J Sign Process Syst (2012) 69:11–21
DOI 10.1007/s11265-011-0640-8
SDR OFDM Waveform Design for a UGV/UAV
Communication Scenario
Christian Blümm & Christoph Heller & Robert Weigel
Received: 30 September 2011 / Revised: 28 October 2011 / Accepted: 7 November 2011 / Published online: 30 November 2011
# Springer Science+Business Media, LLC 2011
Abstract In the course of a national research project,
EADS Innovation Works has developed a waveform on a
self-designed hybrid Software Defined Radio (SDR) plat-
form, consisting of an FPGA (Xilinx Virtex 5) and a GPP
(Intel Atom). The waveform realizes a video link between
an Unmanned Ground Vehicle (UGV) and its Ground
Control Station (GCS). This link is established indirectly
with an Unmanned Aerial Vehicle (UAV) acting as a relay,
in order to enable non-line-of-sight (NLOS) communication
and to increase the communication range. Additionally, a
direct video link from the UAV to the GCS is set up
simultaneously as well as multiple low data rate control
channels between the individual link partners. To cope with
diverse operation areas, accompanied by a heterogeneous
set of requirements, the waveform must offer outstanding
adaptability and flexibility to maximize data rates in each
scenario, while maintaining link robustness. The developed
waveform is based on OFDM with freely customizable
modulation parameters. Time Division Multiple Access
(TDMA) is implemented on medium access layer to switch
between the different users in the scenario (UGV, UAV,
GCS). In this article, we give an overview of the waveform,
its implementation on a hybrid platform and its validation
for the intended field of application.
Keywords SDR . OFDM . Waveform . Data links . UAV .
UGV
C. Blümm (*) : C. Heller
EADS Innovation Works,
Munich, Germany
e-mail: christian.bluemm@eads.net
R. Weigel
University of Erlangen-Nuremberg,
Erlangen, Germany
1 Introduction
Autonomous mobile robot systems with UGVs or UAVs
are increasingly used for military purposes. For civil use,
there is also a wide range of conceivable applications,
where small unmanned platforms are more flexible, higher
available and cheaper than traditional manned systems.
Equipped with adequate sensors, autonomous systems
could for instance detect hazardous substances at an early
stage and monitor their spread. Other examples are Search
and Rescue (SAR) activities, security services, image and
video recording, environmental mapping and -monitoring,
the surveillance of major sport events, etc. In combination
with satellite navigation technology any captured informa-
tion could be georeferenced exactly.
Typically, the UGV/UAV data is sent without much
preprocessing to the GCS. To therefore guarantee a reliable
communication link, the special needs of UAVs/UGVs and
their environmental conditions must be taken into account
[1–3].
This article focuses on an appropriate self-designed
waveform, which is based on an OFDM physical layer in
combination with a TDMA medium access (MAC) protocol.
This work is part of the project SiNafaR (Safe navigation
for autonomous robotic platforms), which was launched by
the Bavarian ministry of state for economy, infrastructure,
traffic and technology to develop methods and processes
for the use of cooperating air and ground robots in civil
applications.
2 The Considered Communication Scenario
The waveform which is presented here can cope with a
broad variety of communication scenarios. Nevertheless, a
12
specific set-up was defined for a proof of concept. It will be
introduced in the following.
UGV communication in rough terrain is often hindered
by a missing line-of-sight (LOS) data link between UGV
and the GCS. A promising way to overcome this problem is
the use of a UAV as relay, optimized for acting as
autonomously as possible. A second advantage of this
approach beside the NLOS compatibility is the increased
available communication distance between UGV and its
GCS.
Figure 1 depicts the considered communication set-up.
The UGV is equipped with scenario-specific sensors and a
camera. The UAV is acting as a communication relay for
the UGV, and in addition, it carries another camera to
monitor the UGV. For this set-up unidirectional video links
are needed in parallel to bidirectional control links from the
UGV to the GCS (indirect link) and from the UAV to the
GCS (direct link). There will be no direct link between the
UGV and the GCS.
The main technical needs within this scenario are given
as follows. For each video link, a minimum data rate of
3 Mbit/s is required, when low-quality MPEG 2 compres-
sion is chosen [4]. However, better video quality is desired,
to be traded off against its need for higher bandwidth.
The control links are low-bandwidth data streams and
thus negligible for throughput considerations. Another
difference between the control link and the video link is
the height of tolerable bit error rate: Video stream errors
lead to a decreased video quality, whereas the worst case of
a defective control link would be a complete system
failure.1 The aspect of different reliability demands of the
two link types must be kept in mind throughout the
complete design process, starting with an adequate system
interface: Both link types provide the Internet Protocol (IP),
but differ in the transport layer. For the video link, the
minimal protocol User Datagram Protocol (UDP) suffices,
whereas the critical control link covers the Transmission
Control Protocol (TCP) to guarantee reliable message
delivery. The difference is that UDP is based on continuous
unidirectional transmission, no matter if the message
reaches its destination or not. In contrast, TCP offers the
concepts of message acknowledgment, retransmission and
timeout [5].
The set-up in Fig. 1 is intended for heterogeneous
operational environments. To comply with their inevitably
varying demands, the waveform needs to behave differently
for maximum performance. Hence, it must be customizable
to allow different configuration parameters tailored for
1 or both. Consequently, the decision was made to design a
Even in case of a system failure, the UAV can be prevented from
crashing by navigating it with a redundant remote control, as used for
model helicopters
J Sign Process Syst (2012) 69:11–21
Radio link
UGV-UAV
0,5-1 km
Radio link
LOS UAV UAV-GCS
5-10 km
LOS
NLOS
UGV scenario
narrowband, high-available control-data
link, bidirectional, about 115 kbit/s (TCP/IP) GCS
broadband mission-data link, unidirectional,
about 10-20 Mbit/s (UDP/IP)
Figure 1 The UGV/UAV communication scenario.
varying environments and varying channel characteristics.
The video quality, for example, can be improved dramat-
ically under good channel conditions, as soon as the
waveform parameters fit the channel quality.
Scenario- and OFDM- specific major channel challenges
include high relative speed between the communication
partners and frequency-selective fading channels, e.g. due
to multipath in urban environments [6].
Beside the waveform, also the SDR hardware, on
which the waveform is implemented, needs to account
for the specific scenario requirements. Beside other, the
most obvious requirement is a small and lightweight
form factor, suitable for mobile robots and flying
platforms.
In summary, the three most important needs to be
covered are the outstanding waveform parameter flexibility
on runtime, the need for two logical links at once (video/
control) and the need for compact SDR hardware.
There are already a couple of promising SDR
platforms available, which address similar goals. In
some cases, even with OFDM reference designs provid-
ed. On the open-source side, two of the most popular
platforms are GNU Radio with its Universal Software
Radio Peripheral (USRP) family [7, 8] and the Rice
University’s Wireless Open-Access Research Platform
(WARP) [9–11]. In the commercial field, the popular
SFF SDR Development Platform from Lyrtech [12] or
Sundance’s SDR ‘baseband’ developers’ platform series
(e.g. SDR 3.0 or SDR 4.0) [13] come close to the
requirements of our communication scenario. However,
among all mentioned platforms and others [14], no
optimal solution was found which could be used without
major adaptations regarding the hardware, the waveform
completely new SDR OFDM waveform from scratch for a
novel hardware platform.
J Sign Process Syst (2012) 69:11–21 13

