DSTO
High Frequency Over-the-Horizon Radar IEEE Lecture Atlanta, GA, May 2012
Dr. Giuseppe A. Fabrizio Senior Research Scientist, High Frequency Radar Branch, Intelligence, Surveillance and Reconnaissance Division, DSTO Australia.
[email protected]
DEPARTMENT OF DEFENCE DEFENCE SCIENCE & TECHNOLOGY ORGANISATION
DSTO
DEFENCE: PROTECTING AUSTRALIA 1
© Commonwealth of Australia 2010
DSTO
Presentation Outline 1.
Fundamental Principles
2.
Sky-Wave OTH Radar
3.
HF Radar Sub-Systems
4.
HF Signal Environment
5.
Conventional Processing
2
DSTO
1. Fundamental Principles Section Outline:
•
Surveillance Radar & Frequency Bands
•
Interest in the High Frequency Region
•
Essential OTH Radar Concepts
3
DSTO
Surveillance Radar & Frequency Bands
Radar Frequencies The choice of frequency band has a pronounced influence on the characteristics and performance of a radar system. Meteorological Effects
Ionospheric Effects Over-the-Horizon Radar
Microwave Radar (0.4 – 40 GHz)
Band
HF
VHF
UHF
L
S
C
X
Ku
K
Ka
Frequency
3-30 MHz
30-300 MHz
300-1000 MHz
1-2 GHz
2-4 GHz
4-8 GHz
9-12 GHz
12-18 GHz
18-27 GHz
27-40 GHz
Wavelength
10-100 m
1-10 m
0.3-1 m
~20 cm
~10 cm
~5 cm
~3 cm
~2 cm
~1.4 cm
~0.8 cm
Physically Larger
Range Coverage
Physically Smaller
Resolution/Accuracy 4
DSTO
Surveillance Radar & Frequency Bands
Target Types Focus on radars used primarily for surveillance of man-made targets. ¾
Conventional & OTH surveillance radars have many common target types
Example: Surveillance radar targets & mission priority: Large Aircraft
Fighter-Sized & Helicopters
Missiles
Large Ships
Destroyers & Patrol Boats
Go-Fast Boats
Aircraft (Primary Mission)
Ships (Secondary Mission) Remote sensing radars (e.g. sea-state mapping) target natural scatterers. ¾ Remote sensing applications of OTH radar not explicitly considered here
5
DSTO
Surveillance Radar & Frequency Bands
Surveillance Functions Conventional & OTH surveillance radars share the main functions. ¾ Target detection, localization & tracking
Example: Main surveillance radar functions: Clutter
Noise
Jamming
1) Target detection-estimation: Discriminate target echoes against disturbance signals and estimate target parameters of interest to infer target geographical position and velocity. Range
Direction
Radial Velocity
Coordinate Registration
2) Target track-while-scan: Establish, maintain and display detected target tracks while continuing to search the coverage area for new targets. 6
DSTO
Surveillance Radar & Frequency Bands
Early HF Radar British “Chain Home” radar, the first used for air-defence in wartime [2]. Æ HF technology was only available means to generate sufficient power (1935) Radar designed for line-of-sight ranges, not for over-the-horizon detection. Æ Echoes from very long distances constituted “interference” for radar operators Later during world war II, microwave radars were successfully employed. Æ Regarded as the most competitive frequency band for line-of-sight applications Example Chain Home radar station (East UK coast). Frequency 20-30 MHz
Robert Watson-Watt (1892-1973).
7
DSTO
Surveillance Radar & Frequency Bands
Conventional Radar Great majority of line-of-sight radars implemented at microwave frequency. Æ main technical reasons (at a glance)
LINE-OF-SIGHT PROPAGATION-PATH
TARGET RCS (OFTEN IN OPTICAL REGION)
(TARGET LOCALIZATION ACCURACY)
PHYSICALLY SMALL HIGH GAIN ANTENNAS (EASIER TO SATISFY SITE CONSTRAINTS)
GREATER USEABLE BANDWIDTHS (FINE RANGE RESOLUTION)
LOW AMBIENT NOISE LEVEL (INTERNAL NOISE LIMITED) CONVENTIONAL RADAR
POTENTIAL FOR CLUTTER REDUCTION (e.g. “UP-LOOKING” GEOMETRY)
8
DSTO
Surveillance Radar & Frequency Bands
Line-of-Sight Coverage Microwave radar coverage is mostly restricted to line-of-sight (LOS). ¾ Propagation is shadowed by mountains & limited by the Earth’s curvature
Earth’s S u r f ac e
Low-flying targets escape early detection
Range increased by raising radar platform (or by anomalous propagation). • Doubling range requires quadrupling the platform height (e.g. airborne radar) • Atmospheric “ducting” is not predictable (and may also degrade performance) 9
DSTO
Interest in the High Frequency Region
Beyond Line-of-Sight High frequency signals (3-30 MHz) propagate beyond the line-of-sight. 1.
A “sky-wave” mode involving reflection(s) from the ionosphere
2.
