1. Evaluation of coded OFDM systems using different type of codes
Dinh Van Linh, Simona V. Halunga, Member, IEEE,
Alexandru Vulpe, Member, IEEE
1Faculty of Electronics, Telecommunications and IT
University POLITEHNICA of Bucharest
Bucharest, Romania
@VulpeAlex

2. Content Short Biography Introduction Performance Analysis Results
Conclusions
@VulpeAlex
D. Van Linh, et.al. (2015)

3. Short Biography (1)
Graduated from the Faculty of Electronics, Telecommunications and Information Technology at the University “Politehnica” of Bucharest (UPB), Romania (
PhD in 2014, thesis on RRM algorithms in LTE- Advanced
Currently Post-doc Researcher focused on the field of 4G/5G communication, Big Data, Cloud Computing, Mobile application development, Ambient Assisted Living
Working in projects related to designing scalable transceivers for critical communication
@VulpeAlex
D. van Linh (2015)

4. Short Biography (2) FP7 (1 on-going) NATO SfP (1 on-going)
Projects –
FP7 (1 on-going)
REDICT : Regional Economic Development by ICT
eWALL : eWall for Active Long Living
NATO SfP (1 on-going)
ORCA: Optimization and Rational use of Wireless Communication Bands
RIWCoS: Reconfigurable Interoperability of Wireless Communications Systems
Danube Transnational Programme (2 under review)
TERES
CAMPINNO
National (more than 10 past projects, 5 on-going)
SARAT-IWSN : Scalable Radio Transceiver for Instrumental Wireless Sensor Networks
CORONA: Investigation of the cosmic radiation in the Universe using advanced techniques
@VulpeAlex
D. van Linh (2015)

5. Introduction(1)
OFDM based on spreading high speed data to be transmitted over large no. of low rate orthogonal carriers
Advantages
High spectral efficiency
Reduces Inter Symbol & Inter Frame Interference
Computationally efficient using FFT/IFFT on chip implemented blocks
Disadvantages
Sensitive to frequency offset
High PAPR.
ORTHOGONAL Frequency Division Multiplexing (OFDM) is a multicarrier transport technology for high data rate communication systems, such as 4th generation LTE. The OFDM concept [1] is based on spreading the high speed data to be transmitted over a large number of low rate carriers.
The carriers are orthogonal to each other and frequency spacing between them are created by using the Fast Fourier transform (FFT).
OFDM technique offers a high spectral efficiency, reduces the Inter Symbol Interference (ISI) and Inter Frame Interference (IFI) by reducing the individual rate on each carrier and by the use of the Cyclic Prefix (CP) and is computationally efficient by using FFT / IFFT on chip implemented blocks for the modulation and demodulation functions.
On the other hand OFDM proves to be sensitive to frequency offset and suffers from high Peak to Average Ratio [2].
In order to improve the system performances, OFDM technique is often used in combination with Forward Error Encoding / Decoding techniques, reducing thus the Bit Error Rates (BER) and improving the overall system performances
@VulpeAlex
D. van Linh (2015)

6. Introduction(2)
OFDM often used in combination with Forward Error Encoding / Decoding techniques
Bit Error Rates (BER) reduced
overall system performances improved
Termed Coded OFDM (COFDM)
Various research performed for improving COFDM
Combinations of 16-QAM, 64-QAM with LDPC coding over Rayleigh channel
Reduced complexity implementations
Investigation of COFDM in different fading channels
D. van Linh (2015)
@VulpeAlex

7. Performance analysis results obtained based on Matlab simulation,
study the effects of
modulation type,
number of subcarriers
length of guard interval
on the BER vs SNR for
BPSK,
QPSK,
16QAM
64QAM
OFDM signal with convolutional encoding
Viterbi decoding
Hamming encoding decoding as well,
D. van Linh (2015)
@VulpeAlex

8. Results CP
1/4
1/8
1/16
1/32
CG(Pe=10-2)
3.8126
3.0676
2.8923
3.1552
CG(Pe=10-3)
4.7977
4.4192
4.3538
2.9549
Coding gain of OFDM systems, BPSK modulation, generator polynomials [171, 130] different CPs
It can be seen that, in both cases, the encoded system reaches a BER=10^-3 around SNR=13 dB. and the coding gains at BER=10^-3 is between 3 and 5 dB.
Fig.
OFDM system over Viterbi decoding, BPSK modulation with generator [171, 130] with different CPs
@VulpeAlex
D. van Linh (2015)

9. Results CP
1/8
1/16
1/32
CG(Pe=10-2)
2.4970
2.5343
2.3480
2.4784
CG(Pe=10-3)
3.3660
2.8602
3.5230
Coding gain of OFDM systems, BPSK modulation, generator polynomials [161, 133] different CPs
Comparing the results one can see that, using the same degrees for the generator polynomials the results are similar.
However the coding gains of generator [171, 130] are slightly higher than the ones obtained with the generator [160, 133].
In both cases, the system performs better without coding for SNR’s up to approximately 5dB, because, with a large number of errors the Viterbi decoder does not perform well [11].
The influence of the cyclic prefix length is less than 1 dB in all cases.
OFDM system over Viterbi decoding, BPSK modulation with generator [161, 133] with different CPs
@VulpeAlex
D. van Linh (2015)

