CSE475
Satellite and Space Networks
ÖMER KORÇAK
omer.korcak@marmara.edu.tr
Marmara University
Department of Computer Engineering
Satellite Networks Research Laboratory (SATLAB)
Department of Computer Engineering,
Boğaziçi University
Istanbul, Turkey
OUTLINE
• Satellites
– GEO, MEO, LEO
• High Altitude Platforms
• Integration Scenario
• Problem Definition & Solution Approach
2
SATELLITES
Distance: 378.000 km
Period: 27.3 days
3
Basics
• Satellites in circular orbits
– attractive force Fg = m g (R/r)²
– centrifugal force Fc = m r ²
– m: mass of the satellite
– R: radius of the earth (R = 6370 km)
– r: distance to the center of the earth
– g: acceleration of gravity (g = 9.81 m/s²)
– : angular velocity ( = 2  f, f: rotation frequency)
• Stable orbit
– Fg = Fc
2
gR
r3
2
(2 f )
4
Satellite period and orbits
24
satellite
period [h]
velocity [ x1000 km/h]
20
16
12
8
4
synchronous distance
35,786 km
10
20
30
radius
40 x106 m
5
Basics
elliptical or circular orbits
complete rotation time depends on distance satellite-earth
inclination: angle between orbit and equator
elevation: angle between satellite and horizon
LOS (Line of Sight) to the satellite necessary for connection
 high elevation needed, less absorption due to e.g. buildings
 Uplink: connection base station - satellite
 Downlink: connection satellite - base station
 typically separated frequencies for uplink and downlink
– transponder used for sending/receiving and shifting of
frequencies
– transparent transponder: only shift of frequencies
– regenerative transponder: additionally signal regeneration





6
Inclination
plane of satellite orbit
satellite orbit
perigee
d
inclination d
equatorial plane
7
Elevation
Elevation:
angle e between center of satellite beam
and surface
minimal elevation:
elevation needed at least
to communicate with the satellite
e
8
Link budget of satellites
• Parameters like attenuation or received power determined by four
parameters:
 sending power
 gain of sending antenna
 distance between sender
and receiver
 gain of receiving antenna
• Problems
L: Loss
f: carrier frequency
r: distance
c: speed of light
 4 r f 
L

 c 
2
 varying strength of received signal due to multipath propagation
 interruptions due to shadowing of signal (no LOS)
• Possible solutions
 Link Margin to eliminate variations in signal strength
 satellite diversity (usage of several visible satellites at the same time)
helps to use less sending power
9
Atmospheric attenuation
Attenuation of
the signal in %
Example: satellite systems at 4-6 GHz
50
40
e
rain absorption
30
fog absorption
20
10
atmospheric
absorption
5° 10°
20°
30°
elevation of the satellite
40°
50°
10
Satellite Orbits
Distance (km)
Period
Low Earth Orbit (LEO)
700 - 2000
~2 hr
Medium Earth Orbit (MEO)
10.000 – 15.000
~6 hr
Geosynchronous Earth Orbit (GEO)
36.000
24 hr
11
Satellite Orbits 2
GEO (Inmarsat)
HEO
MEO (ICO)
LEO
(Globalstar,
Irdium)
inner and outer Van
Allen belts
earth
1000
Van-Allen-Belts:
ionized particles
2000 - 6000 km and
15000 - 30000 km
above earth surface
10000
35768
km
12
GEO Satellites
•
•
•
•
•
No handover
Altitude: ~35.786 km.
One-way propagation delay: 250-280 ms
3 to 4 satellites for global coverage
Mostly used in video broadcasting
– Example: TURKSAT satellites
• Another applications: Weather forecast, global
communications, military applications
• Advantage: well-suited for broadcast services
• Disadvantages: Long delay, high free-space attenuation
13
MEO Satellites
•
•
•
•
•
•
Altitude: 10.000 – 15.000 km
One-way propagation delay: 100 – 130 ms
10 to 15 satellites for global coverage
Infrequent handover
Orbit period: ~6 hr
Mostly used in navigation
– GPS, Galileo, Glonass
• Communications: Inmarsat, ICO
14
MEO Example: GPS
• Global Positioning System
– Developed by US Dept. Of Defence
– Became fully operational in 1993
– Currently 31 satellites at 20.200 km.
• Last lunch: March 2008
• It works based on a geometric principle
– “Position of a point can be calculated if the distances between this point
and three objects with known positions can be measured”
• Four satellites are needed to calculate the position
– Fourth satellite is needed to correct the receiver’s clock.