A “surface-wave” mode guided by a conductive sea-surface
Different physical mechanisms that are essentially unique to the HF band. ¾ Exploited by OTH radar & short-wave communicators since G. Marconi (1901) MICROWAVE
IONOSPHERE
HF Surface-Wave
Guglielmo Marconi
HF Sky-Wave
TX EARTH’S SURFACE
10
DSTO
Interest in the High Frequency Region
About the Ionosphere Ionized gas (plasma) formed by the Sun’s extreme UV radiation [3]. • Electron density distribution with height exhibits local maxima (regions) • No direct radiation at night but plasma in ionosphere never decays fully
Courtesy of http://www.windows.ucar.edu
1864 - 73
James Clerk Maxwell describes theory of electromagnetic radiation and predicts existence of radiowaves
1887
Heinrich Hertz proves existence of radiowaves
1895
Guglielmo Marconi demonstrates wireless (radio) communication in Bologna, Italy
1899
Marconi transmits radio signal across English Channel
Dec. 12, 1901
Marconi transmits radio signal across Atlantic Ocean from Cornwall, England to St. John's, Newfoundland
1902
Oliver Heaviside; Arthur Kennelly propose existence of conducting layer in upper atmosphere
1909
Marconi awarded Nobel Prize
1924
Edward Appleton and others develop ionosonde & begin ground-based soundings; prove existence of ionosphere
1925
Appleton discovers second layer (the F region)
1926
Robert Watson-Watt (later developer of radar) coins word "ionosphere"
1927
Sydney Chapman describes theory for formation of ionosphere
1947
Appleton awarded Nobel Prize
1958
Incoherent Scatter Radar developed
11
DSTO
Interest in the High Frequency Region
Useful Coverage Ray-tracing through a model ionosphere using simulation software. • Escape rays at high elevations produce a “skip zone” Æ Earth not illuminated • Reflected rays at lower elevation Æ useful range extent beyond the skip-zone Escape Rays
CONCEPTUAL REPRESENTATION m 600 k
500
Altitude (km)
fc >
400
φ
300
1st Hop
2000
ESCAPE RAYS
m 300 k
Single 200 Frequency
1500
1000
Range (km) 2st3000 Hop
2500
fp
fc ≤
sin φ
fp
400
sin φ
300 REFLECTED RAYS
fc
100
200 100
Radar Footprint Useful Coverage
Backscattered Power (Two-way path)
Leading Edge Focusing Surface Clutter Target
Skip Zone Range (km) 0
500
1000
1500
2000
2500
3000
12
DSTO
Essential OTH Radar Concepts
Sky-wave OTH Radar Sky-wave OTH radar exploits oblique reflection over a two-way path. • Cost-effective early-warning (long-distance) & wide area surveillance • Monitor strategic areas where it is not possible to install conventional radar
Concept of Operation Receive
Ionosphere
Transmit Transmit Skip-Zone Limit Dwell Time (CPI)
Potential Radar Coverage Radar Footprint
Resolution Cells
Radar Footprint Higher Frequency TX & RX Beam Steering
13
DSTO
Essential OTH Radar Concepts
Radar Equation Noise-limited radar equation for OTH and conventional radar systems. ¾ Same form but range increases by order of magnitude (for all target altitudes) ¾ Target echo received by OTH radar experiences additional 40dB spreading loss Transmit Power (Average)
Output Signal-to-Noise Ratio
Transmit Antenna Gain
Receive Antenna Gain
Target Cross Section
Pave G t TA e σ F p
S = N N o L ( 4π ) 2 R External Noise Power per unit bandwidth
4
=
Effective Integration Time
Pave G t G r T λ 2 σ F p N o L ( 4π )3 R Losses (Path and System)
Operating Wavelength
Propagation Factor
4
Slant Range (Radar-to-Target)
Radar Type
Range Coverage (km)
Surface Coverage (sq km)
Sky-wave OTH Radar
1000-3000
Millions
Ground-Based Microwave
1-300
Tens of thousands 14
DSTO
Essential OTH Radar Concepts
Resolution & Accuracy OTH radar range resolution limited by useable bandwidths in HF spectrum. 1. 2.
User-congestion in HF band limits availability of clear frequency channels Frequency dispersion in ionosphere places a limit on coherence bandwidth
Antenna gain & beamwidth are dependent on aperture size in wavelengths. • HF radar wavelengths are three orders of magnitude greater than microwave • Antenna apertures 3 km long needed for beamwidths in the order of 1 degree
Spatial resolution comparison – “order of magnitude”. Radar Type
Useable Bandwidths
Range Resolution
Aperture Size
Antenna Beamwidth
OTH Radar
5-50 kHz
3-30 km
3000 m
0.2-2.0 deg.
Microwave
500-5000 kHz
30-300 m
6m
0.1-1.0 deg.
Target location accuracy determined by propagation-path knowledge • Propagation through ionosphere is much more uncertain than line-of-sight • Target location accuracy for OTH radar may at times be up to 10-40 km 15
DSTO
Essential OTH Radar Concepts
Target Scattering Target RCS in Rayleigh-resonance scattering regimes for OTH radar. 1.
Influenced mainly by gross target dimension (conductive segments)
2.
Depends on operating frequency, aspect angle & TX/RX polarization
3.
