You are on page 1of 15

ENABLING INTER-SATELLITE COMMUNICATION AND RANGING FOR SMALL

SATELLITES

R. Sun
(1)
, D. Maessen
(1)
, J. Guo
(1)
, E. Gill
(1)

(1)
Chair of Space Systems Engineering, Faculty of Aerospace Engineering, Delft University of
Technology, Kluyverweg 1, 2629 HS, Delft, the Netherlands, +31152786098, r.sun@tudelft.nl

ABSTRACT

As a first step in the development of a fully functional, reliable and efficient inter-satellite
communication and ranging system for small satellites, various key system drivers in terms of
layer-based communication architectures and technologies are analysed based on the different
operational needs for different missions. The lower three layers, being the physical, date link, and
network layers, are of primary concern. After analyzing the current existing inter-satellite ranging
methods, focuses have been put on Global Navigation Satellite System (GNSS)-like ranging
methods with respect to its signal characteristics, achievable ranging accuracy, and the feasibility of
its use to inter-satellite links.

1. INTRODUCTION

To increase mission return, future space missions can utilize multiple miniaturized spacecraft flying
together in a coordinated formation. These spacecraft will likely communicate with each other via
an Inter-Satellite Link (ISL) to enable advanced functions, e.g. relative navigation and formation
control, clock synchronization, science and spacecraft health data exchange. Flying two or more
spacecraft in a precisely controlled formation presents high requirements for ISLs, because inter-
satellite sensors are needed not only to support communication, but also to enable formation
acquisition and maintenance in a precise relative position using inter-satellite tracking. Direct inter-
satellite ranging is one of the tracking strategies. It can be applied both in Low Earth Orbit (LEO) as
an ancillary to Global Positioning System (GPS)-based relative positioning, and in Medium Earth
Orbit/Geosynchronous Earth Orbit (MEO/GEO) or non-Earth orbits that are out of the range of GPS
signals. Therefore, the combination of inter-satellite communication and ranging into a single
package with its flexible use of inter-satellite sensors on small satellites allows a wide range of
applications related to distributed spacecraft systems.
A series of missions have been flown or proposed with high demands on the inter-satellite ranging
accuracy. This includes the GRACE mission with its radiometric K/Ka-band inter-satellite tracking
at micrometer-level accuracy [1], the PRISMA mission with its S-band Radio frequency (RF)-based
metrology at centimeter-level ranging accuracy [2], and their follow-ups such as PROBA-3 [3]. The
success of these missions depends on the ability to achieve formation keeping by autonomous on-
board control, as well as on their accurate relative positioning in a precise manner. Some other
missions like NASAs New Millennium Program missions ST-3 (Starlight) [4], ST-5 [5], and
Techsat-21 [6], although aborted or heavily modified, their technologies regarding ISLs are still
valuable and inspiring. The Autonomous Formation Flying (AFF) sensor developed for ST-3 has
been modified with the intention to reuse it for the future Terrestrial Planet Finder (TPF) mission
[7]. Magnettospheric Multiscale Mission (MMS) [8] and MicroArcsecond X-ray Imaging Mission
(MAXIM) [9] are other examples that rely on ISL-based technologies to ensure mission success.
Apart from mission specific applications, inter-satellite communication and ranging allows for a
reduced ground segment, and with that reduced operation costs, as well as enhanced system
robustness and real-time operations. Therefore, this technology is of high interest both for current
and future formations.
When designing the inter-satellite sensor, it is wise to take advantage of existing hardware and
software used in commercial terrestrial communication networks. However, the space environment
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 1
does pose some challenges, such as multiple mobile nodes forming a dynamic network topology,
intermittent communication links, line of sight ranging, limited on-board power and computing
resources, and bandwidth constraints, which make the design of an inter-satellite sensor a
challenging task. With these considerations, this paper analyzes various key system drivers in order
to provide recommendations for the development of a fully functional, reliable and efficient inter-
satellite communication and ranging system for small satellites. This paper consists of three main
parts. The first one regards the basic inter-satellite communication issues, including Radio
frequency or optical links, frequency allocations, data rate, and power control mechanism for small
satellites. The second one is about inter-satellite communication network architectures and
technologies, which emphasize on how to share the channels and how to route data. The third part
of the paper summarizes the existing inter-satellite ranging methods, focusing on GNSS-like
ranging. The ranging signal characteristics, achievable ranging accuracy, signal acquisition and
tracking, and the feasibility of their use via ISLs are discussed. Concerning the limitations of power,
mass and cost for small satellites, this paper gives recommendations at the end of each part.

2. LAYER-BASED INTER-SATELLITE COMMUNICATIONS

Different operational needs for missions trigger different ISL communication requirements and
architectures. A layer-based Open Systems Interconnection (OSI) model serves as a good reference
for understanding what is needed to make communication work, as depicted in Fig. 1 [10]. As such,
the requirements levied on inter-satellite communications systems will be compatible with the basic
tenets of networking. The physical layer, related to the trade-offs among several system drivers such
as power, frequency, data rate and bandwidth, has been found to be of primary concern, since it has
the largest role to play when it comes to reliable and efficient communication via ISL.

