5G Waveform Competition: Performance

Comparison and Analysis of OFDM

and FBMC in Slow Fading and Fast
Fading Channels

Muhammad Imran(&), Aamina Hassan, and Adnan Ahmed Khan

College of Signals, National University of Sciences and Technology (NUST),

Islamabad, Pakistan
[email protected]

Abstract. An analysis of an emerging physical layer multicarrier (MC),

waveform Filter Bank Multi Carrier (FBMC) with Orthogonal frequency divi-
sion Multiplexing (OFDM) waveform under variety of channels has been per-
formed. Filter banks are the advanced form of MC sub band processing and
promise to deliver better results than OFDM. These filter banks exploit the
shortcomings arising due to the usage of Fast Fourier transforms (FFT) at the
trade-off of adding complexity to the systems i.e. Poly-Phase Filter Networks
(PPN). We investigate transceivers design, out of band Emission (OOB), Power
Spectral Density (PSD), Bit Error Probability (BEP) for these waveform. We
validate our analyses through simulations ascertaining FBMC’s performance
better than OFDM which makes it an ideal candidate for 5G physical layer.

Keywords: OFDM
FBMC
5G
Poly-Phase Filter

1 Introduction

Multi carrier modulation is nowadays widely adapted for broadband communication in

the form of OFDM [1]. The MC divides the spectrum into number of narrow band sub-
carriers aided through FFT. FFT produce number of orthogonal sub-carriers which do
not overlap with each other. Addition of CP to FFT at the transmitter side caters for the
multipath fading effects. Due to these properties OFDM standard is widely adapted in
Digital subscriber Lines, Power Line communication, cellular communication LTE and
Wi-Fi standard WiMAX [2]. Despite of the fact that OFDM is one of the most ideal
technique because of its robustness and accelerated implementation using FFT, many
drawbacks are associated with OFDM especially when its application is considered in
low latency and spectrum efficient environment. 5G applications demands low latency,
SE, robustness and loose synchronization [3]. OFDM suffers a higher OOB. OFDM
has a very high peak to average power (PAPR) ratio [4]. OFDM uses a CP which adds
to the spectral redundancy and it demands for a strict synchronization at the receiver.
Several techniques are proposed to cater for the OOB radiation and PAPR of
OFDM like windowing, companding [5], tone injections, interleaved OFDM, block

© Springer Nature Switzerland AG 2020

K. Arai and R. Bhatia (Eds.): FICC 2019, LNNS 69, pp. 51–67, 2020.
https://doi.org/10.1007/978-3-030-12388-8_4
52 M. Imran et al.

coding, clipping and cosine filters [6], but they actually do not cover all the merits of
5G that are needed. Scientific community has been working to propose another variant
of MC technique which is most promising one. FBMC appeared in 1960’s preceding
OFDM. Chang proposed the preliminary work for FBMC followed by Saltzberg [7]
who proposed the theory for a prototype filter design in such a way to as to meet the
Nyquist criteria, under which the prototype filter was supposed to achieve the ideal
reconstruction without Inter symbol Interference (ISI) [8]. The prototype filter gained
much attention as it claimed for the side lobes as small as possible [9]. This research
paper focuses on the transceiver structure of OFDM and FBMC, enlightening their key
differences, their performance and analysis in AWGN, Pedestrian Channel and
Vehicular Channel.
The already proposed waveforms for the 5G physical layer access spectrum
includes FBMC, Universal filter bank multi carrier (UFMC), Generalized frequency
division multiplexing (GFDM) [10] and OFDM [11].
FBMC applies the filter property to every subcarrier before it is transmitted; this
reduces the OOB radiation to several folds. UFMC applies the filter properties to the
whole symbol after being modulated and added [12] while in GFDM a tail biting
technique for CP is added [10]. All these techniques contribute to low PAPR as
compared to OFDM.
FBMC is one of the most widely researched waveform for 5G. In this paper only
FBMC and its implications are covered. FBMC is an evolved form of OFDM and MC
modulation scheme. It uses the blocks of FFT and inverse FFT (iFFT), adds PPN
structure to it [13] and produce more flexible results than already existing 4G multiple
access spectrum technology OFDM. FBMC do not use CP which adds to its SE. CP in
OFDM is used to cater for multipath fading when delay of the symbol is less than the
CP. The filter banks add features to FBMC that is easy computations, band selectivity,
low PAPR, low OOB radiation and low sampling rates.
The paper begins with the introduction to the 5G challenges and problems to 4G
physical layer scheme. Section 2 covers the OFDM transceiver model supported with
expressions and drawbacks with OFDM. In Sect. 3 FBMC transceiver model is cov-
ered in detail, along with its implementation ways. It also cites the key differences
between the two modulation schemes and proofed analyses. Section 4 is the simula-
tions and proven results section. The paper is concluded with future work, acknowl-
edgment and references.