Stealth by energy absorbing materials and shaping ineffective at HF
σ Optical
• 3 MHz (100 m wavelength) “Rayleigh” Æ RCS falls very rapidly with frequency • 30 MHz (10 m wavelength) “Resonance”
Cross Section
Example: Missile (Length = 10 m)
Rayleigh
Resonance
Æ variable but often higher target RCS Operating Frequency
f
OTH radar “point” targets contained within single resolution cell. • Target physical size << spatial dimensions of radar resolution cell • Radial velocity produces steady phase progression (Doppler shift)
16
DSTO
Essential OTH Radar Concepts
Clutter & Interference-plus-Noise 1. External interference-plus-noise often dominates internal receiver noise. ¾ Atmospheric noise (e.g. lightning) propagated long-distances by the ionosphere ¾ Anthropogenic (man-made) “interference” from other users of the HF spectrum
2. OTH radar prone to high clutter levels (40-70 dB > than target echoes). ¾ “Look down” geometry illuminates Earth’s surface coincidently with targets ¾ Large resolution cell sizes increases effective clutter RCS relative to targets Noise Spectral Density versus Time of Day
Backscattered clutter power versus range and frequency Operating frequency 2000-3000 km range coverage
Skip Zone
17 Internal noise 20-30 dB lower
DSTO
Essential OTH Radar Concepts
Doppler Processing 1. Doppler processing is essential for target detection in OTH radar. • Resolves Doppler shifted target echoes from clutter in same resolution cell • Provides coherent gain (time-on-target) to improve signal-to-noise ratio
2. OTH radar CPI are much longer than microwave radar (ms). • A few seconds for aircraft and tens of seconds for ship detection • To compensate for spreading loss & smaller Doppler shifts at HF
3. Limits on OTH radar CPI length arise from factors including: • Temporal instability of ionosphere or manoeuvring targets in CPI • Region revisit rate for tracking (also trade-off with coverage area) Target
Sea Clutter
Range Cells (one beam only) Doppler Frequency (0 Hz at centre of display)
18
DSTO
Essential OTH Radar Concepts
Power & Waveforms OTH radar transmit power 10-100 X higher than microwave radar. • Average transmit power from 10kW - 1MW Æ sensitivity against noise
• Frequency modulated continuous waveforms Æ reduces peak powers Transmitter & receiver separation ~100 km (continuous waveform). • Referred to as a “quasi-monostatic” configuration (described later) • Relaxes receiver dynamic requirements by attenuating direct wave Quasi-monostatic OTH radar configuration
TX
RX
~ 100 km separation
19
DSTO
Essential OTH Radar Concepts
Propagation-Path Characteristics Propagation via the ionosphere is very complex & challenging to model • Unpredictable variation of path characteristics over a very wide range of scales • Real-time radar management techniques indispensable for successful operation Characteristic
Ionospheric Propagation-Path Variability
Temporal
Dynamic: Intra-CPI, intra-mission, diurnal, seasonal and solar cycle (11 year)
Spatial Frequency Polarization
Heterogeneous: Intra-region, over coverage area, latitudinal variability Dispersive: in time-delay (range), Doppler frequency & ray angle-of-arrival Anisotropic: Magneto-ionic components, Faraday rotation (polarization fading)
Multipath
Ever present: E and F regions over two-way path, variable number of modes
Attenuation
High: D-layer absorption in day time may cause significant signal attenuation
Use of auxiliary sounders to select & update “optimum” radar frequency • Appropriate illumination of the coverage area + minimize interference-plus-noise • Updates to reflect changes in ionosphere Æ over time & different coverage areas 20
DSTO
2. Sky-wave OTH Radar Section Outline:
•
Example Systems
•
Skywave OTH Radar Characteristics
•
The Ionosphere & Propagation Effects
21
DSTO
Early Research & Current Systems
Example Sky-wave Systems OTH radars exhibit significant diversity in architecture (no standard system). US Navy OTH Radar • “ROTHR” (NRL)
French OTH Radar • “Nostradamus” (ONERA)
Russian OTH Radar • “Steel Yard” (NIDAR)
• • • • •
• • • • •
• • • • •
Two-site Æ linear FMCW Maximum power 200 kW Receive aperture 2.6 km 372 elements & receivers Counter-drug application
Mono-static (coded pulse) Maximum Power 50 kW Y-Array, 384 m arm length 288 elements, 48 sub-arrays First reported detection 1994
Two-site (coded pulse) Average Power ~1 MW Vertical array, height 140 m Horizontally polarized dipoles Operational in the mid 1970’s
22
DSTO
Early Research & Current Systems
Australian OTH Radar - Jindalee DSTO Australia “Jindalee Team” (1975) Team Leader: John Strath (circled below)
Dr Malcolm Golley
Dr. Fred Earl
History of OTH radar development in Australia 1952 1974 1979 1986 1986 1987 1990 1991 1998 2002
– – – –
1972 Research to determine ionospheric stability 1978 Jindalee Stage A (Detections in one direction) 1985 Jindalee Stage B (“Track while scan”, 90 deg) 1989 Jindalee Stage C (“Operational capability”) Announcement of JORN Defence White Paper on broad area surveillance JFAS Transferred from DSTO to RAAF JORN Contract Signature Contract to RLM JORN Commissioned
Jindalee “Bare Bones” OTH Radar Receiver Array (Central Australia)
23
DSTO
Early Research & Current Systems
Australian OTH Radar - JORN Australian Jindalee Operational Radar Network Æ two additional radars • Longreach (Queensland), Laverton (Western Australia), Control centre (Adelaide) JORN Laverton OTH Radar
Typical Characteristics
TX & RX Site
Separation ~100km
Transmitter Array
Dual Band, Linear VLPA (~150 m)
Receiver Array
480 monopole pairs (~3 km aperture)
Coverage
+/- 90 degrees
Frequencies
5-30 MHz
Waveform
Linear FMCW
Bandwidth
5-50 kHz
Average Power
250 kW
PRF
4-80 Hz
CIT
1.5-30 seconds
RX Site
TX Site
TX Site
RX Site
24
DSTO
Skywave OTH Radar Characteristics
Configuration & Site Selection Radar configuration refers to relative transmitter and receiver locations. ¾
Monostatic Æ economical (single radar site & no need for inter-site links)
¾
Bistatic Æ Allows use of continuous waveforms (two propagation paths)
¾
Quasi-monostatic Æ High sensitivity but essentially one propagation path
¾
Multi-static: De-couples ionosphere from target localization & tracking
Site selection for OTH sky-wave radar takes several factors into account. 1.
