Optical Wireless Communications

Optical Wireless
Prof. M. Brandt-Pearce
Lecture 6
Ultraviolet Communications
 Introduction
 Sources and Detectors
 Benefits and Challenges
 Applications
 Channel Modeling
 Modulation Techniques
Ultraviolet (UV) Light
 UV region is divided into three parts
 UVA (315 nm - 400 nm)
 UVB (280 nm - 315 nm)
 UVC (100 nm - 280 nm)
 Photons in UV region have higher energies, and therefore, large
numbers of them are harmful for human health
 Large fraction of the UV from the Sun is filtered by the ozone layer
in upper atmosphere
 The filtered UV light is from 200 nm to 280 nm
UV Sources
 LEDs are inefficient in UV and have low power (~0.5 mW).
 Have to use large arrays as optical sources
 Lamps
 UV Lamps are cheap and can generate high power
 The transmitted beam has a significantly large angle
 Are appropriate for networking purposes
 Fluorescent lamps without an internal phosphor coating emits
UV light
 Two peaks at 253.7 nm and 185 nm due to the peak emission of
the mercury: 85%-90% of the produced UV is at 253.7 nm
Slow modulation
UV Sources
 Laser
 UV lasers are divided into two type:
o UV light is directly generated by the lasing process: these kind
of lasers have low output powers
o Third or fourth harmonic generation is used to generate UV light
from visible light: higher output powers can be achieved, but the
size of the laser is large.
UV Detectors
 APDs are immature for UV technology
 Have low gain in UV
 Have small aperture (μm’s)
 Have low responsivity for UV, but huge gain
 Collects background light from a wide spectrum: an optical
filter is required to limit the bandwidth of the incident light
 Still the best option for UV communications
NLOS Optical Communications
 In some situations direct path may not be available.
 Therefore, line-of-sight (LOS) optical communication is not
 Non-line-of-sight (NLOS) communications is the option that would
be interesting for these cases.
 NLOS optical communication can be easily done when refractive
surfaces (buildings, clouds, sea surface) are available
 But what if they are not available or reliable?
 The solution is optical scattering
 In FSO systems, the transmitted optical signal can be scattered in
different directions using aerosols and molecules
UV for NLOS Communications
Why UV is suitable for NLOS communications?
It has higher atmospheric scattering compared to visible
and infrared bands
 The scattering is done via molecules and aerosols
The background radiation in the UV range (200-280nm) is
low due to the filtered sun light by the ozone layer
Unlike the LOS communications, fog, rain, sand storm,
and pollution increase the scattered power and accordingly
the received power level
Challenges of UV Communications
Limiting Factors for Bit-Rate in NLOS UV Links
 Inter-symbol interference (ISI)
 The power scattered from any particle inside the common
area of the transmitted signal and receiver field of view is
Area in which
received by detector
scattering happens
 Scattered energies from particles
inside the common volume travels
different paths to get to receiver
The energy transmitted at a certain time is received in
different times
Transmitted pulses are subjected to a high temporal dispersion
This cause inter-symbol interference (ISI)
Ultraviolet (UV) Communications
Limiting Factors for Bit-Rate in NLOS UV Links
 By increasing the range, transmitted beam angle, or receiver
filed of view, the common volume become larger, and hence,
the ISI becomes worse
 Received power
Since the power is received via scattering, the channel has a
great loss
 The detector receives very weak powers
 By increasing the range, received power reduces
 NLOS UV communication is suitable for short-range links
Applications of NLOS UV Communications
 Used when line-of-sight communication is not possible
 Used for short range (<4km) and low-rate (<5Mb/s) communications
 For example:
 In urban area as a backup network
 For military applications in the battlefield
Applications of NLOS UV Communications
Link Geometry of NLOS UV Systems
Side View
Top View
Channel Modeling: LOS
 In order to get an accurate performance analysis for a LOS
UV system, the channel needs to be modeled
 Loss is limiting issue
