battle of the waveforms for 5g kurt.pdf · itu wcrl battle of the waveforms for 5g gunes karabulut...

ITU WCRL

Battle of the Waveforms for 5G

GUNES KARABULUT KURT, SELAHATTIN GOKCELI

gkur t@i tu .edu. t r, gokce l i s@i tu .edu. t r

W i re less Communicat ions and Research Laboratory (WCRL)

ISTANBUL TECHNICAL UNIVERSITY

ITU WCRL

OUTLINE

h"p://sine.ni.com/cs/app/doc/p/id/cs-17378#

Introduc;on: OFDM/OFDMA

5GChallenges

WaveformDesignTargets

Implementa;onPrespec;ve:

UniversallyFilteredMul;-carrier(UFMC)Systems

ErrorPerformance&Sidelobelevels

FurtherImprovements

Summary

IEEE 5G GREECE SUMMIT 2

ITU WCRL

OFDM/OFDMA

Orthogonal Frequency Division Multiplexing/Orthogonal Frequency Division Multiple Access: Enables transmission of parallel data streams

þ  High data rates

þ  Robustness to frequency selectivity

ý  Sensitivity to time/frequency offsets

ý  High peak to average power ratio (PAPR)

ý  High sidelobe levels

IEEE 5G GREECE SUMMIT 3

ITU WCRL

5G Challenges Solutions to OFDM/OFDMA problems: ê Spectral efficiency ê Energy efficiency

IEEE 5G GREECE SUMMIT 4

ITU WCRL

5G Challenges Solutions to OFDM/OFDMA problems: ê Spectral efficiency ê Energy efficiency

OFDM/OFDMAmaynotaddress5Gconstraints!

IEEE 5G GREECE SUMMIT 5

ITU WCRL

5G Challenges Solutions to OFDM/OFDMA problems: ê Spectral efficiency ê Energy efficiency

OFDM/OFDMAmaynotaddress5Gconstraints!

Newwaveformsmaybeasolu;on

IEEE 5G GREECE SUMMIT 6

ITU WCRL

Design Targets é Spectral efficiency é Energy efficiency ê PAPR ê Sidelobe levels

+ Simpler synchronization

IEEE 5G GREECE SUMMIT 7

ITU WCRL

Candidate Techniques ü  Filter Bank Multicarrier Modulation [SIOHAN, 2002]

ü  Generalized Frequency Division Multiplexing [FETTWEIS, 2009]

ü  Filtered-OFDM [ABDOLI, 2015]

ü  Zero-tail DFT-spread-OFDM [BERARDINELLI, 2013]

ü  Universal Filtered Multi-Carrier [VAKILIAN, 2013]

ü  …

IEEE 5G GREECE SUMMIT 8

ITU WCRL

Implementation Perspective

A good compromise: UFMC þ Better control of sidelobe levels/interference þ Robustness to syncronization sensitivity:

Carrier Frequency Offset/Timing Offset

IEEE 5G GREECE SUMMIT 9

X1 N1-IDFTFiltering

F1

DACX2 N2-IDFTFiltering

F2

XB NB-IDFTFiltering

FB

.

.

.

.

.

.

Cyclic-Prefix

RemovalADC

Synchronization

Filtering/Channel

Equalization

Estimated Bits2N-DFT

TRANSMITTER

RECEIVER

ITU WCRL

SDR Testbed

IEEE 5G GREECE SUMMIT 10

ITU WCRL

Measurement Results

IEEE 5G GREECE SUMMIT 11

ITU WCRL

Sidelobe levels (1/2)

IEEE 5G GREECE SUMMIT 12

ITU WCRL

Sidelobe levels (2/2)

Normalized frequency-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

PSD

(dBW

/Hz)

-100

-80

-60

-40

-20

0

20 32 Subcarriers

OFDMDFT-s-OFDMUFMC

(a)

CCDF0 1 2 3 4 5 6 7 8 9 10 11

PAPR

(dB)

10-5

10-4

10-3

10-2

10-1

100

OFDMDFT-s-OFDMUFMC

(b)

Normalized frequency-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

PSD

(dBW

/Hz)

-100

-80

-60

-40

-20

0

2048 Subcarriers

OFDMDFT-s-OFDMUFMC

(c)

CCDF0 1 2 3 4 5 6 7 8 9 10 11

PAPR

(dB)

10-5

10-4

10-3

10-2

10-1

100

OFDMDFT-s-OFDMUFMC

(d)

Normalized frequency-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

PSD

(dBW

/Hz)

-100

-80

-60

-40

-20

0

2064 Subcarriers

OFDMDFT-s-OFDMUFMC

(e)

CCDF0 1 2 3 4 5 6 7 8 9 10 11

PAPR

(dB)

10-5

10-4

10-3

10-2

10-1

100

OFDMDFT-s-OFDMUFMC

(f)Figure 5.3 : For 16, 24 and 32 subcarriers, PSD results of UFMC, OFDM and

DFT-s-OFDM are shown at (a), (c) and (e), respectively. Moreover,respective combined PAPR results are shown at (b), (d) and (f).