1.6 GHz Xilinx Interface

3 Hardware Platform Intel Atom GPP Virtex 5 FPGA Board

SPI

Control
A compact three-staged hardware stack was developed Linux OS

MAC

FSM
Config Regs.
Serial,
according to the photo in Fig. 2 and the corresponding

LP
Monitor

16 bit DAC
I/O Expansion (Spartan 3)
Application Software

PCIe Interf.
block diagram in Fig. 3. It is used in exactly the same

Dual
Ethernet
Driver

Frontend Board
PCIe
PCIe Interf.
configuration for the UGV, the UAV and the GCS. It DDR

LP
TX
combines a nanoETXexpress-SP Board from Kontron with USB
I/Q

Processing
Baseband
a 1.6 GHz Intel Atom Processor Z530 [15], a Xilinx Virtex

Packet Buffers

ADC
14 bit

LP
5 FPGA, coming as Express Card version from Pico

Waveform
Control
Ethernet DDR
Computing [16] to be inserted in a PCIe-based slot of the RX

ADC
14 bit
GUI

LP
Kontron board and a self-designed interface board. As I/Q

system interface, this platform features Gigabit Ethernet to

connect to the mission control computers of UGV, UAV and
Figure 3 Hardware stack block diagram.
the GCS. Via this interface, video and control data streams
are exchanged in an IP packet format. The Kontron GPP
board serves as router here. 3.2 The Virtex 5 FPGA

3.1 The Interface Board
The hardware platform is combinable with any propri-
etary or commercial-off-the-shelf (COTS) RF transceiv-
er hardware. For this reason, a front-end interface board
was designed to offer analog baseband I/Q signal ports.
The board comprises a dual 16 bit 500 MSPS DAC for
the transmitter path (currently 14 of 16 available bits in
use) and two 14 bit 125 MSPS ADCs for the receiver
path.
A downside of the Virtex 5 FPGA Express Card is
its limited number of free I/O ports. This number would
be too small for the exchange of full 14 bit I and Q
signals. The workaround to this problem is I/O port
expansion with a second FPGA, a low-cost Xilinx
Spartan 3. Therefore, data is exchanged between the
Virtex 5 and the interface board on both edges of the
clock signal. More precisely, I data is exchanged on
rising edges and Q data on falling edges. This double
data rate transfer (DDR) is then resolved to single data
rate transfer using the Spartan 3.
Low pass filters are included at the frontend I/O ports
against aliasing effects.
In order to achieve high signal processing performance, all
the OFDM baseband processing is done within the Xilinx
Virtex 5 FPGA. Beside baseband processing, another major
function block of the FPGA is the TDMA MAC control,
realized as a finite state machine (FSM).
Multiple first-in-first-out (FIFO) buffers are imple-
mented for temporary IP packet storage in order to
decouple the data stream, as it passes from the GPP-
based application layer to the FPGA-based MAC layer.
As already mentioned, this interface is realized with
PCIe, and it serves not only for exchanging payload
data, but also for configuring the waveform: All flexible
waveform parameters are set on runtime according
PCIe-fed signals, which are stored in a configuration
RAM register in the FPGA. Polling this configuration
register for every new data frame allows frame-wise
parameter updates.
Beside payload data and parameter control, the third
signal type which is transmitted via PCIe contains
information about relevant internal FPGA signals. This
information is required to evaluate the suitability of the
current waveform parameter set. Details of these monitor-
ing signals are given in Section 7.