1.1 Challenges to 4G Waveform

Challenges trigger the process of research and development. The cellular generations
are always heading towards new user needs and enhancements towards quality and
services. Since the last four decades’ cellular generation has evolved from 1G to 4G.
1G was the era of Analogue communication, 2G was the era of Digital communication,
3G dealt with the pseudorandom [14] and 4G with the multi carrier frequency mod-
ulation [15]. 5G will cover those aspects which the previous generations have not
achieved so far and will introduce more SE, high data rates, some killer applications
(MTC and IoT), battery efficient communication, target oriented services, high
capacity, low spectrum leakage, self-organizing networks (SON) [3] and much more.
 5G Waveform Competition … 53

Having said that we have OFDMA in current 4G networks. OFDM technologies have a
few shortcomings which make us re-think about a new flexible technique that can heal
the shortcomings of 4G waveform. The CP is added to the OFDM symbol, which is
1/4th of the symbol length itself, it is then concatenated at the beginning of the symbol
as shown in the Fig. 1. CP adds to extra redundancy and hence, reduced SE [16].
The OFDM symbols have higher PAPR and OOB radiations. These shortcomings
forced researchers to look for another solution. Filter banks made their way into these
problems and offer much better solutions along with FFT/iFFT band.

Fig. 1. CP in OFDM

2 OFDM and Its Drawbacks

OFDM is the one of the most widely used MC modulation technique. OFDM divides the
sub-carriers orthogonally by using FFT and iFFT modules which are easy to implement
[17]. Data is modulated using any QAM and higher order modulation and converted into
parallel streams, the parallel streams are of low data rates and easy to process. Then the
inverse Fourier transform block is inserted, iFFT is for the Orthogonality of the subcar-
riers see Fig. 2 for the block diagram. IFFT coverts the signal from frequency domain to
time domain but here in OFDM by convention it is used for Orthogonality purpose only.
(1) shows the IFFT operation and (2) shows the FFT operation for orthogonal sub carrier
generation. The Equations symbols are defined in Table 1.

FBMC TX

FBMC RX

Fig. 2. OFDM and FBMC transeiver

54 M. Imran et al.

Table 1. Equation symbols

Symbol Mapping in equations
M Data symbols
m Symbol index
x(n) Data in serial form
X(t) Data in time domain
X(k) Data in frequency domain
T Symbol period
K Interpolation co-efficient
k Impulse index
L Filter length
N FFT points
Eb Energy per bit
Pb Bit error probability
Ccomp Computational complexity
I Modulation array
No Noise density

N
2 1
X
xð t Þ ¼ X ½kej2pkt=N ð1Þ
k¼N=2

N
2 1
1 X
X ½K  ¼ ej2pkt=N ð2Þ
N t¼N=2

Orthogonality of the two subcarriers is proven by (3).

N
2 1
X j2pkt j2ppt
e N e N ¼ 0; 8 p 6¼ k
ð3Þ
t¼N=2

OFDM is highly sensitive to multipath fading affects so the demodulation of the

symbols become difficult [18]. Multipath fading add inter symbol interference (ISI) to
the symbol. Due to ISI delayed version of a symbol overlaps with the previous adjacent
symbol. To counter the effect of Multipath fading a guard band is introduced in OFDM
in such a way that a copy of the symbol is replicated and added to OFDM symbol. The
symbol is made longer and a copy of tail is inserted to the beginning of the last adjacent
symbol as shown in Fig. 1. In 802.11 standard CP is 1:4 of the original symbol. The
signal is periodic because of the usage of FFT, so FFT (n’) is the delayed version of the
original FFT (n) as shown by (4).
 5G Waveform Competition … 55

FFT ðn0 Þ ¼ eð2jprf ÞFFT ðnÞ ð4Þ

Delay in time division means rotation in frequency domain, so the correct signal is
obtained by anti-rotation in frequency domain. Now without Multipath rotation is
shown in (5).

yðtÞ ! FFT ðnÞ ! y½K  ¼ H ½kX ½k ð5Þ

(6) shows the behavior of the signal with Multipath fading.

 
yðtÞ ¼ FFT ðn þ n0 Þ ! a 1 þ e2jprk  X ½k ¼ H 0 ½kX ½k ð6Þ

There are certain drawbacks with OFDM waveform which include Inter-Carrier
Interference (ICI), PAPR, higher OOB, and low SE. These shortcomings pave ways to
redesign the physical layer for 5G applications.

