ICTON 2017 We.C2.

Simulation Approach of Wireless OFDM and FBMC Signals

Transmission over Fiber Based on Equivalent Electrical Models
Anne-Laure Billabert1, Ali Kabalan1, Salim Faci1, Rostom Zakaria2, Maina Moutaly1,
and Catherine Algani1
1
ESYCOM, Le Cnam, 292 rue saint Martin, 75141 Paris Cédex 3, France
2
Cédric, Le Cnam, 292 rue saint Martin, 75141 Paris Cédex 3, France
e-mail: anne-laure.billabert@lecnam.net
ABSTRACT
In this paper, we present the impact of a Radio over fiber (RoF) link on the transmission of a signal dedicated to
wireless link. Comparison is given between OFDM and FBMC based wireless signal by simulation with the
electrical modelling approach of O/E and E/O devices and optical components. The considered link concerns the
Intensity Modulation - Direct Detection (IM-DD) architecture where the wireless signal is transmitted at the
intermediate frequency. The simulation system includes the noise sources and nonlinearity of each component
making thus the easy study of the impact of each one separately. It is then possible to quantify the degradation of
the transmitted signal in term of error vector magnitude (EVM) or bit error rate (BER). Thus, the system
architecture can be optimized for various applications and furthermore depending on the number of users.
Keywords: RoF link, Remote antenna unit, FBMC and OFDM coding, system simulation.

1. INTRODUCTION
Multi-service data user applications need large bandwidth supported by novel technologies like FTTx (fiber to
the x). In addition, the unlicensed millimeter wave band gives more potentiality of the broadband
communication. The future 5G communication is confronted to network challenges for the transport of the
multiplex signals. The networks must be low cost due to high density of the base stations, supporting high data
rate, simple and flexible. Analog transmission over fiber (RoF, radio over fiber) is a lithe way to migrate the cost
effective of the base stations to a share components in the central office. To realize this new network, high
performance optical devices were developed and various experimental works on RoF performances were given.
These experimental works were supported by simulation with commercial software or numerical computations.
Some limitations of these solutions concern the connection between optical and electrical devices, and
particularly for the integrated components. An alternative way to study the RoF links more efficiently is by
developing an optical devices models and using a large potentiality offers by analog circuit/system design tools
[1], [2].
This paper summarizes the electrical modeling of the optical devices to analyze and compares the transmission
of wireless signal based on OFDM and FBMC over the fiber. These wireless signal are complained following
IEEE 802.15.3c devoted to high data rate communication in home area network. An external modulation RoF
link with Mach-Zehnder Modulator (MZM) is selected because of its better linearity compared to the other
solutions. The signal transmission penalty is discussed by EVM evaluation for both signals.

2. MODELLING OF E/O, O/O AND O/E ELEMENTS OF ROF LINK

The analog feature of the RoF link for wireless signal transmission make the system analysis efficient with an
equivalent electrical models of these components. The integrated optical and electrical circuits can be optimized
to enhance the system performances. The simulation of RoF link is then possible with an electrical simulator
software where resourceful analysis tools are available. In this study, an IM-DD link with external modulation is
modeled to analyze and design RoF link for optical transmission of wireless signal.
2.1 DFB Laser
The electrical model of the laser is based on the rate equations governing the carrier, photon evolution inside the
laser cavity and the optical phase [3]. The direct detection in the IM-DD link makes the use of only the two first
rate equations sufficient. These equation are converted to electrical node equations where the output current of
the model represents the optical power envelope. The Relative Intensity Noise (RIN) is taken into account by
including Langevin noise forces into the rate equations. To have a realistic laser models, all physical parameters
used in the rate equations are extracted from measurements. The electrical impedance mismatch is also
integrated but with no effect in the case of external modulation.
2.2 Mach-Zehnder Modulator
The MZM is chosen for this study thanks to its high linearity and wide bandwidth. The modulation of the biased
voltage permits to modulate the phase shift between its two arms and thus the intensity of the recombined optical
signal at its output. The high linearity is obtained when the modulator is biased in the middle of the linear region

978-1-5386-0859-3/17/$31.00 ©2017 IEEE 1

ICTON 2017 We.C2.2

of its characteristic [4]. The equivalent electrical model of the MZM is based on the power static response
implemented by equation based module available in ADS (SDD, Symbolically Defined Device). The linearity of
the MZM is considered through the sine variation of the optical power with the biased voltage.
2.3 Optical Fiber
The equivalent electrical model of the optical fiber takes into account the optical losses and the chromatic
dispersion effect. Input and output signals of the model, which are an information of current, correspond to the
envelope of the optical power.
2.4 PIN Photodetector
At the output of the optical fiber, the modulated optical power is photodetected by a photodetector like a PIN
diodes. When the incident optical power is low compared to the diode power saturation, the equivalent model
includes a controlled source to generate the electrical current proportional to the optical power envelope and RC
circuits for the bandwidth. The mismatch circuit, packaging parasitic elements and load are added to the model.

