 g.
• Transmit: 8470 MHz
• Receive: 7209.125 MHz
• Gain:
>= 0dBic, 40 to 80 degrees
>= 2dBic, 80 degrees
>= 4dBic, 90 degrees
• 50 Ohm impedance
• Voltage Standing Wave
Ratio (VSWR):
< 1.2 at Transmit Freq
< 1.5 at Receive Freq
• Fit inside a 6” cylinder
CMU
ST5 Quadrifilar Helical Antenna
Prior to our work, a
contract had been
awarded for an
antenna design.
Result: quadrifilar
helical antenna (QHA).
Radiator
Under ground plane:
matching and phasing
network
CMU
ST5 Quadrifilar Helical Antenna
Underside:
Matching and
phasing circuitry
CMU
Ultimate Goal
1 Requirements
• Transmit Frequency: 8470 MHz
• Receive Frequency: 7209.125 antenna input port at Transmit
Freq, < 1.5 : 1 at the antenna input port at MHz
• Antenna RF Input: 1.5W = 1.76 dBW = 31.76 dBm
• VSWR: < 1.2 : 1 at the Receive Freq
• Antenna Gain Pattern: Shall be 0 dBic or greater for angles
40 <= theta <=80; 0 <= phi <= 360
• Antenna pattern gain (this shall be obtained with the antenna
mounted on the ST5 mock-up) shall be 0.0 dBic (relative to
anisotropic circularly polarized reference) for angles 40 <=
theta <=80; 0 <= phi <= 360, where theta and phi are the
standard spherical coordinate angles as defined in the IEEE
Standard Test Procedures for Antennas, with theta=0 to
direction perpendicular to the spacecraft top deck. The
antenna gain shall be measured in reference to a right hand
circular polarized sense (TBR).
• Desired: 0 dBic for theta = 40, 2 dBic for theta = 80, 4 dBic
for theta = 90, for 0 <= phi <= 360
• Antenna Input Impedance: 50 ohms at the antenna input
port
• Magnetic dipole moment: < 60 mA-cm^2
• Grounding: Cable shields of all coaxial inputs and outputs
shall be returned to RF ground at the transponder system
chasis. The cases of all comm units will be electrically
isolated from the mounting surface to prohibit current flow to
the spacecraft baseplate.
• Antenna Size: diameter: < 15.24 cm (6 inches), height: <
15.24 cm (6 inches)
• Antenna Mass: < 165 g.
What you want
2 Design System
CMU
3 Design
Problem
Setup
EC
Optimization
Algorithm
Magic
Physics
Solver
What you asked for
Ultimate Goal
1 Requirements
• Transmit Frequency: 8470 MHz
• Receive Frequency: 7209.125 antenna input port at Transmit
Freq, < 1.5 : 1 at the antenna input port at MHz
• Antenna RF Input: 1.5W = 1.76 dBW = 31.76 dBm
• VSWR: < 1.2 : 1 at the Receive Freq
• Antenna Gain Pattern: Shall be 0 dBic or greater for angles
40 <= theta <=80; 0 <= phi <= 360
• Antenna pattern gain (this shall be obtained with the antenna
mounted on the ST5 mock-up) shall be 0.0 dBic (relative to
anisotropic circularly polarized reference) for angles 40 <=
theta <=80; 0 <= phi <= 360, where theta and phi are the
standard spherical coordinate angles as defined in the IEEE
Standard Test Procedures for Antennas, with theta=0 to
direction perpendicular to the spacecraft top deck. The
antenna gain shall be measured in reference to a right hand
circular polarized sense (TBR).
• Desired: 0 dBic for theta = 40, 2 dBic for theta = 80, 4 dBic
for theta = 90, for 0 <= phi <= 360
• Antenna Input Impedance: 50 ohms at the antenna input
port
• Magnetic dipole moment: < 60 mA-cm^2
• Grounding: Cable shields of all coaxial inputs and outputs
shall be returned to RF ground at the transponder system
chasis. The cases of all comm units will be electrically
isolated from the mounting surface to prohibit current flow to
the spacecraft baseplate.
• Antenna Size: diameter: < 15.24 cm (6 inches), height: <
15.24 cm (6 inches)
• Antenna Mass: < 165 g.
