Transcript Deep-Space Optical Communications

Optical Wireless
Communications
Prof. Brandt-Pearce
Lecture 8
Deep-Space Optical
Communications
1
Outline
 Deep-Space Optical Communications
 Introduction
 Channel Model
 System Performance
 Optical Deep-Space Network
 RF/FSO Hybrid System
2
3
Deep-Space Communications
 Sending and receiving data from space crafts has been a
challenging problem since 1950s
 Communication over deep-space distances is extremely
difficult, much more difficult than satellite communications
 Communications beams spread as the square of the distance
between the transmitter and the receiver
4
Deep-Space Optical Communications
 The distance from Earth to Neptune or Pluto can be on the
order of 4,000,000,000 km. After propagating over such a
distance, the communications beam from a spacecraft will
spread to an area 10 billion times (100 dB) larger in area than
if the beam from the same system traveled from just the GEO
distance (40,000 km).
 A system capable of transmitting 10 Gbps from GEO to the
ground would only achieve 1 bps from Pluto/Neptune
distances.
5
Deep-Space Optical Communications
 Optical communications has lower divergence compared to RF
 Comparison of RF and optical beam spreads from Saturn.
6
Deep-Space Optical Communications
 An important factor for a high data-rate deep-space optical
link is the laser transmitter
 Lasers are required to have
 High output power
 Low divergence
7
Deep-Space Optical Communications
 Another key technology component is a thermally stable
and lightweight optical spacecraft telescope.
 Similar to satellite communications, for a small beam
divergence, tracking and pointing plays an important role
in the reliability of deep-space optical links
 This pointing must be accomplished in the presence of
attitude changes of the host spacecraft that are perhaps a
thousand times larger than the laser beam divergence.
8
Growth of the Deep-Space Comm. Capacity
9
Hamid Hemmati, “Deep Space Optical Communications”, Jet Propulsion Laboratory, California Institute of Technology, 2005
Deep-Space Communications
 Optical deep-space communications can be implemented
in two ways:
 Direct optical link: A direct optical link is set up between
the earth station and space-craft
 Atmosphere disperses and attenuates the transmitted and
received signals
 High power transmitter and large receivers can be used
 Indirect optical link: the optical signal is sent from a
satellite outside the atmosphere
 Atmosphere effect is mitigated
 Transmitter and receiver sizes are limited
10
METOL
MARS-EARTH Terahertz Optical Link
5 W 1.54 micron Laser
1 - 10 Gbps
5W 26 GHz 100
Mbps (RF)
Small Lander UHF:
128 kbps (150 Mb in 20 minutes)
11
Channel Model
 Cloud opacity is an atmospheric physical phenomenon that
jeopardizes optical links from deep space to any single ground
station
 Clearly, when clouds are in the line-of-sight, the link is
blocked
 Ground receiving telescopes need to be located in sites where
cloud coverage is low and statistically predictable
 To guarantee continuity of data delivery from deep space to
ground, while the Earth is rotating, a global network of
telescopes is necessary
 The selection of the sites for telescopes belonging to an optical
deep space network (ODSN) is driven by considerations
based, among other factors, on cloud-cover statistics
12
Channel Model: Atmospheric Transmittance
 Main Gases composing the Earth Atmosphere
13
Channel Model: Atmospheric Transmittance
 Earth atmospheric number density profiles for individual species
14
Channel Model: Atmospheric Transmittance
 Transmittance spectrum at sea level with zenith angle of zero.
15
Channel Model: Sun Irradiance
16
Channel Model: Sky Irradiance
 Sky radiance spectrum experienced at an observation point at sea level for 23 km
of visibility and Sun zenith angle of 45 deg while observer zenith angle varies as
10, 40, and 70 deg
17
Deep Space Optical Communications
 Merits of five deep-space communication link wavelengths.
