Sensing and Communications Using Ultrawideband Random Noise Waveforms

1
Sensing and Communications Using Ultrawideband
Random Noise Waveforms
2005 AFOSR Program Review for Sensing, Imaging
and Object Recognition , Raleigh, NC, May 26, 2005

• Professor Ram M. Narayanan
• Department of Electrical Engineering
• The Pennsylvania State University
• University Park, PA 16802, USA
• Tel [phone]
• Email ram_at_ee.psu.edu

2
Outline

• Introduction
• Why use noise waveforms
• Noise waveform modeling
• Heterodyne correlation approach
• Polarimetric radar applications
• Radar imaging applications
• Covert communications applications
• MIMO network concept
• Conclusions

3
Introduction

• Military operations require low probability of
  intercept (LPI), low probability of exploitation
  (LPE), low probability of detection (LPD), and
  anti-jam characteristics
• Traditional radar and communications systems use
  conventional deterministic waveforms
• Deterministic waveforms (such as
  impulse/short-pulse and linear/stepped frequency
  modulated) do not possess above desirable features

4
Why use noise waveforms?

• Noise waveforms are inexpensive to generate both
  in analog and digital formats
• Noise waveforms have featureless LPI/LPD
  characteristics and are therefore covert
• Noise waveforms are inherently anti-jam and
  interference resistant
• Noise sources are easily obtained using current
  microwave and RF circuit technology
• Noise waveform spectral characteristics can be
  adaptively shaped to suit the dynamic environment
• Noise waveforms are spectrally very efficient and
  can share spectral bands without mutual
  interference
• Noise waveforms exhibit excellent waveform
  diversity characteristics

5
Waveform comparison
6
Simple noise radar architecture using homodyne
correlator
7
Phase coherence injection

• Homodyne correlation noise radar downconverts
  directly to DC and hence loses important phase
  information of returned signal
• There is a way to inject phase coherence in noise
  radar using time-delayed and frequency-offset
  transmit replica
• Heterodyne correlation noise radar downconverts
  to offset frequency and preserves phase
  information of returned signal

8
Noise waveform - stochastic model

• Thermal noise is stochastic and can therefore
  only be described by its statistics
• Noise signal x(t) can described as follows
• PDF px(X) ? Zero-mean Gaussian
• Autocorrelation Rxx(t) ?Impulse at t 0
• PSD Sxx(f) ?White, assumed uniform and
  bandlimited
• Above representation does not permit
  time-frequency equivalence for tracing the signal
  through the system

9
Noise waveform time-frequency model

• where
• a(t) is Rayleigh distributed amplitude that
  describes amplitude fluctuations
• d?(t) is uniformly distributed frequency that
  describes frequency fluctuations -?? d? ??
• average power ½a2(t)/R0, assuming a(t) and
  d?(t) are uncorrelated
• center frequency ?0/2p f0
• bandwidth 2??/2p B

10
Bandwidth descriptors

• Narrowband ? B/f0 10
• Ultrawideband (UWB) ? B/f0 25
• Although time-frequency representation is
  inherently narrowband, we extend it to the UWB
  case owing to its simplicity and ease of signal
  analysis

11
Alternate time-frequency representation
where sI(t) and sQ(t) are zero-mean Gaussian
processes and f0 is the center frequency This
can be recast as
where
Uniformly distributed
Rayleigh distributed
12
Homodyne correlation noise radar
13
Heterodyne correlator noise radar
14
Heterodyne correlation noise radar signal analysis

• Transmit waveform ?
• Received waveform ?
• Time-delayed transmit replica ?
• Time-delayed and frequency-offset transmit
  replica ?
• Low-pass filtered correlator output when both
  delays match (zero otherwise) ?
• where G and T are magnitude and phase of
  target reflectivity, t0 and td are target and
  internal delays, and ?' is the offset frequency

15
Coherent reflectivity extraction

• Output of correlator is ALWAYS at offset
  frequency!!
• UWB transmit waveform collapses to a single
  frequency!
• We can shrink detection bandwidth at correlator
  output to enhance SNR
• Power in correlator output is proportional to G2
• I/Q detector in receiver can measure T
• Doppler, if any, will modulate correlator output
  and can be extracted from the I/Q detector
• Offset frequency usually lies between 10-15 of
  center frequency of transmission

16
What can coherency give us?

• Polarimetry
• Interferometry
• Doppler estimation
• SAR imaging
• ISAR imaging
• Monopulse tracking
• Clutter rejection
• ALL USING INCOHERENT NOISE RADAR!!!