2.1 Sensitivity to Carrier Offset and High PAPR

A small frequency offset results in high ICI in OFDM. Single carrier systems are less
sensitive to frequency and drift offsets. PAPR is the ratio of the maximum power of the
transmission system to the average power [19]. The amplifier of the OFDM must
behave in the linear regions in order to avoid saturation. To compensate the effects of
PAPR many techniques have been introduced which include clipping, windowing,
interleaving and block coding etc. [6]. However, these techniques do not fully qualify
with the results needed for 5G applications.

2.2 Higher OOB and Reduced SE

The side lobes of the sub carriers after modulation in OFDM have very high side lobes
in case of OFDM [20]. OOB radiations affect the data transferred in the main lobe and
increases the PSD outside the main lobe.
CP in OFDM is added at the cost of bandwidth, although its addition reduces
multipath fading but it also reduce SE. The Signal to Noise Ratio (SNR) loss due CP is
given by the (7), where Tcp is the CP time and the overall symbol duration is T which is
the addition of CP and the original symbol ðTcp þ TÞ as shown in Fig. 1.
 
Tcp
SNRloss ¼ 10 log10 1  ð7Þ
T

2.3 Time to Re-innovate Multiple Access Approach

After enlisting disadvantages of OFDM and keeping in mind the 5G requirements and
key drivers, it make the researchers think to redesign the multiple access scheme which
ensures high SE, higher data rates, low latency and should also qualify for the
requirements of IoT and MTC [3]. Several new modulation schemes as mentioned
56 M. Imran et al.

earlier are under research by the research and development sector of various countries.
FBMC is one of the most widely researched and oldest waveform for 5G. It is nearest
to OFDM and was proposed earlier than OFDM. Proposed research will analyse
FBMC as lurking waveform nominee for 5G and will focus on its bit error analysis in
more realistic environments than done before.

3 FBMC Model

FBMC is a derived form of Multi carrier (MC) modulation. It uses FFT and iFFT quite
differently than OFDM see Fig. 2 for detailed structure and difference of OFDM and
FBMC transceivers. Since in the presence of multipath fading, FFT modulated signals
cannot be reconstructed fully to original signal at Receiver (Rx) so in order to
reconstruct the signal at receiver we have two options. First is to add an extended
sequence or guard-band greater than channel impulse response which is followed in
OFDM. Second way is to keep intact the length of the symbol but add some cus-
tomization and processing to FFT block, second approach is used in FBMC Modu-
lation [13].
FBMC model uses the filter characteristics of the FFT block and implements a bank
of filters approach. An FIR filter is represented by (8).

sin pfM
Ið f Þ ¼ ð8Þ
M sin pf

Now Consider FFT (2) with k index and M FFT size we get (9).

1MX1
j2pki
yk ¼ xðn  M þ iÞe M ð9Þ
M i¼0

The coefficients of FFT appears to manifold with an exponential term, which

actually respond to a shift in frequency. With k/M as the shift in frequency, when a
number of FFT blocks combine a filter bank of size M is obtained. Orthogonality
condition is exhilarated at the zero crossings when frequencies are integer multiples of
1/M and it is in accordance with the Nyquist criteria. So there exist a low pass filter in
FFT. Banks of these filters are combined together in the form of PPN in FBMC. A filter
with zero frequency offset is termed as prototype filter [13]. Prototype filter is featured
with a constant K.

Filter Impulse Response Duration

K¼ ð10Þ
MC Symbol Duration T

Here, the Nyquist Criteria which promises for zero Inter symbol interference
(ISI) of symbols ensures the division of the filter into two halves and meet the sym-
metry conditions by taking squares of the frequency co-efficients. The co efficients are
 5G Waveform Competition … 57

derived in [13] (see Table 1) using the same Criteria. The interpolation formula for the
values of the coefficients is given in (11).
 
