High-Speed Digital Signal Processing for

Satellite Communications
Chirag Koli
Department of Computer Engg
Vidyalankar Institute of Technology,
VIT
Mumbai, India
Email-idchirag.koli@vit.edu.in

Sumeet Kamble
Department of Computer Engg
Vidyalankar Institute of Technology,
VIT
Mumbai, India
Email-idsumeetkamble22@gmail.com

Abstract
New digital signal processing (DSP)
technology
opportunities,
wherein
signals of large bandwidth can be
digitally processed and complex digital
signal processing algorithms can be
implemented, make the use of new
techniques that have heretofore been
limited to audio/voice applications now
feasible
for
space
applications.
Although many efforts are being
devoted to the technology aspects of
this work, ongoing work by the
authors is concerned with the unified
consideration of the theoretical and
practical aspects of high-speed digital
signal
processing
and
their
applications in space communications.
In this context, this paper comprises a
review of some basic high-speed
digital signal processing techniques
that are applicable in the satellite
communications domain. Specifically,
this paper treats the subjects of
numerically stable algorithms for highspeed DSP, adaptive signal processing
techniques for demodulation of multiaccess communications, interference
suppression
in
wideband
communications,
on-board
signal
processing for demultiplexing and
transmultiplexing, and oversampling
techniques applied to sigma-delta
modulation and related schemes.
Keywords demultiplexing, transmultiplexing,
oversampling

I.

INTRODUCTION

The field of digital technology is dramatically

evolving due to the development of higher

Rohit Mishra
Department of Computer Engg
Vidyalankar Institute of Technology,
VIT
Mumbai, India
Email-idrohoitmshr9@gmail.com

integration capacities, faster analog-to-digital

(A/D) and digital-to-analog (D/ A) converters, and
other high-speed digital technologies. The
potential applications of this new technology have
a direct impact on satellite communications. First,
they allow one to consider digital solutions for
functional blocks that were traditionally analog in
existing satellite payload. For instance, quite
promising in this context are digital solutions for a
significant part of the radio-frequency (RF) front
end and their associated functional blocks (routing,
downconversion, and filtering). Secondly, new
digital processing possibilities arise for future
satellite payloads, such as beamforming for
example. And finally, complex digital signal
processing (DSP) algorithms for small ground
stations (VSAT type or portable terminals) can
now be implemented, thereby significantly
improving ground-terminal performance. A very
promising possibility is this context is the use of
advanced digital signal processing for mobile
communications terminals (e.g., multipath
cancellation, fading countermeasures, multipleaccess interference suppression, etc.).
These new technology opportunities, wherein
signals of large bandwidth can be digitally
processed and complex digital signal processing
algorithms can be implemented make the use of
new techniques that have heretofore been limited
to audio/ voice applications now feasible for space
applications. Although many efforts are being
devoted to the technology aspects of this work, the
present work is concerned with the unified
consideration of the theoretical and practical
aspects of high-speed digital signal processing and
their applications in space cornmunications. 'The

purpose of this paper is to provide an overview of

key fundamental issues that arise in the use of
high-speed DSP technology for the processing of
wideband communication signals.
II.

CORE IDEA

Small satellite communication systems

are usually power limited rather than bandwidth
limited. The small solar-cell area of these small
satellites provides power in tens of watts rather
than in the hundreds or thousands of watts of
large geosynchronous satellites. consequently,
the overall power efficiency of the entire
spacecraft communication system should be
optimized. This includes not just the efficiency of
the modu1at~on itself but also the efficiency of
the required power amplifier (which must have
the proper linearity), and the power consumption
of the circuitry itself. Cynetics has previously
investigated using DSPs for modulation and
demodulation in small satellite communication
systems. At that time, the DSPs consumed up to
several watts more power than discrete
implementations. In fact, in many cases, the
DSPs would have consumed more power than
was actually being transmitted. Since small
satellites have limited power available, discrete
implementation of the modulation and
demodulation circuitry were preferred. The saved
power was available for the payload, for
improved signal-to noise ratios, for smaller
antennas, or for increased data rates.
However, the decrease in power
consumption in the newer DSPs is obtained
partly from the decreased feature size on the DSP
chips. This decreased feature size causes an
increase in susceptibility to radiation.
Fortunately, DSP products have reached the
maturity where radiation testing has begun on
some of the available chips.
This paper discusses the power
consumption and radiation hardness of DSPs. It
then mentions possible uses for DSPs on small
satellites. The paper concludes that the newer,
power-efficient DSPs are suitable for use in the
communication systems in small satellites, but
that they must be properly shielded to achieve
the necessary radiation hardness.

III.