• Selective Availability
• Glonass (Russian): 24 satellites, 19.100 km
• Galileo (EU): 30 satellites, 23.222 km, under development
(expected date: 2013)
• Beidou (China): Currently experimental & limited.
15
LEO Satellites
•
•
•
•
•
•
Altitude: 700 – 2.000 km
One-way propagation delay: 5 – 20 ms
More than 32 satellites for global coverage
Frequent handover
Orbit period: ~2 hr
Applications:
– Earth Observation
• GoogleEarth image providers (DigitalGlobe, etc.)
• RASAT (First satellite to be produced solely in Turkey)
– Communications
• Globalstar, Iridium
– Search and Rescue (SAR)
• COSPAS-SARSAT
16
Globalstar
•
•
•
•
•
•
•
•
•
Satellite phone & low speed data comm.
48 satellites (8 planes, 6 sat per plane)
and 4 spares.
52˚ inclination: not covers the polar
regions
Altitude: 1.410 km
No intersatellite link: Ground gateways
provide connectivity from satellites to
PSTN and Internet.
Satellite visibility time: 16.4 min
Operational since February 2000.
315.000 subscribers (as of June 2008)
Currently second-generation satellites
are being produced (by Thales Alenia
Space) and 18 satellites launched in
2010 and 2011.
17
Globalstar – Coverage Map
18
Iridium
•
•
•
•
•
•
•
66 satellites (6 planes, 11 sat per plane)
and 10 spares.
86.4˚ inclination: full coverage
Altitude: 780 km
Intersatellite links, onboard processing
Satellite visibility time: 11.1 min
Satellites launched in 1997-98.
Initial company went into bankrupcy
– Technologically flawless, however:
– Very expensive; Awful business plan
– Cannot compete with GSM
•
•
•
•
Now, owned by Iridium satellite LLC.
280.000 subscribers (as of Aug. 2008)
Multi-year contract with US DoD.
Satellite collision (February 10, 2009).
19
COSPAS-SARSAT
(Search And Rescue Satellite Aided Tracking)
•
•
•
•
•
An international satellite-based
SAR distress alert detection &
information distribution system.
4 GEO, 5 LEO satellites
Aircraft & maritime radiobeacons
are automatically activated in case
of distress.
Newest beacons incorporate GPS
receivers (position of distress is
transmitted)
Supporters are working to add a
new capability called MEOSAR.
– The system will put SAR processors
aboard GPS and Galileo satellites
Since 1982, 30.713 persons rescued
in 8.387 distress situation.
20
Satellites - Overview
• GEOs have good broadcasting capability, but long
propagation delay.
• LEOs offer low latency, low terminal power
requirements.
• Inter-satellite links and on-board processing for
increased performance and better utilization of satellites
– From flying mirrors to intelligent routers on sky.
• Major problem with LEOs: Mobility of satellites
– Frequent hand-over
• Another important problem with satellites:
– Infeasible to upgrade the technology, after the satellite is
launched
21
High Altitude Platforms
(HAPs)
•
•
•
•
Aerial unmanned platforms
Quasi-stationary position (at 17-22 km)
Telecommunications & surveillance
Advantages:
– Cover larger areas than terrestrial base
stations
– No mobility problems like LEOs
– Low propagation delay
– Smaller and cheaper user terminals
– Easy and incremental deployment
• Disadvantages:
– Immature airship technology
– Monitoring of the platform’s movement
22
HAP Coverage
23
HAP-Satellite Integration
•
•
•
HAPs have significant advantages.
Satellites still represent the most attractive solution for broadcast and
multicast services
Should be considered as complementary technologies.
24
An Integration Scenario
HAPs
Satellites
• Integration of HAPs and mobile satellites
• Establishment of optical links
25
Optimal Assignment of
Optical Links
CONSTRAINTS:
• A satellite and a HAP should have line of sight in order to
communicate with each other.
– Elevation angle between HAP and satellite should be larger than a
certain εmin value.
• Number of optical transmitters in satellites is limited.
– A satellite can serve maximum of Hmax HAPs.
• One to many relation between HAPs and satellites.
AIMS:
• As much HAP as possible should be served (Maximum utilization)
• Average of elevation angles between HAPs and satellites should be
maximized.
26
Optimal Assignment of
Optical Links (cont.)
SOLUTION APPROACHES
• This optimization problem can be represented as an Integer Linear
Programming (ILP) problem
• ILP solution approaches: Exclusive search, Branch-and-bound
algorithms, etc.
– Exponential time complexity
– Not feasible for large networks
• Optimization algorithm should be applied repeatedly
– In periodic manner: every ∆t time unit
– In event-driven manner: When a link becomes obsolete
• Faster algorithm is necessary.
