International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 07 Issue: 04 | Apr 2020 www.irjet.net p-ISSN: 2395-0072

A New Framework in Filter Bank Multi Carrier Modulation with OQAM

Lakshmi Priya J1, Kumaratharan N2
1 Student & Department of ECE, Sri Venkateswara College of Engineering, Tamilnadu
2 Professor & Department of ECE, Sri Venkateswara College of Engineering, Tamilnadu

---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Major requirement in flexible allocation of In many aspects FBMC behaves similar to OFDM,
available time- frequency resources is seen as a major but some become a little bit challenging because of
difficulty in conventional multi-carrier modulation systems occurrence of imaginary interference [3] or Multiple
with rapid use of wireless communication systems. This lead to
further modifications in windowing or filtering techniques Input Multiple Output(MIMO). Although drawbacks
involved in multi-carrier modulations. In addition, a different occur in combining FBMC and MIMO, the most effective
modulation technique called Filter Bank Multicarrier(FBMC) method has been sought out i.e.) spreading of data
is deployed along with Doubly Selective Channel Estimation
symbols in either time or frequency domain thereby
which offers better spectral properties. A comparison of OFDM
with FBMC is made and thus identified two key observations enhances to eliminate imaginary interference that
for future mobile systems. In addition, Offset Quadrature occur. This might require some additional processing
Amplitude Modulation(OQAM) is mainly focused because of its [4]. For multicarrier systems, one tap channel is been
higher spectral efficiency. A unifying framework discussion
and performance evaluation of FBMC and OFDM systems are
used for modelling transmission over time-variant
provided and necessary comparisons of Bit Error Rate and multi-path channels which often simplifies the
Power Spectral Densities are made using simulation results. equalization. Minimum Mean Square Error[MMSE]
method of estimation is more commonly used to
Key Words: FBMC, OQAM, OFDM, Windowing, filtering.
estimate doubly selective channel estimation. Only one
1.INTRODUCTION OFDM symbol in time is considered which also requires
clustered pilots. To accurately estimate all elements of
The requirement of future mobile systems(5G) doubly selective channel, a compact matrix description
mainly resides on larger range of user cases from low is incorporated with very few pilots.
latency communication to machine type
communication by providing flexible allocation of Thus for proper modulation and multiple access
available time-frequency resources. This allocation is scheme in future mobile systems, two main key
difficult in conventional OFDM systems because of its observations has been kept forward: (i) To more
poor spectral properties. Thus we alternatively employ efficiently support may user requirements and
new waveform derivatives like windowed OFDM, characteristics of channel, allocation of flexible time-
Universal Filtered Multi Carrier (UFMC) [1]. Another frequency is made. (ii) Majorly in very dense
prominent scheme with much better spectral heterogeneous network providing lower delay spread
properties that employs pulses in both time as well as by significantly utilizing MIMO and higher carrier
frequency is Filter Bank Multi Carrier(FBMC) along frequencies.
with Offset Quadrature Amplitude Modulation(OQAM) Hence the above two observations efficiently make
[2]. Although there exist many variants in FBMC, our FBMC as more feasible option for future mobile
main focus is on OQAM as it provides much better systems as it provides better localization in time as
spectral efficiency. More arguments like lower delay well as frequency resources and also guarantees simple
spread or decreasing cell size and spatial filtering one tap equalizer is sufficient to provide lower delay
through beamforming can become even more spread thereby reducing the necessary of complicated
noteworthy for future mobile communication because receiver structures. The important contributions of this
of more densification in network that may lead to paper can be summarized as follows:
increase of propagation path loss.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 6404
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 04 | Apr 2020 www.irjet.net p-ISSN: 2395-0072