What you want
2 Design System
CMU
3 Design
Problem
Setup
GA
Optimization
Algorithm
CEM
Sim
What you asked for
1st Set of Evolved Antennas
Non-branching:
ST5-4W-03
Branching:
ST5-3-10
CMU
Inside the Algorithm
CMU
• Genotype is a tree-structured encoding that specifies the
construction of a wire form
f
f
rx
ry
rz
forward(distance)
rotate_x(angle)
rotate_y(angle)
rotate_z(angle)
rx
f
f
rz
f
rx
5.0cm
f
• Branching in genotype results in
branching in wire form
• Re-evolved (2nd set) antennas had
no branching
Feed
Wire
2.5cm
Fitness Function
• Antenna designs are evaluated by NEC4 running on Linux
Beowulf supercomputer (80 processors)
• 4 evaluations (3 w/added noise) for each design
• Fitness function (to be minimized):
F = VSWR_Score * Gain_Score * Penalty_Score
• VSWRs from both freqs are
scaled and multiplied
VSWR
• Gain is sampled at 5 degree increments between theta=40
and theta=90
f=0
if gain > 0.5 dB
f = 0.5 – gain
if gain < 0.5 dB
• Penalty: proportional to # gain samples less than 0.01 dB
CMU
Measurements
At NASA Goddard Space Flight Center
CMU
Gain for Antenna ST5-3-10
Data measured at GSFC on prototype ST5-3-10, Build LIR-2003-02-08
CMU
Comparison: Pattern Analysis
for Antennas @ 8.47 GHz
Conventional Antenna
CMU
Evolved Antenna
Max and Min Gain vs Theta for 8.47 GHz
10
5
0
0
10
20
30
40
50
60
70
80
90
Gain
-5
-10
Max
Min
-15
-20
-25
-30
-35
Theta (deg)
Shaded Yellow Box Denotes Area In-Spec, According to Original Mission Requirements
ORBIT CHANGE RE-DESIGN
CMU
New Mission Requirements
• Launch vehicle change: spacecraft will go into LEO
(low-earth orbit)
• New requirements:
• Deep null at zenith not acceptable; No way to
salvage original evolved design
• Desire to have wider range of angles covered with
signal:
• Gain:
• >= -5dBic, 0 to 40 degrees
• >= 0dBic, 40 to 80 degrees
• Quadrifilar helical antenna still ok, though not very good
CMU
Rapid Response to
Changing Requirements
• Within 4 weeks we re-evolved two new
antennas & had prototype hardware
• At first it was a bug… then became a feature
• NASA mission requirements frequently change
• Because of the dual antenna configuration, the
evolved antennas:
• yield better coverage
• significantly better efficiency
CMU
2nd Set of Evolved Antennas
EA 1 – Vector of Parameters
CMU
EA 2 – Constructive Process
Best of Gen During Evolution
• Constraints:
–
–
–
–
–
Single wire form (no branching)
Small volume (5 cm3)
Bend angles (min angle 20 degrees)
Number of segments (5-9)
Segment length (4-15mm)
CMU
Evolved Antenna on ST5 Mockup
TOP
DECK
CMU
BOTTOM
DECK
GROUND
PLANE
AXIS OF
ROTATION
Field Pattern
CMU
Gain vs Theta for Phi=90 degrees
QHA-QHA: 38% efficiency
EA104-QHA: 80% efficiency
don’t care
don’t care
desired gain regions
EA6303-QHA: 74% efficiency
EA6303-EA6303: 97% efficiency
CMU
ST-5 Flight Unit
Underside
• Connector
• Gold-plated Groundplane
• Tuning Assembly
Top
• Coated Radome
• Fill Tube
CMU
Evolved
X-Band
Antenna
ST5 Spacecraft Integration
Evolved Antenna
Quadrifilar Helix
Antenna
ST5 Spacecraft #2 in
NASA Goddard Spaceflight Center
Clean Room
CMU
ST5 Spacecraft Integration
CMU
Pegasus Support Structure
CMU
ST5 Mission
CMU
• March 22 – June 30, 2006
• Pegasus Launch Vehicle
• Antennas “flawless”
Benefits
• Lower Power: evolved antenna achieves high gain (24dB) across a wider range of elevation angles:
– allows a broader range of angles over which maximum data
throughput can be achieved.