Land Æ Needs flat wide open spaces with relatively homogeneous surface
2.
Electrically Quiet Æ Avoid strong HF noise near industrial/residential areas
3.
Self-Interference Æ Isolation to protect RX from TX continuous waveform
4.
Skip-Zone Æ Minimum detection range of ~1000 km (surveillance region) 25
DSTO
Skywave OTH Radar Characteristics
Pulsed & Continuous Waveforms Use of pulsed/continuous waveforms depends on OTH radar requirements. Waveform Type
TX-RX Configuration
Spectral Behaviour
Average-to-Peak Power
Pulsed
Single site
Poor out-of-band
Low (sensitivity in noise)
Continuous
Two sites
Better out-of-band
Higher radar sensitivity
Linear Frequency Modulated Continuous Waveform (FMCW). Frequency
B fc + 2
T CIT
Coherent Integration Time (CIT)
Time
fc fc −
B 2
Ambiguity Resolution cT p c c R = = amb Range Δ R = 2 2 fp 2B
T
p
Pulse Repetition Frequency (PRF)
Resolution
Velocity
Δv =
cΔ fd c = 2 fc 2 f c T CIT
f p = 1/ Tp Ambiguity cf p v amb =26± 4 fc
DSTO
Skywave OTH Radar Characteristics
Typical Missions OTH radar surveillance missions broadly classed as air & surface tasks. Æ Task coverage divided into number of “Dwell Interrogation Regions” (DIRs) Task A: “Barrier” Task • wide area surveillance • mainly used for aircraft
C
Task B: “Stare” Task • surveillance of airports • air/ship lanes, missile sites
Task C: Force Protection
A
• Radar steps through DIRs in a scheduled sequence.
D
• air route surveillance • navy fleet protection
DIR’s
Task D: Remote Sensing • sea-state mapping • cyclone tracking
• Time on each DIR = CIT, all DIR’s revisited in turn.
B 27
DSTO
Skywave OTH Radar Characteristics
Dwell Interrogation Region (DIR) Each DIR consists of many radar resolution cells in range & azimuth. Example dimensions: • Rx array aperture (D) = 3 km, TX array aperture 150 m, Range (R) = 2250 km • Carrier Frequency (F) = 15 MHz, Waveform Bandwidth (B) = 10 kHz Transmitter Footprint
ΔR =
Range-Azimuth Resolution Cell
300 km (20 Beams)
c = 15 km 2B
DIR contains 1200 resolution cells
ΔL = RΔθ ≈
Rλ = 15 km D
900 km (60 range cells)
Transmitter D=150m (8 deg at 15 MHz) Receiver D=3000m (0.4 deg at 15 MHz)
1 20 Receiver “Finger Beams” 20
28
DSTO
Skywave OTH Radar Characteristics
Aircraft & Ship Detection Aircraft & ships typically detected against noise & clutter respectively.
fc
• Air Æ Maximize Signal-to-Noise Ratio (SNR) for high velocity targets • Ship Æ Minimize clutter Doppler spectrum contamination for slower targets
B
• Air Æ Low to find clear frequency channels (with adequate range resolution) • Ship Æ High to reduce range cell size and increase sub-clutter visibility (SCV)
fp
• Air Æ High to avoid velocity ambiguities for fast moving aircraft targets • Ship Æ Low to avoid range-folded spread-Doppler clutter (unambiguous targets)
TCIT
• Air Æ Short for rapid region revisit rates (allows tracking over many DIR’s) • Ship Æ Long for fine Doppler resolution to resolve targets from strong clutter
Example waveform parameters
(Assume carrier frequency=15 MHz, detection range=1500 km).
Mode
B
ΔR
D
ΔL
f
p
R amb
v amb
Air
10
15
1000
30
50
3000
Surface
30
5
3000
10
5
Units
kHz
km
m
km
Hz
TCIT
Δv
900
1
10
30000
90
20
0.5
km
km/h
s
m/s 29
DSTO
Skywave OTH Radar Characteristics
Signal & Data Processing Rudimentary OTH radar signal and data processing steps. Signal Processing Beam Forming
Doppler Processing
CFAR
Note:
Peak Detection Tracking
Early-warning allows more time to decide about target presence compared with certain conventional radars.