 We compare bit error probability and power To simplify the notation mainly for CP-OFDM and
spectral density of various existing systems also its derivatives, we simply use same transmit and
with FBMC and put-forth the betterment of receive pulse.
FBMC than other modulation techniques.
 A doubly selective channel estimation is been Some fundamental limitations do occur for
proposed which does not require clustered multicarrier systems as stated in Balian–Low theorem
pilot plots for estimation.
[6] that all the following mathematical properties are
2. SYSTEM MODEL satisfied at the same time i.e.) having maximum symbol
density TF=1, time localization σt<∞ and frequency
Generally, the pulses through which the information localization σf<∞. Generally, in OFDM, frequency
is transmitted commonly occupy only a smaller localization is not satisfied as the prototype filter is
bandwidth which might overlap in time and frequency. based on rectangular pulse function. So in addition,
Mathematically, u(t) the transmitted signal of Cyclic Prefix (CP) is added such that it sacrifices
multicarrier system under time domain is given by, spectral efficiency and thereby gaining high robustness
in frequency selective channels. In contrary, FBMC
satisfies the condition with least real orthogonality
condition such that only real valued symbols xl,m Є R
………………………….. (1) are transmitted.

Where x l,m is the transmitted symbol at l This constitutes for the idea of designing a prototype
subcarrier position and k time position. filter of time frequency spacing TF=2 based on hermite
polynomials [7] and thus time frequency spacing is
The basis transmitted pulse gl,m(t) of the above reduced to TF=1/2. The imaginary interference created
expression is given by, can be eliminated by considering only the real part.

3. MULTICARRIER MODULATION
3.1 CP-OFDM:
…………………… (2) One of the most prominent technique applied in
Wireless LAN and long term evolution(LTE) CP-OFDM
greatly reduces the computational complexity as it
With s(t) as prototype filter of time and frequency employs rectangular transmit and receive pulses. But
shifted version and also T and F denotes time spacing the rectangular pulses are not been localized in
and frequency spacing respectively and θl,k denotes the frequency domain that leads to high Out Of Band (OOB)
phase shift. emission as depicted in fig. (1) which stays as greatest
drawback of CP-OFDM. In addition to that the cyclic
The received symbols after transmission through a prefix tries to simplify the equalization in frequency
channel are decoded by received signal r(t) through selective channels thereby reducing its spectral
basis pulses gl,m(t) is given by, efficiency.

Furthermore, to minimize the OOB emission, the 3rd

Generation Partnership Project(3GPP) in recent
surveys consider windowing and filtering [8]. The
………… (3)
windowed OFDM is always referred as WOLA i.e.)
Which corresponds to matched filter that maximizes Weighted OverLap and Add [8]. Here at the transmitter
the SNR in AWGN channel. side, the edges of rectangular pulses are replaced by
smoother function and the nearby WOLA symbols

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 6405
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 04 | Apr 2020 www.irjet.net p-ISSN: 2395-0072

overlap in time. The receiver side also thereby applies spectral efficiency. The prototype filter for FBMC-QAM
windowing with an exception that overlap and add is based on hermite polynomial [10] Hn(.) given by,
operation is executed in same WOLA symbol thus the
inter-band interference is minimized with the fact that ……….(4)
CP must account for both transmitter and also at
receiver. Where the coefficients are given by,