18
Hamid Hemmati, “Deep Space Optical Communications”, Jet Propulsion Laboratory, California Institute of Technology, 2005
Deep Space Optical Communications
 Data of a NASA optical link between Earth and Mars
 Modulation scheme: 256-ary PPM
 BER: 10-3
 Bit-rate: 1 Mbps
 Range: 3.59 × 108 km
19
Hamid Hemmati, “Deep Space Optical Communications”, Jet Propulsion Laboratory, California Institute of Technology, 2005
Optical Deep Space Network
 To support deep space missions aimed to the exploration of the
universe for the last four decades, NASA has designed and operated
a global network of radio-frequency ground stations termed the
Deep Space Network
 A similar network can be used for optical communications called
optical deep-space network (ODSN)
 Today NASA’s DSN only requires three radio-telescope hubs to
successfully operate the network. The DSN stations (located at
approximately 120 deg of separation around the Earth: Goldstone,
California; Madrid, Spain; and Canberra, Australia) allow
continuous coverage of deep space from Earth
20
Optical Deep Space Network
 Since the laser transmitter beam width from space covers a limited
area on Earth it is necessary that the ODSN consists of a number of
ground stations located around the Earth as a linear distributed
optical subnet (LDOS)
 The idea behind LDOS is to have the spacecraft always pointing at
a visible station belonging to the LDOS
 When either the line of sight is too low on the horizon (20 deg of
elevation) or is blocked by atmospheric conditions (e.g., clouds or
low transmittance), the spacecraft beam is switched to a different
station (or network node) by pointing to the adjacent optical ground
station
21
Optical Deep Space Network
 Example of LDOS (star = telescope) architecture for an optical
deep space network (ODSN)
22
Hamid Hemmati, “Deep Space Optical Communications”, Jet Propulsion Laboratory, California Institute of Technology, 2005
Global Sites for Deep-Space
Optical Communications
23
System Model
 Usually the received photon count is very low
 PMTs are used to detect signal
 The operation temperature of the space-craft is low
 Thermal noise is proportional to the temperature: 𝜎 2 =
4𝑘𝐵 𝑇 Δ𝜈
 Hence, shot noise is the dominating noise
 Poisson statistics should be used for analysis
24
System Model
 For OOK:
 Probability density functions for transmitting “0” and “1” when
 K𝑆 =Data average photon count/pulse
 K 𝐵 =Background average photon count/pulse
 Then
 Pr photon count = 𝑘 𝑏 = 0) =
1 𝑘
K exp(−K 𝐵 )
𝑘! 𝐵
 Pr photon count = 𝑘 𝑏 = 1) =
1
𝑘!
K 𝐵 + K𝑆 𝑘 exp(− K 𝐵 + K𝑆 )
 As discussed before, threshold is where the two pdf’s become equal
 Threshold = K𝑆 log 1 + K𝑆 /K 𝐵
1
2
 BER = −
1
2
𝑇ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑
𝑘=0
−1
Pr 𝑘 𝑏 = 0) − Pr 𝑘 𝑏 = 1)
1
2
 When K 𝐵 =0, Threshold=0 and BER = exp(−K𝑆 )
25
Performance of Deep-Space Optical Communication
 For PPM
 Symbol error probability is
 For Poisson distribution
where
 In the absence of background light
1 −𝐾
𝑃𝑏 = 𝑒 𝑠
2
26
Performance Analysis of OOK
 BER versus signal level for uncoded OOK signaling on a Poisson
channel, for various background levels
27
Performance Analysis of PPM
 BER of uncoded PPM on a Poisson channel, versus Ks
28
Performance Analysis of PPM
 BER of uncoded PPM on a Poisson channel, versus Pav =
Ks /M
29
FEC in Deep-Space Optical Comm.