• There exists a polynomial time solution approach
27
Solution Approach
•
•
•
•
Example scenario
Three satellites & seven
HAPs
Visible pairs are
connected
Elevation angles are
given on the links
28
Solution Approach (cont.)
•
•
•
•
•
Bipartite graph
First group: Node for each
satellite transmitter
Second group: Node for each
HAP
Edge exists between a HAP
and each transmitter of a
satellite, if they are visible to
each other.
Weights: Elevation angles
Maximum weighted maximum cardinality matching
29
Max Weighted Max Cardinality Matching
• Matching: A subset of edges, such that no two edges
share a common node.
• Maximum cardinality matching: Matching with
maximum number of edges.
• Maximum weighted maximum cardinality matching:
Maximum cardinality matching where the sum of the
weights of the edges is maximum.
• Hungarian Algorithm: O(n3)
30
Results
100 HAPs (height: 20 km)
10 MEO Satellites: ICO (height: 10350 km, inclination: 45˚)
Total time: 1 day, Δt=1 minute
31
Increasing the Link Durations
• Matching with maximum “average elevation angle” may
result in frequent optical link switching
– Switching from one satellite to another is an expensive
operation.
• Reduction of switching = Increasing the link durations
• Method: Favor existing links with a particular amount γ
– Weights of utilized edges are incremented by γ
32
Results - 2
εmin=-2
33
Net gain function
G( ) 
L( ) 
LD( )  LD(0)
LD(0)
Aavg (0)  Aavg ( )
Aavg (0)
NGF ( )    G ( )  (1   )  L( )
“Efficient Integration of HAPs and Mobile Satellites
via Free-space Optical Links,” Computer Networks, 2011
More on Satellites
•
•
•
•
Hand-over
Satellite-fixed / Earth-fixed footprints
Network Mobility Management
Routing
35
Handover in satellite systems
•
Several additional situations for handover in satellite systems compared to
cellular terrestrial mobile phone networks caused by the movement of the
satellites
– Intra satellite handover
• handover from one spot beam to another
• mobile station still in the footprint of the satellite, but in another cell
– Inter satellite handover
• handover from one satellite to another satellite
• mobile station leaves the footprint of one satellite
– Gateway handover
• Handover from one gateway to another
• mobile station still in the footprint of a satellite, but gateway leaves the
footprint
– Inter system handover
• Handover from the satellite network to a terrestrial cellular network
• mobile station can reach a terrestrial network again which might be
cheaper, has a lower latency etc.
36
Satellite-fixed vs Earth-fixed
Footprints
Satellite Fixed Footprints
(asynchronous handoff)
Earth Fixed Footprints
(synchronous handoff)
37
Virtual Node
Routing
is corresponds
carried out without
considering
38
A satellite
in a time.
 Physical network
topology
is dynamicto aVN
Virtual
Node topology is fixed
the movement of satellites
Routing
• One solution: inter satellite links (ISL)
 reduced number of gateways needed
 forward connections or data packets within the satellite network as
long as possible
 only one uplink and one downlink per direction needed for the
connection of two mobile phones
• Problems:
 more complex focusing of antennas between satellites
 high system complexity due to moving routers
 higher fuel consumption
 thus shorter lifetime
• Iridium and Teledesic planned with ISL
• Other systems use gateways and additionally terrestrial networks
39
Features of Satellite Networks
• Effects of satellite mobility
–
–
–
–
Topology is dynamic.
Topology changes are predictable and periodic.
Traffic is very dynamic and non-homogeneous.
Handovers are necessary.
• Limitations and capabilities of satellites
– Power and onboard processing capability are limited.
– Implementing the state-of-the-art technology is difficult.
– Satellites have a broadcast nature.
• Nature of satellite constellations
– Higher propagation delays.
– Fixed number of nodes.
– Highly symmetric and uniform structure.
40
Routing & Network MM
• Considering these issues various routing & MM
techniques are proposed. Main ideas are:
– To handle dynamic topology changes with minimum overhead.
– To prevent an outgoing call from dropping due to link handovers
– To minimize length of the paths in terms of propagation delay
and/or number of satellite hops.
– To prevent congestion of some ISLs, while others are idle (Load
balancing).
– To perform traffic-based routing.
– To provide better integration of satellite networks and terrestrial
networks.
– To perform efficient multicasting over satellites.
“Exploring the Routing Strategies in Next-Generation Satellite Networks”
IEEE Wireless Communications, 2007
41
Thank You
Any questions?