The filtered based OFDM proposes two methods m0 =1.412692577 m12 =-2.2611.10-9
namely Universal Filtered Multi-Carrier(UFMC) [9] and
filtered OFDM (f-OFDM). The UFMC implies sub-band m4 =-3.0145x10-3 m16 =-4.4570x10-15
wise filtering and Zero Padding (ZP) or conventional
m8 =-8.8041x10-6 m20 =-1.8633x10-16
CP guarantees the orthogonality. The sub band wise
filtering is employed at both transmitter and receiver These hermite pulses exhibit symmetries as it has same
so that at time-frequency spacing of TF=1.14, the shape in time and frequency.
orthogonality is satisfied. In order to improve the
spectral efficiency, the time-frequency spacing is Another prominent prototype filter is PHYDYAS [11]
reduced to TF=1.09 thereby allowing very least which is given by,
interference. The f-OFDM utilizes all subcarriers
belonging to specific user case. For a better
comparison, the same time-frequency spacing TF=1.09
is considered and thereby increasing the length of
transmit and receive filters thereby minimizing the
…..(5)
self- interference.
The coefficient ni depends on overlapping factor O.
3.2 FILTER BANK MULTI CARRIER MODULATION
considering O=4, the coefficients are given by,
FBMC is an advanced technique of OFDM where no
cyclic prefix is used. This results in greater spectral n1=0.97195983 n2=√2/2 n3=0.23514695
efficiency and increase the capacity of the system. Each
subcarrier of multicarrier signals are filtered and thus Thus on comparison between hermite and PHYDYAS
reducing side lobes. Thus FBMC is analyzed with filter filter, PHYDYAS has much better frequency localization
bank at both transmitter and receiver where it is called but poor time localization and also poor time frequency
as ‘synthetic filter bank’ at transmitter and ‘analytic localization.
filter bank’ at receiver.
3.4 FBMC-OQAM:
Difference between OFDM and FBMC lies in the FBMC-OQAM is more closely related to FBMC-QAM
choice of TF (time frequency spacing). In conventional but has similar spectral density like OFDM which has
OFDM, p(t) is a rectangular pulse of height one and no Cyclic Prefix. In this, the time-frequency spacing is
width T while the pr(t) is also a rectangular pulse of greatly minimized to T=T0/2 and F=1/T0 such that
height one but width is reduced to Tfft<T where TF=0.5 but the fact that only real part of the symbol is
Tfft=1/B. In FBMC, T=Tfft=1/B, however pr(t) and pt(t) transmitted. The name Offset-QAM is only because the
are greater than T. real part of the symbol is matched to the first time slot
and the second slot is matched with the imaginary part.
3.3 FBMC-QAM Loss of complex orthogonality is the major drawback in
Although there is no unique definition for FBMC- FBMC-OQAM. Furthermore, FBMC promotes efficient
QAM, we consider time frequency spacing TF=2 to separation of subsequent blocks by single guard
satisfy all desired properties thereby sacrificing its symbol thereby enhancing better time-frequency

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 6406
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 04 | Apr 2020 www.irjet.net p-ISSN: 2395-0072

localization. This implies particularities for some MIMO 4. DOUBLY SELECTIVE CHANNEL ESTIMATION
techniques, such as space-time block codes or The transmission over doubly selective channel is
maximum likelihood symbol detection as well as for given by the relation [5]
the channel estimation.
y = Dx+n………………………………..(6)
3.5 Coded FBMC-OQAM:
In order to straightforwardly employ all MIMO where x represents the transmitted data symbols, y
methods and channel estimation techniques known in gives the received symbol with n as Gaussian noise.
OFDM, we have to restore complex orthogonality in The transmission matrix D is given by [5]
FBMC-OQAM. This can be achieved by spreading D = QHHG………………………………(7)
symbols in time or frequency. Although such spreading
is similar to Code Division Multiple Access (CDMA), Where H is the time-variant convolution matrix
employed in 3G, coded FBMC-OQAM is different in the
sense that no rake receiver and no root-raised-cosine Generally, in doubly-selective channel the estimation
filter is necessary. Instead, we employ simple one-tap of channel impulse response H seems [13] to be
equalizers which is possible as long as the channel is problematic as the number of active subcarriers is
approximately flat in time (if we spread in time) or in often lower than Fast Fourier transform(FFT) i.e.)
frequency (if we spread in frequency). Because L<NFFT. Computational complexity is also another
wireless channels are highly under-spread, such major aspect. This huge computational burden may be
assumption is true in many scenarios. Furthermore, the caused even if one is accurately able to estimate the
good time-frequency localization of FBMC allows the impulse response. The sampled time- variant transfer
efficient separation of different blocks by only one function is interpolated, delivering an estimate of the
guard symbol and no additional filtering is necessary. full time-variant transfer function. This is possible
Another advantage can be found in the uplink. because of high correlation in time and frequency. The
Conventional FBMC-OQAM requires phase weighting factor has a major influence on the channel
synchronous transmissions θl,k = π/2(l + k)) which is estimation accuracy. Interference cancellation is also
problematic in the up-link (but not in the down-link). very important for the channel estimation process
In coded FBMC, this is no longer an issue because we because the LS channel estimates at pilot positions are
restore complex orthogonality. The main disadvantage, corrupted by interference. By cancelling this
on the other hand, is the increased sensitivity to interference the channel estimation accuracy can be
doubly-selective channels. This, however, was not an improved.
issue in our measurements. In some the authors utilize
Fast Fourier Transform (FFT) spreading, while others 5. RESULTS AND DISCUSSION
employ Hadamard spreading. The latter can be The BER curves for both FBMC and CP-OFDM
implemented by a fast Walsh Hadamard transform, are presented for 200 channel realizations and N=
which reduces the computational complexity and 1000 symbols per-subcarrier for each channel
requires no multiplications; we thus prefer it over FFT realization. The bit error rate performance of OFDM
spreading. Comprehensive overview of such Hadamard and FBMC was analyzed and the results show that
spreading approach and investigate the effects of a coded FBMC-OQAM has better performance than OFDM
time-variant channel is provided. Both, FBMC and systems.
OFDM, have the same Bit Error Ratio (BER), validating
the spreading approach. However, FBMC has much
better spectral properties.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 6407
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 04 | Apr 2020 www.irjet.net p-ISSN: 2395-0072