 The impulse response is short – no ISI
 For LOS channel:
 The free space loss at distance r:
 Atmospheric attenuation:
 Ke : extinction factor
 Receiver gain:
 LOS path loss:
Channel Modeling: NLOS
 Methods for calculating the impulse-response and link loss
 Simulation methods
Analytical approaches
 For small transmitted beam angle and small receiver field of view
the received power can be calculated as follows
 Received power density at distance r1
 Portion of the power scattered from volume V to receiver is
Channel Modeling
 Simulation Methods for calculating the impulseresponse and link loss
Monte-Carlo (impulse-response and link loss):
In this method one photon in each trial is sent and after
simulating the scattering effect if it is in the receiver field of view it
is counted as a received photon.
By repeating this trial for many times the ratio between the
received photons and transmitted ones determine the channel gain.
 Numerical integration (impulse-response and link loss):
 The common volume is divided into smaller differential
 The received energy and its corresponding time via each volume
is calculated.
 The impulse response and link loss are calculated using this
differential received energies
Numerical Integration
for Channel Modeling
Transmitter gain profile:
Receiver gain profile:
Received energy via δV:
Total received energy:
Numerical Integration vs. Monte Carlo
 Model the channel faster than Monte-Carlo method
Numerical Integration
O (N)
O (P)
Error Bound
O (N -2/3)
O (P -1/2)
N : Number of volumes in numerical integration
P : Number of tries in Monte-Carlo simulation
 Able to model the channel in the presence of shadowing
Experimental Results
 Path loss versus distance, for different Tx and Rx elevation angles
G. Chen, et. al. , “Experimental evaluation of LED-based solar blind NLOS communication links”,
OPTICS EXPRESS, Vol. 16, No. 19, 2008
Experimental Results
 Path loss versus Tx elevation angles for different Rx elevation angles
G. Chen, et. al. , “Experimental evaluation of LED-based solar blind NLOS communication links”,
OPTICS EXPRESS, Vol. 16, No. 19, 2008
Modulation Techniques
 On-Off Keying (OOK)
 For NLOS, UV channel is usually time-variant
 Finding optimum threshold may not be easy
 Pulse Position Modulation
 Does not require threshold to make an optimum decision
 Because of the high ISI effect in NLOS UV, is more
susceptible to interference
 Spectral Amplitude Coding
 Can increase the data-rate by providing M-ary transmission
 Symbols are spectrally encoded signals
 Similar to PPM, does not require a threshold
 Since PMTs are used for detection, the receiver is shot noise
 For shot-noise limited system SNR is:
 G : PMT gain
 η : detector efficiency
 h : Planck’s constant
 c: light speed
 Pr: received power
 R: bit rate
 By Gaussian approximation, the BER is
 Gaussian approximation may not be valid
 Can receive very few photons.
 If no background noise or dark current,
1 −
where ns is the mean number of photons received
For ns = 11, BER= 10-5!
 With background noise or dark current:

 −( + ) ( + )
 − ( )
Spectral Amplitude Coding
Transmitter structure
 Using a diffraction grating the spectral content of the UV
LED or laser is divided into small spectral bins
 An encoder mask is used to block or pass the decomposed
spectral components
Spectral Amplitude Coding
Transmitter structure
 If we use harmonic generation as UV source, the spectral
encoding can be done in the following form
 Encoding done in visible domain (blue or green)
 Encoded signal converted to the UV range using a harmonic
Spectral Amplitude Coding
Receiver Structure
APD based receiver
 Let F spectral bins be used to encode
the signal
F APDs are used for detection
 Each APD detects one spectral
 Decision is made using the outputs
of the F APDs
 Small aperture size of the APDs
limit the FOV
 Low gain of APDs limit the receiver
Spectral Amplitude Coding
Receiver Structure
PMT based receiver
PMT has much higher gain compared to APD (G≈106)
PMTs have larger aperture size
Two PMTs used for symbol detection
Decoder mask changed M times in each symbol time
Numerical Results
Maximum attainable bit rate versus the distance for BER of 3× 10-5

similar documents