Coordinate Registration
Higher false detection rates can be tolerated & filtered by the tracker in time before targets declared present.
Display
Data Processing
Pulse Compression
More details on signal and data processing to follow.
30
DSTO
The Ionosphere & Propagation Effects
Ionospheric Regions The ionosphere may be broadly divided into three altitude regions. ¾ where electron-density versus height profile tends reach local maxima Ionosphere
Height (km)
Main Region Characteristics
D Region
50-90
Formed during the day-light hours Ionization too low for HF reflection
Attenuation of radar signals Electron-neutral collisions
E Region
90-140
May contain anomalous Sporadic-E Generally stable propagation layer
One-hop paths to ~2000km Allows signals to penetrate
Highest layer with maximum ionization Splits into F1 and F2 layers in the day F1 peak (140-210 km) is sun following F2 peak (210-400 km) present at night
Fundamental to OTH Radar 1-Hop F1 can reach 3000km 1-Hop F2 can reach 4000km F2 less stable in space & time
F Region
140-400
Relevance to OTH Radar
Ionosphere exhibits significant variability in structure in space & time. ¾ Temporal variations occur diurnally, seasonally and over the 11 year solar cycle ¾ Significant spatial variations occur across mid-latitude, equatorial & polar region 31
DSTO
The Ionosphere & Propagation Effects
Multipath Propagation Simple illustration of two-way one-hop reflections from E and F layers. • Target Æ multiple echoes often resolved in cone angle, range & Doppler shift • Clutter Æ contamination of Doppler frequency spectrum (mode superposition) • Interference Æ a single source can spread over a significant number of beams F-Layer
Simple One-Hop Modes
Mixed or “Hybrid” Modes
1F − 1F
1F − 1E 1E − 1F
1E − 1E E-Layer
TX-RX
Earth
Target
More complex modes involving multi-hop propagation, top-side layer reflections and trans-equatorial (chordal) modes also exist.
32
DSTO
The Ionosphere & Propagation Effects
Propagation-Path Information 1. Synoptic information about ionosphere useful for radar design. Æ Statistical forecasts of diurnal, seasonal, solar-cycle & global variations
2. Real-time information is useful for optimizing radar operation. Æ Mission-to-mission propagation-path data for DIR’s in all radar tasks
Updates from Ionospheric Sounders
Real-Time Propagation-Path Information
Backscatter sounding
Clutter power levels
Carrier frequency
Spectrum surveillance
Noise spectral density
Waveform parameters
Clutter Doppler profile
Spectral purity
Track association
VI & OI Sounders
Mode structure
Coordinate registration33
Radar Parameter Optimization
DSTO
4. HF Radar Sub-systems Section Outline: • Transmitter • Receiver • Radar Management
34
DSTO
Transmitter
Vertical Log-Periodic Array Vertically polarized log-periodic monopole arrays with ground-screen. 1. Simultaneously covers all useful elevation angles at reasonable cost. • Elevations of 5-40 degrees for one-hop illumination to ranges 1000-3000 km • Exploits illumination of very large range depths when the ionosphere permits
2. Broadband operation over required frequency range. • Use of two (or more) VLPA matched to different sub-bands in HF spectrum • JORN VLPA ~ 40 m tall and mechanically stabilized to reduce Aeolian noise
35
DSTO
Transmitter
Ground Screen HF antenna radiation patterns depend heavily on ground properties. Ground mesh-screens provide two main benefits: A.
Increase antenna gain at low elevations for long range coverage
B.
Stabilize ground impedance to reduce antenna pattern distortion
JORN site Æ approximately 300,000 sqm of galvanised earth-mat. JORN Laverton Transmit Site Ground-Screen High-Band Low-Band
36
DSTO
Transmitter
Transmit Aperture Uniform linear arrays containing 8-16 transmitting elements per band. Transmitter aperture length trades off sensitivity with coverage rate. • Larger apertures provide higher antenna gain (to increase radar sensitivity) • Short apertures provide a broader beam (increases coverage & revisit rate)
JORN Longreach Transmitter Site
37
DSTO
Transmitter
Elevation Control Two-D transmit apertures permit the beam to be steered in elevation. Enhanced transmit directivity in elevation has positives & negatives. + Æ Improves sensitivity against clutter, facilitates mode selection & CR
- Æ Can reduce range depth & azimuth resolution for a fixed # channels Elevation control with ground distributed or vertically raised antennas. • Good resolution at low elevations • Expensive and difficult to stabilize
• Less expensive & easier to stabilize • Poorer resolution at low elevations
38
DSTO
Receiver
Element Design 1. Matched antennas less important for externally noise-limited receivers. ¾ Antenna efficiency experienced by targets & noise Æ no SNR improvement ¾ Match elements at high end (lower noise) with graceful frequency response
2. Reduce cost by using small end-fire antenna element doublets. ¾ Antenna heights of 4-6 meters (less susceptible to Aeolian noise effects) ¾ Twin elements Æ combined with time-delay cable for front-to-back ratio Jindalee Rx Antennas (980 installed by Jim McMillan & Wife in 32 days)
JORN Receive Antennas
39
DSTO
Receiver
Array Aperture Uniform linear arrays (ULA) ¾ Best spatial resolution for cost but elevation-azimuth ambiguity (cone angle)
Two dimensional arrays ¾ Elevation control & mode filtering (with 2D TX) plus wider azimuth coverage
Wide receive apertures improve:
Upper limit on RX aperture size:
• Gain (sensitivity) & spatial resolution
• Greater expense & additional land
• Target detection, location & tracking
• Need greater # of coherent beams
Jindalee Æ Uniform linear array (~ 2.8 km, 90 deg. )
JORN Æ L-shaped array ( ~ 3 km apertures, 180 deg. )
40
DSTO
Receiver
Reception Channel Traditional heterodyne receiver and sub-array beamforming architecture. ¾ Fine resolution “finger beams” formed by digital combination of all RX outputs ¾ Jindalee groups 462 elements into 32 over-lapped sub-arrays (of 28 doublets) • Antenna doublet • Front-to-back ratio
• Wide-band RX front-end • Rapid frequency changes
• Network of switched delay-lines • Steers sub-arrays over footprint
• Tuneable local oscillator • Fixed IF filter bandwidth
• High Dynamic range • 16 bit I&Q sampling
• Conversion to base-band • Calibrated freq. response
Limiting factors: Linearity, A/D conversion, reciprocal mixing & image rejection.