Fig.1. Bit Error performance of OFDM vs FBMC

In Fig.2. the power spectral density of both FBMC

and CP-OFDM are depicted for a system with 5 MHz of
bandwidth. FBMC has better spectral properties
compared to CP-OFDM. Windowing(WOLA) and
Fig.3. Spectral efficiency
filtering (UFMC,f-OFDM) can improve the spectral
properties but FBMC still performs much better and The time-frequency efficiency depends on the
has additional advantage of maximum symbol density. number of subcarriers and the number of time symbols

Fig.4. BER over SNR for FBMC using auxiliary symbols

In Fig.4. we employ four auxiliary symbols per

Fig.2. Power Spectral Density
pilot plot, guaranteeing that the channel induced
interference at pilot position is relatively low and also
the auxiliary symbol power is close to zero. Here the
performance is better because the additional pilot
symbol is taken from the auxiliary symbols and not
from data symbols.

Compared to Fig.4. the BER is slightly worse in Fig.5.

because of spreading and the fact that the additional
pilot power is taken from the data symbols. However,
the spreading method gas a lower overhead and thus a
higher data rate.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 6408
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 04 | Apr 2020 www.irjet.net p-ISSN: 2395-0072

REFERENCES
[1] F. Schaich, T. Wild, and Y. Chen, “Waveform contenders
for 5Gsuitability for short packet and low latency
transmissions,” in IEEE Vehicular Technology Conference
(VTC Spring), 2014, pp. 1–5.

[2] B. Farhang-Boroujeny, “OFDM versus filter bank

multicarrier,” IEEE Signal Processing Magazine, vol. 28, no. 3,
pp. 92–112, 2011.

[3] R. Nissel and M. Rupp, “Bit error probability for pilot-

symbol aided channel estimation in FBMC-OQAM,” in IEEE
International Conference on Communications (ICC), Kuala
Lumpur, Malaysia, May 2016.