Physical
Media, Signal, and
Binary Transmission
Data Link
MAC and LLC
(Physical Addressing)
Network
Path Determination and IP
(Logical Addressing)
Transport
End-to-End Connections
and Reliabi lity
Session
Interhost Communicat ion
Presentation
Data Representation
and Encryption
Appli cation
Network Access to
Application
Data
Data
Data
Segments
Packets
Frames
Bits
Data unit Layer
H
o
s
t

L
a
y
e
r
s
M
e
d
i
a

L
a
y
e
r
s

Fig. 1. OSI model [10]

Radio Frequency or Optical Inter-satellite Links

Radio frequency (RF) and optical/laser are the two primary communication media for an ISL.
Optical links have the ability to provide very high speed communications with data rate on the order
of Gbps[11], which is suitable for image transfer. Optical sensors are also used once the involved
spacecraft reach very short distances with respect to each other, guaranteeing very high position
accuracy. The narrow beam and high directivity of the laser affords interference-free signal transfer
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 2
and low possibility of interception. However, optical sensors tend to have a relatively small field of
view. Hence, to obtain spherical coverage one must either have a large number of them or extend
their field of regard by scanning. A more challenging part of an optical sensor is that a highly
accurate Acquisition, Tracking and Pointing (ATP) subsystem is required in order to establish and
maintain the communication link, as depicted in Fig. 2. In the process of scanning for target and
maintaining link, a direct line-of-sight is needed and a rapid relative motion between two
spacecraft will make this process more complex. Although transmission of an image by laser link
from one satellite to another has been demonstrated in SILEX [12] (Semiconductor Laser Inter-
satellite Link Experiment) with a data rate of 50 Mbps, optical communication is still a very new
technology for spacecraft.
Yes
No
No
Scan for
target
Wait for
confirmation
Generate
beacon
laser

Get
confirmation?
Yes
Maintain
link
If need
maintain link?

Fig 2. ATP assembly in optical links
Compared to optical sensors, RF sensors have problems of their own, but they appear more
manageable. The data rates achievable with an RF link are lower than with an optical link, but if in
a distributed spacecraft mission the link message traffic only consists of navigation data, spacecraft
health and status, and some science data (not including imaging information such as in
interferometric and optical mapping missions), data rates less than 10Mbps are more than adequate.
RF links can provide omni-directional coverage when considering multi-antenna combination.
Although RF equipment is subject to co-channel interference, multipath, atmospheric and man-
made noise, careful system design and use of technologies such as spread spectrum modulation can
significantly reduce interference effects in most cases. Furthermore, long term experience with
radio transmission for space-to-ground links makes RF-based inter-satellite communication more
reliable and easier to implement in space. On balance, once there is a need for a very high data rate
or a very accurate positioning requirement, optical sensors can be the solution. Otherwise, an RF
sensor is preferable for a small satellite mission.

Frequency Allocation

The appropriate frequency band (or bands) is an important part of any recommendation for inter-
satellite communications. The choice of frequency bands depends upon the spectrum regulations
specified by the International Telecommunication Union (ITU), technical characteristics and
constraints (including availability of hardware), and mission requirements.
Table 1. Frequency allocation candidates for inter-satellite communications [14]
Band Frequency Range Service Examples
S
2025 - 2110 MHz
2200 - 2290 MHz
SRS
SRS
PRISMA
TPF
Ku
13.75 - 14.3 GHz
14.5 - 15.35 GHz
SRS
SRS

Ka
22.55 - 23.55GHz
25.25 - 27.5 GHz
32.3-33.4 GHz
ISS
ISS
ISS, RNSS
Iridium
GRACE (K/Ka band)
StarLight
W
59 - 64 GHz
65 - 71 GHz
ISS
ISS

International and national spectrum regulations are an important consideration when identifying the
proper spectrum or when making frequency allocations for distributed spacecraft inter-satellite
communications. Based on the service designation by ITU, several frequency allocation options
may be available for inter-satellite communications [13, 14]: allocations to Space Research Service
(SRS); Inter-satellite Service (ISS); Radionavigation and radionavigation-satellite service (RNSS).
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 3
Table 1 provides several frequency bands that are appropriate for implementing ISLs between 1 and
100GHz as recommended by ITU. Examples of several distributed spacecraft missions with
different frequency allocations are also listed in Table 1.
Besides these spectrum regulations, the system designer needs to consider the availability of
hardware and the technical characteristics of the frequency bands. Several technical parameters
influence the selection of frequency bands for ISLs including:

- Available bandwidth and data rates;
It is well understood that the higher the frequency allocation, the wider the bandwidth available.
Based on the Shannon theorem, the maximum theoretical data rate is proportional to the bandwidth.
So, high speed data can be achieved with high frequency bands for which wide band use is
permitted. In the process of frequency selection, the different amount of transmitted data in different
missions determines the required bandwidth, which will consequently influence the frequency
allocation, as shown in table 2.
Table 2 bandwidth and data rates equivalences
Bandwidth Maximum data rate
Recommended
Frequency
allocation
narrow <100 kbps S
100kbps-1Mbps S, Ku
Medium
1Mbps-10Mbps Ku, Ka
10Mbps-100Mbps Ku, Ka
Wide
>100Mbps Ku, Ka, W