41
DSTO
Radar Management
Frequency Management Sub-systems providing real-time frequency advice for main OTH radar.
Backscatter Sounder Returned clutter power in group-range & frequency for different beams
Vertical/Oblique Incidence Sounders Mode content & virtual heights versus frequency for point-to-point links
HF spectrum surveillance Power spectral density of natural & man-made noise across HF band
Mini-radar Clutter Doppler profile in group-range & azimuth at selected frequencies
Channel Scattering Function Mode distribution in time-delay and Doppler for a narrowband HF circuit 42
DSTO
Radar Management
Backscatter Sounder 1. Backscattered clutter power versus frequency, group-range & beam. ¾ Original system resolutions: 200 kHz, 50 km and 8 beams over 90 deg. ¾ Update intervals in order of 5-10 min, and sounder is co-located with radar
2. Concurrent ionograms recorded in early evening ~45 degrees apart. ¾ Note significant azimuth dependence, and possibility of range-folded clutter ¾ Range extent of 2000-3000 km illuminated most by frequencies 17-18 MHz 17 MHz
18 MHz Range Ambiguity
Range-folded clutter
2nd Hop
1st Hop
Range Coverage
Skip-Zone 43
DSTO
Radar Management
Spectrum Surveillance 1. Identify unoccupied frequency channels in real-time at receiver site. ¾ Avoid interference to other HF services (e.g. broadcasting/communications) ¾ Omni-directional antenna Æ measures noise power in 2 kHz wide channels
2. SCV Æ combine spectral surveillance & clutter power measurements. ¾ Both databases acquired at time of radar operation and in the radar location ¾ Sub-clutter visibility (clutter-to-noise ratio) good indicator of radar sensitivity Entire HF Spectrum
1 MHz Wide Zoom Other HF Users
Protected emergency channels forbidden
Background noise level in clear channels & in azimuth
Clear channels (> 100 kHz)
Background noise level 44
DSTO
Radar Management
Vertical & Oblique Incidence Sounders 1. Maintain a real-time ionospheric model (RTIM) of mode structure. ¾ Enables propagation modes to be identified and reflection heights estimated ¾ Propagation-path information for track association & coordinate registration
2. Network of sounders with rapid (5 min) updates near dawn & dusk.
OI Ionogram Æ Darwin-Alice Springs (1260 km path) Frequency Dispersion
VI Ionogram Æ Similar time near path mid-point
F-layer (high rays) X-ray O-ray F-layer (low rays)
Virtual height F
E-layer RFI
Virtual height E
45
DSTO
5. HF Signal Environment Section Outline: • Composite Signal Environment • Land & Sea Surface Clutter • Ionospheric Clutter & Meteors • Noise & Radio Frequency Interference
46
DSTO
Composite Signal Environment
Signal Environment Composite received signal for OTH radar is a superposition of: ¾ Radar waveform echoes and interference-plus-noise.