3. SYSTEM SIMULATION OF ROF LINK

ADS offers the possibility to associate the analog and digital environments by co-simulation method. The Data
Flow Controller (DF) is used to manage digital functions such as QAM, Fourier transform or filtering and link
through specific components with the analog environment managed by an envelope simulator. Wireless signal
generation, modulation and signal detection are then possible in the same software.
3.1 Wireless Signal
The OFDM/FBMC modulation signal is generated according to the IEEE 802.15.3c standard. This signal is
distributed across the RoF link at an intermediate frequency of 3 GHz due to the limited bandwidth of the RoF
link components.
3.1.1 OFDM Signal
OFDM is widely used in ultra-wideband (ULB) wireless communication systems [5]. The digital data is
transmitted at a high rate by spreading them over a large number of subcarriers, each of which is modulated at
low rate. The frequencies of the subcarriers are chosen such that the signals are mathematically orthogonal over
an OFDM symbol period. Thus, even if the subcarriers overlap in the frequency domain, they can always be
separated at the reception. A guard interval called a cyclic prefix is used between OFDM symbols to suppress the
Inter Symbols Interference (ISI). The spectrum and time frame of transmitted OFDM signal are shown on Fig. 1.

(a) (b)
Figure 1. OFDM signal in: (a) frequency and (b) time domains.
3.1.2 FBMC Signal
During the last decade, Filter-Bank Multi-Carrier (FBMC) modulation has attracted a lot of research interest
thanks to the good features that it can offer. Indeed, FBMC is considered to be a viable alternative to OFDM for
emerging wireless networks [6]. The appealing feature of FBMC lies in the fact that subcarrier signals are
filtered to be well localized in both time and frequency domain. This translates into a steep side-lobe decay,
allowing a flexible spectrum usage and offering an increased resilience against time and frequency misalignment
between users [7]. However, the filters applied on each subcarrier introduce inter-symbol interference (ISI) that
makes symbol detection more difficult. To overcome this issue, QAM symbols are staggered by time-shifting the
imaginary components by T/2 with respect to the real ones, and the filter is designed so that the interference is
always orthogonal to the data symbol [7]. That is, when the transmitted data symbol is real-valued, the value of
the accompanied interference is pure imaginary, and vice versa.
At the receiver side, the symbol detection in a given subcarrier is performed at each period of T/2. Then, after
channel equalization, the self-interference terms are removed by simply taking either the real or the imaginary
parts of the equalized symbols. Finally, the real and imaginary values of the equalized symbols are recombined
to form the received QAM symbols. Although the self-interference is always orthogonal to the desired data

2
ICTON 2017 We.C2.2

symbols, its presence severely deteriorates the system performance when considering, for instance, system
nonlinearities or/and channel estimation errors [7]. Indeed, FBMC-based systems are very sensitive to residual
phase rotation errors [8]. Figure 2 shows the spectrum and time frame of FBMC transmitted signal. As predicted,
a higher spectral efficiency is achieved with FBMC modulation technique.

(a) (b)
Figure 2. FBMC signal in: (a) frequency and (b) time domains.
3.2 System Architecture
Fig. 3 shows the block diagram of the MZM based RoF link for EVM measurement of the transmitted OFDM
wireless signal. Electrical amplification and attenuation are placed at the input of the RoF link to control the
input power. LNA amplifier positioned at the output of the photodiode is used to increase the detected current.
The link budget and non-linearity can be analyzed depending on MZM biased voltage and laser optical power as
done in [9].

Figure 3. Measurement bench for Radio over Fiber link with MZM external modulator – AWG: arbitrary wave
generator, Amp: 20GHz electric amplifier, Att: electric attenuator, OSC: 12GS/s digital oscilloscope