- Number of channels and/or number of links in the local distributed spacecraft network and
multiple access techniques;
Multiple spacecraft in the local distributed spacecraft network will share communication links. For
code division multiple access (CDMA), because it employs spread-spectrum technology and a
special coding scheme to allow multiple users to be multiplexed over the same physical channel, the
total data rate can be equivalent to several times the single access bandwidth, depending on the
modulation and the number of links. Therefore, sufficient spectrum is needed to accommodate this
increased data transmitted simultaneously by multiple users. For a frequency division multiple
access (FDMA) system, which assigns a sub-frequency band for each spacecraft, the bandwidth for
each channel contributes to the total spectrum. In addition, the frequency isolation between sub-
frequency bands should also be considered to mitigate mutual interference and facilitate the
bandpass filters to separate channels. When the number of channels is large, lower frequency
allocation bands (e.g. 2025-2110 MHz in S band has only 85MHz bandwidth available) may not
provide sufficient bandwidth alone when using FDMA technology. For time division multiple
access (TDMA), whose strategy is to share a single carrier frequency with multiple users at
different time slots, has the least influence to frequency allocation compared to CDMA and FDMA.

- Link performance, associated with required transmitter power, propagation, and antenna
characteristics, which can vary greatly depending upon frequency;
The free-space loss L
FS
is inversely proportional to the square of the carrier frequency f [16]:

2
( ) ( )
4 4
FS
c
L
d d
2
f

t t
= = (1)
where d is the distance between the transmitter and receiver, c is the speed of light, and is the
wavelength. Thus, lower frequency has a smaller space loss.
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 4
In addition, the RF frequency also affects the satellite transmitter power, antenna size and
beamwidth. In turn, these factors affect satellite size, mass, and complexity. Their relationship can
be expressed by the following equations [16]:
2 2 2 2 2
2 2 t
D D
G
c
t q t q

= =
f
(2)
21
GHz
f D
u =
(3)
l t
EIRP PL G =
(4)
t
G is the transmitter antenna gain, is the half-power beamwidth in degrees (-3dB lower than the
peak gain), D is the antenna diameter, q is the antenna efficiency, P is transmitter power,
l
L is
transmitter-to-antenna line loss, and EIRP is the transmitter effective isotropic radiated power.
When assuming constant EIRP and decreasing carrier frequency, the satellite antennas diameter
increases to maintain the specified beamwidth until it reaches a maximum size (or mass) limit.
Reducing the carrier frequency further requires more transmitter power to compensate for the loss
in antenna gain. On the other hand, going to higher frequencies makes the antenna beam narrower,
which means that higher frequencies are incompatible with having an omni-directional coverage.
The overall link performance can be expressed by the link budget using the received signal-to-noise
power density ratio C/N
0
of the communication system. If we assume identical transmitter power,
antenna gains, and implementation losses in both the lower and the higher frequency bands, the
received signal will be much weaker at higher frequency band.
In general, higher frequency bands enable smaller antenna size, but require more transmitter power
to compensate for all the attenuation effects, such as free-space loss, reduced effective C/N
0
, and
ultimately make the overall communication system more complex. Distributed spacecraft missions,
who usually consist of multiple small satellites with mass, power and cost constraints, are
recommended from link performance perspective to use lower frequency allocation (e.g. S band).

- Ionospheric errors, Multipath, and Doppler shifts effects;
ISLs are normally used both for the distribution of data among spacecraft and for navigation
purposes. Performing pseudorange measurements is a common way to realize relative navigation.
From J.A. Avila-Rodriguez work on the feasibility of using C-band (2-4GHz) as future GNSS
frequencies [17], we conclude that frequency allocation has an effect on the pseudorange error
budget, including ionospheric effects, carrier multipath and the signal acquisition process.
Ionospheric effects are inversely proportional to the square of the carrier frequency, and carrier
multipath is inversely proportional to the carrier frequency. Therefore, a higher frequency allocation
helps to reduce the errors caused by ionosheric path delay, and to mitigate the carrier multipath as
well. However, due to higher maximum Doppler shifts at higher frequency, the Doppler search
region increases, which negatively influences signal acquisition. Assuming identical code length,
signal acquisition takes a longer time at high frequency bands.

- The carrier frequencies for inter-satellite links must be assigned so as to avoid interference
with other onboard communication systems like the TT&C subsystem.
If the inter-satellite sensor and the TT&C subsystem work in the same frequency band (e.g. S band,
sufficient frequency separation between the inter-satellite link, and the TT&C uplink and downlink
is necessary to reduce the risk of disturbance. The PRISMA mission is a good example when
assigning carrier frequencies to its Formation Flying RF (FFRF) sensor [2]. Table 3 shows the
frequency allocation for PRISMA. Except for the consideration of separation between FFRF and
TT&C, the carrier frequency selection for FFRF also guarantees to maximize the isolation between
ISL bandwidth and deep space research bandwidth (2290 2300MHz), and provides a reasonable
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 5
inter-frequency separation (between S1 and S2) in order to optimize the integer ambiguity
resolution functionality in the phase measurement for inter-satellite ranging.