X
k¼1 K
sinðp f  MK MK
Hð f Þ ¼ Hk   k
 ð11Þ
k¼k1
MK sin p f  MK

3.1 Implementation Methods of FBMC

There are two methods to implement FBMC: One is extended FFT or frequency
spreading technique (FS) and other is PPN technique, Fig. 3a, b respectively. In the
first technique if M data is fed to the input of the frequency spreading block 2K−1 size
of input is fed to the adjacent iFFT block after passing through the frequency spreading
block, in this case a size ok KM iFFT block is used. The output is obtained with the
help of weighted spreading method in FS technique. Second approach is most widely
adapted for FBMC implementation and in this technique iFFT block of size M is used,
which is computationally efficient than the earlier method. In PPN-FBMC interleaved
phase shifter is added in parallel (in a filter bank form), through which the FFT block
size is reduced to K only instead of KM.

Fig. 3. FBMC implementation methods: a FS-FBMC b PPN-FBMC

In this paper PPN technique is followed for the implementations of FBMC.

The PPN networks add more computations to the FFT and iFFT blocks, but the results
produce by them are amazing.

3.2 Modulation Scheme with FBMC

After deciding the filter bank Implementation method for FBMC, next task is to define
the modulation scheme to be followed with FBMC. In FBMC Orthogonality in needed
in sub channels only, not in sub-carriers unlike OFDM, so any modulation scheme can
be used with FBMC when we go for half-duplex communication. If we consider full
58 M. Imran et al.

duplex case, a little bit think process is needed as now we ought to adventure with the
neighboring/adjacent sub channels as well.
One way is to use even index of the sub channel with the real part of iFFT and odd
index of the sub channels with the imaginary part of the iFFT. As the filter follows
Nyquist criteria we can get a basis and increased throughput to choose our modulation
scheme. The throughput is achieved due to the symmetry of the filter, the imaginary
part crosses through the time axis of symbol period at the integral multiple of it, and the
real part cross through half of the symbol period at the odd multiple of it. Offset
Quadrature Amplitude Modulation (OQAM) is adapted to fulfill the criteria of a full-
duplex mode in FBMC [21]. The offset corresponds to the sub channel spacing
between the real and the imaginary part of the symbol.

3.3 Key Differences Between the Two Waveforms and Our Analysis
After the research of Tx, Rx models of OFDM and FBMC we list down their key
differences and our research analysis in rationalized form in Table 2. These differences
are the roadmap for the simulations and testing the two schemes. These logical

Table 2. Summary of the Differences between OFDM and FBMC

Parameters Competitor Waveforms Analysis
OFDM FBMC
FFT/iFFT FFT/iFFT used as in Customized FFT/iFFT Processing is added in FFT
their standard form, blocks used blocks of FBMC at the cost
no modifications of larger computations than
OFDM, this increases the
complexity of the overall
system of FBMC
Guard CP required No CP required SE is achieved in FBMC
band/CP
Modulation Higher order of Higher order of OQAM OQAM in FBMC in needed
QAM for full- modulation for full to fully exploit the spectrum
duplex e.g. QAM, duplex. QAM Modulation and to maintain
16QAM, 64 QAM for half-duplex orthogonality between sub
channels
Orthogonality Requires Requires orthogonality in Neighboring channels
Orthogonality in adjacent sub-channels orthogonality is achieved
sub-carriers only using OQAM Modulation
Filters Window filter PHYDAS filter/hermite Localized frequency co-
generally prototype filter/IOTA efficients are added in the
filter form of PHYDAS filter,
which fruits to low OOB
emissions
Frequency Divides the given Divides the given PPN along with OQAM
exploitation frequency into frequency into number of accomplish the task for full
number of sub sub channels throughput in FBMC
carriers
 5G Waveform Competition … 59

differences are the base line for the re-design of a new waveform which meets the
requirement of the 5G applications, out of which low latency, SE and loose syn-
chronization directly demands from the physical layer. Our research analysis followed
by our simulations and testing validate our research.

4 Simulations and Results

FBMC and OFDM modulations are simulated using MATLAB 2015. The Fig. 2
blocks are adapted as simulation Algorithm in the same sequence. Several parameters
of the two waveforms are simulated and compared. Filters comparison, subcarrier
comparison, PSD and the BEP are simulated. BEPs are simulated using the expressions
derived already in [22, 23].