4. SIMULATION RESULTS
The transmission penalties of the OFDM and FBMC signals are carried out in terms of EVM for transmission
rates of 3 Gb/s (QPSK) and 6 Gb/s (16-QAM). The signals are constructed following the compliance of IEEE
802.15.3c standard. Only measurement with OFDM signal was carried out and presented in Fig. 4 for the MZM
biased at quadrature and laser optical power of 10 mW. The photodiode is 25 GHz New Focus’s with
a responsivity of 0.66 A/W. The detected DC current is 1.8 mA because of the 3dB insertion loss of the MZM.
A short length SMF is used making thus a free-effect of optical attenuation and dispersion at a frequency of
3 GHz. The EVM is evaluated for different signal input power modulating the MZM. It is shown high EVM for
low RF power due to noise level at the detection. The minimum value is reached for RF power between -5 and
5 dBm. Beyond 5 dBm, the limitation power of the experimental setup didn’t permit to attain the MZM
saturation.
System simulations are fulfilled by co-simulation method where the time domain wireless signals (Fig. 1b, 2b)
are used as an envelope information for the analog simulation. Good agreement is obtained with measurement
for the EVM curve with OFDM wireless signal for both QPSK and 16-QAM modulations. Simulation results of
EVM with FBMC are close to EVM OFDM with QPSK in the RF power range (Fig. 4a). With 16-QAM, the
EVM is a little better with OFDM and especially for low RF power since the equalization process becomes less
accurate more complex modulation format. But, the main advantage of the FBMC is its optimal spectral
allocation which would be very advantageous for multiband signals transmission.
For wireless signal emission, the EVM limit required at the output of the transmitter is determined by the IEEE
802.15.3c standard according to the transmission rate. They are 20% and 9% for QPSK and 16-QAM,
respectively. Considering no additional effect got from antenna, a dynamic range greater than 22 dB is obtained
with the QPSK format and more than 14 dB with the 16-QAM for both signals. But, this range is higher for both

3
ICTON 2017 We.C2.2

modulation types when experimental setup limitation is cancelled. Figure 4c shows decrease in the EVM when
AWG noise is suppressed from simulation for 16-QAM. The EVM curve intersects with a limit value of 9% at
a power of -10 dBm either than -6 dBm for the previous simulations.
The influence of the wireless channel between access points and users can be added to analyze the global
system. This propagation channel is based on the model of Saleh-Valenzuela taking into account multipath and
free space absorption effects was introduced in this type of system simulation as in [11].

(a) (b) (c)

Figure 4. EVM measurement and simulation results with: (a) QPSK and (b) 16-QAM modulation formats,
(c) effect of AWG noise on EVM with 16QAM.

5. CONCLUSION
Electrical modeling of optical components on ADS has been allowed to fully analyze the RoF links. The analog
characteristics of the optical link can be easily determined from analog circuit design tools. The EVM of the
transmitted Wireless signal can be either approach from the signal to noise ratio or from co-simulation method
once system non-linearities are present. EVM simulations of the transmitted wireless OFDM signal are very
close to measured values when all electrical components and setup impairments are included. This means that
a realistic system can fully be analyzed and optimized with this method. The RoF link effect on the wireless
signals transmission is similar for both OFDM and FBMC modulation methods except that side lobe suppression
ratio of FBMC is much higher than the conventional OFDM which should be benefit for multiband transmission.

REFERENCES
[1] A-L. Billabert et al.: Simulation of microwave optical links and demonstration of noise figure lower than
electrical losses, Int. J. Microw. Wireless Technol., vol. 2, no. 06, pp. 497-503, 2010.
[2] W-E. Kassa et al.: Electrical modeling of semiconductor laser diode for heterodyne RoF system simulation,
IEEE J. of Quantum Electronics, vol. 49, no. 10, pp. 894-900, 2013.
[3] R.S. Tucker and D.J. Pope: Microwave circuit models of semiconductor injection laser, IEEE Trans.
Microw. Theory and Techniques, vol. 29, no. 3, Mar. 1982.
[4] A. Hilt: Microwave harmonic generation in fiber-optical links, in Proc. Microwaves, Radar and Wireless
Communications, MIKON-2000, Wroclaw, 2000.
[5] P. Banelli et al.: Modulation formats and waveforms for 5G networks: Who will be the heir of OFDM?,
IEEE Signal Processing Magazine, vol. 31, no. 6, pp. 80–93, 2014.
[6] A. I. Pérez-Neira et al.: MIMO signal processing in offset-QAM based filter bank multicarrier systems,
IEEE Transactions on Signal Proc., vol. 64, no. 21, pp. 5733-5762, Nov. 2016.
[7] P. Siohan, C. Siclet, and N. Lacaille: Analysis and design of OFDM/OQAM systems based on filter bank
theory, IEEE Transactions on Signal Processing, vol. 50, no. 5, pp. 1170-1183, May 2002.
[8] R. Zakaria and D. Le Ruyet: SER analysis by Gaussian interference approximation for FBMC system in
the presence of phase error, in Proc. 2015 IEEE International Conference on Communications (ICC),
London, 2015, pp. 2662-2667.
[9] A. Kabalan et al.: Direct and external modulation of IF over fiber for 60 GHz wireless applications, Int. J.
of Microw. Wireless Technol., vol. 8, pp. 1-6, 2015.