OTH Radar Signal Environment
Radar Echoes
Interference-plus-noise
Clutter Returns
Target Echoes
(e.g. Land, Sea)
Æ skin echoes
Anthropogenic (Man-Made)
Naturally Occurring
Unintentional
Intentional
(e.g. Electrical machinery)
(e.g. Radio stations)
Atmospherics
Galactic
(e.g. Lightning)
(e.g. Stars) 47
DSTO
Land & Sea Surface Clutter
Clutter Power Received power of clutter “backscattered” from Earth’s surface. ¾ Function of resolution cell area & normalized backscatter coefficient
σ = σ0 × A
Effective Clutter RCS • Single resolution cell
Resolution Cell Area • Aperture, range & bandwidth
Normalized Backscatter Coefficient • Surface properties & grazing angle
General Characteristics: Sea clutter often more powerful than land clutter (order of magnitude) Æ Higher conductivity of sea-surface and resonant (Bragg) scattering mechanism
High seas towards/away from radar significantly increase clutter power Æ More resonant backscatter, while very flat seas produce near “specular reflection”
Spatial RCS variations gradual over sea, but can be very sharp over land Æ Presence of cities and other topographical discontinuities can enhance RCS Received clutter power may be 40-80 dB stronger than target echoes Æ Receiver dynamic range must be sufficiently high to capture both signals
48
DSTO
Land & Sea Surface Clutter
First Order Clutter Resonant clutter two orders of magnitude stronger than higher order. ¾ Advancing & receding Bragg wave-trains Æ Doppler spectrum Bragg lines EM wavefronts
λ 2
ψ
Bragg Wave-Trains Advancing Wave
Receding Wave
Bragg wavelength
Radar wavelength
L cos ψ =
In deep water, the Bragg wave trains move with radial velocity (i.e. gravity waves):
v = ±
⎤ 1 ⎡ gλ = ± 2 ⎢⎣ π cos ψ ⎥⎦
1 / 2
Without surface currents, this imposes a Doppler shift on two clutter “Bragg Lines”
fb =
2 v cos ψ
λ
⎡ g cos ψ ⎤ = ±⎢ ⎥⎦ πλ ⎣
2
Grazing angle
L gL 2π
λ
1/2
49
DSTO
Land & Sea Surface Clutter
Higher Order Continuum Mainly due to double scattered echoes from pairs of wave-trains. • Second-order clutter “continuum” is distributed in Doppler frequency • May impede target detection, especially slow ships in high sea-states Target visibility depends on echo strength & Doppler shift Bragg Lines (first-order clutter) High RCS
Blind speeds (solid lines)
SCR limits target detection Medium RCS Lower RCS
Higher Order Clutter Continuum 50
DSTO
Land & Sea Surface Clutter
Ship Detections Doppler Shift
10 Beams 10 9 8 7 6 Nested Range Cells
Land clutter
5 4
Sea clutter
3
Note Doppler spreading (transit via ionosphere)
2 1 -
0 Hz
+
51
DSTO
Noise & Radio Frequency Interference
Spectral Density Background noise includes atmospheric (i.e. lightning) & galactic noise. • Received noise depends not only on sources, but also propagation conditions • Atmospheric dominates at low frequencies and galactic at higher frequencies • Frequency and (diurnal) time dependencies determined partly by ionosphere Day Time (~Noon) D-layer absorption in the day attenuates long-range noise at lower frequencies (lossy propagation paths) Anthropogenic RFI 50 dB above background
Background Noise Level
Night Time (~Midnight) Powerful RFI & congested lower HF spectrum
No sky-wave paths for higher frequencies at night
Higher Background Noise Mainly Mainly Atmospheric Galactic
52
DSTO
Noise & Radio Frequency Interference
Background Noise Variations Background noise spectral density variations (monthly median figures). • Higher noise levels in geographical areas of strong thunderstorm activity • Many databases recorded by omni-directional antennas; CCIR report • Background noise is directional Æ level depends on radar look direction Median Noise Spectral Density Local Noon
Local Midnight High frequencies penetrate ionosphere at night
Internal receiver noise spectral density 53
DSTO
6. Conventional Processing Section Outline: • Range, Beam & Doppler Processing • CFAR Detection & Peak Estimation • Tracking and Radar Displays
54
DSTO
Signal & Data Processing Stages
Processing Stages Flow chart of OTH radar signal/data processing steps and displays.
Conventional processing traditionally based on FFT
Radar data and target track displays for operators
Time for processing and display in the order ~ CPI
Signal Processing Stages A/D Conversion
Pulse Compression
Beam Forming
Doppler Processing
Tracking
Coordinate Registration
Data Processing Stages CFAR
Peak Detection
Radar Data and Track Displays
55
DSTO
Range/Beam & Doppler Processing
Pulse Compression Separates received echoes on basis of time-delay.
Target group-range estimated for localization
Resolves targets from each other & multipath
Well-known principle of “pulse-compression” Reference Chirp
Taper Function
Returned Echo
Energy
Difference Frequency Δf
FFT
B Δf =
τ
T
c 2B
ΔR =
τB Tp
R =
p
cT p Δ f cτ = 2 2B
Range Bins
Example for sky-wave OTH radar. Task
Pulse Period
Bandwidth
Ambiguity
Resolution
No. Bins
Range Depth
Air
0.02 seconds
10 kHz
3,000 km
15 km
40
600 km
Surface
0.2 seconds
30 kHz
30,000 km
5 km
80
400 km 56
DSTO
Range/Beam & Doppler Processing
Doppler Processing Separates received echoes on basis of Doppler shift.
Isolates targets/clutter in separate frequency bins
Coherent gain to improve target detection in noise
Estimates target radial velocity to improve tracking Chirp 1
Chirp K
Chirp 2 Processed Range Bin
Range-Doppler Range
Range
Taper Function
Range
Range Bins
Doppler Processing FFT Doppler Bins
Example for sky-wave OTH radar. Task
Pulse Period
CIT
Ambiguity
Resolution
No. Pulses
Frequency
Air
0.02 seconds
2 seconds
+/- 900 km/h
5 m/s
100
15 MHz
Surface
0.2 seconds
40 seconds
+/- 90 km/h
0.25 m/s
200
15 MHz 57
DSTO
Range/Beam & Doppler Processing
Beamforming Separates received signals on the basis of angle-of-arrival.