3.3 The Atom General Purpose Processor

The FPGA configuration register which was described in

Figure 2 Photo of the hardware stack.
the last section is updated by the GPP, more precisely by its
central component, the waveform control block. This block
is primarily an easy-to-use handler for changing any
waveform parameter on runtime, implemented in C
language. The user can conveniently control it via a self-
designed QT-based GUI, as depicted in Fig. 4. A second
task of this GUI besides setting parameters is to display the
aforementioned monitoring signals for real-time perfor-
mance checking.
14
Figure 4 GUI for parameter control and monitoring.
For the Ethernet interface, an adequate driver was
implemented also in C language. It offers a standard Linux
network interface for the IP payload data, with the
opportunity to set priorities for the different ports, namely
the video port and the control port. As a consequence, the
complete data link is fully transparent to the application
software. It does not even notice that a wireless link instead
of a cable underlies.
Since the GPP runs under Linux OS, any standard
application software is easily applicable to the SDR
platform. For this project, applications as the camera driver
or video compression software are running on the Intel
Atom. However, since this is no primary focus of this
article, we will not go into details here.
4 Baseband Processing
Coming with the benefits of great spectral efficiency and
robustness in multipath environments, OFDM is the commu-
nication scheme of choice when UAV systems require high
data rates in heterogeneous outdoor scenarios [17, 18].
Further advantages of OFDM are its efficiency in imple-
mentation based on FFT/IFFT processing and its efficiency
in channel equalization within the frequency domain [19].
J Sign Process Syst (2012) 69:11–21
Fortunately, there was no claim for complying with any
common OFDM wireless standard like DVB, IEEE 802.11a
or WiMAX, so there was any freedom to pick up common
ideas and disregard others. Just to name one example for an
adoption from IEEE 802.11a, the synchronization algorithm
and the Automatic Gain Control (AGC) setting was imple-
mented according Fort’s and Eberle’s proposal [20, 21].
The broad parameter flexibility must not only apply to
the waveform in general but for each logical link (video
link, control link) independently of the other. So the idea is
to have two sets of parameter settings, one referring to a
robust channel with low throughput for control data and
another one providing high throughput at the expense of
reliability for video data. Since subsets of subcarriers are
freely assigned to the individual links (video/control), they
can be transmitted simultaneously. This technique is
commonly known as OFDMA (orthogonal frequency
multiple access).
Figure 5 gives a simplified overview of the whole
functionality of the central part of the waveform, which is
the FPGA part. Control signals are illustrated with thin
arrows and data signals with thick arrows. All function
blocks are self-designed, except for the IFFT/FFT, the En-
and Decoder and the PCIe Interface, which are standard
Xilinx IP cores.
J Sign Process Syst (2012) 69:11–21 15

PCIe Domain
PCIe Domain Baseband Domain
Baseband Domain

Frontend Config
(65.2 MHz Clock)
(65.2 MHz Clock) (100 MHz (100 MHz Clock)
Clock or 48 MHz Clock)

Interface
Config Reg. Frontend Config
Frontend Config Interface State Machine
(e.g. SPI)
DMA/Sys Config MAC State Machine

Baseband Config Idle TX_Control

MAC Config AGC

TX RX State Machine
RX_Control

BERT
TX System Generator Model
Video
TXTX
Control
Control
State
State
Machine
Machine
BERT
TX
Control
TX Buf Video

OFDM Symbol Assembler

Config
HDLC

MUX
HDLC Scramble Convol. Interleave QAM I
FIFO Scrambler Interleaver

Sync Symbol

Conditioning
Pilot Symbol
Serializer

Cyclic Prefix

I/Q Samples
Serializer r Encoder r Mapper Q

Windowing

I/Q Sample
Insertion +
Insertion

Insertion

Scaling
I I I I

TX
TX Buf Control IFFT
Config Q Q Q Q
HDLC
MUX

Scramble Convol. Interleave QAM I

FIFO Serializer Scrambler Interleaver
DMA r Encoder r Mapper Q
Write
Interface
PCIe

Calcul
ation
RSSI
RXRX
Control
Control
State
State
Machine
Machine

Detectio
RX Buf Video

Detect
Sync
Sync
Config

Correl
Auto-

ation
FIFO
HDLC
De- De- De- QAM

OFDM Symbol De-

OFDM Symbol De-
RX Buf De- Viterbi De- I

Offs. Correction
RX Buf Control Serializer Scramble Interleave De-

Conditioning
Carrier Freq.

FIR Filtering

I/Q Samples
Control Scrambler Decoder Interleaver Q

I/Q Sample
Assembler
Assembler
Config r r Mapper

Scaling
I II I I I
Channel
Channel

RX
FIFO FFT
DMA HDLC Equalization
Equalization
Q QQ Q Q Q
Read De- De- De- QAM
De- Viterbi De- I
Serializer Scramble Interleave De-
Scrambler Decoder Interleaver Q
r r Mapper
BERT
RX
BER Video I/Q signal I/Q signal
monitor BERT monitor monitor
RX
Control

IP Frames Bits I/Q Constellation Points I/Q Time Domain Signals

Figure 5 FPGA-internal processing.