4.1 Filters Comparison

OFDM and FBMC filters are simulated both in time domain and frequency domain.
The coefficients values used for the PHYDAS filters are taken from Table 1 of FBMC:
A Primer which is the fundamental document for FBMC Model [13]. The parameters
for filters, subcarriers and PSD are enlisted in Table 3. For the filter comparison the
parameter channel number, M is kept 24 to visualize the difference between the side
lobes of the two filters, if M is taken the same as that in the BER simulations the side
lobes difference would not be clear.

Table 3. Parameters for filters, subcarriers and PSD

Parameters OFDM FBMC
M 24 24
Modulation QAM OQAM
Modulation order 64 64
Sub carrier 7 7
k – 4

Figure 4 shows the OFDM window filter and FBMC PHYDAS filter in time
domain. In time domain the filter used for OFDM is generally a window filter which is
applied to a band of subcarriers to remove clipping. In FBMC a PHYDAS filter is
designed and applied to each sub-carrier before it is transmitted.
Figure 5 shows the comparison of filters in frequency domain, it is clear that side
lobes of FBMC filter are much lower than OFDM filter. Due to very low OOB
radiations in FBMC, it undergoes a smaller ICI than OFDM and is ideal to use in
cognitive spectrum.
60 M. Imran et al.

Filters in Time Domain

1.4

1.2
FBMC

OFDM
1

0.8
PSD

0.6

0.4

0.2

-0.2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Normalized Time

Fig. 4. Filters impulse response in time domain

Filters Impulse respnse in frequency domain

40

30 FBMC
OFDM
20

10
Magnitude (db)

-10

-20

-30

-40

-50

-60

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Normalized Frequency

Fig. 5. Filters impulse response in frequency domain

4.2 Subcarriers Comparison

The sub carriers of the two competitors are simulated using the same parameters as in
filters comparison see Figs. 6 and 7 for the simulation results. The x-axis represents the
normalized frequency and y- axis is the PSD. ICI in OFDM is greater than FBMC.
Moreover, the PSD curve of both of the subcarriers show that the side lobes of the
OFDM bear greater PSD in comparison to the side lobes of FBMC. For an energy
efficient system the side lobes must possess very small PSD.
Figure 8 shows a PSD difference of FBMC and OFDM. The curve shows a visible
difference in the PSD of the two waveforms. The PSD results validate our analyses that
when PHYDAS filter is applied in FBMC, the side lobes of the subcarriers are sup-
pressed and don’t interfere with the adjacent sub channel lobes unlike OFDM
subcarriers.
 5G Waveform Competition … 61

OFDM subcarriers
0

-10

-20

PSD
-30

-40

-50

-60
-1 -0.5 0 0.5 1 1.5
Normalized frequency

Fig. 6. OFDM subcarriers

FBMC SubCarriers
0

-20

-40
PSD

-60

-80

-100

-120
-1 -0.5 0 0.5 1 1.5
Normalized frequency

Fig. 7. FBMC subcarriers

4.3 BEP of OFDM and FBMC in AWGN Channel

In AWGN channel OFDM is simulated without CP as AWGN channel has no mul-
tipath fading effects so CP is not required in this channel. BEP expressions for AWGN
channel are taken from [23]. Pb is the probability of error in (12). The parameters for
the BER simulations are cited in Table 4. The parameters for the earlier parameters are
different to as to clarify the curves and results.
pffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
I1 3log2 IEb
Pb  pffiffi pffiffi erfc½  ð12Þ
I log2 I 2ðI  1ÞN0
62 M. Imran et al.

PSD of OFDM and FBMC

0

OFDM
-10
FBMC

Power Spectral Density [dB]

-20

-30

-40

-50
PSD Difference
-60

-70

-80

-90

-100
-50 -40 -30 -20 -10 0 10 20 30 40 50
Normalized Frequency

Fig. 8. PSD of OFDM and FBMC

Figure 9 indicates that in AWGN channel distribution, OFDM and FBMC shows
almost the same BEP, which strengthen the fact that it is the CP which caters for the
multipath fading and multipath fading don’t occur in AWGN channel. Moreover, it
increases the keenness to simulate the waveforms in more realistic channels to
investigate their performance and see a visible difference.