Table 3. Frequency allocation for PRISMA
FFRF
S1: 2275 MHz S2: 2105 MHz
TT&C
Carrier frequency
(S band)
TM: 2214 MHz TC: 2035 MHz

Data type, Data rates and Bandwidth

The required information to be distributed between the satellites falls into several data types or
traffic, each of which has different levels of data rate and bandwidth requirements:
- Navigation data
- Payload data
- Spacecraft health & status

Navigation data can contain the measured absolute position, relative distance, velocity, attitude, and
time information. The volume and transmission frequency of the navigation data is tightly coupled
to the nature of the mission. In case of tightly cooperating spacecraft in close formations (separation
distances < 1 km) with high positioning accuracy and tight control windows, the typical maneuver
cycle for maintenance of the formation may be too short for a ground-controlled formation and thus
many require a fully autonomous on-board control approach with the help of inter-satellite
communication. In this scenario, the frequency of broadcast of navigation data over an ISL can be
on the order of seconds or even continuously, which enables approximately real-time relative
navigation corrections to command the drifting spacecraft back within its control window. The tight
communication link requires significantly higher data rate and greater bandwidth requirements than
those missions requiring kilometer level positioning accuracy.
The need for payload data transmission via ISLs depends on knowledge of how the payload data
will be processed in the distributed spacecraft system [18]. If the common science data are
periodically dumped to ground stations without space segment processing, no science data needs to
be transmitted to other spacecraft. On the other hand, some collaborative missions may require the
exchange of science data to facilitate distributed space-based computing. Therefore, estimates of the
data rates and bandwidth for payload data present the greatest challenge because payload data type
and the degree of onboard distributed computing can be significantly different for different
missions. For missions collecting imaging information, the data rate can be up to 10Mbps, but the
frequency of transmission can be low, on the order of hours. Some technique for reducing the data
rate can be used, e.g. transmitting only the changes of the payload data [16]. The bandwidth
occupation of payload data is also related to the network topology of the communication system.
The mostly used distributed computing topology is the star topology [19], which facilitates the data
collecting, comparison and processing within the master spacecraft, but is not as bandwidth
efficient as a distributed topology.
Spacecraft health & status data is very low volume and would be broadcast less frequently than the
navigation data. The amount of this information depends upon the complexity of the equipment on
the spacecraft and the need to share the information with other spacecraft. In general, 10
3
bit is
sufficient for spacecraft health & status data, and the frequency of broadcast is on the order of
minutes [18].
Taking account of navigation, payload and spacecraft health & status data together, mission
requirements will decide the major contributors for the overall data rate and bandwidth. For
autonomous formation flying like PRISMA, navigation data dominates the inter-satellite
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 6
communication bandwidth; while for the scientific missions, science data can be the main factor
determining the communication strategy.
Furthermore, the date rate determination is also affected by the power constraints and the separation
distance between the spacecraft. Higher data rate or larger distance means higher power
consumption, and therefore higher system costs. Fortunately, for missions with large separation
distances between spacecraft, positioning accuracy and date rate requirements are normally lower
than close formations, because the control windows are getting larger along with the separation
distance. If a specific formation flying mission is divided into several phases during its lifetime,
variable data rate communication and variable position accuracy requirements can be implemented.
Table 4 shows two examples of this situation [8, 20].
Table 4. Ranging accuracy and data rate requirements at different separation distances on
PRISMA mission and MMS mission
Inter-satellite
transceiver
Spacecraft
separation
3 Ranging
Accuracy
Data Rate
10m 500m
1m (LOS*>45)
20cm (6<LOS<45)
1cm (LOS<6)
12 kbps
FFRF on PRISMA
500m 30km 1m 4 kbps
250m 640km 30m 4.096 kbps
640km 1800km 9km 512 bps IRAS on MMS**
1800km 3500km 35km 128 bps
* Line of sight
** Intersatellite Ranging and Alarm System (IRAS) on Magnetospheric Multiscale Mission (MMS)
In general, the overall process of inter-satellite communication data rates and bandwidth
determination is shown in Fig. 3. Multiple access technology should also be considered because
FDMA, TDMA, or CDMA affect bandwidth in different ways as explained in frequency allocation
part.

Data Rates
Relative motion
between Satellites
Data Update
Frequency
Data length
Separation
Distance
Inter-satellite Comm
Data Type Assessment
Control & Navigation
Requirements
Bandwidth
Spread Spectrum

Relative Motion
Between Satellites
Data Type:
Navigation, Payload,
Health & Status Data
Navigation & Control
Requirement
Data Length
Data Rate
Multiple Access
Technology

Fig. 3. Inter-satellite communication data rates and bandwidth determination

Adaptive Power Control Mechanism

For some missions, the separation distance between the satellites can vary in time from ranges of a
few meters to thousands of kilometres. Using traditional specifications, the transmit output power is
fixed and optimized for the worst-case link budget at the maximum range. Thus, power
consumption is always high. In addition, the inter-satellite link also generates excess interference to
the space-to-ground link as the distance between the two spacecraft decreases. As a result, a power
control mechanism is desired to maintain an optimal power level at all foreseen separation
distances.
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 7
Furthermore, if CDMA signal structures are used to support simultaneous communications and
ranging, the well-known near-far interference problem exists. Near-far interference is a form of
multiple access interference (MAI) in which the signals from other users will appear as noise to the
signal of interest and interfere with the desired signal in proportion to the number of users. The
interference of a signal from a transmitter that is in close proximity to the receiver reduces the
effective received signal to interference power density ratio (SIR) several times more than the signal
from a remote transmitter.
The objective of power control for inter-satellite links is to implement dynamically adjustable
power attenuation in the transmitters to provide a minimum SIR

that satisfies the signal acquisition
and tracking requirements at the receiver. Taking into account both the noise and interference, an
equivalent SIR can be expressed as [21]:
0
0
1
( )
Ri Ti i
i N
R Ri
Tx x Ti i
x
P P G
SIR
P P N B
P G P G N B
=
= =
+
+