As depicted at the top of Fig. 5, the entire frontend Since there is no need to change the frontend on runtime,
configuration, namely the AGC and transceiver setting, is there is also no runtime-configurable baseband clock
realized with simple FSMs. At the lower left part of required. Instead, the clock is defined before synthesizing
Fig. 5, the PCIe interface is shown, which serves for the design. With the 100 MHz oscillator, which drives the
video and control data exchange via IP frames on FPGA, a 100 MHz baseband domain and a 48 MHz were
network layer. These IP frames will be serialized to already realized and tested.
bitstreams within the High-Level Data Link Control
(HDLC) blocks on the transmitter side (TX). On the
receiver side (RX), similar HDLC blocks de-serialize the 5 Runtime Flexible TDMA
bitstreams again.
A similar structure of what is shown from those The channel access of the different users (UGV, UAV and
HDLC blocks on to the right-hand edge of Fig. 5 can GCS) is organized with TDMA. As depicted in Fig. 6, the
be found as standard OFDM baseband processing chain in top layer of the time schedule is made up of a periodically
every OFDM book (disregarding the OFDMA aspect) repeated TDMA superframe. This superframe is then sub-
[22]. Within this processing chain, the bitstream is divided into 64 TDMA slots or less, all of the same length.
mapped to I/Q constellation points (frequency domain) Within each TDMA time slot, only one user is transmitting
and finally transformed to I/Q time domain signals using while the rest is listening. This means, a half-duplex
an IFFT. communication system is provided.
Although the PCIe interface runs natively with Once the traffic demand of the different users is
65.2 MHz, the ideal clock for the baseband domain identified, the slots can be portioned accordingly: The
could differ from that for an optimal match with a more traffic a user requires, the more slots are reserved for
specific frontend. For this reason, two internal clock him to transmit.
domains are realized with the aid of a digital clock The users are not equal. One master node must be
manager and buffers, one for PCIe and one for the chosen to set the timing of the superframe by inserting a
baseband domain. superframe start indicator in the first slot of the schedule.
16 J Sign Process Syst (2012) 69:11–21

Superframe
Fr up A
S DM
am e r
Start Indicator one in the superframe schedule2 and a Pilot Symbol for
channel and fine frequency offset estimation [23, 24].
T

Apart from this static preamble structure, the presented

N
ot A

1 2 3 4 5 6
waveform offers a variety of configuration options for the
Sl M

(max. 64)
s
TD

OFDM frame layout:

Pilot Symb
Slot Sync
Preamble

SF Sync
AGC

Data
am M

Data
Data

...
Fr FD
e
O

Action 6.1.1 OFDM Symbols per Frame

Point
Offset TX
AGC Preamb Slot Sync SF Sync Pilot Symbol Data Symbol
Since the preamble is not usable for payload transmission,
WND/2
WND/2

Control Data

Windowing/2
Cyclic Prefix

Cyclic Prefix
Stream
Windowing

Windowing

Windowing

Windowing

Windowing
the OFDM frame should be generally chosen as long as
m DM
ls

Pilot
bo

Video Data
Sy OF

Stream possible in order to maximize the data-to-preamble ratio.