4.4 BEP of OFDM and FBMC in Vehicular Channel A and Pedestrian

Channel A
Vehicular channel Model A is the fast fading channel and Pedestrian Channel Model B
is slow fading channel. The parameters followed for BEP simulations under these
channels are the same as for AWGN channel. The BEP is now simulated using CP for
OFDM with one tap-equalizer. The channel Models are defined by the ITU standard

Table 4. Parameters for BEP simulations

Parameters OFDM FBMC
M 32 32
Modulation QAM OQAM
Modulation order 64 64
CP ¼ carrier spacing 0
Symbols 3 7
K – 6
Carrier frequency 2.5 GHz 2.5 GHz
Sub carrier spacing 15 kHz 15 kHz
 5G Waveform Competition … 63

Fig. 9. BER of OFDM and FBMC in AWGN channel

Table 5. ITU parameters for vehicular channel model A and pedestrian channel model B
Tap Vehicular channel model A Pedestrian channel model B
Relative delay (ns) Average power (dB) Relative delay (ns) Average power (dB)
1 0 0 0 0
2 310 −1 200 −0.9
3 710 −9 200 −4.9
4 1090 −10 1200 −8.0
5 1730 −15 2300 −7.8
6 2510 −20 3700 −23.9

[24], which are also shown in Table 5. The objects for the channel Models are present
in MATLAB under 802.11g Libraries. Equations for the BEP Simulations are taken
from IEEE letter [22].
Figures 10 and 11 are the same BEP curves under Vehicular Channel Model A
except for the fact that in Fig. 10 BEP is plotted against velocity and in Fig. 11 BEP is
plotted against SNR. The results indicate that FBMC performs better than OFDM in
mobile channel. CP-OFDM shows better BEP than FBMC in lower velocities but in
higher velocities FBMC takes the lead. CP-OFDM shows the low BEP in low
velocities because the CP perfectly gratifies for multipath fading but as soon as the
velocity increases the PHYDAS filter in FBMC plays its part. This indicates that when
the channel becomes dominated by Doppler shift (the change in frequency which
produce delay spread) FBMC performs better than CP-OFDM due to its PHYDAS
filter. The performance of OFDM without CP is the poorest one in both Fading
channels.
64 M. Imran et al.

Fig. 10. BEP curves, 500 km/h velocity

BEP in Vehicular Channel

10 0
OFDM
CP-OFDM
FBMC
Bit Error Probability

10 -1

10 -2

10 -3
-5 0 5 10 15 20 25 30 35
Signal-to-Noise Ratio (dB)

Fig. 11. BEP vs. SNR in vehicular channel, 500 km/h

Figure 12 is the BEP curve for Pedestrian Channel Model B. As Pedestrian channel
is a slow fading channel, FBMC and CP-OFDM shows the same BEP curves as in the
case of AWGN channel. Figures 11 and 12 show that FBMC’s performance in slow
fading channel is better than its performance in fast fading channel.
 5G Waveform Competition … 65

Fig. 12. BEP vs. SNR in pedestrian channel, 10 km/h velocity

5 Conclusion

The research started with enlisting the key drivers for 5G and keenness to redesign its
physical Layer. The concept behind FBMC and OFDM modulation schemes was
developed, alongwith their transceivers differences supported with expressions. Fur-
thermore, the respective filters (both in time and frequency domain) were simulated and
visualized their PSD. For the BEP simulations the AWGN channel was used for fast
fading Vehicular channel and a slow fading Pedestrian channel. The results show that
FBMC has much lower OOB radiations, almost comparable BEP in AWGN channel
and a very good BEP in fast (Vehicular) and slow fading (Pedestrian) channels as
compared to OFDM. This makes FBMC an ideal candidate waveform for 5G. Proposed
work achieved very good bit error ratios for FBMC at the cost of adding more com-
putations to the systems in the form of filters and PPN network.

6 Future Work

5G waveform is to be standardized in 2020. The paper concludes with proposed results

that FBMC offered much lower OOB radiations. In cognitive systems where white
spectrum is to be exploited, localized waveform frequencies are desired. FBMC is one
of the best scheme to be considered as it has localized frequency co-efficients and high
spectrum efficiency.

Acknowledgements. This research paper is an outcome of Masters research work of Aamina

Hassan. Authors are grateful to National University of Sciences & Technology (NUST),
Islamabad, Pakistan, for the motivation, encouragement and support throughout the research
process.
66 M. Imran et al.