DETAILED DESCRIPTION

In the application of DSP in wideband space

communications, it is likely (or even necessary)
that implementation will take place at the wide
bandwidths of the front end. Since this front-end
(i.e., RF) bandwidth is much larger than the
information
bandwidths
underlying
the
modulated and interfering signals, some signal
processing difficulties may arise in the
implementation of the required processing
functions.
The delta version of the Levinson
problem offers a number of advantages over the
standard one for high-speed processing.
However, the Levinson formulation has several
useful properties that are not obviously present in
its delta counterpart. Such issues as lattice
implementation for layered adaptivity, Schur
realization for parallelizability, and direct forms
for calculating reflection coefficients with
covariance data, fall within this category of
problems.
Progress has been made on some of these
issues within the delta context.
The ideas developed for the Levinson problem
can be used in many problems of DSP that arise in
space communications.
Demultiplexing and transmultiplexing are
among the main on-board signal processing
functions for satellite communications. For
example, on might consider the demultiplexing of
a frequency division multiplexing (FDM) system
consisting of 625 channels of 30kIIz each, which
yields a total signal processing bandwidth of
18.750 MHz. As envisioned, this demultiplexing
task would be in support of mobile personal
communications services in which many srnall on
ground terminals will access a geostationary
satellite via an agile multibeam antenna system
featuring channel multiplexing/demultiplexing
performed at the individual antenna ports. Because
of the very wide processing bandwidth involved in
this scenario, on-board high-speed signal
processing is a central issue in the development of
suitable DSP algorithms implementation of such a
demultiplexer (DMUX).
Since transmultiplexers (TMUX's) are used for
conversion between FDM and time-division
multiplexing (TDM), the DSP technology issues
involved in TMUX design are quite relevant to the
DMUX problem noted above [94]. From a signal
processing point of view, transmultiplexing

essentially involves the use of filter banks with

precisely controlled phase characteristics. Thus,
carefully chosen DSP techniques are of critical
importance in the design of TMUX's, and TMUX
design for wideband signal groups is a further
problem in which high-speed DSP methods must
play a central role.
IV. DSP USES ON SMALL SATELLITES

satellite could allow communication among a wide

range of previously incompatible terminals.
4.3 Signal Conditioning. DSPs can
perform onboard signal conditioning for control
and instrumentation on the satellite. 4.4
Control. A DSP which is under utilized on other
tasks can be used as a time-shared controller for
onboard tasks, including satellite and payload
control.
V.

4.1 General: Since DSP chips are primarily

microprocessors, they can be used for a variety of
tasks onboard small satellites. Of course since DSP
architectures and instruction sets are optimized for
signal processing, their primary uses will be for
signal
processing
tasks,
1nclud1ng
communications. In addition to communications
several other possible uses are mentioned below.
4.2 Communications. DSP chips can perform
modulation and demodulation at linear-baseband"
frequencies. For communications, the DSP can
also perform:
1. Data Compression. The DSP can perform
straight compression of data of any type using the
appropriate compression algorithm. Image
,compression can be performed on st111 frames.
(Full-motionvideo compression is beyond the
capability of present DSPs).
2. Image Processing. Images can be preprocessed
onboard
the
satellite
before
transmission. This can be viewed as another form
of data compression.
3. Forward Error Correction (FEC). FEC can
be performed by the DSP. This can be done in realtime for low data rates, of off-line or in parallel for
higher data rates.
4. Memory Error Correction (MEC). As
mentioned previously, single and multiple event
upsets can be expected in spacecraft memory.
DSPs can be used for MEC to protect 5 against
these upsets. This could be done as a shared task
by the same DSP used for communications
modulation and demodulation, provided the data
rates are sufficiently low.
5. Trans-Modulation. Since modulation and
demodulation in DSPs are performed by
calculations, any modulation type can be used,
provided that the required calculation rate is within
the capacity of the DSP. Thus, a message can be
received on an FSK uplink and re-transmitted on a
BPSK downlink. In this way, a DSP-based small

V.

CONCLUSION

VI.

Digital signal
Processor integrated circuits,
DSPs,
now
achieve
sufficiently
low
power
consumptions to allow their
use on power-constrained
small satellites. These devices
can be used for several
communications
tasks,
including: modulation and
demodulation,
data
compression,
image
processing, forward error
correction, memory error
correction,
an
transmodulation.

VII.

DSPs can also be

used on a timeshared or
independent basis for signal
conditioning and control.
However,
the
DSPs
themselves,
and
off-chip
memories (especially DRAM,
as opposed to SRAM), are
susceptible to radiation. The
DSPs can be expected to have
total-dose limits in the range
of 5 to 50 Krads for existing
devices. Since satellites in
low-earth
orbits
can
experience total dose radiation
levels of 100 Krads over four
years,
proper
radiation
shielding of DSPs and
memories is required. In
addition, error detection and
correction of memory should
be
considered.
Where
possible, masked-ROM (readonly-memory) should be used
for program memory. (With

masked-ROM
program
memory,
instructions
are
encoded
in
metal
interconnections which are
radiation-hard by their nature.)
with proper radiation shielding
and error correction, DSPs can
provide
enhanced
communications and on-board
processing for small satellites.
VIII.

ACKNOWLEDGEMENT

VII.

Here, we would like to thank the people who

have made this IEEE paper of ours possible.
Firstly, we would like to thank our institute
Vidyalankar Institute of Technology and the
University of Mumbai for providing us with this
excellent opportunity of understanding and
creating this IEEE paper
Secondly, we would like to thank our guide
Prof. Rasika Ransing for her ample guidance
and encouragement which helped us to
complete this paper

IX.
1.

2.

REFERENCES

USING DIGITAL SIGNAL PROCESSORS ON

SMALL SATELLITES by Don K. Lefevre, Daniel
J. Mu1a11y, Mark A. Heiberger.
High-Speed Digital Signal Processing for Satellite
by H. Vincent Poor and Sergio Verdlfi