f However, because of the essential tasks of the preamble,
FFT length: 64 FFT length: variable FFT length: variable
like AGC setting or coarse frequency offset estimation, the
Figure 6 TDMA / OFDM time schedule. preamble must appear repetitively within the transmitted
stream. With other words, a rise of the frame length directly
improves the data throughput, but is limited due to the need
Since the master node must be directly visible to all other for preamble-based calibration.
users, it must be the UAV in our scenario. When the UGV In order to account for specific channel characteristics,
or the GCS join the network, they must wait for the the preamble repetition rate is flexible, as the number of
superframe start indicator to come, before they can OFDM symbols within one frame can be adapted.
synchronize and actively participate. One example for the need to adapt the preamble
Besides synchronizing, this first slot is also used to repetition rate is a varying sample clock offset between
spread information about important TDMA settings like the the sender and the receiver platform. This could e.g. result
number of slots currently in use and their allocation. This in from temperature changes between the UGV and the UAV,
turn means that the master node is able to change the which would directly affect the oscillators.
TDMA slot allocation on runtime. The induced frequency offset could be compensated by
At the start of each slot, a short idle time is hold in order to pilots only up to this point when the offset equals one full
remedy propagation delays and clock drifts and to ensure that clock cycle at the end of the frame [25]. Here, the term
all users are listening. (called Action Point Offset TX in Fig. 6) pilots refers not only to those within the pilot symbol of the
preamble, but also to the scattered pilots, which are
explained in Section 6.2.3.
6 Runtime Flexible OFDM In the above-stated case, the frame length would need to
be reduced in order to perform coarse frequency offset
The key feature of the presented waveform is its parameter estimation and timing synchronization earlier, before the
flexibility. One flexibility aspect was already introduced in offset compensation capability of the pilots is reached.
the last section, namely the TDMA slot allocation.
In the following, all the OFDM waveform parameters 6.1.2 Cyclic Prefix length
which are runtime-configurable are listed and explained.
All of them have in common the trade-off between Without cyclic prefix, transmitting OFDM symbols over a
bandwidth and reliability. fading multipath channel leads to inter-carrier-interferences
(ICI) and inter-symbol-interferences (ISI). The reason is the
6.1 OFDM Frame Layout spectral discontinuity between OFDM symbols, which dis-
perses into the payload part of the symbol and thus causes a
As shown in Fig. 6, each TDMA slot contains an OFDM loss of subcarrier orthogonality. However, the OFDM symbol
frame. It makes no sense to flexibilize every aspect of the can be cyclically extended by copying the last portion of the
frame structure. Specifically, the preamble structure symbol and appending it as cyclic prefix to the front of the
which is used for calibrations at the start of each OFDM symbol. This way, the discontinuity still remains, but is no
frame must be static. It is loosely adopted from the IEEE longer present within the payload part of the symbol.
802.11 standardization [20] and comprises the following For the receiver, the cyclic prefix is useless and will be
four parts: An AGC Preamble for signal detection and discarded. Consequently, it should be set as short as possible,
AGC setting, a Slot Sync for coarse frequency offset but at least as long as the channel impulse response.
estimation and timing synchronization, an SF Sync, which
2
is only included in case the current TDMA slot is the first The SF Sync is not compliant with the IEEE 802.11 standardization
J Sign Process Syst (2012) 69:11–21
6.1.3 Windowing Length
Thanks to the cyclic prefix, the spectral discontinuity between
OFDM symbols is no longer an obstacle for correct data
transmission. Nevertheless, this spectral blurring is still
existent and poses a problem for using adjacent frequency
channels. A solution is windowing, which denotes simply
cross-fading between consecutive OFDM symbols. To attain a
sharp spectral edge of the OFDM signal, the fading must be as
smooth as possible, which means long windowing.
Just like the cyclic prefix, also the windowing portion of
the OFDM signal is discarded without being used by the
receiver and thus it should be short, too.
6.2 OFDM Subcarrier Usage
According to the TX part in Fig. 5, the OFDM Symbol
Assembler allocates the mapped data to the subcarriers,
which are then fed to the IFFT block. On the RX side, this
procedure is reversed with the aid of the OFDM Symbol
De-Assembler. Especially this part of the baseband pro-
cessing is implemented to provide highest flexibility:
6.2.1 Number of Subcarriers / FFT Length
An essential challenge of the presented scenario is the high
speed the UAV is able to reach. The top speed depends on
several aspects and can reach up to 120 km/h. With the
UGV moving in the opposite direction, the relative velocity
would be even higher. Specifically, in OFDM communica-
tion, high speed leads to much shorter coherence time and
larger Doppler frequency spread, especially for the case that
a high RF carrier frequency is used. As a direct result, the
orthogonality among subcarriers gets lost and ICI degrades
the performance [26]. The solution to this problem is to
broaden the frequency spacing between the subcarriers for
increased tolerance against orthogonality loss. This can be
achieved by decreasing the FFT length, because this
parameter is indirect proportional related to the frequency
spacing. So in order to strengthen against Doppler shift, an
adequate FFT length out of the following can be chosen: 8,
16, 32, 64, 128, 265, 512 or 1,024.
6.2.2 Allocating Video, Control or None of Both
All subcarriers can be fully flexible occupied either with
video data, control data or they are just left open. This has
basically three advantages:
First of all, the throughput of the video link can be
dynamically adjusted at the expense of the control link
throughput and vice versa, which allows further
prioritization.
17
Second, each RF transceiver offers a specific band-
width to be usable for transmission. To maximize
throughput, the waveform must fill all subcarriers lying
within this bandwidth up with data and set the rest to
zero as guard bands. With other words, any given
transceiver bandwidth is supported with this option. Of
course, this parameter is strongly related to the FFT
length. Changing one of both parameters requires
changing the other one as well.
Third, beside the guard band, another reason for setting
subcarriers to zero is to mask out strongly faded
frequencies.
Figure 7 gives a simple demonstration of the change of
subcarrier on runtime. The figure refers to RF transmission of