(5)
where P is the power, G is the channel gain (loss), B is the bandwidth, N is total number of
spacecraft, and the subscript R indicates the received, T means the transmitted, and i, x indicate
different links.
Equation (5) shows that the equivalent SIR is a function of various transmitted powers and channel
gains. What's more, the channel gain is inversely proportional to the square of the distance between
transmitter and receiver. Therefore, in order to provide a minimum equivalent SIR, the transmitted
power P
Ti
should be adjusted along with the changing relative distances between any two
spacecraft. For most missions, the formation geometry does not change rapidly during the time of
the two sequential transmission duty cycles. Therefore, the information provided by the prior duty
cycle can be used as reference to estimate the current relative distances. The transmitter power P
Ti

can then be adjusted to an optimal level, as depicted in Fig. 4. Other power control algorithms exist,
such as SIR-based and position-based methods that are proposed for use in CDMA cellular mobile
radio networks [22]. Further research is needed to demonstrate their feasibility for ISLs.
Modulator TX
Data Demodulator RX
d
t=i-1
, v
t=i-1

d
t=i
Desired SIR
Power adjustment
Data
# S/C noise

interference

Fig. 4. Adaptive power control mechanism for inter-satellite links

3. INTER-SATELLITE COMMUNICATION NETWORK ARCHITECTURE

The architectural complexity of the inter-satellite communication network is depends on its multiple
access technologies and topology schemes. The data link layer and the networking layer in the OSI
model support their implementations. For the data link layer, we focus on different multiple access
technologies with respect to their specific advantages and disadvantages. For the network layer,
several topologies are compared, and the need for multi-hop routing is discussed concerning large
separation distance and strict timing constraints.
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 8

Multiple Access Technology

Three basic techniques for sharing links in distributed spacecraft systems are FDMA, TDMA and
CDMA. Table 5 summarizes their capabilities applied to inter-satellite communications, as well as a
brief overview of their advantages and disadvantages with an attempt to find the candidates for
small satellites with limitations on power, mass and cost. FDMA is not an economic choice because
it needs a large frequency bandwidth to guarantee multiple transmitters or channels assigned to
multiple sub-frequencies. TDMA with its simple timing schedule is suitable for small satellites,
especially in the situation that the number of spacecraft is not large (<10). Several near future
formation flying missions (PRISMA, DARWIN and MMS) use TDMA as their inter-satellite
multiple access technology [2, 8, 23]. The greatest challenge of TDMA is time synchronization with
a maximal synchronization error that is equal to the propagation time of the signal (e.g. smaller than
100s when the separation distance is less than 30km). CDMA is also an efficient way, especially
when considering using GNSS-like signals to realize inter-satellite ranging. However, the near-far
interference problem requires a power control scheme at the transmitters. The performance of
CDMA decreases with increasing number of spacecraft because of mutual interference. The same
constraint of limited number of spacecraft also exists in TDMA, but here the reason is that a large
scalable number of spacecraft makes the cycle period of the TDMA sequence too long for data that
is broadcasted with high update frequency (especially navigation data, which may be required for
real-time data processing). Therefore, both CDMA and TDMA can be used for inter-satellite
communications in small scale formations with small satellites.
Table 5. Multiple Access (MA) technologies in inter-satellite communications
MA Advantages Disadvantages
FDMA
Multiple S/C transmissions can occur
simultaneously
No complex timing
The larger the number of S/C, the greater the frequency
band allocation required for the mission
Multiple S/C cannot share the same transmitter without
mutual interference
Frequency isolation between sub-frequency bands is
needed to mitigate the mutual interference
Requires complex bandpass filters to separate channels
Increased cost due to frequency variation
Difficult filtering to separate large power signals from
adjacent spacecraft
May require power control
TDMA
Single frequency needed
Multiple S/C can use the same
transmitters with high efficiency
Simple timing logic easily separates
large numbers of S/C
Prevent interference from other S/C
completely
Inter-satellite communication can only occur at specific
time slots
The overall throughput performance is reduced since each
S/C must wait their turn to access the shared frequency
Time synchronization is needed
Signal transmission delay varies along with the different
separation distances between S/C. Trade-off of the proper time
slots needs to be made to avoid signal collision and guarantee
efficient channel occupation as well
The greater the number of S/C, the longer the duty cycle
of a TDMA sequence that transmits information once among
all the S/C
CDMA
Multiple S/C transmissions can occur
simultaneously
GPS-like relative ranging method can
operate simultaneously with
communication
Relatively immune to transmitter
distortion and interference
Limited number of S/C due to mutual interference (noise
generated by undesired S/C signal)
Near-far interference: different separation distances
between S/C cause various signal power levels at the receiver
Closed-loop power control scheme may be needed to
tightly control transmit power
Less bandwidth efficient than FDMA and TDMA
Complex signal processing needed
When the number of spacecraft is getting large, hybrid versions involving mixes of these three
multiple access techniques are possible. For example, the Crosslink transceiver (CLT) developed by
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 9
the Applied Physics Laboratory (APL) in Johns Hopkins University considers to apply hybrid
FDMA/CDMA for full duplex inter-satellite communications. This combination can support a
scalable number of spacecraft and meanwhile achieve sufficient isolation between the transmit
signal and receive signal [24].