10. Results CP
1/4
1/8
1/16
1/32
16 QAM
4.8532
5.2829
4.9805
4.8691
16 PSK
6.1248
6.4427
6.322
6.1531
Coding gain of OFDM systems at Pe=10-3, 16 QAM and 16 PSK modulation, different CPs
Figures 4 and 5 show the BER versus SNR results for 16-PSK and 16-QAM respectively.
As the modulation order increase the decrease rate of BER with SNR is lower.
The encoded system reaches a BER=10-3 at SNR=25 dB for 16 PSK and 22.5 dB for 16 QAM respectively.
Moreover, the results obtained with 16 QAM are better than the ones obtained with 16 PSK for both coded and uncoded systems.
Still, the uncoded system performs better than the coded one till 15 dB for 16 PSK and 13.5 dB for 16 QAM respectively.
OFDM system over Viterbi decoding for 16-PSK with different CPs
@VulpeAlex
D. van Linh (2015)

11. Results CP
1/4
1/8
1/16
1/32
16 QAM
4.8532
5.2829
4.9805
4.8691
16 PSK
6.1248
6.4427
6.322
6.1531
Coding gain of OFDM systems at Pe=10-3, 16 QAM and 16 PSK modulation, different CPs
The coding gains for BER=10-3 are shown in Table 3, and we can observe that, regardless the value of CP, the coding gains obtained with 16 PSK are around 1dB higher than the ones obtained with 16QAM
OFDM system over Viterbi decoding for 16-QAM with different CPs
@VulpeAlex
D. van Linh (2015)

12. Results CP
1/4
1/8
1/16
1/32
64 QAM
5.6677
5.4659
5.7044
5.9978
64 PSK
6.0444
6.4028
5.9966
6.7372
Coding gain of OFDM systems at Pe=10-3, 64 QAM and 64 PSK modulation, different CPs
Figs. 6 and 7 show the BER versus SNR results for 64-PSK and 64-QAM respectively.
Again, the 64 QAM obtains better results than 64 PSK, the encoded system reaching a BER=10-3 at SNR=35 dB for 64 PSK and 30dB for 64 QAM respectively.
OFDM system over Viterbi decoding for 64-PSK with different CPs
@VulpeAlex
D. van Linh (2015)

13. Results CP
1/4
1/8
1/16
1/32
64 QAM
5.6677
5.4659
5.7044
5.9978
64 PSK
6.0444
6.4028
5.9966
6.7372
Coding gain of OFDM systems at Pe=10-3, 64 QAM and 64 PSK modulation, different CPs
The coding gains for BER=10^-3 are shown in Table 5, and we can observe, regardless the value of CP, that the coding gains obtained with 16 PSK are between 0.5 and 1dB higher than the ones obtained with 16QAM.
OFDM system over Viterbi decoding for 64-QAM with different CPs
@VulpeAlex
D. van Linh (2015)

14. Results Ns
1024
2048
4096
8192
CG(Pe=10-2)
4.3347
2.7867
1.3417
CG(Pe=10-3)
6.1610
3.4547
3.5699
2.2839
Coding gain of OFDM systems, 64QAM modulation, different values of Ns
From the results above we can conclude that high modulation order schemes are considerably less resilient to noise and interference, best performances being achieved with BPSK modulation. Moreover, although QAM type modulations generally perform better than PSK ones, using coding (especially convolutional) can bring better efficiency to PSK modulation schemes.
By changing the length of the cyclic prefix, the performances are not significantly affected. Therefore, if the OFDM system is affected only by AWGN we can set the guard interval at a lower value, since the inter-symbol interferences can be avoided even in this case.
Next we will keep cyclic prefix at its lowest value CP=1/8, and modify the number of subcarriers. Fig. 8 shows the BER versus SNR for 64 QAM, and Ns = 1024, 2048, 4096, It can be seen that, while the uncoded system performances are not significantly affected by the number of subcarrier increase, the coded one obtains lower BERs for smaller number of subcarriers. The coding gains are given in table 5, and they decrease as the number of carriers increases.
OFDM system over Viterbi decoding for 64-QAM, different values of Ns
@VulpeAlex
D. van Linh (2015)

15. Results OFDM system over Hamming codes for BPSK, different CPs
Fig. 9 presents the BER versus SNR results for BPSK OFDM modulation with and without Hamming encoding / decoding, for Ns=2048, CP={1/4, 1/8, 1/16, 1/32} while in table 6 the obtained coding gains are shown.
The results obtained with coded and uncoded systems are similar, so the coding gain is low, the encoded system reaching a BER=10-3 at around SNR=15 dB
OFDM system over Hamming codes for BPSK, different CPs
@VulpeAlex
D. van Linh (2015)

16. Results
Finally, Fig. 10 shows the BER results versus SNR for BPSK-OFDM with and without Hamming encoding, with fixed CP=1/4, and different values of Ns ={2048, 4096, 8192}.
By contrast with convolutional encoding, it can be seen that as the number of OFDM subcarriers increases the encoding effect is more significant and the BER decreases more rapidly with the SNR.
OFDM system over Hamming codes for BPSK, different values of Ns
@VulpeAlex
D. van Linh (2015)

17. Conclusions
Analysis of effect of cyclic prefix length, no. of subcarriers, modulation scheme, types of codes used in COFDM system.
If communications channel affected by AWGN only, length or type of the guard interval does not influence the performances of COFDM.
Convolutional codes with Viterbi decoding more efficient than Hamming codes.
Best choice of OFDM system in AWGN channel is QAM modulation and high number of subcarriers.
@VulpeAlex
D. van Linh (2015)

18. Any questions ? vanlinhvn.fr@gmail.com
University Politehnica of Bucharest
@VulpeAlex
D. van Linh (2015)

19. References
@VulpeAlex
D. van Linh (2015)