the OFDM waveform at 1.5 GHz center frequency. The top
part of the figure shows the currently measured signal
spectrum and the bottom part gives its history with different
colors for different signal power. The older the measured
spectrum, the more it is shifted downwards. This spectrogram
clearly shows that the bandwidth increases over time. This is
because the 512 FFT OFDM signal was changed from using
20 subcarriers, then 40 and finally 60 for payload data.
6.2.3 Scattered Pilots
In the previous section, three options for subcarrier
allocation were presented. In order to account for the
scenario-specific problem of fast changing channels, yet
another option is provided, which is the use of subcarriers
as scattered pilots.
As already mentioned, the first OFDM symbol of each
OFDM frame is filled with pilots, which are known data to
identify the transfer function of the system. The correction
values calculated with these pilots for each subcarrier are
Figure 7 Over-the-air measured spectrogram: occupied OFDM
subcarriers are changed from 20 to 40 to 60.
18
applied to the following subcarriers of the following OFDM
symbols of this frame, assuming that the channel character-
istics do not change meanwhile.
Since this assumption applies less and less with fast
changing channels, the pilots must occur more often within
the data stream in order to update the correction values
more frequently. To meet this goal, two options are
conceivable, which is either reducing the number of OFDM
symbols per frame, or, the more elegant solution, placing
additional pilots among the payload data.
The second solution, the concept of scattered pilots is
implemented as follows. Before checking for a specific
subcarrier in a specific OFDM symbol, whether it should be
allocated with video or control data or none of both, there is a
prioritized request, if the specific subcarrier shall be occupied
by a pilot. This request refers to a systematic placement of the
pilots within the frame, using up to three completely
independent patterns. These patterns are defined in terms of
their spacing over the subcarriers and their spacing over the
OFDM symbols. In addition, the start position of each of the
patterns is defined by the according OFDM symbol number
and the according subcarrier number. Overlapping the patterns
is possible. The pattern placement is completely runtime-
flexible, simply by spreading the information about each
pattern’s subcarrier- and symbol-spacing, as well as its start
position to all communication partners (UGV, UAVand GCS).
Figure 8 gives two simple examples of how the pilots
can be distributed over the frame. Each row refers to an
OFDM symbol and each line to a specific subcarrier (with
DC in the middle and the guard bands at the edges). For the
structure in Fig. 8a, all three flexible pilot patterns are
needed, while for Fig. 8b, one pattern is enough.
The question remains, how and when to change the
scattered pilot structure. This challenge of optimizing pilot-
aided channel estimation/equalization for OFDM systems is
extensively studied in literature [27–30]. In order to reduce
this complex topic to a rule of thumb, one could split up the
problem of fast changing channels into two extreme shapes:
Fast changes over frequency, e.g. due to a long delay spread
and fast changes over time, e.g. due to a high Doppler
spread [31, 32]. For fast changes over frequency, the pattern
in Fig. 8a could be an adequate solution, while for fast
changes over time the pattern of Fig. 8b would be more
adequate.
The classical method to process scattered pilots is the
sample-and-hold approach, meaning that each time, a
scattered pilot subcarrier is received, the initially (or last)
determined correction factor is updated for this subcarrier,
while all other correction factors stay the same. However,
especially in case of fast changing channels over time
according Fig. 8b, the second provided method is better
suited: For this method, the pattern must be chosen in a way
to have more than one scattered pilot within one OFDM
J Sign Process Syst (2012) 69:11–21
a) f
DC
t
f
b)
Unused Carrier
Data Carrier
Fixed Preamble Pilots
DC
Flexible Pilot Pattern 1
Flexible Pilot Pattern 2
Flexible Pilot Pattern 3
t
Figure 8 Examples of scattered pilots patterns: a optimized for
spectrally dynamic channels, b optimized for temporally dynamic
channels.
symbol. Each time, an OFDM symbol with scattered pilot
subcarriers is received, all correction values of the complete
symbol are updated in terms of interpolating linearly over
all subcarriers of this OFDM symbol.
With an additional control signal provided, the sample-
and-hold method can be changed on runtime to the linear
interpolation method and vice versa.
6.3 Modulation Scheme
Lowering the modulation scheme and therewith the resolution
of the I versus Q constellation diagram is the traditional
countermeasure against a variety of channel effects, including
noisy channels. This is done at the expense of the bits-per-
subcarrier ratio. The waveform presented here supports
BPSK, QPSK, 16QAM and 64QAM. This modulation
spectrum is the state-of-the-art selection of modern OFDM
based wireless standards like WiMAX, IEEE 802.11a or
3GPP Long Term Evolution (without BPSK).
6.4 Code Rate of Convolutional Encoder/Viterbi Decoder
A convolutional encoder is typically used by the sender to
enrich the payload with redundancy. This is done in a way
that allows the receiver to identify and to correct transmis-
J Sign Process Syst (2012) 69:11–21
sion errors with the aid of a Viterbi decoder, completely
without further inquiry [33].
For the presented waveform, a Viterbi decoder IP core
from Xilinx was applied [34]. Its code rate, meaning the
ratio of payload to redundancy, is flexible from 1/2 to 1/7.
With puncturing, the code rate can be increased up to 3/4 or
7/8 [35].
7 System Performance Monitor
In the previous section, all run-time flexible OFDM param-
eters of the waveform were introduced. The question remains
how to choose them in an adequate way for optimal
performance in terms of the current nonideal channel
characteristics. For evaluating the temporary system perfor-
mance, two general monitoring tools are provided: A bit error
rate tester (BERT) and different options to analyze the I versus
Q constellation diagram.
7.1 Bit Error Rate Monitor
During normal operation, the control and video data is
exchanged via PCIe as described in Section 4. However,
when the BERT is applied, the system is alternatively fed
from a source for pseudo random bit sequences (PRBS). At
the receiver part, a second PRBS generator is synchronized
with the incoming data stream in order to evaluate the bit
error rate. Since this number is directly transmitted to the
GPP and indicated with the QT-based GUI, real-time
measurements are available.
Of course, error rate testing is also possible with IP
pinging, but this would only provide information about the
IP packet loss rate and not about the bit error rate. The