 Due to the low received power the BER is high
 BER is usually 0.001
 Forward error correction (FEC) is used to decrease BER down to 10-15
 Deep-space optical systems use high order PPM since they have high energy
efficiency
 Reed-Solomon codes are used as FEC
 High-order PPM modulation (256-PPM) with a high alphabet (8-bit
alphabet) RS code
 Accumulator (product) codes:
30
Outline
 Deep-Space Optical Communications
 Introduction
 Channel Model
 System Performance
 Optical Deep-Space Network
 RF/FSO Hybrid System
31
RF/FSO Hybrid System
 Radio-Frequency (RF) Communications
 Low bandwidth
 Stable Channel
 Relatively immune to cloud blocking
 Sometimes affected by heavy rain
 Free-Space Optical Communications
 High Data Rate
 2.5 Gbps commercially available (Tbps demonstrated)
 Bursty Channel
 Must have clear / haze conditions
 Less degradation than RF in rain
32
Combining RF and FSO System
 Enables FSO Communications bandwidth without giving up RF
reliability and “adverse-weather” performance
 Improves network availability: Quality of Service (QoS)
 More options for adapting to weather
 Common atmospheric path effects and compensation (directional links)
 Physical Layer diversity improves jam resistance
 Size, Weight and Power Focus
 Leverages common power, stabilization, etc.
 Economical use of platform volume
 Enables seamless transition of free space optical communications
into RF Environment
33
Average Data-Rate of a Hybrid FSO/RF
AVERAGE DATA TRANSFER RATE
OF HYBRID FSO/RF LINK
AVERAGE DATA RATE (Gb/s)
3
FSO 2.5Gb/s
2
1
RF 10Mb/s
0
0
10
20
30
40
50
60
70
80
90
100
FSO LINK AVAILABILITY (%)
34
Applications
 Short range applications:
 Mesh networks
 Cross-divide links (rivers, canyons, etc.)
 Indoor systems
 Long-range applications:
 Air-to-air links
 Satellite links
 Wireless basestation connectivity
35
Hybrid RF/FSO Point-to-Point Link
 Either switching between technologies or simultaneous use
 Joint modulation/coding across two technologies
 With channel state information, can optimize throughput
 Without channel state information, can use variable-length
codes (fountain codes)
36
Hybrid FSO/RF
 Two different modulations are assumed for RF and FSO links
with constellation sizes of M1 and M2
 The links are assumed to operate synchronously
 R1 and R2 are the data rates
 Let C1 and C2 be the capacity of RF and FSO channel
respectively (Ci is a function of Ri)
 From Shannon capacity we have
 Then the throughput is
37
Optimal Joint Modulation/Coding
38
Short Range Hybrid RF/FSO Network
39
Hybrid RF/FSO Networks
 Considering that FSO link has a higher cost, only a given
number of FSO links can be used in an RF/FSO system
 Assume that an RF network is given
 The problem is to find the best choices for replacing RF with
an FSO link
 This depends on the topology, distances between nodes and
the availability of FSO link (depends on the weather
condition)
40
Hybrid RF/FSO Networks
 Formulate the problem as follows
 The problem is to maximizes the following function
 where
 Network is modeled with a directed graph G=(N,L)
 i ∈ N denote the nodes in the network
 B is the number of demands
 lij ∈ L denote the directed link from node i to node j.
 f (b)ij represent the flow of traffic on link lij
 Dij is an indicator function of an FSO link from node i to node j
 One unit time is divided into fractions represented by λk, k = 1,2, ..., K
41
Hybrid RF/FSO Networks
 The maximization is subject to
 Input and output flow is equal
for intermediate nodes
 Input flow is zero for source nodes
 Output flow is zero for sink nodes
 Flow has to be positive
 Sum of the time fractions is one
 The maximum number of FSO links is M
42
Hybrid RF/FSO Networks
 Here RF capacity is
CRFij=100 Mb/s and
CFSOij represent the
capacity of FSO links
between nodes i and j
 This problem can be
solved using mixed
integer
linear
programming (MILP)
 Optimal
throughput
and bounds for the 16
node grid network and
28-node random.
43