References
1. Wu, Y., Zou, W.Y.: Orthogonal frequency division multiplexing: a multi-carrier modulation
scheme. IEEE Trans. Consum. Electron. 41, 392–399 (1995)
2. Koffman, I., Roman, V.: Broadband wireless access solutions based on OFDM access in
IEEE 802.16. IEEE Commun. Mag. 40, 96–103 (2002)
3. Andrews, J.G., Buzzi, S., Choi, W., Hanly, S.V., Lozano, A., Soong, A.C.K., Zhang, J.C.:
What will 5G be? IEEE J. Sel. Areas Commun. 32, 1065–1082 (2014)
4. Amstrong, J.: OFDM for optical communications. J. Lightwave Technol. 27(3), 189–204
(2009)
5. Mahmud, Z., Hossain, M.S., Islam, M.N., Abdullah, M.I.: Comparative study of PAPR
reduction techniques in OFDM. ARPN J. Syst. Softw. 1 (2011)
6. Bisht, M., Joshi, A.: Various techniques to reduce PAPR in OFDM systems: a survey. Int.
J. Signal Process. Image Process. Pattern Recogn. 8, 195–206 (2015)
7. Saltzberg, B.: Performance of an efficient parallel data transmission system. IEEE Trans.
Commun. Technol. 15, 805–811 (1967)
8. Farhang-Boroujeny, B.: Filter Bank Multicarrier Modulation: A Waveform Candidate for 5G
and Beyond. ECE Department, University of Utah, Salt Lake City, UT 84112, USA (2014)
9. Farhang-Boroujeny, B.: Filter Bank Multicarrier (FBMC): An Integrated Solution to
Spectrum Sensing and Data Transmission in Cognitive Radio Networks, USA (2009)
10. Michailow, N., Matthé, M., Gaspar, I.S., Caldevilla, A.N., Mendes, L.L., Festag, A.,
Fettweis, G.: Generalized frequency division multiplexing for 5th generation cellular
networks. IEEE Trans. Commun. 62, 3045–3061 (2014). IEEE Communications Society
11. Gerzaguet, R., et al.: The 5G candidate waveform race: a comparison of complexity and
performance. EURASIP J. Wirel. Commun. Netw. 13 (2017)
12. Schaich, F., Wild, T.: Waveform contenders for 5G—OFDM vs. FBMC vs. UFMC. In: 6th
International Symposium on Communications, Control and Signal Processing (ISCCSP),
Conference (2014)
13. Bellanger, M., et al.: FBMC Physical Layer: A Primer. http://www.ict-phydyas.org (2010)
14. Karjaluoto, H.: An Investigation of Third Generation (3G) Mobile Technologies and
Services. University of Jyväskylä, October 2006
15. Reddy, M.H., Jaswanth, S., Pramod, N.V.: Evolution of mobile networks: from 1G TO 4G.
Adv. Res. Electr. Electron. Eng. 3(4), 307–310 (2016)
16. Shah, D.C., Rindhe, B.U., Narayankhedkar, S.K.: Effects of cyclic prefix on OFDM system.
In: Proceedings of the ICWET 2010 International Conference & Workshop on Emerging
Trends in Technology, India, January 2010
17. Pandharipande, A.: Principles of OFDM. IEEE Potentials 21, 16–19 (2002)
18. Sadouki, B.R., Chaker, H., Djebbouri, M.: The effect of multipath on the OFDM system. Int.
J. Comput. Appl. 89(13) (2014). ISSN 0975-8887
19. Gangwar, A., Bhardwaj, M.: An overview: peak to average power ratio in OFDM system &
its effect. Int. J. Commun. Comput. Technol. 1, 22–25 (2012)
20. Baltar, L.G., Waldhauser, D.S., Nossek, J.A.: Out-of-band radiation in multicarrier system: a
comparison, Germany (2009)
21. Sahin, A., Guvenc, I., Arslan, H.: A survey on multicarrier communications: prototype
filters, lattice structures, and implementation aspects. IEEE Commun. Surv. Tutor. 16(3),
1312–1338 (2014)
 5G Waveform Competition … 67

22. Nissel, R., Rupp, M.: OFDM and FBMC-OQAM in doubly-selective channels: calculating
the bit error probability. IEEE Commun. Lett. 21, 1297–1300 (2017)
23. He, Q., Schmeink, A.: Comparison and evaluation between FBMC and OFDM systems. In:
19th International ITG Workshop on Smart Antennas, WSA 2015, 3–5 March 2015
24. Draft 802.20 Permanent Document: Channel Models for IEEE 802.20 MBWA System
Simulations. IEEE (2003)