packet loss rate is less significant, because a packet will be
discarded, no matter if there was only one false bit in the
packet or multiple ones.
7.2 I/Q Constellation Monitor
Bit error rate testing is appropriate to evaluate the overall
transmission robustness across the complete communication
system. However, in case the bit error rate increases, its reasons
must be found with more detailed analysis on particular effects,
which result from specific nonideal transmission conditions.
For this purpose, the receiver’s I/Q signal stream is tapped
before and after the channel equalization block according
Fig. 5. These signals are once more directly transmitted to the
GPP and visualized with the QT-based GUI as constellation
diagram (I versus Q signals). Thus, any effect which
influences the amplitude or phase of the complex modulated
OFDM signal is visualized. A comprehensive overview of
potential effects of nonideal transmission is given in [36].
19
The developed GUI not only visualizes the I/Q constella-
tion, but also provides with the error vector magnitude (EVM)
a corresponding performance metric. The EVM describes the
deviation of the measured transmitted signal against a perfect,
theoretical signal [37]. The GUI supports plotting the EVM
versus subcarrier or versus symbol. This can be crucial in
allowing the operator to quickly locate the source of errors,
such as strong faded frequencies (EVM versus subcarrier) or
fast changing channels (EVM versus symbol).
8 Designing and Testing the Waveform
The process of developing the waveform can be divided
into three major steps.
At first, the software/hardware design: The PCIe and
Ethernet drivers, as well as the waveform control were
implemented in C language. The FPGA design was done with
a hybrid approach of pure VHDL and a model driven design
environment based on MathWorks Matlab/Simulink including
the System Generator for DSP from Xilinx. This approach
allows at a very early stage elaborated simulations, e.g. with the
Simulink multipath Rayleigh or Rician fading channel models.
Second, testing on the SDR hardware was performed
within a predefined environment. More precisely, the fading
channel simulator R&S AMU200A was deployed to stress
the waveform with a set of specific static and dynamic
fading scenarios, as well as Doppler and AWGN noise real-
time simulations.
The final step for the waveform to pass is the test with the
UGV, the UAV and a GCS in the open field. There are two
types of UAVs in use. On the one hand this is the self-designed
UAV in Fig. 1 with the maximum speed of 70 km/h and on
the other hand this is a Swiss UAV, NEO S-300 series, which
flies with up to 120 km/h [38]. The applied UGV is the rover
MERLIN, which was designed for robust operations in harsh
outdoor environments. Its maximum speed is 12 km/h [39].
9 Conclusion and Outlook
In this article, we gave an overview over an OFDM-based
waveform, designed for communication links among
autonomous robotic platforms.
Though the waveform is applicable to a broad spectrum
of scenarios, a specific communication set-up was defined
for validation. Its participants are a UGV, a UAV and a
GCS, all accessing the channel by means of runtime-
flexible TDMA. In this scenario, the communication focus
is on a unidirectional video link from the UGV to the GCS,
which is established indirectly, utilizing the UAV as relay.
In addition, further direct and indirect links for video and
control data among the communication partners are needed.
20
The waveform perfectly accounts for the challenges of this
set-up, as it offers two separated logical links for video and
control data concurrently in OFDMA manner. Furthermore, it
provides outstanding flexibility according its waveform
parameters. The user is able to freely optimize for throughput
or for reliability and even more specific problems like fast
changing channels or Doppler spread can be systematically
and efficiently fought with the right parameter setting.
Moreover, not just the waveform in total is adaptable, but
specifically each logical link independent of the other. All
adaption is done dynamically on runtime with a customized
GUI. Besides parameter setting, this GUI also serves as
system performance monitor, as it provides a bit error rate
monitor and an I/Q constellation monitor including EVM
scope. This comfortably helps to decide, which parameter
setting fits the current channel characteristics best.
The SDR hardware on which the waveform is implemented
is also ideal for autonomous robotic platforms, since it features
a very small form factor. Moreover, with TCP/IP and UDP/IP
over Ethernet, it offers a universal system interface.
As soon as the waveform is fully tested, the next big
goal is to improve the system with adaptivity features. Up
to now, the broad set of waveform parameters is freely
adjusted by the user, who must identify the current channel
characteristics with the help of the system performance
monitor. In future releases, the waveform itself is supposed
to measure, to analyze and to adapt to changing channel
characteristics automatically.
References
1. Zhan, P., Yu, K., & Swindlehurst, A. L. (2011). Wireless relay
communications with unmannned aerial vehicles: performance
and optimization. IEEE Transactions on Aerospace and Electronic
Systems, 47(3), 2068–2085.
2. Palat, R., Annamalau, A., & Reed, J. (2005). Cooperative relaying
for ad-hoc ground networks using swarm UAVs. In Proc. IEEE
MILCOM 2005, pp. 1588–1594.
3. Cheng, C., Hsiao, P., Kung, H., & Vlah D. (2007). Maximizing
throughput of UAV-relaying networks with the load-carry-and-
deliver paradigm. In Proc. IEEE WCNC 2007, pp. 4417–4424.
4. Prasad, K. V. (2004). Principles of digital communication systems
and computer networks (p. 56). Hingham: Charles River Media.
5. Tanenbaum, A. S., & Wetherall, D. J. (2010). Computer networks
(5th ed., p. 525). Upper Saddle River: Prentice Hall.
6. Sklar, B. (1997). Rayleigh fading channels in mobile digital
communication systems—part I: characterization. IEEE Commu-
nications Magazine, 35(7), 90–100.
7. http://gnuradio.org. Accessed 27 September 2011.
8. Dhar, R., George, G., Malani, A., & Steenkiste, P. (2006). Supporting
Integrated MAC and PHY Software Development for the USRP
SDR. In Proc. 1st IEEE Workshop on SDR 2006. pp. 68–77.
9. http://warp.rice.edu. Accessed 27 September 2011.
10. Hunter, C., Camp, J., Murphy, P., Sabharwal, A., & Dick, C.
(2006). A Flexible Framework for Wireless Medium Access
Protocols. In Proc. 40th ACSSC 2006, pp 2046–2050.
J Sign Process Syst (2012) 69:11–21
11. Amiri, K., Sun, Y., Murphy, P., Hunter, C., Cavallaro, J. R., &
Sabharwal, A. (2007). WARP, a Unified Wireless Network Testbed
For Education and Research. In Proc. IEEE MSE 2007, pp. 53–54.
12. Belanger, L. (2007). Advanced SDR platform eases multiprotocol
radio development. RFDesign Magazine, 2007(1), pp. 36–42.
13. http://www.sundance.com/products.php?page=rds. Accessed 27
September 2011.