17
Difficulty of stochastic representation
18
Dual-channel polarimetric noise radar architecture
19
Bandlimited noise waveform
Frequency domain
Time domain
20
Measured point spread functions of Channel 1 and
Channel 2
21
Approximate resolutions

• Range resolution
• where c is speed of light and B is the transmit
  bandwidth
• Doppler resolution
• where Tint is the integration time

22
Average ambiguity functions
B 1 GHz Tint 50 (L), 10 (R) ms
B 100 MHz Tint 50 (L), 10 (R) ms
23
Application examples

• Ground penetration imaging
• Arc-SAR imaging
• Polarimetric ISAR imaging
• Foliage penetration (FOPEN) SAR imaging
• Anti-jamming imaging performance

24
Ground penetration imaging
Detection of multiple objects Two metallic
plates, 17.8 cm and 43.2 cm depth
25
Ground penetration imaging
Detection of non-metallic object Distilled water
in 1 gallon plastic container, depth 7.6 cm
26
Ground penetration imaging
Detection of polarization-sensitive object
Metallic pipe, parallel to transmit polarization
and parallel to scan direction
27
Ground penetrating imaging
Detection of polarization-sensitive object
Metallic pipe, parallel to transmit polarization
and perpendicular to scan direction
28
Arc-SAR imaging
SAR image of two corner reflectors using a 1-2
GHz random noise radar
29
Polarimetric ISAR imaging
Geometry of mock airplane
RGB color composite image of mock airplane
(RedHH, GreenHV/VH, BlueVV)
30
FOPEN SAR imaging
Tree-4
Tree-3
Tree-2
Tree-3
Tree-1
Tree-4
Trihedral-2
Tree-2
Trihedral-2
Trihedral-1
Trihedral-1
Tree-1
Target scenario FOPEN SAR
image SVA enhanced image
Images of two trihedral reflectors under foliage
coverage, HH polarization
31
Anti-jamming imaging performance
Simulated ISAR images of a MIG-25 airplane no
jamming (top), LFM radar image with SJR -10 dB
(top right), and noise radar image with SJR -10
dB (right)
32
Covert communications conceptual architecture
Channel 1 is the key
33
Diversity implementations

• Polarization diversity Channels 1 and 2
  transmitted over orthogonal polariztions
• Band stacking (Frequency diversity) Channels 1
  and 2 are made to occupy contiguous spectral
  bands
• Delay diversity (Time diversity) Channel 2
  delayed and transmitted after Channel 1

34
Diversity implementation features
35
Polarization diversityTransmit waveforms

• Noise source output ?
• Horizontally transmitted waveform ?
  ? Noise
• Modulator output ?
• LSB mixer output ?
• Vertically transmitted waveform ?
• 
  ? Noise-like
• where ?0, ?c, ?m are the center frequency,
  modulator carrier frequency, and the modulating
  frequency respectively
• If , then H and V transmit
  signals occupy same frequency band!

36
Polarization diversityReceive waveforms

• Horizontally received signal ?
• Vertically received signal ?
• Amplitude limited horizontally received
• signal ?
• Amplitude limited vertically received
• signal ?
• Mixer difference output ?
• 
  ? Spectrum lies between 0 and 2d?
• Mixer sum output ?
• ? Spectrum is ALWAYS
  centered around ?c !!!

37
Noise and noise-like signal comparison
Frequency (Hz)
Time (s)
38
Amplitude and polarization angle of transmitted
signal
39
Instantaneous polarization vector

• Temporal variation of electric field vector of
  the propagating composite wave

1
4
2
5
6
3
40
BER performance without coding
41
BER performance with coding
42
Channel propagation issues
Four factors that may cause distortion
Atmospheric Absorption
Path Loss (distance)
Rain
Vegetation
TransmittedSignal
Received Signal
43
Spectral efficiency issues

• Since independently generated noise waveforms are
  uncorrelated, they can share same spectral space
• Non-interference feature is useful in MIMO-type
  polarimetric applications to avoid
  cross-polarization contamination
• In MIMO-type radar networking applications, many
  more users can be added when using noise
  waveforms compared to conventional waveforms

44
Noise waveform based networking scheme

• Ultrawideband (UWB) noise used for attaining
  spread spectrum characteristics
• UWB noise radar is used for high-resolution
  covert target detection, tracking, and imaging
• Fragmented slices within noise band can be used
  for network communications (node to node and node
  to base station)
• Camouflaged communications appears noise like
  to adversary