Topology

The communication topology is tightly coupled to the nature of the mission and the number of
spacecraft. Canonical topologies generally used for communications within distributed spacecraft
systems include centralized (star), fully distributed (mesh), and hierarchical forms.
The centralized formation flying topology consists of a center or mothership spacecraft and
several daughter spacecraft, such as in the MAXIM mission [9]. The mothership acts as the
control point or data collector point for the formation, since it normally has a stronger data
processing capability. Information is passed from the daughter spacecraft to the mothership via
ISLs. However, the mothership is a single point of failure due to its unique capabilities and its
centralized role within the formation. The distributed formation flying topology consists of several
spacecraft that have similar capabilities, such as in the MMS mission [8]. A hierarchical topology is
the combination of the previous two topologies and divides the formation into manageable subsets.
Due to this, it works well for larger formations. A comparison of the different topologies is provided
in Table 6 [24].
Table 6 Comparison of different communication topologies used in distributed spacecraft system
Topology Advantages Disadvantages Examples
Centralized
(Star)
Simple design
N-1 links for N nodes;
Relies on the capability of the central
resource, which is responsible for primary
control and information dissemination
Potential faults within the central
resource greatly influences the
implementation of whole mission
DARWIN
MAMIX
TPF

Distributed
(Mash)
Support direct interaction among all
distributed assets
Fault tolerate for each node
Real-time communication to each
node
Topology architecture is N(N-1)/2 for N
nodes
Rapid growth in complexity as N
increases
Resource limitations such as
communication bandwidth and processing
capability.
GRACE
MMS
Hierarchical
Robustness is supported
Control structure complexity
depends on the functional relationships
between S/C
Needs multilevel approach;
Not necessary for small distributed
missions.


Multi-hop Routing

Whereas the aforementioned topologies with direct inter-satellite communications are the primary
and relative simple solutions, some indirect communication strategies such as dynamic network
routing with multiple hops are potentially applicable to communicate over long distances that are
greater than the normal transmit power can support. Data can be relayed through various spacecraft
in order to achieve the desired end-to-end communication objectives that cannot be realized by
direct, line-of-sight inter-satellite communications.
However, both the changing formation geometry and the choice for best paths to relay data pose
challenges to the routing strategy. Distributed missions are consequently taxed with relatively
demanding on-board processing requirements. For some missions with strict time constraints,
especially in a formation where the satellites need to maintain a precise relative position and
orientation, a possible end-to-end latency requirement may be imposed (e.g. an upper bound of
signal transmission latency is required in real-time control systems). This implies constraints on the
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 10
multi-hopping because the signal end-to-end delay depends on the number of traversed satellites
(processing, queuing, and propagation delay) [18].
Other indirect inter-satellite communication links exist, such as using a ground station or via the
common commercial LEO communication constellations (Iridium, Orbcomm, or Globalstar) as a
relay. Their feasibility depends on the communication architectures and differs from mission to
mission.

4. INTER-SATELLITE RANGING

Inter-satellite Ranging Methods

Depends on the mission requirements, inter-satellite ranging can be performed in many ways. Fig. 5
depicts a breakdown of the various options. First, a choice needs to be made whether or not to
perform the ranging in a direct or an indirect manner. The indirect method comprises the subtraction
of two measurements for which the difference yields the inter-satellite range. This can be done
using GNSS measurements from onboard GNSS receivers, tracking data via ground station/
Tracking and Data Relay Satellite System (TDRSS)/ Precise Range and Range Rate Equipment
(PRARE) [15], or two-line elements as provided by North American Aerospace Defense Command
(NORAD).

Inter-satellite
ranging
Indirect Direct
One-way Two-way
GNSS
Ground station
trackin
GNSS-like
Pulse
g phase or code
PRARE
trackin
Range rate
/Do
Transponder
g ppler shift
TDRSS
tracking
Two-line
elements

Fig. 5. Inter-satellite ranging options
If indirect measurements are not available, not exact enough, or if it is desired to obtain the range in
a more autonomous manner, a direct range measurement between the satellites needs to be made.
This can be done using a one-way signal or a two-way signal. The latter can be split into pulsed
(radar or lidar) and transponder-based methods while the former can be divided into GNSS-like or
range rate/Doppler shift methods. Note that if the clocks on the two satellites are perfectly
synchronized, accurate one-way ranging using pulses is possible. However, since this is very hard to
achieve in practice, this method is not considered to be a feasible option.
The one-way ranging methods suffer from ambiguity (GNSS-like, phase or code measurements) or
do not directly provide a range measurement (range rate measurements e.g. interferometric). An
interferometric measurement namely provides the range rate and needs to be processed using an
estimator to yield the inter-satellite range. Due to the repetitive nature of the GNSS-like signals,
phase and code measurements on the other hand suffer from ambiguity, which needs to be resolved
using a proper system design. This can be done by using multiple signals with different frequencies
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 11
to resolve the ambiguity for phase measurements and by using very long codes or by adding
additional timing data to resolve the ambiguity for code measurements.

