2) OFDM modulator: Starting from the symbols a
i
at
its input, the OFDM modulator macro-block generates the
baseband signal corresponding to an OFDM modulation with
N = 64 subcarriers, N
u
= 48 of which are actually used to
transmit modulation symbols (useful subcarriers) whereas the
remaining N
z
= 16 have null amplitude (15 virtual subcarriers
+ 1 DC subcarrier in correspondence of the frequency zero
3
).
More specifically, the N = 64 subcarriers are used as follows:
• 8 virtual subcarriers (from 1 to 8) with null amplitude;
• 24 data subcarriers (from 9 to 32) with 2-ASK modula-
tion;
• 1 DC subcarrier (number 33) with null amplitude (corre-
sponding to the zero frequency);
• 24 data subcarriers (from 34 to 57) with 2-ASK modu-
lation;
• 7 virtual subcarriers (from 58 to 64) with null amplitude.
With the introduction of the virtual and DC subcarriers, the
output sampling rate of the OFDM macro-block is 2400 sam-
ples/s. Data are organized in frames to simplify the real time
management by the hardware.
3) Upsampling: In order to modulate the baseband OFDM
signal, the sampling frequency must be increased. It is conve-
nient, on this regard, to increase it by a factor of 20, taking it
to the 48000 samples/s required by the audio card’s DAC. The
upsampling operation is performed by the sequence of blocks
included in the Upsampling macro-block, that performs an
upsampling by a factor of 2 at first and then by a factor of
10. The two stage procedure is preferable, compared with a
single stage upsampling by a factor of 20, because it reduces
the overall computational burden [6], [7].
4) Baseband to IF: The Baseband to IF macro-block mod-
ulates the signal, translating it from baseband to intermediate
frequency. Such operation is performed through a quadrature
modulator with internally generated carriers at a frequency
f
IF
= 15 kHz. The Baseband to IF macroblock is a classic
quadrature modulator, with two separate paths for the real
(in-phase) and the imaginary (quadrature) components of the
signal. Each component is upconverted by a mixer (represented
by the Product block) driven by a cosine or a sine carrier,
generated by the Cosine Wave and Sine Wave blocks. Both
modulated signals are then summed up and taken out.
5) Automatic Gain Control: The Automatic Gain Control
macro-block adapts the signal’s dynamic range to the require-
ments of the output port. As a consequence of this operation,
the highest value in each frame is equal to 2
14
− 1, consistent
with the highest value required by the output port (2
15
− 1).
6) Raspberry Pi output: The Raspberry Pi output macro-
block represents the system output port. It corresponds, there-
fore, to the sound card’s DAC.
Once the Simulink model is realized and checked through
Simulink simulations, it is possible to carry out the Deploy
to Hardware, which is triggered by Simulink itself. The
3
In order to facilitate the receiver in the research of the band center, the
subcarrier corresponding to the zero frequency (in the baseband) is usually
assigned a null amplitude. The acronym DC means Direct Current.
Fig. 7. OFDM signal spectrum.
implemented system starts as soon as the automatci download
of the corresponding software on the device is completed. Con-
necting the Raspberry Pi to the spectrum analyzer, according
to the scheme in Fig. 5, the signal spectrum appears as shown
in Fig. 7. The expected bandwidth of approximately 2 kHz
and the DC subcarrier centered at 15 kHz can be observed.
VI. CONCLUSION
Despite its popularity, the usage of Raspberry Pi boards as a
signal processing device for SDR applications is an innovative
exception and opens new possibilities to the teaching of signal
processing. Owing to its low cost, it can be massively used
in lab activities that do not require any programming skills,
thanks to the Simulink support. In the next future, even RF
transmitters and receivers could be part of the SDR platform
here presented, thank to the availability of SDR peripherals
such as the HackRF One [8] transmitter/receiver and the RTL-
SDR receiver [9].
ACKNOWLEDGMENT
The authors wish to thank Mirko Mirabella for his great
contribution to the Simulink Defined Radio project.
REFERENCES
[1] Raspberry Pi website. [Online]. Available: http://www.raspberrypi.org
[2] Raspberry Pi Academy website. [Online]. Available:
http://www.raspberrypi.org/picademy/
[3] Raspberry Pi Support from Mathworks website. [Online]. Available:
http://it.mathworks.com/hardware-support/raspberry-pi-simulink.html
[4] Raspberry pi based SDR experiences. [Online]. Available:
http://www.simulinkdefinedradio.com/
[5] Simulink and model based design. [Online]. Available:
http://it.mathworks.com/services/consulting/proven-solutions/model-
based-design.html
[6] R. Crochiere and L. Rabiner, “Interpolation and decimation of digital
signals; a tutorial review,” Proceedings of the IEEE, vol. 69, no. 3, pp.
300 – 331, March 1981.
[7] G. Pasolini and R. Soloperto, “Multistage decimators with minimum
group delay,” in IEEE International Conference on Communications
(ICC), May 2010, pp. 1–6.
[8] Hackrf One. [Online]. Available: https://greatscottgadgets.com/hackrf/
[9] RTL-SDR. [Online]. Available: http://www.nooelec.com/store/sdr/sdr-
receivers/nesdr-mini-rtl2832-r820t.html
2016 24th European Signal Processing Conference (EUSIPCO)
402