Provides coherent gain in surveillance beam direction
Estimates the (cone) angle-of-arrival of target echoes
Attenuates sidelobe disturbance due to clutter & noise
Rx 1
Beam Cell
Range
Beam FFT or DFT
Processed Cell
Receivers Range & Doppler Processing
Doppler
ULA Narrowband Sensors
Range Rx N
Range & Doppler Processing Doppler
Example for sky-wave OTH radar.
Range
θ1 Doppler Beams Range
θN Doppler
Taper Function
Task
Aperture
Frequency
Ambiguity
Resolution
No. Beams
Coverage
Air
1500 m
15 MHz
d / λ < 0.5
~ 1 deg.
10
10 deg.
Surface
3000 m
15 MHz
d / λ < 0.5
~ 0.5 deg.
20
10 deg. 58
DSTO
Range/Beam & Doppler Processing
Processing Example Pulse Compression
Doppler Processing
Beamforming
Æ Receivers, ranges, & pulses
Æ Receiver Range-Doppler maps
Æ Beam Range-Doppler maps
Æ Strong clutter masks target
Æ Clutter confined close to 0 Hz
Æ Target becomes clearly visible
Range
Rx 1
Beams, Nested Range Cells
Receivers, Nested Range Cells
Receivers, Nested Range Cells
Rx 2
Clutter
Target
59 Time (Chirp Number)
Doppler Frequency
Doppler Frequency
DSTO
Range/Beam & Doppler Processing
Multipath Echoes BEAM SPECTRUM
“Zoom” (beam containing target) Group Range
Array Boresight
1F − 1F “Coning Effect”
1F − 1E 1E − 1F 1E − 1E
}
-
0 Hz Doppler Frequency
Hypothesized mode structure F-Layer
Key observations: • Single target Æ multiple echoes • Distinct range, Doppler & beam
E-Layer
TX-RX Earth
Target
F-F mode c.f. E-E mode • Longer group-range • Smaller Doppler shift 60 • Higher coning effect
+
DSTO
Range/Beam & Doppler Processing
Window Functions Importance of window functions to control spectral leakage. Without Doppler window
With Doppler window
Target
Target masked by Clutter sidelobes
61
DSTO
CFAR Detection & Peak Estimation
CFAR Processing Constant false alarm rate (CFAR) processing applied to ARD data. ¾ To reduce the number of false detections made on clutter & noise Variety of CFAR techniques: •
Definition of cell under test (CUT)
•
Window in range-Doppler & beam
•
Cell-averaging or ordered statistics
•
Estimation of a “background level”
•
Normalization of CUT by this level
•
Repeat for all radar resolution cells
Doppler Window
CUT
Guard Cells Range Window
CFAR window dimensions may be changed to suit local disturbance features. ¾ Cell Averaging (CA) or Greatest of Ordered Statistics (GOOS) methods. 62
DSTO
CFAR Detection & Peak Estimation
CFAR Detection Peak detections on the CFAR output are passed onto tracker if: • Cell under test is a local peak in range, Doppler & beam dimensions
• Magnitude of this peak exceeds a pre-set target detection threshold CFAR Output
Possible Clutter False Alarm
Low Threshold
High Threshold
(high false alarm rate)
(low detection probability)
Target
}
Suitable detection threshold range
Target
63
DSTO
CFAR Detection & Peak Estimation
Peak Estimation Peak parameters estimated using ARD data (before CFAR): • Quadratic interpolation using peak and immediate neighbours • Non-integer estimates range, Doppler, beam & SNR to tracker • Step must be repeated for all detected peaks in CPI data cube
Target Peak
Quadratic Interpolation
Target Beam Estimate
64
DSTO
Tracking and Radar Displays
Tracking Target presence may be declared on basis of established target tracks. • Early-warning Æ time for tracker to filter out many false (noise/clutter) peaks • Permits use of low peak detection thresholds to capture weaker target echoes
Single target may produce several distinct echoes due to multipath. • Tracking usually performed in radar coordinates on all propagation modes • Separate processing to associate multipath tracks & covert them to ground
Probability data association (PDA) filter successful for OTH radar. •
Track updated by combined influence of multiple peaks in neighbourhood
•
Simultaneous tracking of multiple targets with multiple hypothesis models
65
DSTO
Tracking and Radar Displays
Coordinate Registration Challenging problem of converting from radar to ground coordinates. Æ Uncertain propagation via the ionosphere Several possible CR techniques: ¾
Ray tracing with real-time ionospheric model (RTIMs)
¾
Transponders at known locations in surveillance area
¾
Detection of sea-land clutter boundaries in coverage
¾
Association of detections with available GPS reports
¾
Detections on commercial aircraft & shipping lanes
¾
Registering airports where tracks begin or terminate
¾
Correlation of clutter RCS enhancements with cities
Æ Effective fusion of different CR techniques 66
DSTO
Tracking and Radar Displays
Radar Data Displays Geographical Track Display
“Whitened” ARD Display
• Detections filtered in time (CPI)
• Target position + Doppler
• Displays multipath target tracks
• For a single DIR and CPI
Tracks for single target
“Stare” Scroll Display
• 3 tracked ionospheric modes
• Localized area (one beam)
• With possible TID presence
• Clearly shows manoeuvres Range x
Target Detections
Range x+1
Manoeuvring Target
Time
67 Doppler
Doppler