57

IEEE 5G GREECE SUMMIT 13

ITU WCRL

Further Improvements

Normalized frequency-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

PSD

(dBW

/Hz)

-100

-80

-60

-40

-20

0

20 32 Subcarriers

ZP-DFT-s-OFDMZP-DFT-s-UFMC

(a)

CCDF0 1 2 3 4 5 6 7 8 9 10 11

PAPR

(dB)

10-5

10-4

10-3

10-2

10-1

100

OFDMZP-DFT-s-OFDMZP-DFT-s-UFMC

(b)

Normalized frequency-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

PSD

(dBW

/Hz)

-100

-80

-60

-40

-20

0

2048 Subcarriers

ZP-DFT-s-OFDMZP-DFT-s-UFMC

(c)

CCDF0 1 2 3 4 5 6 7 8 9 10 11

PAPR

(dB)

10-5

10-4

10-3

10-2

10-1

100

OFDMZP-DFT-s-OFDMZP-DFT-s-UFMC

(d)

Normalized frequency-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

PSD

(dBW

/Hz)

-100

-80

-60

-40

-20

0

2064 Subcarriers

ZP-DFT-s-OFDMZP-DFT-s-UFMC

(e)

CCDF0 1 2 3 4 5 6 7 8 9 10 11

PAPR

(dB)

10-5

10-4

10-3

10-2

10-1

100

OFDMZP-DFT-s-OFDMZP-DFT-s-UFMC

(f)Figure 5.5 : For 16, 24 and 32 subcarriers, PSD results of ZP-DFT-s-UFMC and

ZP-DFT-s-OFDM are shown at (a), (c) and (e), respectively. Moreover,respective combined PAPR results are shown at (b), (d) and (f).

60

X1

M1-DFT

Zero Padding N1-IDFT Filtering

F1

DAC

Precoding

X2

M2-DFT

Zero Padding N2-IDFT Filtering

F2

Precoding

XB

MB-DFT

Zero Padding NB-IDFT Filtering

FB

Precoding

.

.

.

.

.

.

.

.

.

Figure 5.1 : Block diagram of the transmitter implementation.

Figure 5.2 : Block diagram of the receiver implementation.

supported with zero-padding, zeros are added into data symbols in order to implement

ZP-DFT-s-UFMC. This step can be represented as

ui [m] =

(

0, m = 0,1, . . . ,Z �1ui[m], m = Z,Z +1, . . . ,M�1,

(5.1)

where Z is the number of added zero symbols, M is the length of DFT, i = 1,2, . . . ,B is

the index of the subband where total number of subbands is B, ui is the data symbols at

ith subband before the M-DFT operation. Then to complete ZP-DFT-s-UFMC symbol

generation, spreading with DFT-precoding can be implemented. Lets assume that the

output of this operation is Xi[k], where k = 0,1, . . . ,K �1 is the subcarrier index with

total number of subcarriers K. If spreading version of UFMC is not considered, then

weighted correlative coding operation can be applied to Xi[k] as [28]

Xi[k] =p

22

(e jq Xi[k]+Xi[k+1]), k = 1,2, . . . ,K �1, (5.2)

51

TRANSMITTER

Figure 5.1 : Block diagram of the transmitter implementation.

Cyclic-Prefix

RemovalADC

Synchronization

Filtering/Channel

Equalization

Estimated Bits2N-DFT ML

Detection

Figure 5.2 : Block diagram of the receiver implementation.

supported with zero-padding, zeros are added into data symbols in order to implement

ZP-DFT-s-UFMC. This step can be represented as

ui [m] =

(

0, m = 0,1, . . . ,Z �1ui[m], m = Z,Z +1, . . . ,M�1,

(5.1)

where Z is the number of added zero symbols, M is the length of DFT, i = 1,2, . . . ,B is

the index of the subband where total number of subbands is B, ui is the data symbols at

ith subband before the M-DFT operation. Then to complete ZP-DFT-s-UFMC symbol

generation, spreading with DFT-precoding can be implemented. Lets assume that the

output of this operation is Xi[k], where k = 0,1, . . . ,K �1 is the subcarrier index with

total number of subcarriers K. If spreading version of UFMC is not considered, then

weighted correlative coding operation can be applied to Xi[k] as [28]

Xi[k] =p

22

(e jq Xi[k]+Xi[k+1]), k = 1,2, . . . ,K �1, (5.2)

51

RECEIVER

IEEE 5G GREECE SUMMIT 14

ITU WCRL

Summary

1. OFDM&OFDMAareproventechniques2. Densenetworksmayrequiremoreflexiblewaveformdesign

3. UFMCisagoodop;onintermsofitsflexibility

IEEE 5G GREECE SUMMIT 15

ITU WCRL

Selected References:

IEEE 5G GREECE SUMMIT 16

[SIOHAN, 2002] P. Siohan, C. Siclet and N. Lacaille, "Analysis and design of OFDM/OQAM systems based on filterbank theory," in IEEE Transactions on Signal Processing, vol. 50, no. 5, pp. 1170-1183, May 2002. [FETTWEIS, 2009] G. Fettweis, M. Krondorf and S. Bittner, "GFDM - Generalized Frequency Division Multiplexing,” IEEE 69th Vehicular Technology Conference, Barcelona, 2009, pp. 1-4. [ABDOLI, 2015] J. Abdoli, M. Jia and J. Ma, "Filtered OFDM: A new waveform for future wireless systems," 2015 IEEE 16th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Stockholm, 2015, pp. 66-70. [BERARDINELLI, 2013] G. Berardinelli, F. M. L. Tavares, T. B. Sørensen, P. Mogensen and K. Pajukoski, "Zero-tail DFT-spread-OFDM signals," 2013 IEEE Globecom Workshops (GC Wkshps), Atlanta, GA, 2013, pp. 229-234. [VAKILIAN, 2013] V. Vakilian, T. Wild, F. Schaich, S. ten Brink and J. F. Frigon, "Universal-filtered multi-carrier technique for wireless systems beyond LTE," 2013 IEEE Globecom Workshops (GC Wkshps), Atlanta, GA, 2013, pp. 223-228

ITU WCRL

Questions?

Thankyouforyoua"en;on!

gkurt@itu.edu.tr

IEEE 5G GREECE SUMMIT 17