45
Waterfilling waveform optimization

• Waterfilling optimization maximizes mutual
  information between input and output
• MIMO noise radar has many options available for
  optimization
• Waterfilling options in radar include
  polarization, operating frequency range, transmit
  bandwidth (resolution), spectral shaping

46
Waterfilling examples in radar

• FOPEN applications Higher signal losses through
  foliage for vertical polarization (due to
  vertically oriented trees) may imply the need for
  diverting larger fraction of transmit power to
  horizontal polarization
• Imaging applications Higher bandwidth can be
  used to achieve better resolution from aspect
  locations where higher resolution is necessary to
  image finer identifying features of the target,
  while lower bandwidth (thus better spectrum
  usage) may be used from aspect locations where
  finer features may be concealed in the shadow
  region

47
Adaptive beamforming

• Adaptive beamforming has been suggested for
  sensor networks
• Individual nodes respond to commands from base
  station and coordinate their transmissions to
  accomplish coherent beamforming
• MIMO radar can greatly benefit from this approach

48
Adaptive beamforming examples in noise radar

• Noise radar nodes can receive pings from base
  station through the covert spectrally fragmented
  bands
• Standard approach would be an incoherent
  beamforming scheme since different noise
  waveforms are uncorrelated and phase
  synchronization is not possible
• Incoherent beamforming may only improve received
  power advantage by a factor of N instead of N2
• Possible to achieve coherent beamforming if
  pseudorandom noise waveform is used at each node

49
Radar tags

• Radar tag is a wireless device that can embed
  information into radar data acquisition by
  receiving radar pulses, modifying and coding
  these, and retransmitting them back to the radar
• Backscatter modulation is primarily used in
  sensor networks to interrogate remote devices

50
Applications of radar tags in noise radar

• Simultaneous tagging by each noise radar will
  not cross-pollinate other noise radars due to
  uncorrelated nature of the transmissions
• Radar tag can be designed with specific frequency
  dependence to be adaptive to environment
  conditions as viewed by each node
• Radar tags can also be used to covertly
  communicate information about target from one
  radar node to another

51
Noise radar networking advantages

• UWB Noise Radar Technology
• Noise-like transmissions for covert operations
• Large signal bandwidth, hence excellent range
  resolution
• LPI/LPD, anti-jam, and interference-resistant
  characteristics
• Efficient use of the frequency spectrum
• Low cost and compact
• Ad hoc Sensor Networks
• Deployed inside or around scene of interest
• Low-cost, low-power, untethered, multi-functional
  sensing devices
• Data-processing and communication
• Powerful protocol stack
• Fault-tolerant and scalable
• Application dependent

52
Proposed netted MIMO noise radar system
53
Possible field implementation
54
Features of proposed system

• It has LPI/LPD characteristics for detection,
  tracking, and imaging
• It can be used for covert communications and
  signaling
• It is based on a self-organized network-centric
  architecture
• The network can be used for both high and low
  data rate applications
• Network possesses high spectral efficiency

55
Combat Identification (Combat ID)
Take Aways from the Combat Identification Systems
Conference (CISC) held in Portsmouth, VA, May
23-26, 2005

• About 3-5 fatalities in war are due to friendly
  forces mistakenly targeting military targets of
  friendly forces (called fratricide)
• Problem is exacerbated due to adverse
  environmental conditions (fog/rain), harsh ground
  clutter, multitude of benign-looking target types
  (cars, etc.), crowded EM spectrum, and need to
  remain covert/LPD/anti-jam
• Solution requires multiple disciplines, such as
  sensing, communications, networking, image
  processing, fuzzy logic, information management,
  and decision sciences

56
Combat ID defined

• The process of attaining an accurate
  characterization of detected objects in the joint
  battlespace to the extent that high confidence,
  timely application of tactical military options
  and weapons resources can occur

57
Combat ID approaches

• Thermal signatures
• RF tags on vehicle
• Dynamic optical tags (DOTs) using lasers
• Millimeter wave cooperative transponder
• Microwave long range RF tags
• Digital radio frequency tags (DRAFTs)

58
Noise radar RF tag solution to Combat ID

• High spectral efficiency for dense usage
• Covert operation
• LPD capability
• Anti-jam capability
• Adaptable and diverse waveform features

59
Questions ?

• Thank You