GNSS-like Inter-satellite Ranging Signal Characteristics

Inter-satellite ranging is based on locally generated inter-satellite ranging signals. A cost effective
manner to generate these signals is to modify an existing GPS receiver such that it can operate as a
transceiver. This could result in a highly miniaturized and accurate ranging device that performs the
ranging using a one-way signal modulated with a GPS-like pseudo random noise (PRN) code. The
PRN code acts as a noise-like carrier that is used for bandwidth spreading of the signal energy, and
for mitigating the deleterious effects of reflected and interfering signals. The auto-correlation
property of the PRN code, resulting in a sharp peak, allows accurate range measurements. The chip
rate of the PRN code will decide the transmission bandwidth. The Coarse/Acquisition (C/A) code
with chip rate of 1.023Mcps and the P(Y) code with chip rate of 10.23Mcps are two types of PRN
code used by GPS.
Several existing inter-satellite transceivers resemble the GNSS paradigm. They include the FFRF
sensor for PRISMA [2] and Proba-3 [3], the Autonomous Formation Flying (AFF) sensor for ST-3
[4], the Constellation Communications and Navigation Transceiver (CCNT) for ST-5[5], IRAS for
MMS [8], and the Star Ranger for Techsat-21[25].
It is a fact that GPS-like signal can be used for inter-satellite ranging. However, other GNSS code
waveforms exist, such as the signals used for Galileo [26]. The GPS-like signal structure uses
BPSK-R (Binary Phase Shift Keying with rectangular spreading symbols) modulation, while the
Galileo-like signal uses a Binary Offset Carrier (BOC(a,b)) structure, which offers significant
changes in time and spectral domain. The BOC modulation is a square sub-carrier modulation,
where a signal is multiplied by a rectangular sub-carrier of frequency f
sc
(f
sc
= a*1.023Mcps) equal
to or higher than the chip rate f
c
(f
c
= b*1.023Mcps).
Fig. 6 shows the properties of C/A code, P code and BOC code in terms of auto-correlation and
amplitude spectrum. The BOC signal has a sharper auto-correlation peak than the C/A code when
they have the same chip rate. In addition, the C/A and P code place most of the signal power in the
middle frequencies of their bands, while BOC signals split the spectrum into two symmetrical
lobes. These properties make BOC signals have a better ranging performance in many aspects, such
as ranging accuracy, multipath effect, and interference mitigation effect.

Fig 6. Normalized auto-correlation and amplitude spectrum properties of C/A, P and BOC code


Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 12
Ranging Accuracy

Regarding the GNSS-like method, the ranging accuracy is significantly influenced by the selection
of the ranging code in several aspects: chip rate, code length and code modulation. The higher the
chip rate, the wider the signal bandwidth (e.g. the spectral main lobe of the P code is ten times
wider than that of the C/A code). The sharper auto-correlation peak results in smaller multipath and
a more accurate estimate of the signal arrival time. Regarding the code length, when we use a
longer code, we get a lower cross-correlation and less interference among the signals transmitted by
the various spacecraft [27].
Besides the influence of the signal structure, the ranging accuracy is determined by the type of
range measurements. Basically, two available kinds of measurements (code/pseudorange and carrier
phase) are biased. The main error contributors are atmospheric errors, clock offset, multipath and
receiver noise (mainly thermal noise). For formation flying missions in LEO, ionospheric path
errors dominate the atmospheric errors, but fortunately ionospheric models and dual-frequency
measurements can be considered to reduce these errors. Dual one-way ranging can be used for
cancellation of clock offset. In addition, differential measurements take advantage of the fact that
the errors associated with the satellite clock and atmospheric propagation are similar between
receivers or time epochs, which can be used to improve ranging accuracy. By combining code and
carrier phase measurements (e.g. carrier-smoothing of the code measurement), the code multipath
and receiver noise can also be smoothed to a large extent.

Signal Acquisition and Tracking

The acquisition and tracking for a GNSS-like signal proceeds in two stages. The first stage is a
global search over frequency and code in a two dimensional search space for approximate values of
Doppler shift and pseudorange. This process, known as signal acquisition, requires long integration
times to get sufficiently high carrier to noise ratio (C/N
0
). Therefore this process is time and
processing consuming. Fast Fourier Transform (FFT) can be used to create a one-dimensional
search space, but at the cost of having to perform FFT.
The second stage is a local search for accurate estimates of pseudorange and carrier phase. This
process is called signal tracking because it is continuous and the estimates are updated as the
satellites move. It is accomplished as a feedback control loop, called a delay lock loop (DLL),
which continuously adjusts the replica code to keep it aligned with the received code in the
incoming signal. The integration time can be shorter since the C/N
0
does not need to be very high
anymore. Another feedback control loop, the phase lock loop (PLL), is used to track the carrier
phase for highly accurate positioning. Doppler shift or delta pseudorange can also be measured via
the PLL.

5. CONCLUSION

This paper presents various inter-satellite communication and ranging technologies. To this end,
important communication system design drivers are analysed in the lower three layers of the OSI
layer-based communication protocol, and several inter-satellite ranging methods are summarized.
The current analysis is based on the general mission requirements for small satellites. Future work
will follow for a specific mission scenario, especially in the field of autonomous on-board
navigation with high ranging accuracy requirements. The functionality of the future inter-satellite
sensor will be augmented to enable relative navigation and formation control, common time, and
communication in one package for advanced on-board autonomy.




Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 13
6. REFERENCES

[1] Bertiger W., Bar-Sever Y., Desai S., et al. GRACE: Millimeters and Microns in Orbit,
Proceedings of ION GPS, Portland, OR, 24-27 September 2002.
[2] Lestarquit L., Harr J., Grelier T., et al. Autonomous Formation Flying RF Sensor
development for the PRISMA Mission, 19th International Technical Meeting for the Satellite
Division (ION GNSS), Fort Worth, TX, 26-29 September 2006.
[3] Wishart A., Teston F., et al. The PROBA-3 Formation Flying technology Demonstration
Mission, 58th International Astronautical Congress, Hyderabad, India, 24-28 September
2007.
[4] Aung M., Purcell G.. H., Tien J. Y., et al. Autonomous Formation-Flying Sensor for the
StarLight Mission, Proceedings of the International Symposium Formation Flying Missions
and Technologies, Toulouse, 29-31 October 2002.
[5] Bar-Sever, Y., Srinivasan J., et al. From AFF to CCNT: JPL's evolving family of
multifunction constellation transceivers. 2
nd
International Workshop on Satellite Formation
Flying, Haifa, Israel, 2001.
[6] Zenick R. G., Kohlhepp K., GPS Micro Navigation and Communication System for Clusters
of Micro and Nanosatellites, Small satellites conference, UT, Sep. 2000.
[7] Tien J. Y., Srinivasan, J. M., Young, L. E., et al, Formation acquisition sensor for the
Terrestrial Planet Finder (TPF) mission, IEEE Aerospace Conference, Big Sky, Montana, 6-
13 March, 2004.
[8] Heckler G., Winternitz L., Bamford W., MMS-IRAS TRL-6 Testing, 21st International
Technical Meeting of the Satellite Division (ION GNSS), Savannah, GA, 16-19 September
2008.
[9] Cash W. C., MAXIM: The Micro Arcsecond X-ray Imaging Mission, SPIE Proceedings
Interferometry in Space, Feb, 2003.
[10] Kerri L. K., Michael A. P. Intersatellite Links: Lower Layer Protocols for Autonomous
Constellations, First Joint Space Internet Workshop, NASA Goddard Space Flight Center,
Maryland, November 2000.
[11] Pasquale Maurizio De Carlo, Leonardi Roberto, et al. Intersatellite link for Earth
Observation Satellites constellation, http://www.corista.unina.it/Docs/intersatellite_link.pdf
[12] Gilles Planche, Vincent Chorvalli, SILEX in-orbit performances, Proceedings of the 5th
International Conference on Space Optics, Toulouse, France, 30 March - 2 April, 2004.
[13] Bernard L. Edwards, Distributed Spacecraft Crosslink Study Part 1: Spectrum requirements
and allocation survey report and recommendations, Goddard Space Flight Center, May
2002.
[14] National Telecommunications and Information Administration, Manual of Regulations &
Procedures for Federal Radio Frequency Management, U. S. Government Printing Office,
Washington, D.C., May, 2003.
[15] Montenbruck O., Gill E., Satellite Orbits: Models, Methods, Applications, First edition, pp
193-199.
[16] James R. Wertz, Wiley J. Larson, Space Mission Analysis and Design, Third edition, pp 535-
585.
[17] Avila-Rodriguez J.A., Wallner S., Hein G.W., Eissfeller B., A vision on new frequencies,
signals and Concepts for future GNSS systems, Proceedings of the International Technical
Meeting of the Institute of Navigation (ION GNSS), Fort Worth, Texas, 25-28 September,
2007.
[18] Bernard L. Edwards, Distributed Spacecraft Crosslink Study Part 2: Distributed Spacecraft
Crosslink Communications System Requirements Report, Goddard Space Flight Center, May
2002.
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 14
Small Satellite Systems and Services Symposium (4S), Funchal, Portugal, 31 May 4 June 2010 15
[19] Abdul-Halim Jallad, Vladimirova T., Distributed Computing for Formation Flying Missions,
Proceedings of the IEEE International Conference on Computer Systems and Applications,
Dubai/Sharjah, UAE, 08 - 11 March, 2006.
[20] Harr J., Delpech M., Grelier T., Seguela D., The FFIORD Experiment: CNES' RF Metrology
Validation and Formation Flying Demonstration on PRISMA, 3rd International Symposium
on Formation Flying Missions and Technologies, Noordwijk, the Netherlands, 23-25 April
2008.
[21] Thompson W. L., Israel D. J., Adaptive Power Control for Space Communications, IEEEAC
paper #1188, 2007.
[22] Das S., Ganu S., Rivera G., et al. Performance analysis of downlink power control in CDMA
system, IEEE Sarnoff Student Symposium, Princeton, NJ, 2004.
[23] Bourga C., Mehlen C., et al. A Formation Flying RF Subsystem for DARWIN and SMART-2,
International Symposium Formation Flying: Missions & Technologies, Toulouse, France,
29-31 October 2002.
[24] Stadter P. A., Heins R. J., Chacos A. A., et al. Enabling Distributed Spacecraft Systems with
the Crosslink Transceiver, AIAA Space Conference and Exposition, Albuquerque, NM, 28-
30 August 2001.
[25] Zenick R. G., Kohlhepp K., GPS Micro Navigation and Communication System for Clusters
of Micro and Nanosatellites, Small satellites conference, UT, Sep. 2000.
[26] Betz J. W., Binary Offset Carrier Modulations for Radionavigation, Journal of the Institute
of Navigation, vol. 48, no. 4, pp. 227-246, 2001.
[27] Misra P., Enge P., Global Positioning System: Signal, Measurements and performance,
Second edition, pp 68-78.