14. Gustafsson, O., et. al. (2010). Architectures for Cognitive Radio
Testbeds And Demonstrators—An Overview. In Proc. 5th
CROWNCOM 2010, pp. 1–6
15. Kontron (2009). Kontron User’s Guide nanoETXexpress-SP. Rev. 134.
16. Pico Computing (2009). E17 Hardware Technical Reference,
Release: 1.00.00.02, Hardware Revision: B.
17. Wu, Z., Kumar, H., & Davari, A. (2005) Performance evaluation
of OFDM transmission in UAV wireless communication. In Proc.
37th SSST 2005, pp. 6–10.
18. Jian, H., & Yue-tong, Z. (2009) Impact of Doppler on High Speed
UAV OFDM System. In Proc. ICCSN 2009, pp.742–745.
19. Ozdemir, M. K., & Arslan, H. (2007). Channel estimation for
wireless OFDM systems. IEEE Communication Surveys &
Tutorials, 9(2), 18–48.
20. (1999). (R2003) IEEE 802.11a, Part 11: Wireless LAN Medium
Access Control (MAC) and Physical Layer (PHY) specifications:
High Speed Physical Layer in the 5 GHz Band. Piscataway, NJ.
21. Fort, A., & Eberle, W. (2003). Synchronization and AGC Proposal
for IEEE 802.11a Burst OFDM Systems. In Proc. IEEE GLOBE-
COM 2003, pp. 1335–1338.
22. Prasad, R. (2004). OFDM for wireless communications systems
(p. 118). Norwood: Artech House.
23. Fazel, K., & Kaiser, S. (2008). Multi-carrier and spread spectrum
systems: From OFDM and MC-CDMA to LTE and WiMAX (2nd
ed., p. 138). Chichester: John Wiley & Sons.
24. Li, Y. G., & Stüber, G. L. (2006). Orthogonal frequency division
multiplexing for wireless communications (p. 113). New York:
Springer US.
25. van de Beek, J. J., Sandell, M., & Borjesson, P. O. (1997). ML
estimation of time and frequency offset in OFDM systems. IEEE
Transactions on Signal Processing, 45(7), 1800–1805.
26. Engels, M. (2002). Wireless OFDM systems: How to make them
work? (p. 26). Boston: Kluwer.
27. Chen, W. (2005). Time-Frequency-Selective Channel Estimation
of OFDM Systems. PhD dissertation, Drexel University.
28. Sanzi, F., Jelting, S., & Speidel, J. (2003). A comparative study of
iterative channel estimators for mobile OFDM systems. IEEE
Transactions on Wireless Communications, 2(5), 849–859.
29. Garcia, M. J. F.-G., Zazo, S., & Paez-Borrallo, J. M. (2000). Pilot
patterns for channel estimation in OFDM. Electronics Letters, 36
(12), 1049–1050.
30. Morelli, M., & Mengali, U. (2001). A comparison of pilot-aided
channel estimation methods for OFDM systems. IEEE Trans-
actions on Signal Processing, 49(12), 3065–3073.
31. Sklar, B. (1997). Rayleigh fading channels in mobile digital
communication systems—part II: mitigation. IEEE Communica-
tions Magazine, 35(7), 102–109.
32. Pearce, D. (2008). Gsw time variant multipath channels. Getting
Started with Communications Engineering.
33. Clark, G. C., Jr., & Bibb Cain, J. (1981). Error-correction coding
for digital communications. New York: Plenum Press.
34. Xilinx (2011). LogiCORE IP Viterbi Decoder v7.0 Product
Specification DS247.
35. Haccoun, D., & Begin, G. (1989). High-rate punctured convolu-
tional codes for Viterbi and sequential decoding. IEEE Trans-
actions on Communications, 37(11), 1113–1125.
36. Speth, M., Fechtel, S. A., Fock, G., & Meyr, H. (1999). Optimum
receiver design for wireless broad-band systems using OFDM—Part
I. IEEE Transactions on Communications, 47(11), 1668–1677.
J Sign Process Syst (2012) 69:11–21
37. Shafik, R. A., Rahman, M. S., Islam, A. H. M. R., & Ashraf, N. S.
(2006). On the error vector magnitude as a performance metric
and comparative analysis. In Proc. IEEE ICET 2006, pp. 27–31
38. http://www.swiss-uav.com/uav_systems.php. Accessed 27 Sep-
tember 2011.
39. Eck, D., Stahl, M., & Schilling, K. (2007). The small outdoor
rover MERLIN and its assistance system for tele-operations.
Springer Tracts in Advanced Robotics, 2008(42), 277–286.
Christian Blümm received his Dipl.-Ing. degree in mechatronics
engineering from the Institute for Electronics Engineering at the
University of Erlangen-Nuremberg, Germany in 2009. After research
experience in the field of smart sensors at Siemens Corporate Technology
and Areva N.P., he is currently working towards his PhD degree in the
communications research department at EADS Innovation Works in the
field of waveform design for software-defined radio.
Christoph Heller has a degree in communications engineering from
the RWTH Aachen University, Germany and received his PhD from
Stuttgart University in 2010. He joined the European Aeronautic
Defence and Space Company (EADS) in 2005 and is employed in the
communications research department of EADS Innovation Works.
Christoph Heller is heading the development of next-generation
wireless data links for aeronautic applications, focusing on digital
21
baseband processing. He is involved in several EADS internal,
national and international research projects dealing with software-
defined radio and cognitive radio.
Robert Weigel was born in Ebermannstadt, Germany, in 1956. He
received the Dr.-Ing. and the Dr.-Ing.habil. degrees, both in electrical
engineering and computer science, from the Munich University of
Technology in Germany, in 1989 and 1992, respectively. From 1982
to 1988, he was a Research Engineer, from 1988 to 1994 a Senior
Research Engineer, and from 1994 to 1996 a Professor for RF Circuits
and Systems at the Munich University of Technology. In winter 1994/
95 he was a Guest Professor for SAW Technology at Vienna
University of Technology in Austria. From 1996 to 2002, he has
been Director of the Institute for Communications and Information
Engineering at the University of Linz, Austria. In August 1999, he co-
founded DICE – Danube Integrated Circuit Engineering, Linz,
meanwhile an Infineon Technologies Design Center which is devoted
to the design of mobile radio circuits and systems. In 2000, he has
been appointed a Professor for RF Engineering at the Tongji
University in Shanghai, China. Also in 2000, he co-founded the Linz
Center of Competence in Mechatronics, meanwhile also a company.
Since 2002 he is Head of the Institute for Electronics Engineering at
the University of Erlangen-Nuremberg. In 2010, he co-founded the
company eesy-id.
He has been engaged in research and development on microwave
theory and techniques, integrated optics, high-temperature supercon-
ductivity, SAW technology, digital and microwave communication
systems, and automotive EMC. In these fields, he has published more
than 750 papers and given about 300 international presentations. His
review work includes international projects and journals. In 2002, he
received the German ITG Award, and in 2007 the IEEE Microwave
Applications Award.
Robert Weigel is a Fellow of IEEE, a member of the Institute for
Components and Systems of The Electromagnetics Academy, and a
member of the German ITG and the Austrian ÖVE. He serves on
various editorial boards and has been editor of the Proceedings of the
European Microwave Association (EuMA). He has been member of
numerous conference steering committees. Currently he serves on
several company and organization advisory boards. He is an elected
scientific advisor of the German Research Foundation DFG.