– may require less power from the solar array and batteries.
• More Uniform Coverage: Very uniform pattern with small
ripples in the elevations of greatest interest (40 – 80
degrees):
– allows for reliable performance as elevation angle relative to
the ground changes.
• Inexpensive Design Cycle: 3 person-months to run
algorithms and fabricate the first prototype as compared to
5 person-months for contractor team.
• Rapid Re-Design: common in NASA missions
• First Evolved Hardware in Space, March-June 2006
CMU
Side by Side
CMU
Antenna Examples
• Omni antennas: (a) ST-5 single-arm,
(b) multi-arm antenna, (c) low-profile
wideband high-power comm antenna,
(d) slant-45 pol wideband field sensor
array
• High-gain antennas: (e) dual-pol Sband high-gain Yagi, (f) twisted Yagi
• UWB antennas: (g) 2-18 GHz CWGA
• Reconfigurable antennas: (h) antennaforming network of switches
• Planar antennas, including patches,
spirals and vivaldi notches: (i) 2-18
GHz compact evolved spiral
• Antenna arrays: (j) 8x8 dual-polarized,
wideband (~2:1), electronically
steerable array for spaceborne
platform
• There are many, many more, e.g.:
–Reflector antennas
–Planar UWB antennas
–Patch antennas
–In-situ optimization
d
a
b
c
g
e
h
j
f
i
Latest Prototype
6-inch aluminum ground plane
CMU
Impedance Bandwidth Measurement
CMU
• VSWR is below 2 from
2100-2375MHz, a
bandwidth of 275MHz
VSWR
• VSWR=1.6 at 2271
(downlink)
<2.0 VSWR
275MHz bandwidth
2100-2375MHz
3
1.95
2
1.60
1
1900
1960
2020
2080
2140
2200
2260
frequency
2110
2271
2320
2380
2440
2500
Gain Patterns
CMU
Trends in Antenna Design Automation
• More demanding requirements
• Increasing time-to-market pressures
• Explosive growth: wireless & mobile
• Solid growth: aerospace, RFID, sensor
networks, data networking
• Cheap compute power: multi-core CPUs;
scale w/Moore’s Law
• Reduced supply of antenna engineers
2015
2014
2013
2012
2011
2010
2009
2008
Estimate
2007
Need for Optimization
Growth of Wireless & Mobile
(subcribers)
Upward Pressures
CMU
Traditional
Design Process
Semi-Automated
Design Process
Fully Automated
Design Process
Requirements
Requirements
Requirements
Prior Design Search
Prior Design Search
Create Initial Design
Create Initial Design
Compare to
Requirements
Automated
Run Setup
GA Optimizer
GA Optimizer
CEM
Tool
eg, IE3D
Run Optimization
standard design
evolved design
•Slow, labor intensive
•Expensive
•Can be limited by design bias
•Requires significant expertise
•Faster, yet still requires much
labor to configure
•Requires substantial GA &
computer programming
•Current system is not
commercial-quality
Automatic
Simulate & Analyze
Performance
Refine & Iterate
Program Optimization
Software
Automated
Design
Selection
CEM
Tool
eg, IE3D
standard and/or evolved design
•Fast, automatic, powerful
•Less expensive
•Easy to use
•Optimize across multiple
objectives: RF specs, cost
•Minimal design bias
•Requires less expertise
•Finds untraditional antennas
•Augments engineer’s abilities
•Commercial-quality CAD tool
Software Tool Under Development
• Developing very powerful, next-generation antenna
synthesis and optimization tool
Antenna specs
AADOS  Antenna Design
CMU
AADOS Software Tool
CMU
Antennas for Emergency Comms
• GATR Technologies: Inflatable, 45 min set-up,
rolls up like sleeping bag, weighs 90 lbs
CMU
Antennas for Emergency Comms
• LTA Projects: Portable inflatable towers, 37 ft,
mount your antenna, handles 35 mph winds
CMU
Antennas for Emergency Comms
• Study showed that 4 randomly placed transmitters
acting as an array gave 7 dB vs single transmitter
CMU
Thank You
Questions?