[4] R. Nissel and M. Rupp, “Pruned DFT spread FBMC: Low

Fig.5. BER for FBMC based on data spreading approach PAPR,low latency, high spectral efficiency,” IEEE Trans.
Commun., 2018, to appear.
6. CONCLUSION
[5] R. Nissel, S. Schwarz, and M. Rupp, “Filter bank
OFDM based multicarrier techniques like WOLA,
multicarrier modulation schemes for future mobile
UFMC and f-OFDM provide relatively higher spectral communications,” IEEE Journal on Selected Areas in
efficiency if they have higher number of subcarriers. Communications, vol. 35, no. 8, pp. 1768–1782, 2017.
Though not all possible use cases tend to provide
higher number of subcarriers in future mobile [6] J. G. Andrews, S. Buzzi, W. Choi, S. V. Hanly, A. Lozano, A.
communication systems there evolved FBMC systems C. Soong, and J. C. Zhang, “What will 5G be?” IEEE Journal on
Selected Areas in Communications, vol. 32, no. 6, pp. 1065–
where for a small number of subcarriers, it becomes
1082, 2014.
much more efficient than OFDM, mostly if the
transmission band is shared between different use [7] G. Wunder, P. Jung, M. Kasparick, T. Wild, F. Schaich, Y.
cases. Many challenges like MIMO and channel Chen, S. Brink, I. Gaspar, N. Michailow, A. Festag et al.,
estimation have been dealt with FBMC and thus in “5GNOW: nonorthogonal, asynchronous waveforms for
addition one tap equalizers are deployed which are future mobile applications,” IEEE Communications Magazine,
much more sufficient in many practical cases if the vol. 52, no. 2, pp. 97–105, 2014.
subcarrier spacing is matched to the channel statistics. [8] P. Banelli, S. Buzzi, G. Colavolpe, A. Modenini, F. Rusek,
In highly doubly selective channels, we can switch from and A. Ugolini, “Modulation formats and waveforms for 5G
an FBMC-OQAM transmission to an FBMC-QAM networks: Who will be the heir of OFDM ?: An overview of
transmission, thus deliberately sacrificing spectral alternative modulation schemes for improved spectral
efficiency but gaining robustness. This leads to an even efficiency,” IEEE Signal Process. Mag., vol. 31, no. 6, pp. 80–
higher SIR than in CP-OFDM. While it is true that the 93, 2014.
computational complexity of FBMC is higher than in
[9] 3GPP, “TSG RAN; study on new radio access technology;
windowed OFDM, both methods require the same basic
(release 14),”
operations, allowing us to reuse many hardware http://www.3gpp.org/DynaReport/38912.htm, Sept. 2016.
components. Thus FBMC based on auxiliary symbol
method outperforms OFDM in terms of BER and FBMC [10] “TSG RAN; study on new radio access technology; radio
data spreading method has an additional advantage of interface protocol aspects; (release 14),”
higher data rate. http://www.3gpp.org/DynaReport/38804. htm, Mar. 2017.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 6409
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 04 | Apr 2020 www.irjet.net p-ISSN: 2395-0072

[11] “TSG RAN; study on new radio access technology;

physical layer aspects; (release 14),”
http://www.3gpp.org/DynaReport/38802. htm, Mar. 2017.

[12] M. Bellanger, D. Le Ruyet, D. Roviras, M. Terre´, J.

Nossek, L. Baltar, Q. Bai, D. Waldhauser, M. Renfors, T.
Ihalainen et al., “FBMC physical layer: a primer,” PHYDYAS,
January, 2010.

[13] B. Farhang-Boroujeny, “OFDM versus filter bank

multicarrier,” IEEE Signal Processing Magazine, vol. 28, no. 3,
pp. 92–112, May 2011.

[14] A. Sahin, I. Guvenc, and H. Arslan, “A survey on

multicarrier communications: Prototype filters, lattice
structures, and implementation aspects,” IEEE
Communications Surveys Tutorials, vol. 16, no. 3, pp. 1312–
1338, December 2012.

[15] H. Bo¨lcskei, “Orthogonal frequency division

multiplexing based on offset QAM,” in Advances in Gabor
analysis. Springer, 2003, pp. 321–352.

[16] B. Farhang-Boroujeny, “Filter bank multicarrier

modulation: A waveform candidate for 5G and beyond,”
Advances in Electrical Engineering, vol. 2014, December
2014.

[17] H. Asplund, K. Larsson, and P. Okvist, “How typical is the

typical urban” channel model?” in IEEE Vehicular
Technology Conference, 2008, pp. 340–343.

[18] L. Bernado´, T. Zemen, F. Tufvesson, A. F. Molisch, and C.

F. Mecklenbra¨uker, “Delay and doppler spreads of
nonstationary vehicular channels for safety-relevant
scenarios,” IEEE Transactions on Vehicular Technology, vol.
63, no. 1, pp. 82–93, 2014.

[19] S. Payami and F. Tufvesson, “Delay spread properties in

a measured massive MIMO system at 2.6 GHz,” in IEEE
International Symposium on Personal, Indoor, and Mobile
Radio Communications (PIMRC), Sept 2013, pp. 53–57.

[20] S. Caban, J. A. Garc´ıa-Naya, and M. Rupp, “Measuring

the physical layer performance of wireless communication
systems,” IEEE Instrumentation & Measurement Magazine,
vol. 14, no. 5, pp. 8–17, 2011.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 6410